Journal of Applied Physiology
Vol. 83, No. 3, pp. 867-874, September 1997
EXERCISE AND MUSCLE

Efficiency of human skeletal muscle in vivo: comparison of isometric, concentric, and eccentric muscle action

T. W. Ryschon, M. D. Fowler, R. E. Wysong, A.-R. Anthony, and R. S. Balaban

Laboratory of Cardiac Energetics, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892

ABSTRACT

INTRODUCTION

METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Ryschon, T. W., Fowler, R. E. Wysong, A.-R. Anthony, and R. S. Balaban. Efficiency of human skeletal muscle in vivo: comparison of isometric, concentric, and eccentric muscle action. J. Appl. Physiol. 83(3): 867-874, 1997.The purpose of this study was to estimate the efficiency of ATP utilization for concentric, eccentric, and isometric muscle action in the human tibialis anterior and extensor digitorum longus in vivo. A dynamometer was used to quantitate muscle work, or tension, while simultaneous ³¹P-nuclear magnetic resonance data were collected to monitor ATP, phosphocreatine, inorganic phosphate, and pH. The relative efficiency of the actions was estimated in two ways: steady-state effects on high-energy phosphates and a direct comparison of ATP synthesis rates with work. In the steady state, the cytosolic free energy dropped to the lowest value with concentric activity, followed by eccentric and isometric action for comparative muscle tensions. Estimates of ATP synthesis rates revealed a mechanochemical efficiency [i.e., ATP production rate/work (both in J/s)] of 15.0 ± 1.3% in concentric and 34.7 ± 6.1% in eccentric activity. The estimated maximum ATP production rate was highest in concentric action, suggesting an activation of energy metabolism under these conditions. By using direct measures of metabolic strain and ATP turnover, these data demonstrate a decreasing metabolic efficiency in human muscle action from isometric, to eccentric, to concentric action.

phosphorus 31-nuclear magnetic resonance; tibialis anterior; extensor digitorum longus; phosphocreatine; magnetic resonance imaging; adenosine 5-triphosphate

INTRODUCTION

LOCOMOTOR ACTIVITY depends on variable contributions from isometric, concentric, and eccentric muscle actions. Activities that predominantly use a single mode of action can be cited. For example, tension development during muscle shortening (concentric action; see Ref. 6) enables the thigh muscles to elevate body mass during stair ascent. Eccentric muscle action, involving tension development during muscle lengthening (6), occurs in the same thigh muscle during step descent, providing a method of decelerating body mass. Isometric muscle actions involve tension development but no change in muscle length (6) and are the basis of posture maintenance.

In addition to the differential implementation of these muscle actions during locomotion, other prominent differences exist between them. Muscles acting in eccentric mode can produce greater peak tension than in other modes (8, 35, 40, 42). In addition, in humans, eccentric muscle action is associated with estimates of whole body energy cost (oxygen uptake) that are lower than for concentric activity at a similar intensity (1, 2). The classic human study is that of Abbott et al. (1), in which the "positive" working cyclist used much less oxygen than the "resisting" cyclist, despite the fact that both generated the same force on opposing bicycles. The differences in the force-velocity relation of eccentric and concentric muscle action increases the discrepancy in efficiency with increasing contraction velocity (1, 17). The lower energy cost for eccentric action could be explained by recruitment of more efficient fibers or by an alteration in the efficiency of converting high-energy phosphate bonds into measurable work [so-called mechanochemical efficiency (24)].

Alterations in the mechanochemical efficiency are of particular interest to muscle biochemists in their understanding of muscle contraction. However, these explanations may have important clinical implications as well. In particular, eccentric action of muscle is associated with activity-related damage to muscle involving delayed onset muscle soreness (38), appearance of muscle cytosolic enzymes in the blood plasma (5, 30), and temporary impairment of tension production (14, 18, 32, 38). To date, the role of mechanical and metabolic influences on these observations remains unclear.

The purpose of this study was to determine the metabolic response and estimate the mechanochemical efficiency of human skeletal muscle to isometric, concentric, and eccentric action. Toward this goal, two noninvasive ³¹P-nuclear magnetic resonance (NMR) spectroscopy methods and dynamometry were used to assess mechanochemical efficiency. In the first method, mechanochemical efficiency was calculated directly from the ratio of mechanical power output to ATP synthesis rate during steady-state concentric and eccentric action. ATP synthesis rate was inferred from the initial rate of phosphocreatine (PCr) resynthesis immediately after exercise (20, 21), as measured by ³¹P-NMR. Mechanical power was measured directly with a dynamometer. After correcting for the free energy in ATP, both the ATP synthetis rate and work could be expressed in the same units (J/s).

