Skeletal muscle fatigue is most often studied in vitro at room temperature and is classically defined as a decline in maximum force production or power output, exclusively linked to repeated isometric contractions. However, most muscles shorten during normal use, and we propose that both the functional correlate of fatigue, as well as the fatigue mechanism, will be different during dynamic contractions compared with static contractions. Under isoflurane anesthesia, fatigue was induced in rat soleus muscles in situ by isotonic shortening contractions at 37°C. Muscles were stimulated repeatedly for 1 s at 30 Hz every 2 s for a total of 15 min. The muscles were allowed to shorten isotonically against a load corresponding to one-third of maximal isometric force. Maximal unloaded shortening velocity (V0), maximum force production (Fmax), and isometric relaxation rate (−dF/dt) was reduced after 100 s but returned to almost initial values at the end of the stimulation protocol. Likewise, ATP and creatine phosphate (CrP) were reduced after 100 s, but the level of CrP was partially restored to initial values after 15 min. The rate of isometric force development, the velocity of shortening, and isotonic shortening were also reduced at 100 s, but in striking contrast, did not recover during the remainder of the stimulation protocol. The regulatory myosin light chain (MLC2s) was dephosphorylated after 100 s and did not recover. Although metabolic changes may account for the changes of Fmax, −dF/dt, and V0, dephosphorylation of MLC2s may be involved in the fatigue seen as sustained slower contraction velocities and decreased muscle shortening.
- MLC phosphorylation
- oxidative metabolism
- isotonic contractions
- aerobic exercise
fatigue develops in skeletal muscle during exercise. This is a much-studied phenomenon, and the various possible mechanisms of fatigue have been extensively reviewed (1, 7, 14). Even so, there is no general agreement about the mechanisms of muscle fatigue during normal locomotion at moderate intensity. The manifestations of muscle fatigue are not well defined, and accordingly, the molecular mechanism of muscle fatigue is not understood.
The majority of fatigue studies have been done with isometric contractions, and fatigue has variably been defined as an inability to maintain required or expected force (12) or as a fall of maximum isometric force (Fmax) or power not reflected in performance before the muscle is exhausted (6). Also, contraction rates (±dF/dt) have been used as fatigue parameters. More recently, however, it has been emphasized that the degree of fatigue will be underestimated by these parameters if muscle movement is included in the protocols (43). Several authors have pointed to the disparate muscle responses to isometric and isotonic contractions (2, 3, 9, 15, 16, 30, 32). When allowing the muscle to shorten, it becomes feasible to assess muscle function by the velocity of shortening and relaxation (±dL/dt), the extent of shortening (Smax), and work and efficiency (43). For many locomotive muscles, isotonic contractions, during which the muscle performs work, are also more physiological than isometric conditions. To date, only a few experimental studies specifically describe “isotonic fatigue,” and it seems that the classical definitions of fatigue may not reflect deteriorating muscle function during these conditions (8, 10, 38, 43, 47).
Muscle fatigue has been extensively studied in humans. It seems clear that fatigue mechanisms during dynamic exercise are highly dependent on intensity (44). Apart from early studies on glycogen depletion (20), few studies have addressed the molecular mechanism of fatigue in slow-twitch muscle that carries out concentric movements and performs work at a moderate intensity under aerobic conditions. Also, it is very difficult to compare experimental studies with studies done in exercising humans, since the in vitro conditions that are used differ so much from the in vivo conditions. First, introducing shortening will raise the energy demand, which makes it difficult to maintain aerobic conditions in isolated muscles unless temperature is dramatically reduced (5, 13, 42). However, reducing temperature will affect muscle metabolism in several ways. For instance, it is now clear that several effects of low pH on muscle contractility disappear as temperature is raised (35). Also biochemical reactions have different Q10. It is, therefore, quite probable that the rate-limiting step of a chain of reactions can vary with temperature. Therefore, also the “point of control” relevant for fatigue may differ with temperature (41).
Several studies on fatigue have focused on substrate availability [glycogen, creatine phosphate (CrP), ATP] or accumulation of certain metabolites (ADP, Pi, H+), and in recent years, slowing of intracellular Ca2+ cycling has been proposed to be an important cause of fatigue (45). However, again, these changes do not seem to provide an explanation for reduced muscle performance with prolonged dynamic muscle activity. Few studies have addressed the possibility of altered posttranslational regulation of essential proteins that participate in the cross-bridge cycle. Several of the proteins of the contractile apparatus may be modified for instance by phosphorylation. Troponin I is important in regulating cardiac function (24), but how it regulates skeletal muscle function is not known. Also with regard to myosin binding protein C, little is known about the skeletal muscle isoform (17). However, it is well known that altered phosphorylation level of MLC2 seems to be important for the Ca2+ sensitivity and possibly also kinetics of the contractile apparatus in skeletal muscle (26).
