Am J Physiol Regul Integr Comp Physiol 291: R1040 –R1048, 2006;
doi:10.1152/ajpregu.00116.2006.
Endurance exercise differentially stimulates heart and axial muscle
development in zebrafish (Danio rerio)
T. van der Meulen,1 H. Schipper,1 J. G. M. van den Boogaart,1
M. O. Huising,2 S. Kranenbarg,1 and J. L. van Leeuwen1
1
Experimental Zoology Group, Wageningen University, Wageningen; and 2Department of
Animal Physiology, Radboud University Nijmegen, Nijmegen, The Netherlands
Submitted 15 February 2006; accepted in final form 16 May 2006
role in the development and maintenance of vertebrate tissues (2, 4, 20).
Particularly for muscle, bone, and connective tissue, mechanical loads are considered to modulate the genetically
determined tissue architecture. To investigate the role of
mechanical load, it can be altered in many different ways. A
common and effective way to impose increased mechanical
loads is endurance training, where fish offer several advantages over mammalian models. As many species of fish
control buoyancy via their swim bladder, the effects of
gravitational forces are limited compared with those of
terrestrial animals. Therefore, the main loads on the tissue
arise from the forces generated by the animal and the
reactions of the surrounding medium. Another advantage of
fish is that most of their locomotory musculature is relatively simple, as it is located predominantly around the axial
skeleton in a series of myomeres. Furthermore, in most fish
species, slow and fast muscle fibers are spatially segregated:
the bulk of fast muscle is located close to the axis, and a
small strip of slow muscle is located under the skin at the
level of the horizontal septum, which facilitates the analysis.
The cyprinid zebrafish (Danio rerio) is a popular model
animal in developmental biology research (26, 31, 36),
which has resulted in a wealth of information on the species,
in particular on gene sequences and early developmental
processes, mainly during the first 5 days postfertilization
(dpf).
Whereas muscle growth in larger species of fish, such
as salmonids, is characterized by extensive hyperplasia (cell
division), small and slow-growing species such as zebrafish (maximum size 50 mm) display only limited muscle
hyperplasia (22, 34, 35). In zebrafish, the number of red
muscle fibers increases into adulthood, whereas growth of
white and intermediate muscle by hyperplasia stops at a
total length of ⬃20 mm (34). The limited contribution of
hyperplasia to muscle development in small fish species
such as zebrafish might preclude radical hyperplastic muscular adaptations to endurance training and suggests other
means of adaptation.
One of these means of adaptation is growth through
hypertrophy (cell growth), which often also accompanies
training in larger fish species (9, 10). Myotomal muscle
hypertrophy starts in zebrafish ⬎15 mm total length (TL)
and continues into adulthood (34). A synergistic adaptation
to endurance training is an increase in cardiac output. In this
respect, the cardiac muscle of zebrafish is of special interest,
as it retains the ability to grow through hyperplasia, as well
as hypertrophy (15) and is capable of regenerating from
injury without the formation of scar tissue (28), suggesting
a great potential for adaptation to increased physical demands. The increase of muscle capillarization, which improves the local supply of oxygen and transport of metabolic
waste products, is yet another property that may benefit the
adaptation to endurance training (11, 27).
To investigate the role of mechanical load in muscle
development, we increased mechanical load by subjecting
juvenile zebrafish to endurance training for 6 h/day from 14
dpf. During a 10-wk training period, axial and heart muscle
composition was compared with that of untrained (control)
siblings in an integrated analysis of endurance training. Our
results indicate that endurance training has profound and
differential effects on axial and heart muscle in zebrafish
development.
Address for reprint requests and other correspondence: T. van der Meulen,
Wageningen Univ., Experimental Zoology Group, Marijkeweg 40, 6709 PG
Wageningen, The Netherlands (e-mail: [email protected]).
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
endurance exercise; mechanical load; aerobic muscle development
MECHANICAL LOAD PLAYS AN IMPORTANT
R1040
0363-6119/06 $8.00 Copyright © 2006 the American Physiological Society
http://www.ajpregu.org
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
Van der Meulen, T., H. Schipper, J. G. M. van den Boogaart,
M. O. Huising, S. Kranenbarg, and J. L. van Leeuwen. Endurance
exercise differentially stimulates heart and axial muscle development
in zebrafish (Danio rerio). Am J Physiol Regul Integr Comp Physiol
291: R1040 –R1048, 2006; doi:10.1152/ajpregu.00116.2006.—Mechanical load is an important factor in the differentiation of cells and
tissues. To investigate the effects of increased mechanical load on
development of muscle and bone, zebrafish were subjected to endurance swim training for 6 h/day for 10 wk starting at 14 days after
fertilization. During the first 3 wk of training, trained fish showed
transiently increased growth compared with untrained (control) fish.
Increased expression of proliferating cell nuclear antigen suggests that
this growth is realized in part through increased cell proliferation. Red
and white axial muscle fiber diameter was not affected. Total crosssectional area of red fibers, however, was increased. An improvement
in aerobic muscle performance was supported by an increase in
myoglobin expression. At the end of 10 wk of training, heart and axial
muscle showed increased expression of the muscle growth factor
myogenin and proliferating cell nuclear antigen, but there were major
differences between cardiac and axial muscle. In axial muscle, expression of the “slow” types of myosin and troponin C was increased,
together with expression of erythropoietin and myoglobin, which
enhance oxygen transport, indicating a shift toward a slow aerobic
phenotype. In contrast, the heart muscle shifts to a faster phenotype
but does not become more aerobic. This suggests that endurance
training differentially affects heart and axial muscle.
