The functional role of the skeletal muscle intermediate filament system was investigated by measuring the magnitude of muscle force loss after cyclic eccentric contraction (EC) in normal and desmin null mouse extensor digitorum longus muscles. Isometric stress generated was significantly greater in wild-type (313 ± 8 kPa) compared with knockout muscles (276 ± 13 kPa) before EC (P < 0.05), but 1 h after 10 ECs, both muscle types generated identical levels of stress (∼250 kPa), suggesting less injury to the knockout. Differences in injury susceptibility were not explained by the different absolute stress levels imposed on wild-type versus knockout muscles (determined by testing older muscles) or by differences in fiber length or mechanical energy absorbed. Morphometric analysis of longitudinal electron micrographs indicated that Z disks from knockout muscles were more staggered (0.36 ± 0.03 μm) compared with wild-type muscles (0.22 ± 0.03 μm), which may indicate that the knockout cytoskeleton is more compliant. These data demonstrate that lack of the intermediate filament system decreases isometric stress production and that the desmin knockout muscle is less vulnerable to mechanical injury.
- intermediate filaments
- muscle injury
the desmin intermediate filament protein network is suggested to mechanically integrate the myofibrillar matrix in the radial and, to a lesser extent, in the longitudinal direction (7). A previous study reported that desmin knockout muscles showed greater evidence of mechanical injury and lower isometric force production relative to wild types when subjected to mechanical stress (9). Increased susceptibility to mechanical injury was also reported in the murine mdx model of muscular dystrophy (16) in which the subsarcolemmal protein dystrophin is absent. On the basis of these studies and the presumed role of desmin in normal muscle, we hypothesized that desmin knockout muscles would demonstrate a greater injury response to high-stress eccentric contractions (EC) where the muscle is forcibly lengthened while activated. This treatment could be used to probe the mechanical integrity of muscle. The fact that muscle injury occurs after cyclic EC is already well documented in a variety of experimental models (1, 12). Cyclic EC may result in loss of isometric force-generating capacity, disruption of the excitation-contraction coupling mechanism (21), and structural disruption of the myofibrillar apparatus (4). One of the earliest structural changes noted in rabbit fast muscles subjected to cyclic EC is the rapid and widespread loss of immunostaining for the muscle-specific intermediate filament protein desmin (11). Because the desmin intermediate filament protein network mechanically stabilizes myofibrils, loss of desmin due to EC may mechanically destabilize myofibrils, resulting in greater functional and morphological injury. Because the specific role of desmin in normal muscle is still not completely understood and has been implicated in a muscle injury mechanism, the purpose of this work was to quantify the magnitude of muscle injury that occurs in both wild-type and desmin knockout extensor digitorum longus (EDL) muscles subjected to cyclic ECs. We hypothesized that a lack of desmin in the desmin knockout muscles would render them more susceptible to mechanical injury compared with wild-type muscles.
Experimental model and design.
The model used for this study was the fifth toe muscle of the multibellied mouse EDL. This belly was chosen on the basis of its fiber length homogeneity, distinct origin and insertion tendons, mixed fiber-type distribution, and relatively small number of muscle fibers (3). Mice were anesthetized with a cocktail composed of (in mg/kg) 10 ketamine, 5 rompum, and 11 acepromazine given by intraperitoneal injection. All procedures were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the University of California and Veterans Affairs Committees on the Use of Animal Subjects. Muscles were dissected in a Ringer solution composed of (in mM): 137 NaCl, 5 KCl, 1 NaH2PO4, 24 NaHCO3, 2 CaCl2, 1 MgSO4, and 11 glucose containing 10 mg/l curare. Ringer solution was equilibrated with a 95% O2–5% CO2 gas introduced by bubbling, and the pH was adjusted to 7.5 at 25°C. After muscles were removed, mice were euthanized by intracardiac injection of pentobarbital sodium.
