Muscle damage is not a function of muscle force but active muscle strain

RICHARD L. LIEBER AND JAN FRIDfiN Department of Orthopedics and Biomedical Sciences Graduate Group, University of California and Veterans Affairs Medical Centers, San Diego, California 92161; and Departments of Anatomy and Hand Surgery, University of Urned, S901 87 Ume&, Sweden

LIEBER,RICHARD L., ANDJANFRID~N. Muscledamageis not a function of muscle force but active muscle strain. J. Appl. Phys- iol. 74(Z): 520-526, 1993.-Contractile properties of rabbit ti- bialis anterior muscles were measured after eccentric contrac- tion to investigate the mechanism of muscle injury. In the first experiment, two groups of muscles were strained 25% of the muscle fiber length at identical rates. However, because the timing of the imposed length change relative to muscle activa- tion was different, the groups experienced dramatically differ- ent muscle forces. Because muscle maximum tetanic tension and other contractile parameters measured after 30 min of cy- clic activity with either strain timing pattern were identical (P > 0.4), we concluded that muscle damage was equivalent despite very different imposed forces. This result was sup- ported by a second experiment in which the same protocol was performed at one-half the strain (12.5% muscle fiber length). Again, there was no difference in maximum tetanic tension after cyclic 12.5% strain with either strain timing. Data from both experiments were analyzed by two-way analysis of vari- ance, which revealed a highly significant effect of strain magni- tude (P < 0.001) but no significant effect of stretch timing (P > 0.7). We interpret these data to signify that it is not high force per se that causes muscle damage after eccentric contrac- tion but the magnitude of the active strain (i.e., strain during active lengthening). This conclusion was supported by morpho- metric analysis showing equivalent area fractions of damaged muscle fibers that were observed throughout the muscle cross section. The active strain hypothesis is described in terms of the interaction between the myofibrillar cytoskeleton, the sar- comere, and the sarcolemma.

eccentric contraction; muscle injury; stress; cytoskeleton; inte- grins; intermediate filaments; desmin

IT IS DELL KNOWN that eccentric contractions (EC) re- sult in high forces and are a normal part of the gait cycle, especially in the extensor muscles (8, 11). Despite their common occurrence, the underlying basis for the behav- ior of muscle during EC is not as well understood as mus- cle behavior during concentric (shortening) contractions. For example, Harry et al. (10) recently performed frog sartorius muscle EC at high velocities and concluded that the shape of the lengthening portion of the force-velocity curve could not be explained on the basis of simple modi- fication of the cross-bridge theory proposed by Huxley (14). The experimental data of Harry et al. were accompa- nied by an intriguing explanation of muscle lengthening behavior based not on cross-bridge properties but on in-

teraction between sarcomeres along the fiber length (21). These two reports outlined the difficulty in understand- ing EC of skeletal muscle.

Chronic studies have also demonstrated that EC is se- lectively associated with muscle injury and muscle sore- ness in humans and in various animal models (see Ref. 4 for review). That injury occurs after EC is intriguing in light of the observation that EC is very common during normal gait. Because high muscle tension and muscle injury can both be associated with EC, it has been pro- posed that high muscle force is responsible for the mus- cle injury observed. However, this proposal has not been explicitly tested. The purpose of this study, therefore, was to determine the effects of force and strain on the tension generated by rabbit tibialis anterior (TA) mus- cles. These experiments were performed to test the hy- pothesis that muscle damage is due to high force. A brief account of this work has been presented elsewhere (20).

METHODS

The muscle chosen for study was the TA of the New Zealand White rabbit. This muscle was chosen primarily because the muscle fibers are oriented with a pennation angle of only 3O (16) and thus demonstrate negligible angular rotation during lengthening. Pilot experiments with more highly pennated muscles, such as the gastroc- nemius, revealed significant shear stresses that were nonlinear and highly dependent on strain magnitude. Contractile properties were measured and experimental treatments were performed essentially as previously de- scribed (17).

Preparation and contractile measurements. Briefly, rabbits were anesthetized with a subcutaneous injection of a ketamine-xylazine-acepromazine cocktail (50,5, and 1 mg/kg body mass, respectively) and mai ntained on hal- othane anesthesia. Heart and respiratory rate were mon- itored during muscle isolation and testing (model 78O7C, Hewlett-Packard, Palo Alto, CA), and anesthesia level was adjusted as needed. All experimental procedures were performed in accordance with the guidelines set forth by the National Institutes of Health “Guide for the Care and Use of Animals.”