ENDURANCE TRAINING IN ZEBRAFISH
R1041
MATERIALS AND METHODS
Fish
Embryos were generated from the gol-1 strain, characterized by
reduced body pigmentation (32) and high fecundity, by natural mating
and maintained at 28.5°C. When the larvae were free swimming, they
were fed Paramecium three times a day. From 7 dpf, increasing
amounts of live brine shrimp were added to the feed, and the amount
of Paramecium was reduced proportionally. At 14 dpf, the larvae
were transferred to the training and control tanks. From this age, the
larvae were fed only brine shrimp twice daily (15 min before and
immediately after training). From 35 dpf, the pretraining feed was
supplemented with CypriCo Crumble Excellent Pellets (Coppens
International).
Training Setup
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
A training setup for zebrafish was custom-built (Fig. 1A). To
linearize the flow, water in the flow tank was led through a compact
stack of straws and a tapering width of the channel. Grids placed in the
tank kept the fish in a partition of the 1-m-long, 0.15-m-wide,
0.28-m-high flow tank. Copper-free tap water, which was filtered
biologically, was used. Salts were added to the water to a conductivity
of ⬃650 ␮S/cm. The temperature, 26.8°C (SD 0.5), was measured
daily. Nitrate [21.0 mg/l (SD 6.8)], nitrite [0.0 mg/l (SD 0.0)],
ammonia [0.0 mg/l (SD 0.1)], and pH [8.25 (SD 0.07)] were measured
weekly. The water speed in the flow chamber could be varied
gradually between 0 and 17 cm/s.
Training Regimen
Fish from six clutches were randomly assigned to the control or the
trained group. Each day, starting at 14 dpf, larvae were trained
continuously for 6 h by enforcement of a flow rate of three total body
lengths per second (TL/s). Flow rate was adjusted after every sampling day. After 1 wk of training (at 21 dpf), flow rate was increased
to 5 TL/s and increased daily on the basis of the extrapolated growth
data from previous samples to closely match 5 TL/s. This training
regimen represents about twice the normal activity of developing
zebrafish (14). At 50 dpf, the maximum flow rate of the system was
reached at 17 cm/s; thus the speed gradually declined toward 4.2 TL/s
at 84 dpf because of growth of the fish. At the beginning of each
training session, the flow rate was gradually increased over the course
of 2–3 min to minimize stress. Three fish were removed from the
experiment: two trained fish that stopped swimming during training
and one control fish that was found dead early in the experiment. The
training protocol was approved by the Dierexperimenten Commissie
(Experimental Animal Committee) of Wageningen University (under
number 2004020.d) and in accordance with the Dutch Wet op de
Dierproeven (Law on Experiments With Animals).
Sampling
Fish were sampled at 14 (before training), 17, 21, 24, 28, 35, 45,
63, and 84 dpf. Sampling of trained fish started immediately after the
end of 6-h training session of that day. The control fish were sampled
immediately before the trained fish. Both groups of fish were irreversibly anesthetized by 1.0% tricaine methane sulfonate buffered
with 1.5% NaHCO3. The animals to be used for analysis of gene
Fig. 1. A: schematic representation of the flow chamber. After water enters the
aquarium, large particles are removed by a grid. For creation of laminar flow,
water is led through a parallel array of tightly packed straws and a tapering
width of the channel. Zebrafish are prevented from escaping by the 2nd and 3rd
grids. Inset: aquarium dimensions. B: in near-stagnant water, zebrafish cover
the entire aquarium and swim in all directions. C: during training, zebrafish
swim against the direction of flow and preferably maintain a school near the
front-end grid. Scale bars, 10 cm.
AJP-Regul Integr Comp Physiol • VOL
291 • OCTOBER 2006 •
www.ajpregu.org
R1042
ENDURANCE TRAINING IN ZEBRAFISH
expression or determination of individual muscle fiber area (n ⫽ 8)
were immediately snap frozen in liquid nitrogen and stored at ⫺80°C
until RNA isolation. The other animals (n ⫽ 14) were photographed
in lateral view with a digital camera (model DP50; Olympus) mounted
on a microscope (Stemi SV11; Zeiss) with AnalySIS software (Soft
Imaging System). Because ⬎45-dpf larvae were too large for this
approach, a camera (model D-100; Nikon) with a 105-mm 1:2.8 lens
(AF NIKKOR) and 0.5-cm graph paper were used for image calibration. Four of these fish were fixed overnight at 4°C in Bouin’s fixative
(125 ml of 40% neutral formaldehyde, 375 ml of saturated picric acid,
and 50 ml of acetic acid, filtered before use) for histological analysis.
The remaining 10 fish were weighed individually on a scale (model
PE 360; Mettler-Toledo) after careful removal of excess liquid. After
10 wk of training, at 84 dpf and in addition to the normal sampling,
individual organs were isolated from eight fish from each group.
Length measurements were taken from the calibrated photos. TL is
measured from the tip of the snout to a point halfway between the
dorsal and ventral tips of the tail. Data were plotted over time as
means with SD. A sigmoidal curve {y ⫽ 1/[b1 ⫹ b2 ⴱ exp(⫺x/b3)]}
was fitted through the length data in Matlab 7.0 (Mathworks) using
least squares. The interval is a 95% nonsimultaneous confidence
interval around the regression line for the true value of the function at
the specified input values.
Determination of Total Red Muscle Fiber Area
For determination of total red muscle fiber area, the fish were
dehydrated and embedded in paraffin (Paraclean) according to standard histological protocols. Sections (5 ␮m thick) were cut on a
microtome (model 2040; Reichert-Jung) and collected on proteinglycerin-coated slides. For histological staining, paraffin was removed
from the sections, which were rehydrated and subjected to Crossmon
staining (5). Two trained and two control fish at 21, 24, 28, 35, 45, and
63 dpf were used. For each fish, the size of the total red muscle area
in each of the four body quarters (delineated by the horizontal and
vertical septa) was individually assessed and then averaged for a more
accurate estimate of cross-sectional area. These averages were used in
Determination of Individual Muscle Fiber Area
To accurately determine individual muscle fiber diameter, we used
7-␮m cryostat cross sections of fish (45 and 84 dpf) that were flash
frozen in liquid nitrogen. The sections were collected on polysine
slides (Menzel-Gläser, Braunschweig, Germany) and stained without
prior fixation for the presence of slow muscle fibers with an antibody
against zebrafish slow muscle [S58 supernatant, a generous gift of Dr.
