INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

FORCED LENGTHENING OF ACTIVATED skeletal muscles [i.e., eccentric contraction (EC)] results in injury to both animal and human muscles (2, 3, 22) and causes subjective muscle soreness in humans (1, 4, 7). Numerous investigators have also demonstrated that even a single EC exercise bout provides a protective effect against subsequent EC-induced muscle injury (26, 29). However, neither the cellular mechanism of injury nor the adaptive response that protects muscle against further injury is fully understood. Determination of both of these mechanisms will provide a better understanding of muscle structure and plasticity and enable development of rational exercise and therapeutic interventions needed to strengthen muscle.

Since the demonstration, over 20 years ago, that myofibrillar disruption was associated with EC (10), several other reports of cellular disruption have led to varying hypotheses regarding the mechanism of EC-induced muscle injury. Of course, not all of these hypotheses are mutually exclusive. For example, the demonstration of increased membrane permeability in both wild-type (6, 9, 23) and genetically altered (20, 28) muscles subjected to EC suggests that membrane breakdown is a consequence of EC-induced injury. Yet in a rabbit model, cytoskeletal disruption, evidenced by the loss of immunostaining of the intermediate filament protein desmin, preceded increased membrane permeability (17), suggesting that the primary disruption is intracellular and that membrane disruption follows. The fact that cytoskeletal disruption occurred as fast as 5 min after initiation of EC (18) suggests that direct mechanical or biochemical events are responsible for this disruption rather than events that require gene regulation. One such mechanical mechanism invokes the rapid, undamped elongation of sarcomeres (i.e., sarcomere "popping" as described in Ref. 24) that may lead to other cellular events such as sarcotubular perforation and, ultimately, morphological disruption (25, 29). Alternatively, we proposed a proteolytic mechanism whereby sarcomere strain (16) resulted in loss of calcium homeostasis and activation of the intracellular calcium-activated neutral proteases known as the calpains (8). Activation of these enzymes is already implicated in exercise (5) and muscular dystrophy (31, 32), but their role, if any, in EC-induced muscle injury remains unknown.

To identify the dynamics of desmin after EC-induced injury, including potential proteolysis by calpain, we performed quantitative analysis of immunoblots of muscle 6, 12, 24, 48, 72, 120, and 168 h after a single EC bout. While no such proteolytic fragments were observed, we discovered a tremendous increase in the relative amount of desmin present in the muscle cells 72-168 h after the EC bout. Based on the relatively small fraction of muscle fibers injured, we suggest that desmin content in individual cells may have increased as much as 15-fold. We further suggest that this may provide a structural basis for cellular remodeling of the myofibrillar complex and protection from subsequent EC-induced muscle injury.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animal care. Experimental subjects for this study were untrained male Sprague-Dawley rats (Harlan, Indianapolis, IN) with an average mass of 435.2 ± 3.8 g (n = 34). Rats were housed two or three per cage at 20-23°C with a 12:12-h dark-light cycle. All procedures were approved by the University of California and Veterans Affairs Medical Center Committees on the Use of Animal Subjects in Research. Animals were anesthetized with 2% isoflurane during the eccentric exercise, and after terminal experiments, animals were euthanized with an intracardiac injection of pentobarbital sodium (0.5 ml of 390 mg/ml solution).

ECs. A more detailed description of the experimental protocol can be found elsewhere (27). Briefly, dorsiflexors were subjected to a single exercise bout consisting of 30 ECs delivered at 2-min intervals by stimulating the peroneal nerve for 650 ms at 100 Hz with a small nerve cuff (subminiature electrode, Harvard Apparatus, Holliston, MA). After achieving a fused tetanic contraction, 38° of ankle plantarflexion (corresponding to a fiber strain of ~10%) was imposed on the ankle using a dual-mode servomotor (model 6650, Cambridge Technologies, Cambridge, MA). Ankle isometric torque, which is dominated by the contractile properties of the tibialis anterior (TA), was measured before eccentric exercise, immediately after eccentric exercise, and after a specified recovery time of 6 (n = 4), 12 (n = 4), 24 (n = 5), 48 (n = 6), 72 (n = 4), 120 (n = 4), or 168 h (n = 5) (27). The contralateral leg served as the control and was only tested for isometric torque. Animals were killed after the final torque measurement, and the TA muscle was excised, immediately frozen in liquid nitrogen-cooled isopentane (159°C), and stored at 80°C for further processing.

Immunohistochemistry. Muscle cross sections (10-µm thick) taken from the midbelly of the TA were stained with monoclonal antibodies to desmin (NCL-DER11, 1:1,000, Novocastra, Vector Laboratories, Burlingame, CA and D33, 1:1,000, Dako, Carpenteria, CA) and visualized using the indirect immunoperoxidase technique (Vectastain Elite ABC Kit, Vector Laboratories). Sections were viewed at ×25 magnification, and negative fibers were counted directly. Fibers were considered desmin negative if the outline of the fiber could still be seen but little or no internal staining occurred. Qualitatively, the lack of staining was consistent with previous reports, and fibers that were negative for desmin still demonstrated positive staining for myofibrillar proteins (17, 18).