Great care was taken to minimize system compliance, ensuring that the imposed deformation was taken up by the muscle itself and not the apparatus or the external tendon. The distal TA tendon was secured to a dual- mode servomotor (model 310, Cambridge Technology,

520 0161-7567193 $2.00 Copyright 0 1993 the American Physiological Society

MUSCLE DAMAGE IS NOT DUE TO HIGH FORCE 521

nominally 55 mm, absolute strain magnitude and strain rate were -13 mm and 65 mm/s, respectively.

In the LS group (late stretch), muscle stretch was de- layed for 200 ms while the muscle developed tension. Then the muscle was stretched with the identical pattern as the ES group. Thus stretch rate and magnitude were identical between groups. H owever, because of the stretch timing, the peak force reached in the LS group was significantly greater than that of the ES group(Fig. 1). In this way, both groups received similar deformation patterns but experienced dramatically different forces. It should be noted that the 200-ms delay is probably longer

Moderate 25 Very high 25

Early stretch Late stretch

125 125

Early stretch Low Late stretch Moderate

12.5 12.5

63 63

L,, fiber length.

Cambridge, MA) and aligned with the motor’s measuring and translation axis. System compliance, including the than the normal 60- to SO-ms delay usually observed be-

tween muscle activation and lengthening (9). However, we increased the delay to produce dramatically different forces in the two experimental groups.

Variation of muscle strain. The experi ments described

transducer, was 1.3 pm/g. The peroneal nerve was isolated for direct muscle acti-

vation. Muscle temperature was then maintained at 37°C with radiant heat, mineral oil, and a servo-tempera- ture controller (model 73A, Yellow Springs Instruments, Ye1 .low Springs, OH). Under computer control (19), mus- cle length was adjusted to the length at which twitch tension was maximum (L,), and contractile properties were determined before experimental treatment .(see be- low). Contractile properties measured included time to peak twitch tension, the rate of rise of twitch and tetanic tension (dP/dt), twitch half-relaxation time, maximum twitch tension, passive force in response to 12.5 or 25% strain at 63% muscle fiber length (L,) /s or 125% L,ls (see below), and contractile tension at stimulation frequen- cies of 5, 25, 50, 75, and 100 Hz. Maximum tetanic ten- sion (P,) was defined as the tension measured while stim- ulated at 100 Hz, the peak of the force-frequency rela- tionship. Half-fusion frequency was defined as the frequency at which the tension was 50% P,. Measure- ments of L, were made on each muscle from th .e tibia1 tubercl .e to the most distal muscle fiber insertion. On the basis of previous architectural studies (16), L, was calcu- lated for each muscle as 0.67L,.

above were repeated at one- ,half the total strain (12.5% L,) in two separate groups of animals (n = Wgroup). Thus, combining the results of both experiments, high and low strains were imposed at high and low forces to test the effect of force and strain independently (Table 1).

It was imperative that the two groups have nearly the same metabolic energy requirements so that the experi- ment tested only mechanical difference s between groups and not differen .ces based on cellular metabolism (2). Therefore, all experimental activation cycles consisted of 400-ms trains of 40-Hz pulses. A separate e xperimental group that only experienced cyclic isometric activation (n = 8) was also studied for comparison.

Treatment was performed on-line while the computer synchronized muscle activation and length change and stored contractile data in real time at specified intervals. In addition to data acquisition and storage, the computer also performed real-time integration of force with re- spect to time for each contraction according to

Experimental treatment. Deformation patterns were imposed on muscles at two strain magnitudes and using two timing patterns relative to muscle activation (Table 1). The experimental design was that of a two-way analy- sis of variance (ANOVA), with two levels of strain (12.5 and 25%) and two levels of timing (“early” and “late” stretch).

rt=400 ms

MI (gas) = I P( t)dt

Variation of muscle force using different stimulation timing. To test the hypothesis that muscle damage is di- rectly related to fiber force, we imposed cyclic linear length changes of identical magnitude and velocity on the TA muscle. Cyclic length changes were completed in 400 ms and repeated, along with muscle activation, every 2 s for 30 min for a total of 900 stretches. The only differ- ence between groups (n = ll/group) was the timing of the stretch. In the ES group (early stretch), the muscle was stretched coincident with the onset of muscle activa- tion (Fig. 1). The magnitude of the stretch was 25% of the TA L, (determined individually for each muscle), and the strain rate was 125% LJs. Both the magnitude and rate of stretch were within the physiological range based on cat kinematic data, which demonstrate TA strains of 10 and 50% during the E, phase of the gait cycle that are complete within 0.1 and 0.3 s during walking and gallop- ing, respectively (cf. Fig. 12 of Ref. 8). Because L, was

where P(t) is the muscle force during the mechanical impulse calculated.