The second approach provided a qualitative confirmation of the direct method, along with an estimate of the efficiency of isometric action. This method relied on the measurement of tension, or work, and the available free energy for contraction in the phosphorylation potential of the muscle in the steady state. This is analogous to measuring the voltage of a battery as a function of the load resistance (28). During muscle action of submaximal intensity, energy metabolism is the "battery" providing energy through a source resistance of metabolic pathways. The load resistance represents the hydrolysis of ATP by myosin during work or tension development. As power output increases, ATP hydrolysis rate increases, reducing the load resistance and decreasing the battery voltage [i.e., the free energy of ATP hydrolysis (G_ATP)]. In this paradigm, G_ATP reflects the efficiency of the ATP utilization as long as the battery and source resistance (i.e., ATP synthesis pathways) are similar under the conditions examined.

It is hypothesized that for a given mechanical power output, relatively efficient action would result in a slower ATP resynthesis rate (resulting in a higher calculated mechanochemical efficiency) and less of a reduction in G_ATP.

METHODS

(1)

Table  1.   Mechanical and metabolic responses during final minute of isometric, concentric, and eccentric foot dorsiflexion in vivo Mode TTI, N · m  GATP, kJ/mol ADP, µM pH Rest  47.2 ± 0.3* 5.1 ± 0.4 7.02 ± 0.01  Isometric 54.1 ± 4.3   44.6 ± 0.6  10.4 ± 1.4  7.06 ± 0.02  Concentric 52.9 ± 5.0   39.7 ± 0.7 25.8 ± 3.2 6.97 ± 0.03  Eccentric 55.7 ± 5.3   42.6 ± 0.6  15.2 ± 2.0  7.04 ± 0.02 Values are means ± SE; no. of experiments for each mode = 8. TTI, tension-time integral; GATP, free energy of ATP hydrolysis. * P < 0.05 vs. isometric, concentric, and eccentric actions; P < 0.05 vs. rest, isometric, and eccentric actions; P < 0.05 vs. concentric and eccentric.

(2)

RESULTS

The torque values obtained at MVC in this study (women, 28.1 ± 2.9 N · m; men, 44 ± 2.5 N · m) are comparable to published values (12). Torque measurements from the final five strokes of isometric, concentric, and eccentric action are shown for a representative subject in Fig. 1. Overshoot of target torque occurred more often with dynamic modes than with isometric action. Nonetheless, TTI was comparable between modes (Table 1), and mechanical power output was not different between concentric (1.1 ± 0.1 W) and eccentric modes (1.2 ± 0.1 W; P = 0.739).

¹H MRI images were collected to confirm the localization of the surface coil by using the dual-tuned surface coil as well as to evaluate muscle recruitment by using the larger birdcage coil. Muscle recruitment was evaluated by imaging the lower leg immediately after exhaustive exercise. In the course of these studies, it was noted that at maximum exertion, it took roughly twice as long for each subject to reach exhaustion during the eccentric action. Figure 2 shows two series of multislice data from the same subject before and after concentric or eccentric activity on two different days. Note the selective enhancement of the anterior compartment of the leg. No significant changes in signal intensity were observed in the other muscle groups when normalized to fat-signal intensity in the bone marrow. These data suggest that the large majority of muscle activation was occurring in the anterior compartment in these studies. These studies were randomized and conducted on separate days to minimize any effects of the previous study. This was confirmed by the control image collected just before each trial.

³¹P-NMR spectra, accumulated to 1-min resolution, from each mode of action are shown in Fig. 3. Isometric and eccentric muscle actions reduced [PCr] and G_ATP less than concentric action (Figs. 3 and 4; Table 1). In five separate studies, the order of the muscle action in the study was reversed, with no statistical difference (P = 0.69) in TTI or steady state [PCr] or G_ATP. These studies suggest that the larger decrease in [PCr] in the concentric action is independent of the order of the trials. The replicate studies were not included in the summary data. These data suggested that isometric activity, followed by eccentric and concentric action, was the most efficient in generating TTI.