Frank E. Stockdale (6)]. Fast muscle fibers were visualized with
standard Crossmon staining. Red muscle fibers were photographed
near the horizontal septum. White muscle fibers were photographed in
an epaxial area at equal distances from the horizontal septum, midline,
and skin. Individual muscle fibers were manually traced on the
photographs, and their cross-sectional areas were measured using
AnalySIS software (version 3.1). Differences in average fiber size
between control and trained fish were evaluated with an independentsamples t-test (SPSS version 12.0.1). P ⬍ 0.05 was accepted as
significant.
RNA Isolation and cDNA Synthesis
RNA was isolated essentially as described elsewhere (33) with
some minor modifications. Briefly, RNA was isolated with an RNeasy
kit (Qiagen) according to the manufacturer’s instructions. After elution in milliQ water, RNA was concentrated (but not dried) in a Speed
Vac. Up to 1,000 ng of total RNA were used as the starting material
for random hexamer-primed single-strand cDNA synthesis (Invitrogen). A non-RT control was incorporated for each sample. Samples
were filled with milliQ water to a final concentration corresponding to
1 ng of total RNA used as starting material for cDNA synthesis per
microliter and stored at ⫺20°C.
Real-Time Quantitative PCR
Primer express software (Applied Biosystems) was used to design
primers (Table 1). Five microliters of cDNA and forward and reverse
primer (300 nM each) were added to 12.5 ␮l of Quantitect Sybr Green
Table 1. Primers used in RT-PCR experiments
Primer
Gene
Accession No.
Forward
Reverse
40S
␤-Actin
Epo
Gdf8
IGF-1 Ea2
IGF-Ra
IGF-Rb
myhz2
myhz5
Myogenin
Myoglobin
NADHd
PCNA
PFK-m
SDHa
Titin
TropC
Fast
Slow
CA472846
AF025305
Dr.11767
NM_131019
AH010825
AF400275
AF400276
NM_152982
AY333451
NM_131006
AY337025
NM_200189
AF140608
BC081652
BC045885
AY081167
AAACAGCCCACCATCTTCCA
CAACAGGGAAAAGATGACACAGAT
CGCGTCCTCGACCATTTC
CGGCTGATGCCCGTTAAG
TGGGCATAGCCACTCTTCCT
ATATGTAAGGATCAAGCACACTCATACTCT
GTTTCAGCCACGGGTTTGA
ACAGTTTTTCAACCACCACATGTT
GCTGGAGAATGAGGTGGAGTTG
AACCAGCAGGAGCATGAACAG
GCTGGGCGAACTGCTGAA
TGCGGATCGGAGATGTGA
ACGCTGGCACTGGTCTTTG
GGTCTTTACACTGGAGCTAAAGTCTTCT
TGCAGCGTTCGGTCTGTCT
GATGCAACTGATCGTCCTAAGGTA
CTGTGATAACGAGGGAGCTTTTC
CAGCCTGGATGGCAACGT
TGCAATCGTCCTTACAAGTTCTCA
AGACGTGACTCCTGCGTTCA
TTCTGCCCCCTGTGTTTCC
GCGTATGTATTATCCACAAGCTCTTC
GGTTTTGTCTCGTCCTCCTGTT
AATGCAAGCGGCCAAGTC
AGTCTGGTAGGTGAGCTCCTTGA
AGAGGACGACACCCCAGTATG
TTTGTGAATGTTGGCGTGTGT
CGACCTGCGCCCTTGTACT
TGCCAAGCTGCTCCACATC
CGGATGTGATCCCCTCCAT
GTTGCCTAACGCAGCATTAATG
GGCATCCAATCTGATTTCTTCAC
NM_131563
NM_181498
GAAGCTAAACCACAATCATGACTGA
GCAGCGGCAGAGCAACTC
TCGAAGGCAGCCTTGAACTC
GCGTCCTGAACAAAAATGTCAA
Epo, erythropoietin; Gdf8, growth and differentiation factor 8; IGF, insulin-like growth factor; myhz2, fast myosin heavy chain; myhz5, slow myosin heavy
chain; NADHd, NADH dehydrogenase; PCNA, proliferating cell nuclear antigen; PFK-m, muscle phosphofructokinase; SDHa, succinate dehydrogenase subunit
A; TropC, troponin C.
AJP-Regul Integr Comp Physiol • VOL
291 • OCTOBER 2006 •
www.ajpregu.org
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
Length Measurements
a Wilcoxon signed-rank test over time to assess statistical differences
between trained and control fish. P ⬍ 0.05 was accepted as significant.
ENDURANCE TRAINING IN ZEBRAFISH
R1043
PCR Master Mix (Qiagen), with addition of milliQ to a final volume of
25 ␮l. Real-time quantitative PCR (95°C for 15 min, 45 cycles at 94°C
for 15 s, 60°C for 30 s, and 72°C for 30 s, followed by 60°C for 1 min)
was carried out on a Rotorgene 2000 real-time cycler (Corbett Research,
Sydney, Australia). After each run, melt curves were collected for
detection of fluorescence from 60°C to 90°C at 1°C intervals. The level
of expression was determined at a threshold of 0.04. Target gene expression was corrected for reference gene expression and for primer efficiency and plotted relative to the control group as described previously
(19). Dual internal reference genes (40S and ␤-actin) were used in all
real-time quantitative PCR experiments, and results were confirmed to be
very similar after standardization to either gene. Only results standardized
for ␤-actin expression are shown. Differences were evaluated with a
Mann-Whitney U-test with SPSS (version 12.0.1) and were considered
significant when P ⬍ 0.05.