To test the role of desmin in normal and desmin knockout muscles, we obtained EDL muscles from one of the following four experimental groups: 1) young wild-type mice (n = 7, age 8–12 wk), 2) young desmin homozygous knockout mice (n = 7, age 8–12 wk) in which the desmin gene had been deleted by homologous recombination (14),3) adult wild-type mice (n = 7, age 37–41 wk), and 4) adult desmin knockout mice (n = 7, age 37–41 wk). In terms of muscle age, the 50% survival point is often considered as the onset of the senescent period (19). For mice, this point is somewhat dependent on strain and generally occurs at ∼24 mo of age. Thus the young mice examined in this experiment should be considered “adolescents,” and the adult wild-type mice should be considered “mature adults.” However, because the life span of the knockout animals is only ∼1 yr (14), the knockouts are relatively older compared with wild types despite the fact that they are chronologically age matched.
Eccentric injury protocol.
After dissection, the fifth toe muscle belly was transferred to a custom muscle chamber filled with Ringer solution that was superfused with the O2–CO2 mixture throughout the experiment. With the use of 8-0 black-braided silk suture (Deknatel, Fall River, MA), the muscle origin was secured to a rigid post and the insertion was secured to the arm of a dual-mode ergometer (model 300B; Aurora Scientific, Richmond Hill, ON, Canada) that could impose rapid length changes. At both the origin and insertion sites, the muscle was tied as closely as possible to the fiber insertions to minimize series compliance of tendons and ensure that deformation applied to the muscle-tendon unit was indeed delivered to the muscle fibers. System compliance, including the ergometer, was ∼5 μm/g. Muscle activation was provided by an electrical stimulator (model S88; Astro-Med, West Warwick, RI) via platinum plate electrodes that extended the length of the muscle. Ergometer arm movement was induced by a computer-controlled function generator (model 3314A; Hewlett-Packard, Cupertino, CA), and the entire experiment was synchronized and recorded by an acquisition program written in the LabWindows environment (National Instruments, Austin, TX) for the particular data acquisition board used (model 512; Gage, Montreal, PQ, Canada).
Muscle belly length was adjusted to a length that was barely taught (slack length), and this length was measured by using a calibrated reticule mounted onto the eyepiece of a dissecting microscope (model 8 Wild; Leitz, New York, NY). The muscle was elongated by 15% of its slack length, and fiber length was measured again. This length was used as the starting fiber length (L f) for all experiments and corresponded to a resting sarcomere length of ∼3 μm (3.02 ± 0.03 μm) as measured by laser diffraction.
Approximately 30 min after the muscle was mounted in the chamber, passive mechanical properties were measured by imposing a 10%L f stretch at a rate of ∼0.7L f/s (0.72 ± 0.011L f/s). This stretch was repeated three times at 3-min intervals. Maximum isometric tension was then measured by applying a 400-ms train of 0.3-ms pulses delivered at 100 Hz while muscle length was held constant. This measurement was repeated twice at 10-min intervals. Each muscle then underwent a series of 10 ECs, one every 3 min, to minimize the effects of fatigue. For each EC, the muscle was first activated isometrically until tension stabilized (∼200 ms), and then a 15% L f change was imposed at a rate of 2 L f/s, resulting in a rapid tension rise (Fig.1). Tension increased in two phases, a reflection of the muscle short-range stiffness (5) and low system compliance. Muscle length was held fixed for a time during which tension declined due to active stress relaxation. Stimulation was then ceased and muscle length was returned to its starting value.
Isometric contractile experiments (n = 3 experimental subjects per group), in which no length change was imposed during stimulation, were performed on young wild-type and desmin knockout muscles to control for the effects of cyclic activation alone. Stimulation parameters and duration were identical to those described above. Passive stretch experiments (n = 3 experimental subjects per group), in which no activation was performed but in which length perturbations identical to those described above were included, were imposed on muscles to control for the effects of length change alone. These control experiments were performed because, whereas isometric contraction or passive stretch alone has no deleterious effects on normal muscle, no such data were available for knockout muscle samples. Fifteen minutes after the last EC, maximum isometric tension was measured five times every 15 min. Finally, passive properties were again measured by imposing a 10%L f stretch at a rate of ∼0.7L f/s, a step that was repeated three times at 3-min intervals.
Tissue preparation for electron microscopy.