activation and MI is

A : : : :. . : . !

i ;

500 ms

FIG. 1. Sample contractile data from early stretch (A) and late stretch (B) experimental groups. Note that both groups receive identi- cal deformation patterns (bottom). However, due to timing of applied deformation, late stretch group experiences much higher forces (top). Stippled area beneath force record represents stimulation duration. L,! muscle fiber length.

522 MUSCLE DAMAGE IS NOT DUE TO HIGH FORCE

TABLE 2. Contractile properties of muscles tested at 25% strain

Parameter Pre

Early Stretch

Post Pre

Late Stretch

Post

Maximum tetanic tension, g Tetanic dPldt, g/ms Time to peak tension, ms Twitch force, g Half-relaxation time, ms

1,325&103 514t56 1,474?64 59Ok52 47.3t4.0 7.47t1.08 54.8t2.60 9.01+1.01 25.4+0.62 227g22.2

22.1k2.32 25.8t0.66 22.9t2.26 31.9k4.6 276.7529.6 44.3t7.4

44.8k2.6 35.7t4.0 43.8k2.6 37.9t4.0

Values are means + SE; n = Wgroup. No significant difference was observed between any pre- (Pre) or postcontractile (Post) parameters measured. dP/dt, rate of rise of twitch and tetanic tension.

After the treatment period (30 min), animals were maintained under anesthesia for 1 h to permit the early inflammatory response. Then contractile properties were again measured. Animals were killed, and the TA was excised and submitted for light-microscopic investi- gation.

Light microscopy. The TA was frozen in isopentane cooled by liquid nitrogen (-159°C) and stored at -8OOC for histochemical processing. Muscle cross sections (8 ,urn thick) taken from the TA midbelly were stained with hematoxylin and eosin to observe overall fiber appear- ance, location of nuclei, and appearance of connective tissue. Enzyme assays were performed on selected mus- cles to demonstrate oxidative enzyme activity (succinate dehydrogenase) (ZZ), glycolytic activity (a-glycerophos- phate dehydrogenase) (ZZ), and myofibrillar adenosine- triphosphatase activity (1). Muscle fibers were classified as fast oxidative glycolytic, fast glycolytic, or slow oxida- tive, according to the classification scheme of Peter et al. (23). The relative area (area fraction) of damaged fibers (see below) was determined by the stereometric point- counting technique of Weibel (26). Sampling protocol consisted of fiber damage measurements from 2 tissue blocks/muscle, 3 sections/block, and 5-10 fields/section.

Statistical analysis. For each muscle, pre- and postex- ercise contractile parameters were obtained. Only post- exercise morphological parameters were obtained from these specimens, although contralateral muscles (un- tested and untreated) were used as controls. For each muscle, the difference between the pre and post value of a parameter was expressed as a percentage change in that parameter. First, it was determined whether the percent change was significantly different from zero us- ing a one-sample t test, i.e., whether treatment signifi- cantly changed the parameter. For the overall experi- ment, a two-way ANOVA was used with strain (12.5 vs. 25%) and stretch timing (early vs. late) as the grouping factors. Assumptions of the ANOVA and t tests (normal-

ity and equality of variances between groups) were explic- itly tested using the diagnostic software in SuperAnova (Abacus Concepts, Berkeley, CA) by plotting group means vs. variance and by visual inspection of the AN- OVA residuals. To quantify the relative effects of strain and timing, data were submitted to a stepwise regression model where the dependent variable was PO after treat- ment and the independent variables were strain and peak force achieved during treatment. Data are means t SE. Significance level was chosen as cy = 0.05. Power analysis revealed that the statistical power (1 - p) for these ex- periments exceeded 80% for all parameters except area fraction of damaged fibers (see RESULTS).