Behavior
In the control tank, zebrafish covered the entire aquarium in
their exploration and foraging behavior (Fig. 1B). In the training tank, on initiation of water flow, the fish momentarily
swam more actively in all directions before they began to swim
against the direction of the current (Fig. 1C). At all water
speeds, the fish retained an intermittent swimming pattern:
after small bursts of activity, the fish coasted. Coasting is
evident from a lack of tail movement and slacking of the
dorsal, pelvic, and tail fins. The fish actively pursued particles
and continued this behavior during training. When the water
speed was returned to normal at the end of a training session,
the trained fish immediately assumed the swimming behavior
of the control fish.
Size and Mass
For determination of the effects of training on growth, the
length and mass of the fish were determined. At the start of the
experiment, TL of the fish was 8.3 mm (SD 1.2). At the end of
the experiment TL was 40.3 mm (SD 2.2) and 40.7 mm (SD
1.7) for control and trained fish, respectively (Fig. 2A). Although trained fish were similar in size to the control fish at the
end of the experiment, they were larger during much of the
training period. The difference is significant (P ⬍ 0.05) from
21 to 45 dpf. The trained fish initially grew faster than the
control fish, but after 35 dpf, the control fish grew faster (Fig.
2A, inset) until, at 63 dpf, both groups are again indistinguishable in terms of size. Mass of the control fish increased from
0.0055 g (SD 0.0021) at 14 dpf to 0.68 g (SD 0.19) at 84 dpf
(Fig. 2B). The trained fish tended to be heavier at all time
points after 14 dpf; however, this difference in mass reached
statistical significance (P ⬍ 0.05) only at 21 and 24 dpf.
Muscle Fiber Cross-Sectional Area
To establish whether training increases the red muscle fiber
mass, the total cross-sectional area of all red muscle fibers was
determined over time. When the trained and control fish were
compared over the entire training period using a Wilcoxon
signed-rank test over time, the trained fish had a larger (P ⬍
0.05) total cross-sectional red muscle fiber area (Fig. 3). In
addition to a larger total cross-sectional red muscle fiber area,
the control and trained fish showed ample fat deposition.
To discover whether training affects hypertrophic muscle
fiber growth, the cross-sectional area of individual white and
AJP-Regul Integr Comp Physiol • VOL
Fig. 2. Trained fish are transiently larger and heavier than control fish. A: total
length (TL) correlated with time. Gray sigmoidal curve with 95% confidence
interval was fitted through control data; black sigmoidal curve was fitted
through data from trained fish. Inset: 1st-order differentiation of sigmoidal
curves, which equals growth speed. Arrow, time from which growth rate of
control fish exceeds that of trained fish. B: pairwise comparisons of the mass
of control and trained fish over time. *Significant difference (P ⬍ 0.05).
red muscle fibers was determined at 45 and 84 dpf. At 45 dpf, the
mean cross-sectional area of the individual white muscle fibers
was 1,571 ␮m2 (SD 839) and 1,801 ␮m2 (SD 744) in control (n ⫽
4 fish, 374 fibers) and trained (n ⫽ 4 fish, 566 fibers) fish,
respectively (Fig. 4A). Fibers from trained fish were not significantly larger than fibers from control fish (P ⫽ 0.602). At 84 dpf,
the mean cross-sectional area of the white fibers was 3,331 ␮m2
(SD 1,231) and 3,228 ␮m2 (SD 1,250) in control (n ⫽ 4 fish, 489
fibers) and trained (n ⫽ 4 fish, 455 fibers) fish, respectively (Fig.
4B), which is not statistically significant (P ⫽ 0.553).
At 45 dpf, cross-sectional area of the individual red muscle
fibers was 732 ␮m2 (SD 264) and 730 ␮m2 (SD 242) in trained
(n ⫽ 4 fish, 277 fibers) and control (n ⫽ 4 fish, 288 fibers) fish,
respectively (Fig. 4C), which is not statistically significant
(P ⫽ 0.372). At 84 dpf, cross-sectional area of the red fibers
was 1,298 ␮m2 (SD 478) and 1,308 ␮m2 (SD 393) in control
(n ⫽ 4 fish, 313 fibers) and trained (n ⫽ 4 fish, 249 fibers) fish,
respectively (Fig. 4D), which is not significantly different (P ⫽
0.935).
Kinetics of Muscle Gene Expression
The expression of muscle-specific genes is regulated over
time as a result of growth and development and was predicted
291 • OCTOBER 2006 •
www.ajpregu.org
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
RESULTS
R1044
mm
ENDURANCE TRAINING IN ZEBRAFISH
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
Fig. 3. Total red muscle fiber area transiently increases as a result of endurance training. Cross-sectional area of the 4 quarters of total red muscle fiber
area were measured and averaged to obtain values for individual fish.
to be regulated by endurance training. We selected a panel of
representative muscle metabolism processes for investigation:
sarcomere construction, energy generation, growth, cell proliferation, and oxygen transport. The expression kinetics of the
selected genes were established in trained and control fish over
the initial 3 wk of training and compared. Because of agerelated changes in gene expression, all expression values are
given relative to the initial samples at 14 dpf, so that control
and trained values may be compared.