The aim of this study is, therefore, to define proper manifestations of muscle fatigue, following dynamic exercise in an experimental model under conditions as close to physiological as possible, and to investigate possible molecular mechanisms that can be linked to this fatigue.
All animals were handled according to the Norwegian Animal Welfare Act and approved by the Norwegian Animal Research Authority. Adult male Wistar rats (Wistar Hannover, Taconic, Skensved, Denmark) were kept in a temperature-, humidity- and light-controlled (12:12-h light-dark cycle) environment. The animals were fed standard rat chow (B & K Universal, Oslo, Norway; standard rat/mouse) and had water available ad libitum. They were caged for ≥1 wk after arrival from the supplier before included in studies.
In situ muscle preparation.
The rats were anesthetized with 1:3 O2-N2O with 2–2.5% isoflurane (Forene Abbott model 506949). A pressure-sensitive catheter (Cardiovascular catheter SPR-407, Millar Instruments, Houston, TX) was lead retrograde through the right arteria carotis communis and positioned in the aortic arch, making it feasible to continuously measure blood pressure during the experiments. Because the anesthetics we used can be cardio-depressant, special attention was paid to any change in systolic blood pressure. Any larger fluctuation led to an adjustment in the isoflurane level, keeping the blood pressure and heart rate at a physiological level through the protocol.
The right leg was skinned from the knee down, and the soleus muscle was dissected free from surrounding tissue. The blood supply was left intact. A combined force and length transducer (model 305B; Aurora Scientific, Ontario, Canada) was connected to the distal tendon using noncompliant surgical thread. Platina electrodes were positioned at the proximal and distal ends of the muscle. The ankle and the middle part of the tibial bone were clamped, leaving the leg immobile and stable. The soleus's core temperature was kept at 37°C by heated 0.9% NaCl constantly flowing over the epimysium covering the muscle. As the NaCl does not make direct contact with the muscle fibers, only the heat is transferred from the solution to the muscle. Additionally, the fluid ensured better transmission of current upon stimulation. The sciatic nerve was cut in the middle of the thigh and isolated with water-free vaseline to prevent retrograde transmission of current. The rats were kept on a heated (37°C) table during the experiment.
The soleus muscle [94% slow-twitch fibers (39)] was stretched to the length that resulted in maximal stimulated force production (preload) and directly stimulated (Pulsar 6bp, FHC Brunswick, ME, USA) isometrically and supramaximally (8 V) to obtain maximum isometric force at 1 Hz and at 100 Hz. Increasing voltage did not increase force or shortening at any stage of the protocol, and 100-Hz stimulation was tested to trigger maximal force (Fmax). During the stimulation protocol, soleus was allowed to shorten after reaching an afterload of one-third of the maximal tetanic force (P0) for that individual muscle (Fig. 1). At this load, muscles are able to produce maximum power and fatigue most rapidly (43). It was stimulated intermittently with stimulation trains (1 s on/1 s off) at 30 Hz for 100 s or 15 min, with 1-ms pulse duration. Thirty hertz is close to the physiological firing rate in soleus muscle during activity (19). By stimulating intermittently, perfusion is probably maintained in the same way that rotation between motor units can ensure perfusion in vivo. Before and after 100 s and 15 min of exercise, isometric force at 30 Hz stimulation was obtained. Force, shortening, muscle surface temperature, aortic blood pressure, and stimulation pulse were sampled at 2000 Hz during the protocol. The following parameters were calculated from the force and shortening tracings: ±dF/dt, ±dL/dt, Tbl, and Smax (see Table 1). The isometric relaxation tracings were fitted to a double exponential decay curve using SigmaPlot (Systat Software, San Jose, CA). Two time constants (tau1 and tau2) were obtained at time 0, 100 s, 5 min, and 15 min of exercise.
Recovery was measured as the ability to perform a new exercise protocol after 15, 30, or 60 min of rest. During the resting period, the muscle was kept at physiological levels by heated saline. After the recovery period, 5 min of the dynamic exercise protocol was performed. The preload was adjusted to equal the original tension, and the afterload was kept at the same level as the first exercise bout.
