Voluntary cage wheel exercise has been used extensively to determine the physiological adaptation of cardiac and skeletal muscle in mice. In this study, we tested the effect of different loading conditions on voluntary cage wheel performance and muscle adaptation. Male C57Bl/6 mice were exposed to a cage wheel with no-resistance (NR), low-resistance (LR), or high-resistance (HR) loads for 7 wk. Power output was elevated (3-fold) under increased loading (LR and HR) conditions compared with unloaded (NR) exercise training. Only unloaded (NR) exercise induced an increase in heart mass, whereas only loaded (LR and HR) exercise training induced an increase in skeletal (soleus) muscle mass. Moreover, unloaded and loaded exercise training had a differential impact on the cross-sectional area of muscle fibers, depending on the type of myosin heavy chain expressed by each fiber. The biochemical adaptation of the heart was characterized by a decrease in genes associated with pathological (but not physiological) cardiac hypertrophy and a decrease in calcineurin expression in all exercise groups. In addition, transcriptional activity of myocyte enhancer factor-2 (MEF-2) was significantly decreased in the hearts of the LR group as determined by a MEF-2-dependent transgene driving the expression of β-galactosidase. Phosphorylation of glycogen synthase kinase-3β, protein kinase B (Akt), and p70 S6 kinase was increased only in the hearts of the NR group, consistent with the significant increase in cardiac mass. In conclusion, unloaded and loaded cage wheel exercise have a differential impact on cage wheel performance and muscle (cardiac and skeletal) adaptation.
- cardiac adaptation
- cardiomyocyte signaling
endurance exercise training elicits numerous physiological changes in skeletal muscle, including metabolic/aerobic adaptations and vascular changes (21, 28, 52). Similarly, the heart responds to a chronic endurance exercise stimulus with an increase in cardiac mass, aerobic capacity, and activation/deactivation of several cardiomyocyte signaling pathways (1, 24, 25, 28). The importance of elucidating the mechanisms that underlie these physiological adaptations is underscored by the increasingly prominent role that exercise plays during cardiac rehabilitation in a setting where cardiac and skeletal muscle function are dramatically compromised. Clinically, the physiological adaptations associated with exercise in patients with coronary artery disease, hypertension, or congestive heart failure result in a reduction of cardiovascular morbidity and mortality and improvements in the quality of life (16, 44, 53).
In addition, resistance (strength) training has gained acceptance as an effective therapeutic intervention to attenuate the age-mediated decline in muscle strength (13). Despite studies showing that exercise training under an increased load leads to an increase in skeletal muscle mass (8, 22, 39), the impact of resistance training on cardiac morphology and function is unclear, and conflicting reports exist. In addition, very little is known about the signaling pathways that mediate cardiac and skeletal muscle adaptation to resistance training. For example, some studies showed that resistance training increases left ventricular (LV) cavity dimensions, septal and posterior wall thickness, and cardiac mass (12, 19); other studies in short- and long-term resistance-trained athletes demonstrated no changes in LV morphology or mass (19, 20). The inconsistency of these data probably reflects the fact that the cardiac response to resistance exercise is likely to depend on the duration (long term vs. acute) and type (bodybuilding, power lifting, or high intensity) of training regimen. Moreover, although rodent models of resistance training have been used for the study of skeletal muscle adaptation (8, 22, 39), the majority of published studies examining the impact of resistance training on cardiac adaptation have focused on human subjects, and of course there is human genetic variability. Thus the impact of exercise training under varying loads and intensities on cardiac and skeletal muscle adaptation needs to be further investigated in inbred mouse strains.
In this study, we used the C57Bl/6 mouse strain to determine the impact of different load-bearing voluntary training regimens on cardiac and skeletal muscle adaptation. This type of exercise training mostly involves low to moderate resistances with higher repetitions and, thus, can be characterized as resistance-endurance training. We chose to impose a known load on a voluntary running wheel, rather than alternative resistance models, for two reasons: 1) our laboratory has previously used the voluntary cage wheel paradigm to demonstrate cardiac and skeletal muscle adaptation to exercise in the mouse (1, 17, 25), and 2) the voluntary cage wheel eliminates physical and psychological stressors associated with forced exercise paradigms. As a major component of this study, we wished to explore the signaling pathways that underlie the cardiac adaptation in response to loaded and unloaded exercise. Previous studies suggested that the signaling pathways leading to pathological cardiac adaptation are distinct from those leading to exercise cardiac adaptation (11, 32, 35, 36, 47, 54, 57, 59). For example, transgenic overexpression of calcineurin induced pathologically enlarged hearts that transition to heart failure (36, 58), whereas downstream targets of calcineurin were not activated after two different exercise (physiological) regimens (57). Similarly, myocyte enhancer factor-2 (MEF-2) transcription factor was strongly activated in pathological (31), but not physiological (25), models of cardiac hypertrophy. Considering that cardiac disease (pathological) and voluntary cage wheel exercise (physiological) activate distinct pathways in the cardiac myocyte, we were interested in whether loaded voluntary exercise would operate through similar signaling components or represent a third, independent signaling pathway. We hypothesized that unloaded and loaded exercise would elicit unique responses in the mouse as determined by exercise performance (distance, duration, speed, work, and power), cardiac and skeletal muscle gravimetric analysis, and molecular and biochemical adaptations in the heart.