Genes involved in sarcomere construction. A decrease (P ⬍
0.05) was observed in the expression of fast myosin heavy
chain 2 (myhz2) after 21 dpf (Fig. 5A) and in slow myosin
heavy chain 5 (myhz5) after 14 dpf (Fig. 5B). Fast musclespecific troponin C showed a slightly elevated expression at 21
dpf, and slow muscle-specific troponin C remained constant
(not shown).
Genes involved in energy generation. Muscle phosphofructokinase (PFK-m) is involved in glycolysis. Its expression
remained constant over time. NADH dehydrogenase (NADHd)
and succinate dehydrogenase complex subunit A (SDHa) are
involved in mitochondrial metabolism. NADHd increased
(P ⬍ 0.05) after 14 dpf (Fig. 5C). SDHa expression remained
constant (not shown).
Genes involved in growth. Growth and differentiation factor
8 (gdf8 and myostatin), an inhibitor of muscle growth, was
elevated after 14 dpf (Fig. 5D). Expression of myogenin, a
muscle growth promoter, showed much variation but appeared
to increase as well (Fig. 5E). Insulin-like growth factor (IGF),
which also regulates muscle growth, was upregulated after 14
dpf (Fig. 5F), whereas expression of IGF receptors a and b was
not changed (not shown).
Genes involved in cell proliferation. Proliferating cell nuclear antigen (PCNA), an indicator of cell division activity,
showed elevated expression levels after 14 dpf (Fig. 5G).
Genes involved in oxygen transport. Expression of myoglobin, which is important in the uptake and delivery of oxygen to
muscle, was upregulated after 14 dpf (Fig. 5H). Expression of
erythropoietin (Epo), which regulates red blood cell production
and hemoglobin synthesis, was also upregulated over time
(Fig. 5I).
AJP-Regul Integr Comp Physiol • VOL
Fig. 4. Muscle fiber cross-sectional area increases over time, and white muscle
cross-sectional area is transiently increased as a result of training. White and
red muscle fiber areas were grouped in 400- and 200-mm2 groups, respectively.
dpf, Days postfertilization.
291 • OCTOBER 2006 •
www.ajpregu.org
ENDURANCE TRAINING IN ZEBRAFISH
R1045
During the first 3 wk, training effects on gene expression
included an increase in the expression of PCNA and myoglobin
(Fig. 5, G and H). Expression of the other above-mentioned
genes was not significantly affected by training.
Differential Effects of 10 Weeks of Training on Axial Muscle
and Heart
The effects of prolonged training were investigated after 10
wk of training in axial muscle and heart. In axial muscle,
training significantly elevated expression of slow myhz5 and
AJP-Regul Integr Comp Physiol • VOL
slow troponin C, myogenin, Epo, myoglobin, and PCNA (Fig.
6A). Expression of fast myhz2, fast troponin C, and myostatin
did not respond to training. In the heart, training significantly
elevated expression of fast myhz2, slow myhz5, fast troponin
C, myogenin, and PCNA (Fig. 6B). Expression of slow troponin C, myostatin, Epo, and myoglobin did not respond to
training.
To put the observed changes in gene expression into perspective, we compared their relative expression levels in axial
muscle and heart from control fish. Figure 7A shows the
291 • OCTOBER 2006 •
www.ajpregu.org
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
Fig. 5. Kinetics of muscle gene expression as obtained by quantitative RT-PCR. Expression was first normalized to ␤-actin expression. Then expression at 14 dpf,
immediately before initiation of training, was set to 1 (dashed lines) so that expression could be followed over time and differences induced by training could be observed.
Muscle constituents fast and slow myosin heavy chain (myhz2 and myhz5, respectively; A and B) decreased over time, irrespective of training. NADH dehydrogenase
(NADHd; C), a measure of mitochondrial activity, increased over time but was not affected by training. Muscle growth factors growth and differentiation factor 8 [gdf8,
D (inhibitor)] and myogenin (E) and insulin-like growth factor (IGF)-Ea2 [F (activators)] increased over time but were unaffected by training. Proliferating cell nuclear
antigen (PCNA; G), an indicator of cell proliferation, increased over time and was further enhanced by endurance training, as was myoglobin (H), which increased
aerobic capacity of muscle. Expression of erythropoietin (Epo, I), which stimulates red blood cell formation, increased over time but was unaffected by training.
*Significantly different from 14 dpf (P ⬍ 0.05). #Significant difference between trained and control (P ⬍ 0.05).
R1046
ENDURANCE TRAINING IN ZEBRAFISH
Fig. 6. Differential effects of training on relative
gene expression levels in axial (A) and heart (B)
muscle after 10 wk of endurance training, i.e., 84
dpf. TropC, troponin C. Values are means with SD
of 8 animals. *Significant difference (P ⬍ 0.05).
lower (P ⬍ 0.001) in heart than in axial muscle, whereas that
of slow troponin C {[2.2 (SD 0.6)] ⫻ 103} was higher (P ⬍
0.0005) in heart. The muscle growth factors myostatin [1.3 (SD
0.5)] and myogenin [0.60 (SD 0.76)] were expressed at similar
levels in heart and axial muscle. Epo [47 (SD 24)] and
myoglobin [35 (SD 13)] were more abundant (P ⬍ 0.0005) in
heart than in axial muscle. Basal proliferation levels also
appeared similar in trained fish, as indicated by PCNA expression levels [3.2 (SD 3.6)].
DISCUSSION
We addressed the effects of increased mechanical load on
developing zebrafish. In the present study, we increased mechanical loads by imposing an endurance-training regimen. In
large salmonid fish species, training induces effects that are
comparable to, although less profound than, those observed in
humans (9, 10). These effects include an enhanced aerobic
Fig. 7. Relative abundance of selected muscle-specific genes in axial and cardiac muscle of control (A) and trained (B) fish at 84
dpf. Fast myhz2 and fast troponin C are
expressed more abundantly in axial muscle.