Unloaded shortening velocity, slack test.
The unloaded shortening velocity (V0) was established by stimulating the muscle isometrically at 100 Hz fully saturating troponin C with Ca2+ and measuring the time until the muscle regained a pull on the force transducer after a sudden, variable shortening of the muscle (Fig. 2). Unloaded shortening velocity was established before the exercise protocol and after 100 s and 15 min of exercise. For each test, the muscle length was set to shorten at intervals of 1 mm, from 2 to 6 mm (8–24% of muscle length). The stimulation started from optimal sarcomere length, and the release was done after the force had stabilized on tetanic force level.
Isometric maximum force and contraction rate.
To calculate exercise afterload, maximal isometric force (Fmax) was measured in every muscle. Fmax was also measured after 100 s and 15 min exercise, as part of the slack test. The rate of isometric force development (100 Hz) was analyzed from these trials, using TableCurve (TableCurve 2D v5.01, Systat Software).
Metabolites and oxygenation.
Analyses of metabolites and oxygenation were done after 100 s and 15 min of exercise. Working muscles were harvested after the appropriate exercise time and frozen in liquid nitrogen within 10 s. The nonworking soleus muscle of the contralateral leg served as control. ATP, ADP, AMP, NAD, and CrP were determined by HPLC. Increase in Pi during exercise is stoichiometrically coupled to the breakdown of CrP (25) and was calculated as the difference between resting and exercise levels of CrP for individual animals. To this value was added to the resting level of Pi (11) to determine the [Pi] at 100 s and 15 min of exercise. Lactate was determined by a fluorometric enzymatic coupled assay (27). Five animals underwent the operation and stimulation protocol with the blood supply to the soleus severed. These animals served as ischemic controls (ISCH). Inividual samples that were analyzed for tissue oxygenation were carefully thawed in ice-cold paraffin oil. Diffuse reflectance near infrared spectra were collected in the spectral range of 400–2500 nm (32 scans) using a FOSS XDS near infrared analyzer (Foss NIRSystems, Silver Spring, MD) equipped with a OptiProbe module featuring a reflectance probe. Spectra were obtained at a spectral resolution of 0.5 nm, and reference scans were obtained using the built-in internal reference. Five replicate spectra were collected for each muscle. They were obtained by making contact between the reflectance probe and the muscle that was placed on aluminium foil and kept at 0°C. The probe was thoroughly washed between every sample measurement. HbO2, HbR, cytochrome aa3, and tissue oxygenation (SO2) were measured according to the method of Wray et al. (46). SO2 level in resting skeletal muscle tissue and in exercising ISCH muscle was set to 70 and 10%, respectively (4).
Myosin light-chain analysis.
Myofibrillar proteins were separated by glycerol/SDS polyacrylamide gel electrophoresis, and gels were stained sequentially by Pro Q Diamond and SYPRO Ruby (both M33305; Invitrogen, Oslo, Norway) for phosphorylated and total proteins, respectively. The fluorescence was detected by Typhoon laser scanner (9410; GE Healthcare, Oslo, Norway) and quantified by Imagequant program (GE Healthcare, Oslo, Norway). Phosphorylation was quantified both during exercise, at time points 100 s and 15 min, and after 15 and 60 min of recovery. Phosphorylation level was calculated in individual legs by dividing the staining intensity reflecting phosphorylation level by the staining intensity of the myosin light chain (MLC2s) protein band and presented as a level of phosphorylation relative to the control muscle. MLC2s bands were identified by Western blot analysis (ALX-BC-1150-S, Clone F109.3E1, Alexis, AH Diagnostics, Oslo, Norway).
Values are expressed as means ± SE. Whenever appropriate, ANOVA supplemented with post hoc Newman-Keuls test was applied. For all tests, the level of significance was 0.05. Changes over time between pre- and postrecovery protocols were tested by comparing means of area under the curve (31). The statistical analyses were performed with Statistica v8 (StatSoft, Tulsa, OK).
On the day of the experiment, the animals weighed 370 ± 4 g (n = 37). Maximal tetanic force (Fmax) varied little between animals and was 2.34 ± 0.04 N (n = 32), which means also that afterload varied little. Also, systolic blood pressure showed little variability (122 ± 4 mmHg; n = 21) and did not change significantly during the experiments.