Experimental animals and voluntary cage wheel exercise.
All experiments were performed using procedures approved by the University of Colorado Institutional Animal Care and Use Committee. Thirty-five inbred male C57Bl/6J mice were randomly assigned to one of four experimental groups: 1) sedentary (n = 10), 2) no resistance (NR, n = 7), 3) low resistance (LR, n = 9), and 4) high resistance (HR, n = 9). Animals were individually housed in a cage (47 × 26 × 14.5 cm) containing an exercise wheel; sedentary control animals were housed in identical cages without a wheel. The exercise wheels have been previously described (1). Briefly, the system consists of an 11.5-cm-diameter wheel with a 5.0-cm-wide running surface (model 6208, Petsmart, Phoenix, AZ) equipped with a digital magnetic counter (model BC 600, Sigma Sport, Olney, IL) that is activated by wheel rotation. In addition, each wheel was engineered with a resistance mechanism allowing adjustment of the load (Fig. 1A). This was accomplished by attaching stainless steel fishing line to the cage top and wrapping the wire around an immovable pulley that was secured to the cage wheel at the axis of rotation so as to not contribute to the wheel load. The wire was, again, secured to the cage top with a spring and screw. This design permitted fine adjustments of the wheel load, which was evenly distributed throughout the rotation of the wheel. Daily exercise values for time and distance run were recorded for each exercised animal throughout the duration of the exercise period. All animals were given water and standard hard rodent chow ad libitum.
Voluntary running (cage wheel exposure) began at an average age of 83.7 ± 2.0 days (∼12 wk) for all groups, with no differences between experimental groups. Each group continued running under varying resistance, depending on experimental group, for 50 days until the animals were 132.6 ± 1.3 days (∼19 wk) of age. The resistance protocols for each group are illustrated in Fig. 1. The load on the wheel was determined by hanging known weights on the wheel until the wheel was slightly displaced. All exercise groups began with no load on the cage wheel for the first week (week 1). However, the “no-load” condition was actually 2 g, which was determined as the load necessary to maintain wheel inertia and frictional load.
Considering a wheel acclimatization period of ∼1 wk (1, 25), we changed wheel loads at 1-wk intervals, except for higher loads, which we changed after 2 wk. The load on the wheel for each experimental group was as follows (Fig. 1): week 1 at no load (all groups); week 2 at 4 g (LR and HR); week 3 at 5 g (LR and HR); weeks 4 and 5 at 7 g (HR); weeks 6 and 7 at 9 g and then, finally, 12 g (HR). The exposure at 12 g in the HR group was maintained for the final 4 days of the exercise protocol; the LR group was maintained at a load of 5 g for the remaining protocol duration. Exercised and sedentary control animals were euthanized by cervical dislocation under inhaled anesthesia immediately after the end of the specific exercise period. Body mass was measured, and hearts were rapidly excised and washed with a modified ice-cold PBS solution (in mmol/l: 136.9 NaCl, 3.35 KCl, 12 NaH2PO4, and 1.84 KH2PO4, pH 7.4). Hearts were dissected longitudinally in half (to include equal amounts of the LV and right ventricle) and frozen in isopentane cooled with liquid nitrogen or used fresh for biochemical assays.
Skeletal muscle immunohistochemistry and fiber cross-sectional area.