Three muscles from each experimental group were randomly selected for electron microscopy. After the experiment was completed, muscles were pinned on cork along with their contralateral counterparts and submerged into phosphate-buffered Karnovsky's fixative, where they were allowed to incubate overnight at 4°C. Specimens were washed three times in cold (4°C) sodium cacodylate buffer (0.1 M adjusted to pH 7) and then incubated in 2% osmium tetroxide for 1 h at room temperature. After three buffer washes of 5 min each, muscles were dehydrated in graded ethanols and propylene oxide. Specimens were then cut into approximately equal pieces and oriented to obtain longitudinal views embedded in Scipoxy 812 Resin (Energy Beam Sciences, Agawam, MA). Micrographs were photographed from longitudinal sections by using a diagonal random sampling algorithm with a grid-fiber orientation angle of ∼40° (22). The end points of all Z disks from each of 10 micrographs from each muscle were digitized between adjacent myofibrils to quantify Z-disk angle as well as the horizontal displacement (Δx) of adjacent Z disks. The 10 micrographs from each muscle were obtained in different regions of different sections and were thus presumably from different muscle fibers.
A computer algorithm was written to process the force-time records from each contraction of each muscle specimen (version 40, Mathematica; Wolfram Research, Champaign, IL). This algorithm used a combination of linear regression, breakpoint analysis, and the known timings of activation and deformation to yield the following mechanical parameters (where applicable): initial baseline force, maximum isometric tetanic tension, peak eccentric tension, mechanical energy absorbed (phase II slope), and yield point during EC (as defined in Ref. 23) (Fig.1). For isometric and passive stretch experiments, baseline and peak force were the only relevant parameters obtained from the force-time record. Muscle stress was calculated by normalizing muscle force to muscle physiological cross-sectional area (PCSA), calculated by using the equations provided by Sacks and Roy (17) and assuming a muscle density of 1.056 g/cm3 (13). PCSA was calculated individually for each muscle on the basis of dissection of small fiber bundles from fixed tissue as previously reported (2). Mechanical energy absorbed was calculated according to the equation where ɛ is energy absorbed (in joules), Fi(t) is muscle force as a function of time for the ith contraction,Vi (t) is muscle velocity as a function of time for the ith contraction (which was 0 everywhere but during the stretch), t 0 is the start of the contraction, and t 1 is the end of the contraction.
All data were screened for normality (skew ≤ ‖0.7‖, kurtosis ≤ ‖1.5‖) to justify the use of parametric statistics. With the use of these criteria, no mathematical transformations were required for parametric analysis of any of the contractile parameters. However, because the morphological data on horizontal displacement (Δx) of adjacent Z disks were highly skewed (skew > 1.4), these data were log transformed before parametric analysis by one-way ANOVA.
For simultaneous comparison of physiological and contractile parameters across muscle type (wild type vs. desmin knockout) and age (young vs. mature), two-way ANOVA was used with a significance level (α) set to 0.05. For comparison of contractile parameters with time, repeated-measures ANOVA was implemented. Results in the text are presented as means ± SE unless otherwise specified.
Animal and muscle masses varied by age and muscle type.
Animal and muscle masses were significantly lower in desmin knockout compared with wild-type specimens for both age groups (P < 0.05, Table 1). Two-way ANOVA revealed a significant effect of type (P< 0.01) and age (P < 0.05) on animal mass with no significant interaction (P > 0.3). Two-way ANOVA also revealed a significant effect of type (P < 0.0001) but no significant effect of age (P > 0.2) on muscle mass with no significant type × age interaction (P > 0.5). These measurements suggest that lack of desmin is a greater determinant of animal and muscle mass than is age and that the age effect on these properties is independent of the muscle type. It is important, however, to consider the difference in muscle mass, and thus PCSA, as a normalizing factor when comparing muscles across types. Architectural analysis of the muscles revealed a significant decrease in fiber length and muscle PCSA in knockout compared with wild-type muscles (P < 0.02) but no significant effect of age (P > 0.5) or age × type interaction (P > 0.3; Table 1).
Isometric stress differences among groups before EC treatment.