Figure 3A shows a 5-min recording from a typical experiment. The muscle was stimulated with trains at 30 Hz for 1 s followed by 1-s pause before the next stimulation period. At the onset of each train of stimulation, force rose rapidly until the preset afterload was reached (1/3 of Fmax; Fig. 3, E–G). Subsequently, the isotonic part of the contraction began (Fig. 3, B–D), and the muscle shortened. At cessation of the 1-s stimulation period, the muscle was passively stretched to resting length (L0). This isotonic restretch was followed by an isometric relaxation, which concluded the 2-s exercise cycle time (see also Fig. 1). The stimulation was carried out for 15 min, during which period, muscle performance deteriorated significantly, but various performance parameters followed different temporal patterns.
Because the isometric contraction rate changed during the exercise, as described in more detail below, the time available for isotonic shortening (from start of shortening to end of stimulation) also varied. It fell from about 951 ± 2 ms in the first contraction cycle (Fig. 3, B and E) to 880 ± 12 ms after 100 s (Fig. 3, C and F) and fell further to 850 ± 17 ms after 5 min and remained constant for the rest of the protocol (Fig. 3, D and G). Similarly, the changes in isotonic relaxation velocity will influence the time available for isometric relaxation (from end of isotonic relaxation to the time the next stimulation is applied). Initially, this period was 943 ± 4 ms (Fig. 3, B and E), fell to 642 ± 54 ms after 100 s (Fig. 3, C and F), and increased again to 929 ± 33 ms after 5 min (Fig. 3, D and G).
Slack test, isometric force, and velocity of shortening.
Following 100 s or 15 min of exercise, the slack test procedure was performed within 10 s of the last stimulation train, to minimize any recovery that might interfere with the results. The procedure was also done before exercise (pre). Figure 3, H–J, shows force records during 100-Hz stimulation at optimum length followed by a sudden slack of 2 mm. V0 and Fmax both decreased to 83% of the prevalue after 100 s (P < 0.05 vs. rest for the two parameters) but were then partially restored at 15 min (Table 2). Isometric force at 30-Hz stimulation decreased to 79% during the first 100 s (P < 0.05 vs. rest) and decreased further to 68% after 15 min of exercise (P < 0.05 vs. rest and 100 s). Accordingly, V0 and Fmax do not follow the isometric force at 30 Hz but seem to parallel the alterations described for the rate of isometric relaxation (see below).
Isometric part of contraction.
The rate of isometric force development (30 Hz) (Fig. 4A) fell rapidly by 56% during the first 40 s (P < 0.001) and fell further by 21% until 100 s of exercise (P < 0.001 vs. first contraction and P = 0.38 vs. 40 s). Subsequently, the rate of force development (30 Hz) tended to rise, but this was not statistically significant. The rate of isometric force development at 100 Hz decreased to 64% of the initial value after 100 s and was then partially restored at 15 min (P < 0.001, Table 2). Thus, the rate of isometric force development at 30 Hz and 100 Hz was affected differently, as the former did not show significant transient alterations. whereas the latter seemed to follow the same pattern, as described for rate of isometric relaxation (see below).
The rate of isometric relaxation (Fig. 4A) also slowed rapidly by 45% during the first 40 s (P < 0.001) and then slowed further, reaching the slowest rate after 100 s (38% of first contraction, P < 0.001 vs. initial contraction and 40 s). This is also reflected in a pronounced increase in the two time constants throughout the initial 100 s, from 2.9 ± 0.1 ms to 17.4 ± 1.5 ms for tau1, and 38.6 ± 2.1 ms to 328.4 ± 17.6 ms for tau2. During the initial phase, rates of force development and rates of relaxation followed each other closely. During the rest of the exercise protocol, the two parameters did not follow each other. The relaxation rate was restored to 62% of initial value at 5 min (P < 0.001 vs. first contraction and 100 s). At this time point, both tau1 and tau2 were still significantly higher than initial values (7.9 ± 2.3 ms and 79.2 ± 11.1 ms). At 15 min of exercise, the relaxation rate was almost completely restored (94% of initial value). Tau1 was, in fact, significantly faster than the initial value (2.3 ± 0.1 ms), whereas tau2 was not different from the initial value (37.7 ± 2.0 ms).