Skeletal muscles (soleus, plantaris, gastrocnemius, and tibialis anterior) were dissected and weighed. Muscles were then embedded in Tissue-Tek and snap frozen in liquid nitrogen-cooled isopentane. Frozen sections (8 μm) were cut and stained with antibodies specific to each myosin heavy chain (MyHC) isoform. The following antibodies were used: anti-MyHC-I (1:50 dilution; Novocastra Laboratories), MyHC-IIa (undiluted, prepared from hybridoma SC-71, deposited in the American Type Culture Collection by S. Schiaffino), MyHC-IId/x (undiluted, prepared from hybridoma developed by Joseph F. Y. Hoh), MyHC-IIb (1:2 dilution, from hybridoma BF-F3, deposited in the American Type Culture Collection by S. Schiaffino), antilaminin (catalog no. L9393, Sigma), goat anti-mouse IgG-FITC (1:100 dilution; catalog no. 115-095-071, Jackson ImmunoResearch), goat anti-mouse IgM-FITC (1:100 dilution; catalog no. 115-095-075, Jackson ImmunoResearch), and anti-rabbit Texas red (catalog no. 711-075-152, Jackson ImmunoResearch). Sections were preincubated for 1 h at room temperature in blocking solution [5% normal goat serum and 2% antilaminin antibody in PBS with 0.12% bovine serum albumin (P/BS; catalog no. A7906, Sigma), 0.12% nonfat dehydrated milk, and 0.01% Triton X]. Sections were rinsed three times in P/BS and incubated in primary antibody diluted in P/BS for 1 h at room temperature. Sections were then rinsed twice in P/BS and incubated three times for 5 min in P/BS before secondary antibodies in P/BS were added and incubated for 1 h at room temperature. Sections were again rinsed in P/BS and mounted in 1,4-diazobicyclo(2,2,2)octane (catalog no. D2522, Sigma)-supplemented glycerol gelatin (catalog no. GG-1, Sigma). Fluorescent microscopy images were obtained with a Zeiss microscope equipped with digital image capture, and fiber area was analyzed using the NIH Image software.
Total RNA was extracted from one-half of the frozen myocardium using TRIzol reagent (GIBCO-BRL) according to the manufacturer’s protocol. The level of mRNA expression of β-MyHC, α-skeletal actin (S-actin), and atrial natriuretic factor (ANF) mRNA was determined using slot-blot analysis with previously described oligonucleotide probes (23, 55).
MEF-lacZ mice and activity assays.
A subset of mice were backcrossed with a transgenic line of C57Bl/6J mice that harbored a lacZ transgene under the transcriptional control of three MEF-2 consensus DNA-binding sites (7), allowing direct monitoring of MEF-2 activity in vivo (7, 26, 41). To determine MEF-2 activity, freshly excised or frozen hearts from MEF-2-lacZ indicator mice were mechanically disrupted in an ice-cold buffer solution (in mmol/l: 40 Tris, 150 NaCl, and 1 EDTA). The homogenized tissue was centrifuged at 10,000 g (Beckman J2-HS centrifuge) for 5 min at 4°C. β-Galactosidase assays (Galacto-Star, Tropix, Bedford, MA) were performed on cardiac extracts from MEF-2 indicator mice as described previously (34) under conditions of linearity with respect to time and protein concentration.
Western blot analysis and gel electrophoresis.
Preparation of heart samples for SDS-PAGE and subsequent Western immunoblotting for detection of proteins and particular sites of phosphorylation on each protein began by homogenization of hearts (minus atria) in a protein extraction buffer for Ca2+/calmodulin-dependent protein kinase (CaMK) assay [in mmol/l: 50 Tris, 0.5 EGTA, 1 EDTA, and 0.5 dithiothreitol, pH 7.0 (for CaMK assay) and 40 Tris, 150 NaCl, and 1 EDTA (for β-galactosidase assay)]. The buffer also contained leupeptin, pepstatin, and phenylmethylsulfonyl fluoride (0.1 mmol/l each) to prevent nonspecific proteolysis and sodium pyrophosphate and sodium vanadate (1 mmol/l each) to prevent nonspecific phosphorylation and dephosphorylation, respectively. The homogenized tissue was then centrifuged at 12,000–14,000 g (Beckman J2-HS centrifuge) for 10 min at 4°C. The supernatant was removed, and protein concentration was determined using the Bradford method. SDS-PAGE was performed on the heart extracts, which were then transferred to a polyvinylidene difluoride membrane. The membranes were probed with antibodies specific for calcineurin, Akt, glycogen synthase kinase-3β (GSK-3β), and p70 S6 kinase, including the phosphorylated forms of Akt, GSK-3β, and p70 S6 kinase. All antibodies were obtained commercially: calcineurin from BD Transduction Laboratories (San Diego, CA), Akt and p70 S6 kinase from Cell Signaling Technology (Beverly, MA), and GSK-3β and phosphorylated GSK-3β from Santa Cruz Biotechnology (Santa Cruz, CA).
Data and statistical analysis.
Values are means ± SE. The weekly values were determined by averaging the values in a 24-h period over a given 7-day period. To calculate average daily external work and power output, the torque necessary to maintain wheel speed at a given load was calculated as τ = mg × r, where τ is torque, m is wheel load, g is Newton’s conversion factor (9.81 m/s2), and r is radius of the cage wheel. Next, work (W) was calculated as W = τ × θ, where θ is angular displacement (2π rad/revolution, 17,952 rad for 1 km). It follows that power (P) was calculated as P = W/t, where t is time. External work and power were adjusted for each individual animal body weight (in kg). The percent change in cardiac mass with exercise was determined by comparing the mass of each exercised animal with the mean cardiac mass of the sedentary group. The difference in cardiac mass was then expressed as a percent change from sedentary animals for each respective animal. The differences between experimental groups were analyzed with a one-way ANOVA followed by Student’s t-test with post hoc Bonferroni’s correction to assess differences among mean values. P < 0.05 was considered significant.