Absolute isometric tension generated before EC treatment was significantly different between mature and young specimens (P < 0.0001) and wild-type and knockout specimens (P < 0.0001) with no significant interaction (P > 0.5). Isometric force before EC treatment was greatest in the young wild-type EDLs (14.6 ± 0.4 g), followed by that in the mature wild-type EDLs (13.4 ± 0.9 g), young knockouts (11.0 ± 0.6 g), and mature knockouts (9.4 ± 0.5 g). When data were normalized to stress, young wild-type EDLs still generated the greatest stress (313 ± 8 kPa) and mature knockouts generated the lowest stress (227 ± 17 kPa), but mature wild types and young knockouts generated isometric stresses that were not significantly different (281 ± 13 and 276 ± 13 kPa, respectively, P > 0.4). This fortuitous result permitted testing of the effects of stress on muscle injury independent of the presence of intermediate filaments. No change in isometric force-generating capacity or peak passive stress was observed for control muscles undergoing cyclic isometric activation or cyclic passive stretch (data not shown).
Unique time course of isometric stress for each experimental group.
Isometric stress measured just before each EC demonstrated a unique time course as a function of muscle type and age (Fig.2 A). All four experimental groups significantly decreased isometric stress-generating capacity as a function of time (P < 0.001), but the magnitude of the decrease was characteristic of each muscle age and type. For example, whereas mature wild-type and mature knockout muscles generated isometric stresses that were significantly different before eccentric exercise (Fig. 2 A), after 10 ECs, the isometric stress generated by the two groups was nearly identical (184 ± 11 and 186 ± 21 kPa, respectively, P > 0.8). Similarly, whereas young wild-type and young knockout muscles also generated isometric stresses that were significantly different before eccentric exercise (Fig. 2 A), after 10 ECs, the isometric stress generated by these two groups was also nearly identical (246 ± 6 and 252 ± 8 kPa, respectively, P > 0.8), and both were significantly greater than those of their mature counterparts. Thus the presence of the intermediate filament system significantly affected maximum isometric stress before any EC treatment, but after 10 ECs, no difference in isometric stress generated was seen among muscle types, a finding that may reflect preferential damage to the wild-type intermediate filament system. This effect was independent of the generally increased vulnerability of the aged muscles from both types. These data provide strong support for the physiological importance of the intermediate filament network for normal muscle force generation and also in response to injury. No sign of recovery after the end of the eccentric period was seen, because only a small transient isometric force change after EC was observed over a 75-min recovery period (Fig. 2 A). Cyclic isometric contraction alone had no effect on either the wild-type or desmin knockout muscle stress (Fig. 2 B), demonstrating that the force changes observed were not confounded by differences in fatigability among muscle types. We thus attribute the force changes measured during EC to impairment of the contractile apparatus rather than simple consequences of metabolic fatigue.
Based on the unique behavior of each experimental group before and after EC treatment, the drop in maximum isometric stress was very characteristic for each group. The largest decrease in isometric stress was observed for the mature wild-type muscles, which generated moderate stress before EC treatment but generated only 66 ± 5% of their initial stress after ECs (Fig.3 A). At the other extreme were the young knockout muscles, which also generated moderate stress before EC treatment and still generated 91 ± 2% of this stress after 10 ECs. The young wild types and mature knockouts demonstrated intermediate results (79 ± 1% and 81 ± 5%, respectively). However, in all cases, the desmin knockout muscles demonstrated significantly greater relative force-generating capacity after ECs compared with their wild-type counterparts.
Passive mechanical properties change after EC treatment.
Whereas no significant differences in passive stress were observed between wild-type and knockout specimens (P > 0.2), a significant difference was observed among specimens as a function of age (P < 0.0005) and timing relative to EC treatment (P < 0.01; Fig. 3 B). Specifically, passive stress decreased significantly for the older specimens, and there was a tendency for a passive stress decrease after EC treatment in the young knockout specimens. These results were not simply due to stress-relaxation of the tissues involved because no such passive stress decreases were observed in muscles subjected to either cyclic isometric contraction or cyclic passive stress. In addition, no correlation was observed between peak eccentric stress and change in passive stress (P > 0.6,r 2 = 0.11) after EC treatment, which would be expected if stress-relaxation were a significant factor.
Energy absorbed during EC treatment does not explain knockout effect.