At the start of the exercise, the length of the muscle was adjusted to the sarcomere length, L0, which gave maximum twitch force. During the 1-s period between stimulation trains, the muscle was first passively stretched to L0 (isotonic relaxation) and subsequently, during the isometric phase, force dropped (isometric relaxation). The baseline tension at L0 (Tbl) recorded just before the next stimulation train varied throughout the protocol. The main reason for this was the slowing of isometric relaxation, and a minor contributing factor was the transient reduction of the time available for isometric relaxation. From an initial value of 64.8 ± 0.8 mN, Tbl increased in a characteristic stepwise fashion. This is illustrated by the wave-like appearance of Tbl in Fig. 3A and represented by mean values in Fig. 4B. Tbl reached its maximum value at 100 s (253.0 ± 31.0 mN, P < 0.001 vs. Tbl at start), coinciding with the slowest rates of isometric force development and relaxation. Subsequently, Tbl declined to initial values after 5 min (69.5 ± 4.0 mN, P < 0.001 vs. 100 s, P = 0.85 vs. first contraction), without any further change for the remainder of the exercise period (64.1 ± 1.9 mN at 15 min).
Isotonic part of contraction.
The velocity of isotonic shortening (Fig. 5A) closely followed the changes seen in the rate of isometric force development throughout the stimulation protocol. There was an initial rapid drop of shortening velocity over the first 40 s by 52% (P < 0.001), and a slower (34%) drop during the next 60 s (P < 0.001 vs. first contraction and 40 s). After 100 s of exercise, there was no further change in shortening velocity, which stabilized at 33% of the initial level (P = 0.6 vs. 100 s).
The isotonic relaxation velocity (Fig. 5A) fell more rapidly than the isotonic shortening velocity during the first 40 s (by 81% compared with 52% for the shortening velocity, P < 0.001). There was a slight further slowing of the isotonic relaxation velocity the next 60 s, to a nadir of 87% of the initial value at 100 s. Subsequently, isotonic relaxation velocity recovered slightly but was still only 24% and 31% of the initial value at 5 min and 15 min of exercise, respectively.
Mainly, as a consequence of the reduction in isotonic shortening velocity, muscle shortening (Smax) fell during the stimulation protocol, almost monotonically (Fig. 5B), reaching a minimum of 29% of the initial value at 5 min. Smax remained at the same level for the rest of the exercise.
Recovery was recorded after 15, 30, and 60 min in three separate sets of experiments. Already 15 min after the end of the exercise protocol, both the rates of isometric force development (30 Hz) and relaxation and the velocities of isotonic shortening and relaxation were not significantly different from initial values recorded at the start of the exercise protocol (nominally between 79 ± 9 and 87 ± 3% of initial values). The variation was quite large, which means that some muscles still had not recovered. However, after a 30-min rest, variation was small and recovery was close to complete for all four rates/velocities (between 87 ± 3 and 93 ± 2%). After a 60-min rest, velocity of shortening had deteriorated again so that Smax was reduced, but all other parameters were close to normal.
Recovery was also recorded as the ability to perform another exercise protocol identical to the initial test protocol, but lasting only 5 min. Strikingly, overall performance was improved after 15 min of rest compared with the initial exercise protocol. Especially the rates/velocities during the first 100 s after onset of exercise were better maintained. Fig. 6, A–C shows an example of the rates of isometric force relaxation throughout the protocol in three separate experiments. A quantitative expression of performance was obtained by calculating the area under the curves. The gray-shaded area is therefore an integrated expression of the isometric relaxation rate throughout the first 5 min of the exercise protocol. The sum of the gray-shaded and the black areas represents a similar expression for isometric relaxation rate throughout the 5 min exercise after a recovery period. The black area gradually disappears after 30 and 60 min of recovery, which means that the transient reduction of isometric relaxation rate reappears gradually and is fully present after 60 min of restitution. The average values for areas under the curve have been plotted in D normalized to the initial fatiguing protocol. Interestingly, the rate of isometric force development (30 Hz) and the velocity of isotonic shortening were both quite significantly better maintained after 15 min rest (more than doubled). Hence, the muscle performed much better than during the initial fatigue protocol. Also, the rate of isometric relaxation and the velocity of isotonic relaxation were faster, but this effect was not as prominent. The characteristic stepwise appearance of the Tbl curve was undetectable after 15 min of recovery, but the steps at 40 and 100 s were easily detectable again after 30 min. After 1 h of rest, the recovered muscle behaved not differently from the unexercised muscle.
Finally, it should be mentioned that although the isotonic shortening velocity was highly improved after 15 min of recovery, the integral of Smax throughout the exercise protocol was not significantly different from the fatigue protocol during any of the recovery exercise periods.
Metabolites and oxygenation.