Voluntary cage wheel performance.
The experimental design is outlined in Fig. 1. Measures of voluntary cage wheel performance were distance and time over a 24-h period (1, 17). These daily values were averaged over each 7-day period, and the results are displayed in Fig. 2. The addition of 4 g, 5 g (LR and HR), and 7 g (HR) of load on the wheel did not affect cage wheel performance compared with the NR group, as indicated by distance traveled (Fig. 2A) and time spent (Fig. 2B) on the cage wheel. Only when the load on the cage wheel was increased to 9 g in the HR group did the mice run significantly less distance than the NR and LR groups; exercise duration, although markedly less in the HR group than in the NR and LR groups, was not significantly different. When the load was increased to 12 g, mice in the HR group were outperformed by the NR and LR groups as determined by distance and time on the wheel. Within the HR group, cage wheel activity was significantly reduced beginning in the 2nd wk after the addition of 7 g of wheel load compared with cage wheel activity before the addition of a wheel load (Fig. 2). The significant reduction in activity persisted with the addition of 12 g.
We previously demonstrated that speed increases throughout the duration of cage wheel exposure (1, 25). We predicted that, with each addition of an increased load on the wheel, the mice would experience an acclimatization period characterized by a decrease in wheel-running speeds. However, the addition of moderate (4–7 g) wheel loads had no impact on wheel-running speed compared with the NR group (Fig. 2C). However, the addition of high-resistance loads (9 and 12 g) significantly reduced wheel-running speeds by ∼40% (Fig. 2C). Again, within the HR group, wheel-running speeds under added load (9 and 12 g) were significantly lower than wheel-running speeds under no load (Fig. 2C). Thus mice can tolerate up to ∼25% of their body weight in resistance load without an effect on cage wheel performance as measured by distance, time, and speed on the cage wheel. With larger loads, exercise performance begins to diminish. Interestingly, NR and LR groups continue to increase wheel-running speeds, perhaps because of continued acclimatization and improved aerobic capacity.
Exercise performance and aerobic adaptation by the mice were indexed by the amount of work performed and power output (22). We did not determine metabolic changes in energy production in each animal; therefore, the work and power measured in this study were external work and external power. For each experimental group, work and power averaged over the week were calculated with adjustment for the body weight of each animal (Table 1); the pooled data are displayed in Fig. 3. The amount of load required to maintain unloaded wheel friction was 2 g. Thus, in the work and power calculated for the NR group and for week 1 of all other groups, a wheel load of 2 g is taken into consideration. There were no significant differences in work performed (Fig. 3A) or power output (Fig. 3B) among the groups studied during the week (week 1) of “unloaded” cage wheel exposure. Although increasing wheel load to 4 g (week 2) and 5 g (week 3) in the LR and HR groups showed a trend toward augmented work performed, there was no significant change. During this same exercise period, however, power output was significantly elevated in the LR and HR groups compared with the NR group. By week 4, mice in the LR group responded to the 5-g wheel load with improved exercise performance as demonstrated by increased work and power compared with the NR group, a significant difference that was preserved throughout the remaining duration of the exercise protocol. When the wheel load was increased to 7 g (weeks 4 and 5) in the HR group, work performed and power output were significantly higher than in the NR group. However, the addition of 9 g in the HR group reduced the amount of work performed, whereas the amount of power remained greater than the NR group. A 12-g wheel load in the HR group resulted in a significant decrease in work performed compared with the LR group but not the NR group. At a 12-g wheel load, power output remained unchanged. Interestingly, the amount of power generated by each mouse at a wheel load above the inertial load (i.e., not including the NR group) was not different between the experimental groups studied. Thus it appears that the limiting factor for exercise performance in mice is power output.
Cardiac adaptation to voluntary cage wheel exercise under different loads.
We previously described the adaptation of the mouse heart to voluntary unloaded cage wheel exercise by an increase in cardiac mass (1, 25). Therefore, at the time of animal death, muscle mass was measured (Table 2). Despite significant differences in work performed and power output among the groups studied, there was no difference in body weight between the exercise groups and the sedentary controls (Table 2). The absolute heart mass after cage wheel exposure was elevated in all groups studied compared with their sedentary counterparts but achieved statistical significance only in the NR group. This was also true after the absolute heart weight was adjusted for the body mass of each animal (Table 1) or after the percent difference from sedentary controls was calculated for each group (Fig. 4). To summarize, the animals in the NR group exhibited the most pronounced increase of cardiac mass (11.8 ± 4.1%) from sedentary controls, while cardiac mass was greater in the LR and HR exercise groups than in sedentary controls: 7.3 ± 3.2% and 8.3 ± 4.4%, respectively.