SDS-PAGE and immunoblots. A 25-µm-thick cross section, serial to that used for immunohistochemistry, was immediately extracted on ice in 100 µl SDS-PAGE sample buffer (100 mM dithiothreitol, 2% SDS, 80 mM Tris base, 10% glycerol, and 0.012% wt/vol bromophenol blue; 300 µl) for at least 30 min and then boiled for 2 min. (The fact that the section used for SDS-PAGE and immunohistochemistry was serial ensured that the same muscle region was studied using both techniques.) Five microliters of a 1:3 dilution of each sample were loaded onto the gel, along with two lanes of desmin standards on opposite sides of the gel (0.013 µg/lane; catalog no. RDI-PRO62005, Research Diagnostics, Flanders, NJ) and two lanes of actin standards on opposite sides of the gel (0.525 µg/lane; product no. A2522, Sigma).

Analysis. Quantitative analysis of the amount of desmin normalized to actin in each sample was performed on the immunoblots using densitometry and NIH Image software (version 1.62). To account for variations in the amount of protein loaded per lane as well as the generally unknown relationship between antibody concentration, affinity, and band optical density, the intensity of each band was calibrated based on the mean of the intensity of the actin or desmin standards on either side of the gel. The mass ratio of desmin to actin was then calculated using these values for each sample, and each sample was run in triplicate on separate gels.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Mechanics. Dorsiflexion torque decreased from an initial value of 0.052 ± 0.001 to 0.030 ± 0.001 N · m immediately after the eccentric bout and increased until full recovery 168 h after the exercise (Fig. 1). Control muscles receiving 30 isometric activations demonstrated no torque decline at any point either during stimulation or up to 2 days later. Both results demonstrate that muscle injury, not fatigue, caused the force loss (27). It is possible that a small portion of the torque change observed resulted from slight shifts in muscle optimal length, which has been reported after eccentric exercise (33).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Dorsiflexion torque after a single eccentric contraction (EC) bout. A: representative EC torque records from the first and last in a series of 30 ECs separated by 2 min: torque record (top) and representation of the rotation imposed on ankle joint (bottom). Muscles are returned to resting length soon after termination of muscle stimulation (not shown). Hatched bar represents time during which muscle was stimulated. B: quantification of maximum isometric dorsiflexion torque measured before (pre-EC), immediately after (post-EC), and at various time periods after the EC bout. * P < 0.02, *** P < 0.0001.

Immunohistochemistry. Desmin immunostaining showed a pattern consistent with previous reports, with desmin loss occurring at the earliest time point measured (6 h) and returning to control levels shortly thereafter (17, 18). Control muscles showed complete desmin staining throughout the muscle cross section of all fibers. Within 6 h after the EC bout, a significant number of fibers had lost desmin staining (P < 0.01), reaching a maximum number of negative fibers at 12 h (255 ± 113 fibers, P < 0.0001) and returning to control levels within 72 h (Fig. 2). On the basis of an average number of fibers per muscle of ~10,000 (14), this suggests that <5% of the fibers were injured to the extent that they demonstrated loss of desmin immunolabeling. Immunostaining results were qualitatively similar regardless of which anti-desmin monoclonal antibody (NCL-DER11 or D33) was used as the primary antibody.

View larger version (94K): [in this window] [in a new window]   Fig. 2.   Loss of desmin immunostaining after EC. A: control muscle showing all fibers staining positively for desmin. B: sample muscle 12 h after an EC bout showing that some fibers have lost desmin immunostaining (representative fibers denoted with arrows). C: no. of fibers that have lost desmin immunostaining as a function of time after EC. Total no. of fibers in the rat tibialis anterior (TA) is ~10,000. * P < 0.01, *** P < 0.0001.

Immunoblots. No desmin proteolytic fragments were visible on any gel at any time point after the EC bout (Fig. 3A). On overexposure of the film, no smear was observed at any location on the gel, and all time points showed similar nonspecific bands.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Immunoblots of rat TA muscles. A: sample blot showing desmin (top band) and actin (bottom band) at select time periods. Blots were incubated simultaneously with primary monoclonal antibodies to desmin and actin and therefore show bands for both desmin and actin. Each lane represents either a desmin (D) or actin (A) standard, a control muscle sample (C), or a representative sample from each time point after EC. No proteolytic fragments are visible. Also shown are molecular mass markers. B: quantitative demonstration of desmin-to-actin mass ratio determined using densitometry. Each sample was run in triplicate. * P < 0.01, ** P < 0.001, *** P < 0.0001.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The purpose of this study was to quantify alterations in desmin immunostaining and content within muscle cells after a single EC bout. Previous studies had implicated desmin remodeling in the injury process (18, 30), and thus our primary interest was in testing the hypothesis that EC resulted in desmin proteolysis and, therefore, loss of immunostaining. We found no evidence for desmin proteolysis but did find a decrease in desmin immediately after the EC (as demonstrated by immunostaining) followed by a significant threefold increase in muscle desmin content (as demonstrated by immunoblot analysis). As desmin has only been found in muscle cells and other myogenic precursors (15), it is unlikely that the desmin increase came from inflammatory or other infiltrating cells. We suggest that the desmin increase may be involved in protecting muscle from subsequent EC-induced muscle injury.