After 100 s of exercise, CrP content of the exercising muscle had fallen by 87% and ATP by 55% compared with the contralateral resting muscles (Table 3). At the same time, the lactate concentration had increased tenfold, and the NAD level had increased by 34% in the exercising muscle. After 15 min, CrP content had increased and was not significantly different from either the contralateral control leg or the 100 s value. Also, lactate fell significantly in spite of continued exercise, but after 15 min, there was still a small lactacidosis. The level of NAD decreased after 100 s, and at 15 min, the level was not different from resting values. The ATP concentration was still lower than initial values after 15 min. The content in the muscle tissue of all measured metabolites harvested from resting legs (100 s and 15 min) was stable and not significantly different at the two time points (referred to as “Rest” values In Table 3).
Compared to the assumed 60% reduction in SO2 during exercise in ISCH soleus muscle, exercising muscle with normal blood supply dropped 14 and 12%, depending on exercise time (100 s and 15 min, respectively, Fig. 7B). SO2 at these time points was not different from resting muscle oxygen saturation (70%). The oxygen level in complex IV in the mitochondrial respiratory chain was 9% higher in the muscles with normal blood supply compared with ISCH. The oxygenation status of exercising normoxic muscle was not different from resting control.
Myosin light chain analysis.
The MLC2s band was verified by immunoblot analysis and analyzed for phosphorylation after 100 s and 15 min of exercise, as well as after 15 and 60 min of recovery (Fig. 8). Relative to resting leg (100%), the MLC2s was dephosphorylated (51% of control) after 100 s (P < 0.05). At 15 min the phosphorylation status of the MLC2s was still at the same level (55%) as after 100 s (P = 0.76 and P < 0.05 vs. control). Thus, it seems that isotonic exercise in slow-twitch skeletal muscle caused a decrease in MLC2s phosphorylation. After 15 min of rest, MLC2s was fully rephosphorylated (84%, P = 0.36 vs. control). A full hour of resting did not lead to further alterations (86%, P = 0.23 vs. control). The protein level of MLC2f was low at all time points.
The main findings of this study can be summarized as follows: 1) After 15 min of isotonic muscle activity, fatigue is clearly seen as a reduction of shortening (Smax), which is in striking contrast to the maintained Fmax and V0. 2) Smax and MLC2s phosphorylation are both reduced during exercise. 3) During exercise, there is a transient decline in the rate of isometric relaxation and in unloaded shortening velocity with a maximum effect at 100 s paralleled by lactacidosis and a transient reduction of CrP increase in Pi. 4) Following 15 min of recovery, contractile parameters are almost normalized; however, during a repeated exercise protocol, muscle performance is significantly superior to the initial fatigue protocol. This effect is lost after 1 h.
Surprisingly, few investigators have examined muscle performance during controlled shortening contractions, and we suspect that important causes of fatigue in slow-twitch muscle during aerobic conditions at physiological temperature seem to have escaped detection and could be related to phosphorylation status of contractile proteins in addition to altered metabolite levels or intracellular calcium cycling.
Correlates of fatigue.
Interestingly, initial Fmax was not different from Fmax after 15 min of dynamic exercise. This is remarkable in light of the significant reduction of Smax and hence the more than 70% reduction in actual work performed by the muscle. During exercise involving movement, shortening capacity seems to reflect fatigue in a better way than Fmax. This is in line with the conclusion of Vedsted et al. (43).
Reduced Smax could be related to reduced velocity of shortening or simply to reduced shortening capacity, in which Smax would not increase even by increasing stimulation duration. In the present setting, also time available for shortening was a limiting factor, but accounted only for a small fraction of the reduction (<10%). Smax and isotonic shortening velocity were highly correlated throughout the experiment, as shown by the almost identical shape of the shortening curves (Fig. 9). Hence, reduced velocity of isotonic shortening probably reflects the intracellular changes responsible for reduced performance.
Contraction and shortening.
Velocity of isotonic shortening and rate of isometric force development (30 Hz) followed each other closely. During the first 100 s of the exercise protocol, they were almost linearly correlated and fell in parallel. Thereafter, these two parameters changed very little but still seemed to be correlated. This is in contrast to both the velocity of unloaded shortening (V0) and rate of force development during maximal isometric contraction (100 Hz), which were restored during the period following the 100-s nadir values and reached values not significantly different from control at the end of the exercise protocol. This could indicate another point of of control during physiological muscle activity, compared with the situation when either force or movement occurs in the absence of the other.