The decreased vulnerability of knockout muscles compared with wild types could not be explained on the basis of differences in energy absorbed during EC treatment (calculated as described inmethods; Table 2). For both mature and young specimens, as expected on the basis of simple size differences, the wild-type muscles absorbed more absolute energy than their knockout counterparts because the absolute force levels were higher. However, when energy was expressed as energy absorbed per unit muscle mass, there was no significant difference between energy absorbed by wild-type and knockout specimens of matched ages, with the younger muscles absorbing ∼600 J/kg and the mature muscles absorbing ∼500 J/kg (Table 2).
Passive stretch or activation alone does not explain differential effect.
As a control for the effect of activation or stretch alone, isometric contraction and passive stretch groups were also studied (n = 3 experimental subjects per group). Because the drop in muscle force can often be explained, at least in part, by muscle fatigue (10), fatigability due to repetitive activation was measured across the same time period but with 10 isometric contractions instead of ECs. Average maximum tetanic tension after 10 isometric contractions was 99.4 ± 0.9% of the pre-isometric contraction level in knockout muscles and 99.7 ± 0.3% in wild-type muscles (Fig. 2 B). Similarly, the 10-passive contraction protocol resulted in an average maximum tetanic tension of 100.5 ± 0.9% of the pre-isometric contraction level in knockout muscles and 99.8 ± 0.6% in wild-type muscles. None of these values were significantly different from 100% (P > 0.4). Because force decline did not result from either passive stretch or isometric contraction, the effects observed were not confounded by potential differences in passive properties or metabolic fatigability of muscles from knockout animals.
Morphological appearance of Z disks.
Displacement of adjacent Z disks was significantly greater in the injured knockout muscles (0.36 ± 0.03 μm) compared with wild-type muscles (0.22 ± 0.03 μm, P < 0.05), and a significantly greater proportion of Z disks was slanted >30° in wild-type (10.4 ± 1.5%) compared with knockout muscles (5.8 ± 0.9%, P < 0.05; Fig.4).
The intermediate filament system of skeletal muscle is believed to be responsible for the mechanical integration of the myofibrillar lattice in both the longitudinal and radial directions (7). The only mechanical studies available on this system come from partial extraction experiments of rabbit skeletal muscle demonstrating that longitudinal intermediate filaments bear significant passive tension only at sarcomere lengths above ∼5 μm (20). Because this was not in the physiological range, the finding seemed to imply that the intermediate filament network was not important for bearing normal loads in skeletal muscle. However, we have demonstrated that the intermediate filament system plays a significant role in the development of isometric stress under physiological conditions, independent of mechanisms that may be related to maturation of the system. In addition, the lack of intermediate filaments actually rendered the mouse EDL muscles less susceptible to mechanical disruption, a result in contrast to previous reports in another desmin knockout system (9) and to reports from studies of the naturally occurring mutant in which the subsarcolemmal cytoskeletal protein dystrophin is absent (16). Thus the simple lack of a particular cytoskeletal protein should not be considered an a priori reason to expect increased mechanical vulnerability to injury.
Stereological analysis of wild-type and knockout muscle revealed no significant difference in volume fraction of contractile filaments, which could have provided an anatomic basis for the differences in isometric stress development. Longitudinal images of knockout muscle typically revealed a striking regularity of myofibrillar structure along the muscle fiber length in noninjured muscles. This is in contrast to previous micrographs demonstrating muscle derangement in the diaphragm, heart, and soleus muscles from desmin knockout animals (8, 18). We conjecture that disparity among these findings represents true differences related to the variability in the amount and type of use experienced by the mouse diaphragm, heart, and soleus muscles compared with the less frequently used EDL muscle studied here.
Morphological measurements demonstrating increased Z-disk displacement and a decreased percentage of slanted Z disks support the mechanical concept that adjacent myofibrils in knockout muscles were allowed to “slide” relative to one another in response to the ECs, perhaps placing less stress on adjacent myofibrils. Conversely, the increased stress of adjacent desmin-based interconnections may be responsible for the increased area fraction of slanted disks in the wild-type muscles. This finding demonstrates that the muscle intermediate filament system does not protect a fiber from muscle to injury but may even permit the injury and associated damage to be more severe (Fig. 4).