Skeletal muscle adaptation.
After completion of the exercise-training regimens, hindlimb skeletal muscles (soleus, plantaris, gastrocnemius, and tibialis anterior) were excised and weighed to determine the impact of each exercise protocol on skeletal muscle adaptation. These muscles have been previously shown to increase muscle mass or cross-sectional area (CSA) in response to resistance-training regimens (8, 22, 39). There was a significant increase in soleus muscle mass in LR and HR groups only (Fig. 5A). In addition, the mean CSA of muscle fibers expressing each MyHC subtype was determined from calf sections in each animal (Fig. 5B). The CSA of type IIa muscle fibers was significantly increased in all groups of exercised animals, whereas the CSA of type IId/x fibers was significantly increased only in animals exercising under an increased load (LR and HR groups). On the other hand, the mean CSA of type I fibers was reduced in the LR group, and the mean CSA of type IIb fibers was reduced in the HR group.
Cardiac mRNA expression.
Pathological cardiac hypertrophy is accompanied by the induction of genes normally expressed during fetal development (33, 40). In an attempt to determine the overlap between the cardiac response to disease vs. exercise, we examined the induction of three hypertrophic markers. We and others previously showed that unloaded exercise is accompanied by a very modest elevation in the mRNAs for two hypertrophic markers, atrial natriuretic factor (ANF) and brain natriuretic peptide, but only after 4 wk of voluntary cage wheel exposure (1). In addition, although pathological hypertrophy is accompanied by induction of β-MyHC (33, 40, 56), exercise induces expression of α-MyHC (49). Therefore, using RNA slot-blot analysis, we aimed to determine the impact of resistance training on mRNAs encoding ANF, S-actin, and β-MyHC (56). The raw data from one example from each experimental group and the pooled results are shown in Fig. 6. None of these hypertrophic markers were elevated by 2 mo of exercise. Instead, the mRNA levels of the indicated markers appeared to be reduced in the NR, LR, and HR groups. Specifically, there was a significant reduction in the amount of S-actin mRNA in all groups (Fig. 6, bottom). In addition, cage wheel exposure under a moderate load (LR) significantly suppressed all the hypertrophic markers.
MEF-2 transcriptional activity.
The MEF-2 family of transcription factors has been implicated as a mediator of cardiac hypertrophy under pathological conditions in mice (37, 43). Although it has been demonstrated that MEF-2 activity increases in skeletal muscle after a physiological stimulus (60), we recently demonstrated that MEF-2 activity increases moderately, at best, in response to voluntary cage wheel exposure (25). Therefore, we explored the potential role of MEF-2 in the heart under conditions of resistance exercise training. To measure MEF-2 activity, a subset of animals from each experimental group (5 animals in each group) harbored a lacZ transgene under the transcriptional control of three MEF-2 consensus DNA-binding sites (7). This allowed direct monitoring of MEF-2 activity in vivo (7, 26, 41). MEF-2 activity as measured by the level of β-galactosidase is displayed in Fig. 7A. MEF-2 activity in sedentary animals was 22.4 ± 9.2 pg β-galactosidase/mg cardiac tissue. In animals subjected to exercise protocols with no added load (NR) or with increasing load (HR), the average level of MEF-2 activity was not different from that in sedentary animals: 12.0 ± 11.5 and 40.8 ± 13.2 pg β-galactosidase/mg cardiac tissue in the NR and HR groups, respectively. Animals exposed to a moderate load (LR) for the duration of the exercise period of 8 wk had no detectable MEF-2 activity (0.4 ± 0.1), even compared with sedentary controls.
A primary mechanism whereby cardiac muscle increases myocyte size is the Ca2+/calmodulin-dependent system. When Ca2+ binds calmodulin, this complex can activate CaMK (15, 43, 46) and a CaMK-dependent phosphatase, calcineurin (35, 36, 59). Yet, the role of calcineurin in physiological hypertrophy has remained unresolved, as evidenced by the contradictory nature of the published studies (11, 47, 57). Therefore, we performed Western blot analysis using a calcineurin-specific antibody to determine whether long-term exercise training with or without resistance had an effect on the level of calcineurin expression (Fig. 7B). In this study, all groups that were exercised for ∼2 mo demonstrated a significant reduction in the level of calcineurin expression from that of sedentary controls: 0.37 ± 0.07-, 0.59 ± 0.24-, and 0.54 ± 0.12-fold for NR, LR, and HR, respectively.
Akt, GSK-3β, and p70 S6 kinase regulation.