The accumulation of H+ and reduction in ATP during the first 100 s could affect the Ca2+ handling by the sarcoplasmic reticulum (SR). Both the ryanodine receptor and the Ca2+ ATPase are sensitive to pH and ATP levels. Thus, Ca2+ release may decrease during exercise. The reduced release of Ca2+ will depress Fmax, but also V0, rate of isometric force development (100 Hz) and isometric force at 30 Hz leading to a right-shifted force-frequency curve after 100 s. Dephosphorylation of MLC2s may also contribute to the reduced tetanic force. At 15 min, the level of Pi and lactate has decreased and are not different from resting levels. One could speculate that this would partly restore SR Ca2+ release rate. Since Fmax increases from 100 s to 15 min of exercise, but isometric force at 30-Hz stimulation frequency does not, it could be that Ca2+ release is only partly restored. Hence, at both frequencies, [Ca2+] would still be below control level, but sufficiently high to saturate binding sites at 100 Hz. Similarly, if dephosphorylation lowers the Ca2+ sensitivity, the resulting reduced isometric force production would be more prominent at 30 compared with 100 Hz. Also, the finding that the rate of isometric force development at 30 Hz and the force production at 30 Hz still are reduced after exercise compared with initial levels, while the rate of isometric force development at 100-Hz stimulation and maximal force production are restored supports the hypothesis that a rightward shift of the force-frequency curve has occurred after 100 s of exercise.
Elevated Pi and lactate and reduced ATP also affect cross-bridge interactions. During isometric contractions, the cross-bridge cycling (attachment, force development, detachment, and reattachment) is halted or cycles slowly, whereas during shortening, the reaction velocity is high (7). When the two modes of operation occur in succession, the rate-limiting step of the cross-bridge reaction cycle might shift so that Fmax, V0, and the rate of isometric force development (100 Hz) are affected in parallel. This could explain the parallell decline of these parameters during the first 100 s of exercise.
Thus, the Ca2+ release during the first 100 s could be reduced concurrent with a slowing of the cross-bridge cycling and a shift in the force-frequency curve. However, during the remainder of the exercise, Fmax and V0, were restored but isometric force at 30 Hz remained depressed. This could reflect an increase in Ca2+ release during stimulation, which does not exclude that, in addition, the force-frequency curve might be shifted to the right due to MLC2s dephosphorylation. This would also explain that the velocity of isotonic shortening and rate of isometric force development at 30-Hz stimulation frequency remained suppressed and only recovered after cessation of exercise.
The most prominent finding was that the initial decline of the rate of isometric relaxation was substantially reversed after the initial drop by more than 60% over the first 100 s. This was also apparent by the inability to reach baseline tension between stimulus trains, so that Tbl was transiently and conspicuously elevated. Because CrP and ATP also were lower and Pi and lactate were higher at 100 s compared with 15 min, it is possible that these events are associated. Furthermore, because also V0, Fmax, and the rate of isometric force development (100 Hz) were transiently depressed, as discussed above, one may ask whether all of these events can be linked to the same mechanism.
There is a plethora of articles discussing the effects of various metabolites on the contractile machinery and on Ca2+ cycling [see Allen et al. (1) for a review]. Ca2+ reuptake rate and the detachment rate seem to be reduced in parallel, and slowing of relaxation can thus not be ascribed to either one process or another. A fall in ATP concentration is probably important in the initial phase of fatigue development, and could thus be of importance during the first 100 s of dynamic contractions but increased levels of H+, ADP, and Pi are also possible candidates for slowing of both processes. However, the detrimental effect of acidosis at physiological temperatures has recently been questioned (36). Further, muscle function is remarkably well preserved in adenylate kinase knockout mice that accumulate large amounts of ADP (18). Because Fmax was transiently depressed after 100 s, it could be that Ca2+ was precipitated by Pi in the SR at this time point. This mechanism is, however, unlikely to play any role after 15 min of exercise, as Fmax by then has increased toward resting level. Hence, accumulation of Pi may be the most important factor for transiently reducing cross-bridge detachment rate and rate of relaxation by reducing the rate of the lever arm movement.
Interestingly, the velocity of isotonic relaxation fell much faster than all other contractile parameters following onset of exercise and recovered only slightly during the remaining exercise period. Currently, we have no explanation for why this parameter seemed so sensitive to exercise.