RESULTS

Contractile properties after high vs. low force. Two-way ANOVA with repeated measures revealed no significant difference between experimental groups for any contrac- tile properties measured before and after treatment for either the 25% strain group (P > 0.4; Table 2) or the 12.5% strain group (P > 0.6; Table 3). The 25 and 12.5% experiments yielded qualitatively similar results in terms of the tension time course during treatment. Results for the 25% strain group revealed that the initial peak force of the LS group was 40% greater than that of the ES group and remained significantly higher throughout the treatment period (Fig. ZA). Interestingly, the mechanical impulse (i.e., the integrated force-time record) experi- enced by the two groups was also significantly different (P < 0.01) but in a way that was not expected. Whereas peak force was always greater in the LS group, the initial mechanical impulse of the ES group was 520 go s, whereas that of the LS group was only 460 g l s (Fig. ZB). The relative difference between groups also changed sig- nificantly over time (Fig. 3). Peak force in the LS group was significantly greater than in the ES group, but the magnitude of this difference decreased rapidly during the

TABLE 3. Contractile properties of muscles tested at 12.5% strain

Parameter

Maximum tetanic tension, g Tetanic dPldt, g/ms Time to peak tension, ms Twitch force, g Half-relaxation time, ms

Early Stretch Late Stretch

Pre Post Pre Post

1,274+54 749t36 1,244t34 741k38 52.0t3.90 15.2k1.75 49.3A3.11 13.8k2.30 25.020.46 24.1k1.12 23.9t0.61 26.3k0.94 233t42.9 47.5t8.2 253k35.3 49.2k9.4

37.9tl.O 37.1k1.9 38.1k1.3 38.8k1.4

Values are means & SE; n = 8/group. No significant difference was observed between any pre- or postcontractile parameters measured.

MUSCLE DAMAGE IS NOT DUE TO HIGH FORCE 523

Early Stretch Late Stretch

0

0 a

600

'0"

0 0 0

o! 1 I 1

0 10 20 30

Treatment Time (min)

FIG. 2. Mechanical environment of muscles during experimental treatment. A: peak force during experimental treatment. Note that late stretch group experiences higher forces throughout treatment period. B: mechanical impulse during experimental treatment. Note that de- spite the fact that late stretch group experiences higher peak forces, early stretch group experiences a greater mechanical impulse. Data shown for 25% strain.

first few minutes (Fig. 3A). Thereafter, the ratio between the groups slowly increased for the next 15 min. The mechanical impulse of the ES group was greater than that of the LS group, and the magnitude of this differ- ence increased as a function of time (Fig. 3B). This ap- peared to be due to the rapidly decreasing peak force in the LS group. Again, most of the change occurred in the first few minutes of treatment, rapidly rising to 1.5 within 3 min and increasing in a relatively linear fashion to 4.0 over the remaining 27 min. This was because the LS group impulse decreased more rapidly during the first few minutes than did the ES group (Fig. 2B).

Thus the peak and integrated tensions experienced by the two experimental groups were different, yet their contractile properties after 30 min of such treatment were identical. As an example, P, and tetanic tension dPldt after cyclic 25% strain decreased by the same amount in both groups, suggesting an identical decrease in performance and speed, independent of treatment tension (Fig. 4).

Contractile properties after high vs. low strain. On the basis of the observation that varied force did not cause a change in contractile properties, we tested the effect of altered strain on contractile properties. Ideally, we would have applied several different strains to the muscle at identical forces. However, to accomplish this, the strain rates required were well out of the physiological range (200-400%/s). We thus tested the effect of strain per se by performing the identical experiments described above

but at 12.5% L, strain. The results were qualitatively simi- lar to the 25% strain experiments; there were no signifi- cant differences between strain timing groups for all con- tractile parameters measured (P > 0.6; Fig. 5). In addi- tion, maximum tetanic tension from all EC groups was significantly lower (P < 0.05) than that measured after treatment with only isometric contractions (Fig. 5).

Differential effects of force and strain. Two-way AN- OVA of all experimental data grouped by strain (12.5 vs. 25%) and timing (early vs. late stretch) revealed no signif- icant effect of timing (P > 0.7), a significant effect of strain (P < O.OOl), and no significant strain X timing interaction (P > 0.9) on P, measured after EC treatment. To determine the dependence of posttreatment P, on strain and force, a multiple regression model was devel- oped that demonstrated that strain accounted for ~50% of the experimental variability in P, measured after EC treatment, whereas peak tension achieved during treat- ment accounted for only an additional 8% of the variabil- ity. This analysis reinforces the idea that strain rather than force is more important in determining the magni- tude of the tension decrease after EC.