Increases in insulin-like growth factor I (IGF-I) have been observed in pathological (51) and physiological (42, 50) cardiac hypertrophy. Phosphoinositide 3-kinase (PI3K), a downstream target of IGF-I, regulates the activity of Akt (protein kinase B). Farther downstream of the IGF-I-Akt signaling cascade is p70 S6 kinase (27). In addition, recent studies have reported that Akt can inhibit (by phosphorylation) the ubiquitous serine-threonine kinase GSK-3β, a negative regulator of cardiac hypertrophy (2, 3, 38). Therefore, we chose to investigate these signaling components in the heart after resistance training. Representative immunoblots and summarized data are displayed in Fig. 8. None of the different exercise protocols had an effect on the total protein expression of Akt, GSK-3β, or p70 S6 kinase (Fig. 8A). There was a significant elevation in the phosphorylated forms of Akt, GSK-3β, and p70 S6 kinase in the NR group over sedentary controls: 1.86 ± 0.09-, 2.02 ± 0.10-, and 1.63 ± 0.06-fold, respectively (Fig. 8B). There were no significant increases in the phosphorylated proteins in the LR or the HR group.
In this study, a voluntary wheel-training apparatus for mice was adapted to permit the addition of different loads, similar to the apparatus described by Ishihara et al. (22) for rats. The response of each measured parameter to each exercise protocol is summarized in Table 3. When the NR, LR, and HR groups are compared (before the addition of larger loads), the addition of a moderate-resistance load did not affect exercise performance when measured by distance, duration, or speed. Furthermore, there was a gradual, significant decline in exercise performance within the HR group with increasing wheel loads, and when the load exceeded 9 g, the animals in the HR group failed to significantly perform equivalently to those in the other experimental groups.
Given comparable exercise distances and durations among the exercise groups, the amount of work performed by the NR group was less than that performed by the other groups (Fig. 3). Workload was inversely proportional to the wheel load, such that at high loads (HR group), workload fell with the fall in exercise distance (Figs. 2 and 3). At a load of 9 g, the amount of work performed by the HR group was no longer different from that performed by the NR group; at a load of 12 g, the amount of work performed by the HR group was significantly less than that performed by the LR group and equal to that performed by the NR group. When the amount of work performed was determined over a given time period, i.e., the calculation of power output, there were no measurable differences between the HR and the LR group. In addition, the power output exhibited by each of these groups was significantly greater than that exhibited by the NR group. The similarity in average power output of the groups exercising under a given load indicates that the limiting component of exercise performance in mice is power output. Thus the suggestion is that mice will adjust other parameters of exercise performance (distance, duration, speed, and work) so as to not exceed the sustainable workload over a period of time for a given animal.
Many physiological factors determine the level of sustainable workload, such as muscle strength and endurance, as well as cardiopulmonary parameters such as anaerobic (lactate) threshold and maximal oxygen uptake (5, 9, 29). It has been demonstrated in elite cyclists that peak power output was highly correlated with the ability to perform (4, 18), while other studies have indicated that power output and cycling performance were more closely related to blood lactate concentrations (5, 6). Moreover, cardiac output and peripheral oxygen capacity contributed to the attainment of maximal workload (29). Although these parameters were not measured in this study, the aforementioned investigations indicate that exercise capacity and power output are tightly coupled.
Although contractile performance of the heart after exercise was not measured, the hearts of animals in all three exercise groups adapted to the exercise stimulus. However, only the mice in the NR group showed a significant increase in absolute and normalized cardiac mass. A previous report from our laboratory demonstrated that, in response to voluntary cage wheel exposure, mice exhibit a significant increase in cardiac mass by 3 wk of cage wheel activity and that this increase in cardiac mass persisted for ≥4 wk (1). The animals in the NR group exhibited a quantitative increase in cardiac mass similar to that reported in this previous study (1), despite an exercise duration that was twice as long. Thus cardiac mass in C57Bl/6 mice appears to reach a plateau after 3 wk of exercise and is maintained for ≥2 mo. Although there are no data for the impact of resistance-endurance training, such as in this study, on the heart in mice, meta-analysis of human exercise data demonstrated that athletes engaged in an analogous exercise regimen that combined dynamic and strength training such as cycling exhibited a significant increase in LV internal dimension and LV wall thickness (45). In the present study, we were unable to perform a detailed analysis of ventricular morphology and were limited to measurement of heart weight only. Nevertheless, it would be difficult to relate these differences to human studies given enormous species differences and the behavioral component of the voluntary exercise paradigm. Future studies must be aimed at determining how the morphological adaptation of the heart (such as changes in ventricular chamber dimensions or wall thickness) affects ventricular function after loaded exercise training in mice.