The initial fall in CrP and ATP and the accumulation of lactate must be ascribed to a lag of stimulation of oxidative metabolism. An initial fall of especially CrP has been recorded by many investigators, and in most reports CrP remain depressed throughout exercise (21, 34). There are, however, a few studies that report no change in the CrP concentration at the end of exercise (28). Of great importance for the transient change in CrP is probably that our protocol mimics a submaximal exercise only moderately and transiently reducing Fmax, which will empty energy stores more slowly than high-intensity setups. Also, during exercise, the muscle core temperature was 37°C, and it was fully perfused, providing the necessary substrates for synthesis of new high-energy phosphates. Lastly, the muscle studied is the slow-twitch soleus muscle that is fatigue resistant and has lower resting concentrations of CrP compared with fast-twitch skeletal muscle.
NIR analysis showed that skeletal muscle SO2 and cytochrome oxidation were, in fact, normal after 100 s and not different from the values at 15 min. Therefore, the reason for the lag of ATP resynthesis by the muscle seems to reside upstream of the respiratory chain. Recent reports have focused on the activation of oxidative phosphorylation, for example, of pyruvate dehydrogenase (PDH) involving phosphorylation (23). Because the respiratory chain seems to be fully oxygenated, lack of acetyl-CoA could explain the buildup of NAD after 100 s. Hence, it may be that in the rat soleus muscle, activation of oxidative metabolism, possibly PDH, requires at least 100 s of exercise. With activation of PDH, increased supply of substrates for oxidative phosphorylation will increase the level of reduced coenzymes, and the level of NAD decreases. We cannot explain why this seems to contrast findings of other studies of submaximal aerobic exercise, in which CrP remained depressed.
It is unclear what leads to the dephosphorylation of MLC2s. Early studies reported that MLC2s phosphorylation was increased during exercise (29, 33) and could explain the posttetanic twitch potentiation in fast-twitch skeletal muscle (33, 48). On this background, we were surprised to find that phosphorylation of MLC2s was reduced in our experiments. Recently, it has been shown that phosphorylation does not seem to be associated with reduced force development (37) but rather inhibits shortening velocity (22). This seems to contradict our results, but the effects on shortening velocity are demonstrated in fast-twitch fibers, and after isometric stimulation. Given that phosphatases and kinases seem to be distributed differently depending on fiber type (33) and that muscle dynamics indeed are dependent on mode of contraction (40), it remains a possibility that MLC2s in slow-twitch fibers that are allowed to shorten will be differently regulated in our experiments. Further, it is tempting to relate MLC2s dephosphorylation to the maintained depression of Smax and the rate of isometric force development at 30 Hz, as well as the velocity of isotonic shortening. It is known that phosphorylation of MLC2s will affect the force-pCa relationship, and as indicated above, a right shift of this relationship could explain our results.
The complete recovery of all mechanical parameters as well as the reversal of the dephosphorylation of MLC2s underscore the observation that the exercise protocol did not damage the muscle. The better performance after 15 min, especially the attenuated decline of the rate of isometric force development (30 Hz) and the velocity of isotonic shortening was striking, but for some reason, this was not reflected in the larger capacity for work. We have no explanation for this. Again, the close relationship between these contraction parameters indicates a common point of control, which is different from the rate-limiting step for relaxation. For future studies, we hypothesize that oxidative metabolism was more rapidly activated during this early recovery test protocol possibly because of a maintained activation of PDH, so that breakdown of ATP and CrP was less, and dephosphorylation of MLC2s not so marked. This would explain why rates of isometric and isotonic contraction were better preserved throughout the first recovery exercise.
Perspectives and Significance
In conclusion, it is quite clear that shortening contractions at body temperature and with maintained blood perfusion cause other responses to exercise than isometric contractions at lower temperatures and with suboptimal perfusion. Even in the fatigue-resistant slow-twitch soleus muscle of the rat, fatigue was easily detected as reduced ability to perform work under these physiological conditions (stimulated at 30 Hz), whereas traditional parameters of fatigue, for instance Fmax (measured at 100 Hz), were not detectably reduced after exercise. Although metabolic changes that occurred during the initial phase of the exercise (decrease of CrP and ATP and increase in lactate and Pi) could explain several of the changes observed in contractile parameters, dephosphorylation of MLC2s may be related to decreased ability of the muscle to shorten and to perform work.
The study was funded by grants from Norwegian Foundation for Health and Rehabilitation, University of Oslo and Ulleval University Hospital, Oslo, Norway.
- Copyright © 2009 the American Physiological Society