A previous report of desmin playing a critical role in maintaining muscle tensile strength and integrity (9) was partly confounded on the basis of the significant size difference between muscles that occurs secondary to the homologous recombination itself (Table 1). The absolute strength differences reported can easily be explained on the basis of size differences between the knockout and wild-type muscles. These limitations were overcome in the current study by explicitly including muscle dimensions in all stress calculations and matching stresses among groups by using mature muscles in which isometric stress drops due to the aging process itself. The decreased vulnerability of knockout muscles to mechanical injury could not be explained on the basis of simple size differences because mechanical energy absorbed was the same between wild-type and knockout muscles (Table 2). In addition, when muscles of similar stress-generating capacity (young knockout vs. mature wild type, Fig. 2 A) were compared, injury was much greater to the wild-type muscle compared with the knockout muscle. Support for the concept that stress and injury are not associated is provided by the finding that there was no significant correlation between peak muscle stress and percentage drop in maximum tetanic tension after EC (r 2 = 0.0005,P > 0.7).
Whereas a significant change in passive stress was observed after EC treatment for all but the young wild-type experimental groups (Fig.2 B), the underlying basis for this result is not clear. A trivial explanation for passive stress differences could be that the resting sarcomere lengths of the experimental groups were different as a function of age and type. Thus passive stress differences would simply be due to muscle fibers operating on different regions of their sarcomere length-passive stress relations. However, this was not the case. One-way ANOVA among the four experimental groups revealed no significant difference among groups in resting sarcomere length (Table1; P > 0.5). In fact, resting sarcomere among all groups demonstrated a coefficient of variation of only 7.5%, probably indicating similar splice variants of the intrasarcomeric protein titin (6). The structural basis for the selective decrease in passive force in three of the four experimental groups probably represents a complex interaction among the titin protein, other cytoskeletal proteins, and extracellular collagen, all of which may change in skeletal muscle as a function of age and in response to creation of the desmin null experimental model. Detailed reorganization of such proteins has not been studied in great detail.
On the basis of these findings, we hypothesize that normal interconnections between sarcomeres along and across the muscle fiber permit efficient transfer of mechanical stress from the myofibril to the fiber exterior. In the absence of desmin, the intermyofibrillar connections may be more compliant, permitting a greater degree of internal sarcomere shortening and motion and thereby decreasing the efficiency of force transfer across and along the fiber. Similarly, on the basis of the lack of these connections, the opposite situation, in which strain is transmitted from the fiber exterior to the myofibrils, could also be less efficient. The result would be a decreased sarcomere strain during a fixed degree of lengthening in knockout compared with wild-type muscles and, thus, less injury despite the fact that the energy delivered to the desmin knockout muscle is identical. This particular argument is predicated on the observation that sarcomere strain is directly related to muscle injury, whereas fiber stress is not. We recently reported real-time sarcomere length measurements in frog skeletal muscle using a paradigm similar to that used here and demonstrated a strong correlation between sarcomere strain and muscle injury with no significant correlation between muscle fiber stress and injury (15).
Of course, an alternative interpretation of these data is that the muscles from knockout animals were already injured simply due to normal animal activity. If this were the case, the decreased stress generated by muscles from knockout animals would reflect their injury, and the lack of further decrease would reflect their already injured state. However, we do not believe that this was the case because muscles from knockout animals did not demonstrate morphological abnormality or disruption that would imply prior injury.
Finally, these data have implications for understanding the mechanism of muscle injury in skeletal muscle. The fact that the age-matched muscles generated the same amount of stress after the injury protocol, independent of the presence or absence of the intermediate filament system, provides support for the selective functional loss of desmin in wild-type muscle and its functional importance. At the end of the EC protocol, the younger muscles generate significantly greater stress compared with the mature muscles (∼250 vs. ∼200 kPa), but there was no difference between knockout and wild-type muscles. These data support the proposal that the time course of loss in desmin immunostaining previously observed (11) would have functional consequences and might even provide a protective effect on the muscle to further mechanical injury. Further studies are required to resolve the precise mechanical events that accompany EC and the loss of stress in wild-type and knockout muscles.
This work was supported by the Department of Veterans Affairs, National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AR-40050, and Swedish Medical Research Council Grant 11200.
Address for reprint requests and other correspondence: R. L. Lieber, Dept. of Orthopaedics (9151), Univ. of California, San Diego School of Medicine and Veterans Affairs Medical Center, San Diego, CA 92161 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society