Muscle morphology after eccentric exercise. To investi- gate the structural basis for the contractile properties measured after 25% strain, multiple serial sections were obtained from the midbelly of the treated muscles. On the basis of previous studies that documented signifi- cantly enlarged, pale staining, and rounded fibers of the fast glycolytic fiber type that were selectively associated with muscle injury (6,17), we examined the area fraction of these enlarged fibers.

Comparisons between the ES and LS groups of the

1.4

5

A

1.3

t 0 0

1 1.2

i

l

I

l o.e 0.

1.1

1.0 I r 1 1

0 10 20 30 40

0 .- 5 III 1.8

2 2 1.6

E -

6 0

l l

1.0 ! 1 I I 1 0 10 20 30 40

Treatment Time (min) FIG. 3. Peak force (A) and mechanical impulse (B) ratio between

early and late stretch groups of 25% strain. Note that ratio continually changes throughout time course of treatment.

524 MUSCLE DAMAGE IS NOT DUE TO HIGH FORCE

2000

1 A 0 Pm-Exercise q Post-Exercise

Early Stretch

Early Stretch

Late Stretch

Late Stretch

Group FIG. 4. Contractile properties after experimental treatment at 25%

strain. A: maximum tetanic tension. B: maximum rate of rise of tetanic tension (dPldt). Despite widely different mechanical treatments, there was no significant difference between groups for any contractile proper- ties measured.

first experiment revealed no significant difference in the area fraction of damaged fibers (P > 0.6). Unfortunately, there was a great deal of variability in the morphometric data. Whereas for the contractile parameters, coeffi- cients of variation ranged from 10 to 35%, morphometric values for area fraction had a coefficient of variation of -80%. This of course made it extremely difficult to de- tect significant differences between groups. Power analy- sis revealed that to have an 80% chance of detecting a 3% difference between groups in area fraction of damaged fibers (given the actual variability of the data), a sample size of ~60 would be required. Statistical power for the given data set was thus only 40%.

DISCUSSION

The purpose of this study was to investigate mechani- cal factors contributing to muscle damage after EC. Pre- vious investigators have suggested that the high forces associated with EC may be responsible for the damage observed. Although this is certainly an attractive hy- pothesis, the current experiments do not support this idea.

Our data demonstrate that large differences in force and mechanical impulse during EC did not result in mus- cle damage differences, as evidenced by identical con- tractile and morphological parameters. However, large strain differences did result in contractile and morpho-

logical differences. If one accepts the first experiment as showing that muscle damage is not a function of force during EC, then the conclusion from the second experi- ment is that muscle damage is a strong function of the muscle fiber strain that occurs during lengthening of an activated muscle (i.e., active strain). There is conceptual support for this idea in the literature based on known sarcomere structure.

The myofibrillar array is embedded in a complex ex- trasarcomeric cytoskeletal framework (24). The cytoskel- eta1 matrix is joined to the muscle cell basal lamina and sarcolemma via adhesive connections made by talin, vin- culin, and a-actinin, among other proteins, and the inte- grin superfamily of adhesion receptors (25). It is likely that active strain that exceeds the limits of these connec- tions may result in cytoskeletal damage. We previously documented the longitudinal extensions that interrupt the normal desmin periodic pattern in muscles subjected to EC. We also recently described subtle changes in fiber integrity accompanying EC-induced injury (7). It is plau- sible that cytoskeletal damage may be the first structure to yield after EC, as has been previously suggested (5). This could then result in the myofibrillar disruption at the level of the A band and Z disk, which has been demon- strated in the literature (5, 15, 17). The mechanism for the damage would thus initially be cytoskeletal disrup- tion followed by myofibrillar derangement. This could, in principle, be similar to the myofibrillar disruption re- ported by Horowits and Podolsky (l3), who selectively irradiated the intermyofilamentous protein titin (which, by itself, did not result in myofibrillar alterations) and, on contraction, observed significant misalignment of the A band and “smearing” of the Z disk. The similarity be- tween their micrographs and those reported by other in- vestigators after EC (e.g., see Refs. 5, 15) is intriguing.

The implications of these findings for exercise involv- ing EC are not clear. It is not appropriate to conclude

2ooo

0

El Early Stretch q Late Stretch

Control Isometric 12.5%

Experimental Group 25%

FIG. 5. Summary of maximum tetanic tension generated by various treatment groups. No significant difference in maximum tension gen- erated was observed between groups that experienced early stretch compared with late stretch at either 25 or 12.5% strain. However, a significant difference in tension generation was observed between groups strained 25 vs. 12.5% of muscle fiber length. Isometric data from Lieber et al. (17). Control, mean + SE of normal rabbit tibialis anterior muscles.