Slow myhz5, myostatin, myogenin, and
PCNA are expressed at similar levels in heart
and axial muscle. Slow troponin C, Epo, and
myoglobin are expressed more abundantly in
heart muscle. Axial muscle gene expression
was set to 1. Note logarithmic scale of y-axis.
Values are means with SD of 8 animals.
*Significant difference (P ⬍ 0.05).
AJP-Regul Integr Comp Physiol • VOL
291 • OCTOBER 2006 •
www.ajpregu.org
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
relative abundance of messengers for the selected genes in
axial muscle and heart of control fish, in which the expression
in muscle was set to 1. The relative abundance of fast myhz2
[0.0072 (SD 0.0095)] and fast troponin C [0.0056 (SD 0.0077)]
messengers was lower (P ⬍ 0.0005) in heart than in axial
muscle, that of slow myhz5 [0.46 (SD 0.60)] was similar in
heart and axial muscle, and that of slow troponin C {[39 (SD
8.6)] ⫻ 103} was higher (P ⬍ 0.0005) in heart than in axial
muscle. The muscle growth factors myostatin [0.54 (SD 0.56)]
and myogenin [1.1 (SD 1.3)] were expressed at similar levels
in heart and axial muscle. Epo [136 (SD 42)] and myoglobin
[24 (SD 7)] were more abundant (P ⬍ 0.0005) in heart than in
axial muscle. Basal proliferation levels appeared similar, as
indicated by PCNA expression levels [2.3 (SD 2.7)]. In trained
fish, the pattern was similar (Fig. 7B), but there were scale
differences. The relative abundance of fast myhz2 [0.016 (SD
0.018)] and fast troponin C [0.012 (SD 0.007)], P ⬍ 0.005 and,
in contrast to control fish, slow myhz5 [0.18 (SD 0.15)], was
ENDURANCE TRAINING IN ZEBRAFISH
AJP-Regul Integr Comp Physiol • VOL
(Fig. 4); whereas, this is a common response to training in
larger fish (9, 10). The total area of red muscle fibers, however,
did increase (Fig. 3). Because this process was not accompanied by red muscle fiber hypertrophy, a hyperplastic response,
consistent with a continued hyperplastic capacity of red muscle
fibers (34), is suggested. This is supported by the increased
PCNA expression throughout the 10-wk training period (Figs.
5 and 6). After 10 wk of training, in the axial muscle, the
increased PCNA expression is accompanied by an increased
expression of slow myhz5 and slow troponin C. Furthermore,
the aerobic capacity of the axial musculature was increased
through increased expression of Epo and myoglobin. The
increase in aerobic capacity is not necessarily limited to red
muscle, as white muscle tissue may also show an increase in
aerobic capacity, e.g., via an increase in capillarization after
training (8, 27, 29). In conclusion, the axial muscle responds to
endurance training by a shift toward a slow aerobic phenotype,
which is accompanied by a hyperplastic increase in red muscle
fibers.
The heart is a continuously active muscle and, as such,
performs endurance exercise in control fish. In control fish,
slow troponin C, Epo, and myoglobin are much more abundant
in heart than in axial muscle at 84 dpf (Fig. 7). This suggests
that, in control fish, heart muscle has a slow, red muscle-like
phenotype, which is consistent with the obligatory aerobic
metabolism of cardiac muscle fibers. Similar to red muscle
fibers in fish, the heart muscle retains the ability to grow and
fully regenerate through hyperplasia (15, 28). Although heart
size may be increased by training (13, 16, 18), it generally is
not (for review see Refs. 9 and 10). Furthermore, training for
11 days in zebrafish did not result in any postexercise changes
in cardiac activity (27). In the present study, we compared the
effect of 10 wk of training on gene expression levels in heart
and axial muscle. Both muscles showed increased expression
of PCNA and myogenin, indicating hyperplastic growth of
heart and axial muscle after prolonged endurance training.
However, training also differentially affected heart and axial
muscle. Axial muscle shifted toward a slower, more aerobic
phenotype, as described above. In heart muscle, expression of
fast troponin C and fast and slow myosin heavy chain increased, whereas expression of Epo, myoglobin, and slow
troponin C did not change (Fig. 6). This indicates that the heart
shifts to a faster phenotype and does not increase aerobic
capacity.
In control hearts, the above-mentioned high levels of myoglobin and Epo expression (Fig. 7) are necessary for extraction
of oxygen from the oxygen-depleted venous blood that supplies the spongy myocardium (7, 12). In the pufferfish Fugu
rubripes, the heart is the major Epo-producing organ (3), in
contrast to mammals, in which the kidney performs this function (21, 23). The lack of increase in expression of these genes
indicates that the existing capacity for oxygen extraction was
large enough to manage the increased demand of endurance
training. The increase in myogenin, PCNA, and, mainly, fast
contractile proteins suggests, instead, that the heart responded
by hyperplastic growth and shifted to a faster phenotype. This
indicates that the heart becomes more powerful.
We compared basal transcription levels between heart and
axial muscle in control fish. The relatively high transcription of
slow troponin C, Epo, and myoglobin in the heart correlates
with a lack of transcription modulation after training, whereas
291 • OCTOBER 2006 •
www.ajpregu.org
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
potential of white and red muscle, a relative increase in the size
of the red muscle compartment, increased fiber cross-sectional
area, and improved heart performance (9, 10). In the small
zebrafish, effects of short periods of training during development on aerobic and anaerobic metabolism have been reported
(1, 27), but a comprehensive picture of the effects of endurance
training is lacking. We therefore trained zebrafish for 10 wk
and studied the effects of endurance training on muscle at
macroscopic, histological, and molecular levels.