Determinations of the kinetics of [PCr] at the onset and end of muscle action were made by using spectra that were accumulated to 13-s resolution (Fig. 5) to evaluate whether a metabolic steady state was established during exercise. As noted above, the drop in [PCr] with concentric action exceeded that found in isometric and eccentric modes. Thus concentric action induced the greatest metabolic strain of the three modes and was least likely to result in a metabolic steady state during exercise. In all studies, [PCr] decreases to a steady-state level for the last 2-3 min of the exercise period. To illustrate this point, [PCr] is plotted for all of the subjects over the concentric exercise period (Fig. 5). [PCr] decreased in the first 2 min of muscle action, reaching a steady state that was maintained for the remainder of the trial. Thus the steady-state requirement of the study design was satisfied.

At the termination of exercise, [PCr] increased, with the first ~30 s of the response being essentially linear (linear regression analysis, r > 0.9; Fig. 6). This initial linear period was assumed to provide adequate temporal resolution to determine the initial rates of PCr resynthesis. The linearity of the data over this time course for eccentric or concentric activity was not significantly different. The initial rate of PCr resynthesis showed that the concentric rate (190 ± 16 µmol/s) was significantly higher than either isometric (62 ± 10 µmol/s; P < 0.05) or eccentric (98 ± 20 µmol/s; P < 0.05), while the rates for isometric and eccentric action were not significantly different. The corresponding mechanochemical efficiency was 34.7 ± 6.1% for eccentric action and 15.0+ 1.3% for concentric action (P = 0.017). Because no external shortening is measurable with isometric action, efficiency cannot be directly calculated.

To ensure that changes in spin-lattice relaxation times were not contributing to the initial rates, the pulse width was reduced by a factor of three from the optimal pulse width in three subjects. This effectively alters the Ernst angle relationship, resulting in less T1 weighting of the NMR data. Despite a decrease in the signal-to-noise ratio of the PCr resonance, no difference was observed in the initial rates of recovery from eccentric or concentric action as a function pulse width (i.e., magnetization flip angle).

The estimated Q_max in concentric action (391 ± 22 µM/s) was significantly higher than either isometric (235 ± 43 µM/s) or eccentric action (271 ± 44 µM/s; P < 0.05). The Q_max was not statistically different between isometric and eccentric modes. These data suggest that the predicted higher metabolic rate of concentric action is associated with an activation of aerobic metabolism. It should also be pointed out that the assumptions that [ADP] is the sole kinetically limiting factor under these conditions could also result in a systematic error.

DISCUSSION

In this study, the same muscle mass was voluntarily activated to similar submaximal levels of torque output, at a fixed duty cycle, using isometric, concentric, and eccentric modes of muscle action. Under these comparable mechanical conditions, significant differences were found for several metabolic variables between the different modes of action. These metabolic changes are consistent with an increasing metabolic strain for a given TTI, going from isometric to eccentric to concentric action.

The higher rate of ATP hydrolysis associated with the changes in muscle length (24) likely explains the lower G_ATP for concentric and eccentric action compared with isometric action. However, between eccentric and concentric action, the muscle is changing the same length in the same time (i.e., same velocity), resulting in comparable power output. If the mechanochemical efficiency of cross-bridge formation were identical for both modes, the metabolic requirements should be identical. However, the metabolic strain, or drop in G_ATP, was greater with concentric action, consistent with a lower mechanochemical efficiency. This conclusion was confirmed by the direct calculation of mechanochemical efficiency (i.e., ATP production rate/work) from the [PCr] resynthesis rates where the eccentric action was found to be ~35% more than twice as efficient as the ~15% observed for concentric activity. It is interesting to note that the concentric and eccentric efficiency values found in these studies are well below the 55-60% values reported for isolated amphibian muscle at low temperatures (~1.5°C) (10, 24).