MUSCLE DAMAGE IS NOT DUE TO HIGH FORCE 525

simply that low-amplitude joint excursions will result in low-strain muscle movements. This is because, for a given joint rotation, length change varies considerably among muscles. For example, the rabbit TA and extensor digitorum longus (EDL) have approximately the same moment arm at the ankle joint. However, because TA fibers are nearly twice as long as EDL fibers (16), EDL fiber length changes twice as much as TA fiber length for a given amount of joint rotation. Further studies of mus- cle-joint interaction are required before isolated muscle strain measurements can be extrapolated to joint angle rotations in the intact individual.

Finally, potential difficulties in interpretation of the present data should be detailed. First, it is extremely dif- ficult to define the strain that is experienced by the various experimental groups at the sarcomere level. This is because the deformation patterns are imposed on mus- cle-tendon units that are operating at different forces. Thus, in the case of the ES group where force and stiff- ness are lower, one might expect that the more compliant muscle might be strained to a greater extent than the LS group where force and stiffness are higher. However, this effect is opposed by the fact that the tendon is also more compliant in the ES group and thus tends to absorb more of the deformation. Pilot experiments in which load-de- formation curves were generated for rabbit TA tendons (as in Ref. 18) revealed large tendon stiffness differences for the two conditions. For example, at low forces similar to the ES group, tendon stiffness was 300-400 MPa, whereas at high forces similar to the LS group, stiffness was l-2 GPa. It is therefore plausible that the two com- peting effects (increased muscle stiffness at high force tending to decrease muscle strain in the LS group vs. increased tendon stiffness at high force tending to in- crease muscle strain in the LS group) might converge to make muscle strain in the two groups nearly identical. Real-time sarcomere length measurements with laser diffraction might specifically answer this question. Fi- nally, because EC-induced injury is more commonly as- sociated with the antigravity muscles (e.g., quadriceps and plantarflexors) than the pretibial flexor studied here, it may be inappropriate to generalize the current findings to all skeletal muscles. Future studies on the ankle extensors will specifically address this issue.

In summary, skeletal muscle injury after cyclic EC with the use of various deformation paradigms suggests that muscle damage is not simply a function of peak muscle force but rather is due to the magnitude of the strain experienced by the muscle during contraction. Fur- ther studies are underway to characterize the specific cellular structures affected by the EC-induced damage.

The authors acknowledge Cindy Brown, Lena Carlsson, Anna- Karin Nordlund, Abbe Zaro, Chris Giangreco, Christy Trestik, and Mary Schmitz for technical assistance.

This work was supported by the Veterans Affairs, the University of California, San Diego Academic Senate, National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AR-40050, the Research Council of the Swedish Sports Federation, the Swedish Society of Medi- cine, and the University of Ume&.

Address for reprint requests: R. L. Lieber, Dept. of Orthopedics (V-151), Univ. of California San Diego School of Medicine and Vet-

erans Affairs Medical Center, 3350 La Jolla Village Dr., San Diego, CA 92161.

Received 11 May 1992; accepted in final form 3 August 1992.