Bagatto et al. (1) reported that a training regimen of 2–5
TL/s for zebrafish larvae at 4 dpf for 24 h/day or at 21 dpf for
15 h/day resulted in high mortality. Mortality was higher
among younger fish and increased with higher flow rate and
longer duration of training (1). There is a risk that any differences in such a setup between trained and control fish are
confounded by high mortality and are not necessarily the result
of adaptation. Therefore, we designed a less severe training
regimen for the present study. Training commenced at 14 dpf,
when larvae had switched to a diet of artemia, and consisted of
a daily period of 6 h at 5 TL/s, instead of 15 TL/s, and lasted
for 10 wk; 5 TL/s represents about twice the normal average
swimming speed (14). Zebrafish retained their intermittent
swimming pattern, with bursts of activity alternating with short
periods of coasting, at all times. The fish were able to keep up
with the training regimen and actively pursued small particles
that floated by. These results indicate that maximal sustainable
swimming speed was not reached. We also did not experience
more mortality in our trained fish than in controls. Together,
these observations suggest that the imposed training regimen
was sustainable. During training, the fish swam close to the
front grid in a school. This type of behavior may serve to
reduce the energy cost of swimming against a current, by
exploiting the vortices that are generated by objects in the flow
(25), such as the grid or, in this case, other animals. Nevertheless, the trained fish were more active than the control fish.
Zebrafish are slow-growing fish that reach a maximum TL of
40 –50 mm over a 3-yr life span. The growth rate decreases
when they are about one-third of adult size (30), which occurs
at ⬃30 days of age. We observed this decrease in growth rate
in our control fish (Fig. 2, inset). Trained fish initially showed
a faster growth rate than control fish, but growth declined to a
rate slower than that of control fish after 40 dpf. A transient
difference in TL between trained and control fish disappeared
at 60 dpf (Fig. 2). In larger fish species, training results in
lasting differences in body size (9, 10); however, the maximum
size in zebrafish is probably genetically determined. Alternatively, a forced swimming speed that is constant in TL per
second might prove increasingly more intensive as body length
increases (17), causing an energy deficit, along with a decrease
in growth rate, in older stages. However, growth rate decreased
in control and trained fish. In addition, both groups were fed ad
libitum and retained fat reserves, indicating that the preferred
energy reserves for endurance training (24) are still present.
Therefore, this intensity hypothesis is implausible.
Whole body growth and hyperplastic growth in white and
intermediate muscle are limited after zebrafish reach a TL of 20
mm. Red and heart muscle hyperplasia continues into adulthood, and hypertrophic growth can be found in all muscle fiber
types (15, 34). Despite a potential for hypertrophic growth,
endurance training did not increase the muscle fiber hypertrophy, as judged by cross-sectional area in white or red muscle
R1047
R1048
ENDURANCE TRAINING IN ZEBRAFISH
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
ACKNOWLEDGMENTS
The authors thank Anja Taverne-Thiele for advice on and assistance with
sampling and De Haar Vissen for animal husbandry.
25.
26.
REFERENCES
1. Bagatto B, Pelster B, and Burggren WW. Growth and metabolism of
larval zebrafish: effects of swim training. J Exp Biol 204: 4335– 4343,
2001.
2. Buchanan CI and Marsh RL. Effects of exercise on the biomechanical,
biochemical and structural properties of tendons. Comp Biochem Physiol
A 133: 1101–1107, 2002.
3. Chou CF, Tohari S, Brenner S, and Venkatesh B. Erythropoietin gene
from a teleost fish, Fugu rubripes. Blood 104: 1498 –503, 2004.
4. Cowin SC. Bone Mechanics Handbook. Boca Raton, FL: CRC, 2001.
5. Crossmon G. A modification of Mallory’s connective tissue stain with a
discussion of the principles involved. Anat Rec 69: 33–38, 1937.
6. Crow MT and Stockdale FE. Myosin expression and specialization
among the earliest muscle fibers of the developing avian limb. Dev Biol
113: 238 –254, 1986.
7. Davie PS and Farrell AP. The coronary and luminal circulations of the
myocardium of fishes. Can J Zool 69: 1993–2001, 1991.
8. Davie PS, Wells RM, and Tetens V. Effects of sustained swimming on
rainbow trout muscle structure, blood oxygen transport, and lactate dehydrogenase isozymes: evidence for increased aerobic capacity of white
muscle. J Exp Zool 237: 159 –171, 1986.
9. Davison W. Training and its effect on teleost fish. Comp Biochem Physiol
A 94: 1–10, 1989.
10. Davison W. The effects of exercise training on teleost fish, a review of
recent literature. Comp Biochem Physiol A 117: 67–75, 1997.
11. De Graaf F, Van Raamsdonk W, Hasselbaink H, Diegenbach PC, Mos
W, Smit-Onel MJ, Van Asselt E, and Heuts B. Enzyme histochemistry
of the spinal cord and the myotomal musculature in the teleost fish
Brachydanio rerio. Effects of endurance training and prolonged reduced
locomotory activity. Z Mikrosk Anat Forsch 104: S593–S606, 1990.
12. Farrell AP and Clutterham SM. On-line venous oxygen tensions in
rainbow trout during graded exercise at two acclimation temperatures. J
Exp Biol 206: 487– 496, 2003.
13. Farrell AP, Johansen JA, Steffensen JF, Moyes JF, West TG, and
Suarez RK. Effects of exercise training and coronary ablation on swim-
AJP-Regul Integr Comp Physiol • VOL
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
ming performance, heart size, and cardiac enzymes in rainbow trout,
Oncorhynchus mykiss. Can J Zool 68: 1174 –1179, 1990.
Fuiman LA and Webb PW. Ontogeny of swimming activity and performance in zebra danios (Teleostei: Cyprinidae). Anim Behav 36: 250 –261,
1988.