The observed differences in mechanochemical efficiency between eccentric and concentric action may be attributable to altered energetics in the working muscle fibers. This implies that the coupling between ATP demand and power output is adjustable within the working fibers, assuming the same population of fibers was activated. The same observations could occur if a population of more efficient fibers were activated. In this regard, electromyography (EMG) suggests that eccentric action activates fewer fibers for a given tension output than concentric action, according to some (2, 4, 36, 40) but not all studies (22, 23, 36). In studies showing lower EMG activity during eccentric action, the EMG-to-torque ratio is approximately twofold higher in concentric action (40), indicating a 100% increase in eccentric efficiency. This is consistent with our estimates of mechanochemical efficiency for these muscle actions. However, if fewer fibers are recruited during eccentric compared with concentric action, then each activated fiber must produce a greater fraction of the work and a higher associated metabolic strain.

Two lines of evidence suggest that this did not occur. First, if a selective pool of fibers were maintaining the workload in the steady-state during eccentric action with the same overall mechanochemical efficiency as during concentric action, then the calculated rate of ATP resynthesis should be the same for the two modes, because this measure is independent of pool size. In contrast, eccentric action had a lower calculated ATP synthesis rate, inconsistent with the compartmentation hypothesis but consistent with the notion that total ATP requirements are lower in eccentric action. Second, no significant change in intracellular pH or splitting of the P_i peak was observed in any of these submaximal workload studies. This is also inconsistent with a small, highly activated fiber population. A highly active fiber population would potentially dominate the overall muscle P_i resonance because of a high concentration of P_i in these fibers. These data are inconsistent with the notion that a small, efficient fiber population is recruited in eccentric action. On a larger scale, ¹H images suggest that only the anterior compartment was significantly involved in the work performed at maximal workloads. However, recruitment below the T2 threshold of this study or outside of the field of view (i.e., upper leg) cannot be completely discounted, nor can these data be used to estimate the percentage of fiber activation in these muscles. It was interesting to note that the eccentric action required roughly twice as much time to reach exhaustion as concentric activity in the present study. This result is consistent with the notion of higher metabolic efficiency in the eccentric action.

These results suggest that the higher mechanochemical efficiency of eccentric compared with concentric action is due to an alteration of the actino-myosin-ATP stoichiometry. We speculate that some component of eccentric muscle action (the direction of length change or energy added to the muscle through the lengthening process) lowers the requirement of ATP per second per watt of mechanical power. Studies of single fibers during shortening have demonstrated a larger sliding distance between actin and myosin than expected from the quantity of ATP hydrolyzed (16). This could represent multiple power strokes per molecule of ATP hydrolysis (27) or a longer power stroke (>5-10 nm) than predicted from in vitro studies (16). The preferential occurrence of either of these phenomena during eccentric action could explain the higher efficiency. An alternative explanation may be that in the eccentric mode the muscle is actually dissipating a potential energy rather than generating it, as in the other modes. Under these conditions, the external load applied to the muscle could overcome weakly bound actin-myosin states or other passive elements, resulting in a resistance to the motion with little ATP hydrolysis. This latter model, where the tension is generated by stretching or destroying passive elements, may explain the higher potential for muscle damage in eccentric action, as discussed above. In any event, the higher efficiency of the eccentric action may contribute to the lower eccentric EMG activity seen in some studies, because less muscle activation is required to provide a given level of work.

The estimated mechanochemical efficiency found in this study for concentric action was much lower than observed in human exercise protocols (13), which are limited by potential errors in the measurement of both metabolic energy cost and mechanical power output (37). In general, studies on humans are difficult because local metabolic rates and muscle power are difficult to determine noninvasively, muscle group recruitment is hard to control, and subject compliance is mixed.

In the present study, we have attempted to minimize many of these problems, although a perfectly controlled study is always extremely challenging in humans. The local metabolic rate in the muscle of interest was directly assessed by using ³¹P-NMR. Care was taken in the positioning and coaching of subjects to isolate the anterior compartment in these studies. MRI confirmed that the anterior compartment was the dominant group recruited. However, complete exclusion of other muscle groups contributing to the generated force cannot be assured. Subject compliance, outside of positioning, was not as critical in these studies, because muscle power was directly measured, and each individual value was used in the efficiency calculations. This study thus reduces many of the limitations of prior studies by using ³¹P-NMR in combination with a quantitative dynamometer.