As mentioned previously, an array of genes, such as ANF, brain natriuretic peptide, and S-actin, is expressed in the heart as a result of pathological stimuli (33, 40). In response to physiological stimuli such as exercise, the induction of these genes is modest at best (1, 10, 24, 48, 49). Although the reduced expression of β-MyHC mRNA was predicted on the basis of previous data (10, 48, 49), the suppressed expression of ANF and S-actin in the LR group was not expected. The critical factors contributing to differences from the present study compared with previous studies most likely involve the differences in species (rat vs. mouse), exercise paradigm (voluntary cage wheel vs. treadmill vs. swimming), and exercise duration (2 vs. 4 vs. 8 wk). If we consider the slight elevation of these hypertrophic markers after 3–4 wk of voluntary cage wheel exercise previously demonstrated in our laboratory (1), the implication is that the myocardium responds to an exercise stimulus initially with a modest increase in these genes and that the change in expression may be attenuated with continued exercise. Thus these factors may be important during the induction of hypertrophy but not the maintenance of an enlarged heart. The absence of hypertrophy and the reduced expression in the LR group suggest that the signals required for the induction of these specific markers during hypertrophy are distinct from the signals that suppress the expression of these markers. The suggestion is that exercise under an increased wheel load represents a third, independent physiological stimulus in the heart. Again, future studies must be aimed at determining the precise timing of expression and/or induction of specific signaling components of the exercise-stimulated heart.
As mentioned above, the ability of Ca2+ handling in the myocyte to vary in amplitude and frequency in response to increases in functional demand, such as exercise, makes Ca2+-dependent enzymes candidate molecules for intracellular myocyte signaling in response to exercise. When Ca2+ binds calmodulin, this complex can activate CaMK (15, 43, 46) and a Ca2+/calmodulin-dependent phosphatase, calcineurin (35, 36, 59). When CaMK is inactive, MEF-2 associates with histone deacetylases (HDACs), namely, HDAC-4 and HDAC-5, thereby suppressing MEF-2 activity. On activation, CaMK directly or indirectly phosphorylates HDACs, allowing their export from the nucleus. Thus MEF-2 transcriptional activity is increased by CaMK by removing the inhibitory action of HDACs (30, 31, 43). In addition, the role of calcineurin in physiological hypertrophy has remained unresolved, as evidenced by the contradictory nature of the published studies (11, 47, 57).
To address this pathway, a subset of animals was included in each experimental group harboring a lacZ transgene under the transcriptional control of three MEF-2 consensus DNA-binding sites, allowing monitoring of MEF-2 activity in vivo (7, 26, 41). Long-term moderate exercise (LR group) was the only protocol to have any statistically significant impact on MEF-2 activity. MEF-2 activity in this group was significantly reduced to levels that were below the levels of all groups, including sedentary controls. We previously demonstrated a modest elevation and then a subsequent attenuation of MEF-2 activity during voluntary cage wheel exposure (25). Results of these longer-term runners are consistent with previous results. As with the hypertrophic markers, long-term training under a moderate load affects those factors that suppress MEF-2 activity compared with those that activate them. Nevertheless, on the basis of these and previous data (25), MEF-2 transcriptional activity plays a minor role, if any, in the maintenance of an exercise stimulus.
Calcineurin and its downstream effector nuclear factor of activated T cells (NFAT) have been shown to be important mediators in cardiac hypertrophy (36, 58). Inasmuch as these transgenic-overexpressing animals develop enlarged hearts that transition to heart failure, the role of calcineurin and NFAT in pathological hypertrophy is strongly implicated. Yet, Eto and co-workers (11) demonstrated an increase in calcineurin activity after voluntary exercise in rats. In contrast, Rothermel et al. (47) reported that overexpression of myocyte-enriched calcineurin-interacting protein 1, an inhibitor of the calcineurin pathway, attenuated the cardiac hypertrophic response to an exercise stimulus. Supporting a negligible role of calcineurin in physiological hypertrophy, a recent study used NFAT-luciferase reporter mice to measure calcineurin activity in vivo and found no increase in NFAT activity associated with cardiac hypertrophy after two different exercise regimens (57). Although the measurements of calcineurin protein do not necessarily correlate with activity, our study demonstrated that 2 mo of voluntary cage wheel exposure reduced the abundance of calcineurin protein relative to that in sedentary controls. Consistent with these findings, mice expressing the NFAT-luciferase reporter gene showed reduced luciferase activity at certain time points during the training regimen (57). On the basis of these data, calcineurin expression in response to an exercise stimulus is most likely transcriptionally regulated by elements that have yet to be identified.