Gamperl AK and Farrell AP. Cardiac plasticity in fishes: environmental
influences and intraspecific differences. J Exp Biol 207: 2539 –2550, 2004.
Greer Walker M and Emerson L. Sustained swimming speeds and
myotomal muscle function in the trout, Salmo gairdneri. J Fish Biol 13:
475– 481, 1978.
Hammer C. Fatigue and exercise tests with fish. Comp Biochem Physiol
A 112: 1–20, 1995.
Hochachka PW. The effect of physical training on oxygen debt and
glycogen reserves in trout. Can J Zool 39: 767–776, 1961.
Huising MO, van der Meulen T, Flik G, and Verburg-van Kemenade
BML. Three novel carp CXC chemokines are expressed early in ontogeny
and at nonimmune sites. Eur J Biochem 271: 4094 – 4106, 2004.
Hutson MS, Tokutake Y, Chang MS, Bloor JW, Venakides S, Kiehart
DP, and Edwards GS. Forces for morphogenesis investigated with laser
microsurgery and quantitative modeling. Science 300: 145–149, 2003.
Jelkmann W. Erythropoietin: structure, control of production, and function. Physiol Rev 72: 449 – 489, 1992.
Koumans JTM and Akster HA. Myogenic cells in development and
growth of fish. Comp Biochem Physiol A 110: 3–20, 1995.
Lacombe C, Da Silva JL, Bruneval P, Casadevall N, Camilleri JP,
Bariety J, Tambourin P, and Varet B. Erythropoietin: sites of synthesis
and regulation of secretion. Am J Kidney Dis 18: 14 –19, 1991.
Lauff RF and Wood CM. Effects of training on respiratory gas exchange,
nitrogenous waste excretion, and fuel usage during aerobic swimming in
juvenile rainbow trout (Oncorhynchus mykiss). Can J Fish Aquat Sci 54:
566 –571, 1997.
Liao JC, Beal DN, Lauder GV, and Triantafyllou MS. Fish exploiting
vortices decrease muscle activity. Science 302: 1566 –1569, 2003.
Nüsslein-Volhard C and Dahm R. Zebrafish. In: Practical Approach,
edited by Hames BD. Oxford, UK: Oxford University Press, 2002.
Pelster B, Sanger AM, Siegele M, and Schwerte T. Influence of swim
training on cardiac activity, tissue capillarization, and mitochondrial density in muscle tissue of zebrafish larvae. Am J Physiol Regul Integr Comp
Physiol 285: R339 –R347, 2003.
Poss KD, Wilson LG, and Keating MT. Heart regeneration in zebrafish.
Science 298: 2188 –2190, 2002.
Sänger AM and Pötscher U. Endurance exercise training affects fast
white axial muscle in the cyprinid species Chalcalburnus chalcoides
mento (Agassiz, 1832), cyprinidae, teleostei. Basic Appl Myol 10: 297–
300, 2000.
Schilling TF. The morphology of larval and adult zebrafish. In: Zebrafish,
edited by Nüsslein-Volhard C and Dahm R. Oxford, UK: Oxford University Press, 2002, p. 59 –94.
Sprague J, Doerry E, Douglas S, and Westerfield M. The Zebrafish
Information Network (ZFIN): a resource for genetic, genomic and developmental research. Nucleic Acids Res 29: 87–90, 2001.
Streisinger G, Walker C, Dower N, Knauber D, and Singer F. Production of clones of homozygous diploid zebra fish (Brachydanio rerio).
Nature 291: 293–296, 1981.
Van der Meulen T, Kranenbarg S, Schipper H, Samallo J, van
Leeuwen JL, and Franssen H. Identification and characterisation of two
runx2 homologues in zebrafish with different expression patterns. Biochim
Biophys Acta 1729: 105–117, 2005.
Van Raamsdonk W, Mos W, Smit-Onel MJ, van der Laarse WJ, and
Fehres R. The development of the spinal motor column in relation to the
myotomal muscle fibers in the zebrafish (Brachydanio rerio). I. Posthatching development. Anat Embryol (Berl) 167: 125–139, 1983.
Weatherley AH and Gill HS. Growth dynamics of white myotomal
muscle fibres in the bluntnose minnow, Pimephalus notatus Rafinesque,
and comparison with rainbow trout, Salmo gairdneri Richardson. J Fish
Biol 25: 13–24, 1984.
Westerfield M. The Zebrafish Book, a Guide for the Laboratory Use of
Zebrafish (Brachydanio rerio). Eugene, OR: University of Oregon Press,
1995.
291 • OCTOBER 2006 •
www.ajpregu.org
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.2 on June 14, 2017
expression of genes that are expressed at a similar or lower
level than in axial muscle, e.g., myogenin, fast myhz2, and
PCNA, may be increased. Vice versa, expression of genes with
lower transcription in axial muscle was modulated as a result of
training. This may be indicative of optimal transcription of the
highly expressed genes, whereas genes with lower expression
can still be upregulated in transcription.
In this study, fish were fed ad libitum, and we observed a
transiently increased growth rate in the trained fish, which
could have resulted from greater food intake by trained fish.
This should also be regarded as a result of the training regimen.
The differences in fish size may have had an effect on the size
of the total red muscle fiber area; however, feeding is not
expected to affect muscle composition with respect to fast-toslow fibers ratios. In addition, because the RT-PCR analysis
uses normalization to internal reference genes, the different
sizes of trained and control fish are not relevant in the analysis
of gene expression over the first 3 wk of training.
Collectively, our results indicate that endurance training
enhances slow aerobic muscle development in the axial musculature and shifts the heart to a faster phenotype. Thus,
although the different organs and tissues are still in the process
of maturation and differentiation, they show profound plasticity and respond to increases in mechanical load.