These studies are consistent with our previous report of the rapid and transient loss of immunostaining in rabbit skeletal muscle after a single EC bout (18). The EC paradigm performed here on the rat dorsiflexors was similar to that performed previously on the rabbit dorsiflexors except that, in the rabbit model, a much greater number of contractions (960) was imposed on the dorsiflexors compared with only 30 in the current study. Thus it is unlikely that metabolic fatigue, present in our rabbit model, was directly or indirectly involved in the loss of immunostaining observed in this study.

The basis for the loss of desmin immunostaining soon after ECs is not clear. The loss of a positive antibody reaction may indicate true loss of the desmin molecule or simply loss of desmin antigenicity. We observed the same loss of desmin immunostaining using two different desmin monoclonal antibodies with epitopes in different regions, so that if epitope loss explains the result, the changes in the desmin protein appear to occur on a scale large enough to include both epitopes. It does not appear that the desmin intermediate filament system is subjected to proteolysis, as no desmin fragments were observed on any of the immunoblots, even after heavily overexposing the blots. We previously hypothesized that the rapid loss of desmin immunostaining implicated the calpain enzyme system (18), but on the basis of the activity of these neutral proteases (8), we would have expected to observe discrete fragments at various locations on the immunoblot itself. Finally, it is possible that the desmin intermediate filament system demonstrated loss of immunostaining based on a global change within the filament system itself. Such a large-scale change could be the result of depolymerization due to desmin phosphorylation (12) or the result of covalent modification such as ADP-ribosylation (11), both of which have been observed in other muscle systems. Further experiments are required to define the precise structural basis for the loss of desmin immunostaining observed.

It should not be surprising that the qualitative immunohistochemistry and the quantitative immunoblots do not show the same result in the time points immediately after EC. Clearly, the loss of immunostaining represents a significant event at the protein level but one that is not simply reflected in proteolytic fragments on an immunoblot with a corresponding decrease in band intensity. The relatively small number of fibers (~250) that lost their immunogenicity would represent only ~3% loss in band intensity on the immunoblot, which would not be detectable on the immunoblot. Moreover, even though the muscle sections analyzed for immunohistochemistry and for immunoblots were serial to one another (so that each method analyzed the same muscle area), immunostaining was analyzed on only a single 10-µm cross section, whereas the immunoblot was taken from a 25-µm section. It remains to be determined whether the entire fiber would show a negative immunostain for desmin, or if desmin loss is more localized. It is also possible that desmin was disassembled from the intermediate filament network and did not immunostain but was retained within the cell. Thus the lack of detection of a desmin decrease on the immunoblots compared with the clear loss of immunostaining in sections at 72 h is best explained by differences in the sensitivities of the two techniques rather than true biological differences.

The increased desmin content observed on immunoblots at the later time points was not reflected as an increase in desmin immunostaining intensity. This is probably due to the fact that the secondary labeling method used here (as opposed to the immunoblots) is notoriously nonlinear. Indeed, in a recent study of the myosin heavy chain in single muscle cells, significant immunostaining of single cells was observed even in the near absence of a myosin heavy chain band on SDS gels (19). It should thus not be surprising that data obtained using the two methods are not completely congruous.

The most striking result of this study was the almost threefold increase in desmin content in relation to actin 168 h after a single EC bout. Such an increase is even more dramatic in light of the fact that we estimate that no more than ~20% of the muscle fibers within the cross section were actually injured based on either loss of desmin immunostaining (Fig. 2), inclusion of plasma fibronectin, expression of embryonic myosin, or cellular infiltration by macrophages (27). If only 20% of the fibers were injured and forced to remodel their intermediate filament lattice, they would have to increase desmin content by ~15-fold to increase the overall muscle desmin content by threefold. Such an increase would represent a profound remodeling of the intermediate filament system. It is tempting to hypothesize that the increase would serve as a protective mechanism against subsequent EC-induced injury. Newly synthesized desmin could serve to reinforce existing sarcomeres or be added to newly synthesized sarcomeres (21). The muscle cell's intermediate filament system is well developed in the radial direction, providing stout interconnections among Z-disks of adjacent myofibrils (15), but it is poorly developed in the longitudinal direction, only weakly interconnecting serial Z-disks within a myofibril (34). It is possible that the majority of the newly synthesized desmin is integrated into the longitudinal intermediate filament system to provide mechanical reinforcement against excessive sarcomere strain (16) or is upregulated to maintain sarcomere length homogeneity (24) that could also prevent injury. Future studies are required to determine the specific structural and functional role of the newly synthesized desmin protein.