The IGF-I signaling axis is believed to be important during developmental and physiological cardiac growth (27). Activation of the IGF-I receptor is coupled to PI3K, which in turn promotes activation of Akt (protein kinase B) via a kinase intermediate. The active, phosphorylated form of Akt targets mammalian target of rapamycin and subsequent downstream phosphorylation of p70 S6 kinse, leading to an increase in protein synthesis and, presumably, cardiac growth. In support of the role of this pathway in physiological cardiac hypertrophy, expression of the catalytic subunit of PI3K-α (p110α) in an activated form in the heart results in increased phosphorylation of Akt, significant hypertrophy, and normal contractility, consistent with physiological hypertrophy (54). Similarly, expression of a dominant-negative p110α prevented cardiac hypertrophy in response to an exercise stimulus (32). In this study, activation of Akt and p70 S6 kinase, as demonstrated by elevated levels of the phosphorylated forms of each protein compared with sedentary controls, was associated with a significant increase in cardiac mass after unloaded exercise (NR group). Hearts from mice exercising under moderate (LR) or high (HR) resistance did not exhibit hypertrophy or elevated levels of activated proteins. In addition, only hearts from animals in the NR group showed an increase in phosphorylated GSK-3β, which is a negative regulator of cardiac hypertrophy and is inactivated by Akt (2, 3, 38). Phosphorylation of GSK-3β is permissive to cardiac hypertrophy and, thus, is consistent with the observed increase in heart mass in the NR group, whereas the absence of activated GSK-3β (phosphorylated) in the other experimental groups is predicted from the lack of significant cardiac hypertrophy.
The differential response of the heart to each type of exercise stimulus was paralleled by a differential response in skeletal muscle. Whereas loaded cage wheel exercise did not increase cardiac mass, exercising under an increased load increased muscle mass, although this increase reached statistical significance only in soleus muscle. The lack of hypertrophy in the skeletal muscles from animals exercising under unloaded conditions was accompanied by a relative increase in the CSA of fibers expressing MyHC type IIa, consistent with previous data examining unloaded voluntary cage wheel exercise (1, 17). The larger CSA of type IIa fibers was also observed in the other experimental groups, whereas the CSA of type IId/x fibers was increased only after the imposition of a moderate (LR group) or a large (HR group) load on the cage wheel. Moreover, animals in the LR group displayed an increase in CSA of type IIb fibers, whereas animals in the HR group exhibited a significant decrease in the CSA of these fibers. Despite a small but significant decrease in the LR group, the relative CSA of type I fibers remained unaffected by cage wheel exercise, consistent with the lack of effect of voluntary, spontaneous cage wheel exercise on type I fiber number (47). These findings may reflect a decrease in the ratio of type II to type I fibers in bodybuilders compared with power lifters in resistance-trained humans (14). Although preliminary studies indicate a shift to more oxidative fibers in animals exercising under a moderately loaded than an unloaded cage wheel (unpublished observations), future studies need to address the fiber type interconversions under each exercise protocol. In addition, the lack of significant cardiac hypertrophy may be explained by these peripheral adaptations, indicating a predominance of peripheral, skeletal muscle adaptation over central, cardiac muscle adaptation in resistance-trained mice, as previously suggested (8, 22, 39).
The initial goal of this study was to determine whether loaded exercise training represented a third, independent stimulus for cardiac and skeletal muscle adaptation. In support of this contention, Table 3 clearly illustrates that each exercise protocol had a unique impact on the parameters measured in this study. Most notably, loaded exercise had a greater impact on skeletal muscle adaptation than unloaded exercise, as determined by muscle mass and CSA. In contrast, unloaded exercise had a greater impact on the IGF signaling pathway within the cardiac myocytes than resistance training. Future investigations need to address the time dependency of signaling factors, including those studied in this present study, to help elucidate the mechanisms underlying the observed differences between exercise protocols. The activation of signaling components at the initiation of exercise training will most likely be phenotypically distinct from those at the time of completion of the exercise protocol, as in this study. Preliminary results from animals that were exposed to a voluntary cage wheel with a moderate load for a short period of time (3 wk) support this idea, with a very modest elevation in MEF-2 transcriptional activity and calcineurin protein level. Nevertheless, the data presented on the differential impact of loaded exercise training on cardiac and skeletal muscle adaptation provide a foundation for future comparisons.
This work was supported by National Heart, Lung, and Blood Institute Grant HL-56510 (to L. A. Leinwand) and National Research Service Award F32 HL-70509 (to J. P. Konhilas). U. Widegren was supported by postdoctoral fellowships from the Sweden-America Foundation, the Erik and Edith Fernström Foundation, the Blancheflor Foundation, and the Swedish Society of Medicine. D. L. Allen is supported by a research fellowship grant from the Muscular Dystrophy Association.
We thank Dana Boucek and Todd Horn for monitoring and analyzing mouse activity.
Present address of U. Widegren: Dept. of Surgical Sciences, Sect. of Integrative Physiology, Karolinska Institutet, Stockholm, SE-171 77, Sweden.
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.
- Copyright © 2005 by the American Physiological Society