REFERENCES

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

BROOKE, M. H., AND K. K. KAISER. Muscle fiber types: how many and what kind? Arch. Neural. 23: 369-379, 1970. CURTIN, N. A., AND R. E. DAVIES. Chemical and mechanical changes during stretching of activated frog skeletal muscle. Cold Spring Harbor Symp. Quant. Biol. 37: 619-626, 1973. EISENBERG, B. R., AND R. MILTON. Muscle fiber termination at the tendon in the frog’s sartorius: a stereological study. Am. J. Anat. 171: 273-284,1984. EVANS, W. J., AND J. G. CANNON. The metabolic effects of exercise- induced muscle damage. Exercise Sport Sci. Reu. 19: W-125, 19%. FRIDBN, J., U. KJ~RELL, AND L.-E. THORNELL. Delayed muscle soreness and cytoskeletal alterations: an immunocytological study in man. Int. J. Sports Med. 5: 15-18, 1984. FRIDBN, J., AND R. L. LIEBER. The structural and mechanical basis of exercise-induced muscle injury. Med. Sci. Sport Exercise 24: 521- 530, 1992. FRID~N, J., R. L. LIEBER, AND L.-E. THORNELL. Subtle indications of muscle damage following eccentric contractions. Acta Physiol. &and. 142: 523-524,199l. GOSLOW, G., JR., R. REINKING, AND D. STUART. The cat step cycle: hind limb joint angles and muscle lengths during unrestrained lo- comotion. J. Morphol. 141: l-42, 1973. GREGOR, R. J., R. R. ROY, W. C. WHITING, R. G. LOVELY, J. A. HODGSON, AND V. R. EDGERTON. Mechanical output of the cat soleus during treadmill locomotion: in vivo vs. in situ characteris- tics. J. Biomech. 21: 721-732, 1988. HARRY, J. D., A. W. WARD, N. C. HEGLUND, D. L. MORGAN, AND T. A. MCMAHON. Cross-bridge cycling theories cannot explain high-speed lengthening behavior in frog muscle. Biophys. J. 57: 201-208,199O. HOFFER, J. A., A. A. CAPUTI, I. E. POSE, AND R. I. GRIFFITHS. Roles of muscle activity and load on the relationship between muscle spindle length and whole muscle length in the freely walking cat. Prog. Brain Res. 80: 75-85, 1989. HOROWITS, R., E. S. KEMPNER, M. E. BISHER, AND R. J. Po- DOLSKY. A physiological role for titin and nebulin in skeletal mus- cle. Nature Land. 323: 160-164, 1986. HOROWITS, R., AND R. J. PODOLSKY. The positional stability of thick filaments in activated skeletal muscle depends on sarcomere length: evidence for the role of titin filaments. J. Cell Biol. 105: 2217-2223, 1987. HUXLEY, A. F. Muscle structure and theories of contraction. Prog. Biophys. 7: 255-318, 1957. JONES, D. A., D. J. NEWHAM, J. M. ROUND, AND S. E. J. TOLFREE. Experimental human muscle damage: morphological changes in relation to other indices of damage. J. Physiol. Land. 375: 435-448, 1986. LIEBER, R. L., AND F. T. BLEVINS. Skeletal muscle architecture of the rabbit hindlimb: functional implications of muscle design. J. Morphol. 199: 93-101, 1989. LIEBER, R. L., T. M. WOODBURN, AND J. 0. FRIDBN. Muscle dam- age induced by eccentric contractions of 25% strain. J. Appl. Phys- iol. 70: 2498-2507, 1991. LIEBER, R. L., M. E. LEONARD, C. G. BROWN, AND C. L. TRESTIK. Frog semitendinosis tendon load-strain and stress-strain proper- ties during passive loading. Am. J. Physiol. 261 (Cell Physiol. 30): C86-C92,1991. LIEBER, R. L., D. E. SMITH, AND A. R. HARGENS. Real-time acqui- sition and data analysis of skeletal muscle contraction in a multi- user environment. Comput. Methods Programs Biomed. 22: 259- 265,1986. LIEBER, R. L., C. L. TRESTIK, M. C. SCHMITZ, AND J. FRIDBN. Damage following cyclic eccentric contraction is not a function of muscle stress (Abstract). Trans. 38th Ortho. Res. Sot. 38: 259,1992. MORGAN, D. L. New insights into the behavior of muscle during active lengthening. Biophys. J. 57: 209-221, 1990.

526 MUSCLE DAMAGE IS NOT DUE TO HIGH FORCE

22. PEARSE, A. G. E. Histochemistry; Theoretical, and Applied (4th ed.). cells in relation to function. Biochem. Sot. Trans. 19: 1116-1120, London: Churchill Livingston, 1961, vol. 2. 1991.

23. PETER, J.B.,R.J. BARNARD, V.R. EDGERTON,~. A. GILLESPIE, AND K. E. STEMPEL. Metabolic profiles on three fiber types of skele-

25. TIDBALL, J. G. Myotendinous junction injury in relation to junc- tion structure and molecular composition. Exercise Sport Sci. Reu.

tal muscle in guinea pigs and rabbits. Biochemistry 11: 2627-2733, 19: 419-446, 1991. 1972. 26. WEIBEL, E. R. Stereological Methods. Practical Methods for Biologi-

24. THORNELL, L.-E., AND M. G. PRICE. The cytoskeleton in muscle cal Morphometry. New York: Academic, 1980, vol. 1.