The estimated Q_max was highest in concentric muscle action. This metabolic recruitment could be a response to the larger metabolic strain associated with this relatively inefficient muscle action. Increases in Q_max with the larger metabolic strain associated with concentric activity could be mediated by increases in dehydrogenase activity, induced by increases in cytosolic Ca²⁺ or ADP levels, with subsequent increases in NADH delivery (3). Consistent with this notion is the activation of dehydrogenase activity observed with near-maximum work in human skeletal muscle (31). Another activation site may include the F₁-F₀ adenosinetriphosphatase directly (9, 34). This type of metabolic activation with work may cause the PCr/P_i "overshoot" observed with recovery after heavy exercise (7) if the metabolic activation persists longer than the exercise period. The increased metabolic capacity persisting after heavy exercise may result in a higher set point for the cytosolic G_ATP, resulting in the overshoot. It is unclear whether the apparent larger metabolic activation with concentric action was specifically caused by the muscle action or by the larger metabolic strain compared with eccentric activity. Studies with matched metabolic strain between eccentric and concentric action, rather than workload as performed in the present study, may resolve this issue.

The Q_max is estimated by assuming that [ADP] alone is rate limiting for oxidative phosphorylation (20). In the present study, however, significant changes in P_i were also observed in proportion to the decreases in PCr. P_i could also be rate limiting for oxidative phosphorylation with a K_m of ~1 mM (3). In the present study, it was estimated that P_i increased to almost 12 mM with concentric activity, from a starting concentration of ~1 mM. Thus changes in P_i could be contributing to the differences in observed in Q_max. To test this hypothesis, we used a bireactant model for ADP and P_i control of respiration (19). By using this model, the Q_max equation becomes

(3)

where K_Pi is the K_m for P_i (1 mM). All other parameters are as described in Eq. 2. By using this approach, the Q_max for concentric action was 410 and 304 µM/s for eccentric action. These results are consistent with the simple [ADP] model presented earlier and support the notion that the calculated maximum rate of respiration is greater with the concentric metabolic strain.

In the present study, the change in Q_max affects the steady-state measurements, because the metabolic ATP production capacity, or the battery and source resistance in the electrical analog model, is not constant with different muscle actions. However, Q_max was highest in concentric action, which would predictably result in a smaller decrease in G_ATP with work, not the larger drop observed in this study. Thus the relative efficiency differences between eccentric and concentric action are most likely underestimated by using the steady-state approach because of the apparent increase in metabolic recruitment occurring in concentric muscle action (i.e., higher Q_max).

Several limitations to using the PCr resynthesis rate as a measure of ATP production are critical for the quantitation of the efficiency and estimation of Q_max. An accurate initial rate must be determined with adequate temporal resolution. On the basis of the linearity of the time-course data, the 13-s time resolution seemed to be adequate for these purposes. This measure is aided in skeletal muscle by its relatively low resting metabolic rate compared with active states. The initial rate of PCr resynthesis is useful for these calculations if it is solely dependent on the metabolic production of ATP. Clearly, some contributing factors may include rapid changes in pH, Mg²⁺ concentration ([Mg²⁺]), spin-lattice relaxation rates, or intra- or intercellular metabolite compartmentation. No significant changes in intracellular pH occurred at these workloads. Changes in intracellular [Mg²⁺] can be determined from the relative positions of the ATP resonances due to Mg²⁺ binding by using difference spectroscopy (25). No spectral dispersion in the ATP resonance was observed (see Fig. 3) in the different spectra, suggesting that intracellular free [Mg²⁺] was also constant (i.e., chemical shift of ATP was constant). The potential effects of rapid changes in spin-lattice relaxation rates were evaluated by using variable flip angles, and no significant effect was observed. Severe intra- or intercellular compartmentation of the metabolites could influence these results to some extent; however, as discussed above, these effects would appear to be minor. Finally, these studies were limited to a single workload and contraction velocity. It is possible that different workloads or velocities could result in alterations in the relative metabolic efficiencies observed.

ACKNOWLEDGEMENTS

We thank Drs. David Wiesler and Chris Combs for their help in the MRI experiments, the Biomedical Engineering and Instrumentation Program in the National Center for Research Resources for continuing help on the dynamometer, and Dr. Andrew Arai for many useful discussions.

FOOTNOTES

Address for reprint requests: R. S. Balaban, LCE, NHLBI, NIH, Bldg. 10, Rm B1D-161, Bethesda, MD 20892-1061 (E-mail:rsb{at}zeus.nhlbi.nih.gov).

Received 26 March 1996; accepted in final form 6 May 1997.

REFERENCES