Journal Logo


Human Muscle Proteome Modifications after Acute or Repeated Eccentric Exercises


Author Information
Medicine & Science in Sports & Exercise: December 2011 - Volume 43 - Issue 12 - p 2281-2296
doi: 10.1249/MSS.0b013e318222edf3

It is well established that unaccustomed or intense exercise, especially if eccentric (lengthening) muscle actions are involved, frequently induces delayed-onset muscle soreness (DOMS). Muscle pain or discomfort generally develops 24 h after exercise, peaks at 48 h, and can subsist for 5–7 d after muscular exercise (14,21,36). Besides these nociceptive sensations, active lengthening contractions results in a range of symptoms that can be functionally disabling such as prolonged impairment of muscle function, swelling, and reduced range of motion (4). Several microscopic studies (22,39,40) have reported the presence of morphological damages after eccentric protocols. Muscle injuries led to the leakage of some intramuscular proteins into blood circulation in such a way that released creatine kinase (CK) is commonly used as biological markers of exercise-induced skeletal muscle damage (3,7,30). While there is no doubt that muscle soreness and muscle and/or connective tissue damage are induced by eccentric exercise, these consequences of unusually intense lengthening muscle actions may not be causally related (32).

An intriguing phenomenon is the skeletal muscle adaptation to repeated bouts of eccentric exercises. When a muscle has been “conditioned” by prior eccentric contractions, the magnitude of changes in muscle damage markers after later bouts of eccentric contractions is attenuated. This muscle adaptation process, commonly called “the repeated bout effect,” is an exciting research area because it demonstrates the remarkable long-term resiliency as well as adaptability of skeletal muscle to merely one intense bout of exercise. Although several studies showed a reduction of the symptoms resulting from the second of two successive maximal eccentric bouts (4,5,25,26), Croisier et al. (8) demonstrated that training with submaximal eccentric exercises that progressively intensified could represent the most useful preventive strategy. In their study, subjects were exposed to two maximal eccentric bouts of the knee extensor and flexor muscles separated by 3 wk, with five submaximal eccentric training sessions in between for half of the volunteers. A significant decrease of CK and muscle soreness after the second maximal eccentric bout was only reported for the trained subjects. McHugh (26) reviewed studies investigating the repeated bout effect and described potential adaptations categorized as neural, mechanical, and cellular theories. Other theories include adaptations in excitation–contraction coupling and reduced inflammatory response. The author concluded that a unified theory explaining the mechanism(s) for this protective adaptation remains elusive, as it is probably multifactorial.

Despite considerable amount of data available on this topic at the clinical and histological levels, the molecular and cellular mechanisms underlying the DOMS phenomenon and the adaptative response that protects muscle against further injury are not known. Therefore, in an attempt to improve the understanding of these mechanisms, we have applied a proteomic analysis approach that enables us to study, without a preestablished hypothesis, the expression of thousands of proteins. During the last few years, the number of skeletal muscle biology studies using proteomic techniques has been increasing (9–11). In particular, one recent proteomic study investigated the effects of exercise training on human skeletal muscle (19). In the present study, we applied the two-dimensional difference gel electrophoresis technique on human muscle samples obtained in different conditions of eccentric exercise to investigate the molecular and cellular consequences of strenuous eccentric physical exercises and of a specific training aimed to prevent DOMS.


Experimental Design


Twelve sedentary or moderately active healthy male volunteers completed the study. All participants read and signed a written informed consent document approved by the Ethics Committee of the University Hospital Centre of Liège. They had no personal history of lower limb injury, and none was involved in competition sports; they were requested not to perform any vigorous physical activity or unusual exercise and to be free from any form of medication during the study period. Six participants assigned to the untrained group (mean age = 24.2 ± 1.2 yr; mean body mass = 71.8 ± 4.8 kg) followed the “untrained group” exercise protocol and the six other participants assigned to the trained group (mean age = 21.8 ± 0.5 yr; mean body mass = 72.7 ± 2.4 kg) followed the “trained group” exercise protocol.

“Untrained group” exercise protocol

Eccentric exercise bouts were performed on an isokinetic dynamometer (Cybex Norm, Medway, MA) at 60°·s−1 angular velocity throughout a constant range of motion (100° of flexion from the maximal active extension). During these experiments, the participants were placed in a supine position to induce a maximal lengthening of the rectus femoris, the only biarticular muscle of the knee extensor muscles, as muscle damage increases at longer muscle lengths (36). Volunteers were strapped across the distal extremity of the femur to avoid compensation from the hip joint. Subjects were submitted to two identical bouts (named Test1 and Test2) separated by a 6-wk interval. A standardized warm-up protocol consisting of 10 min of cycling on a bicycle ergometer and stretching of the quadriceps was applied before each isokinetic test. After familiarization with quadriceps muscles contraction on the isokinetic device, the subjects performed three sets of 30 maximal eccentric contractions of the quadriceps muscles of the right leg. Each set was separated by 30 s of rest. During each contraction, the subjects received spoken encouragement to produce their maximal force, and the examiner checked the temporal evolution of the isokinetic curves on the screen to ensure that the subjects worked at maximal intensity. Peak torque and work parameters were recorded by the dynamometer.

“Trained group” exercise protocol

This protocol was applied to study the effects of a specific training program known to confer a protective effect against eccentric exercise–induced muscle damage. The six volunteers assigned to the trained group performed two isokinetic tests, which consisted of three sets of 30 maximal eccentric contractions by the quadriceps muscles of the right leg, as described. The two bouts were 6 wk apart, and during the last 2 wk of this interval, volunteers underwent five training sessions with a rest period of 48–72 h between each session. The training session consisted of a standardized warm-up protocol followed by five sets of 10 submaximal isokinetic eccentric contractions of the quadriceps muscles separated by 1-min rest phases. The first training session was performed at 50% of the individual peak torque developed during the first eccentric test. The training intensity was increased by 10% per session, with the last training session at 90% peak torque. During the training exercises, the subjects were asked to follow on-screen torque curves that provided visual feedback of the intensity of contractions. Finally, 4 d after the last training session, the subjects performed the maximal eccentric test for the second time (see Figure, Supplemental Digital Content 1, which illustrates the testing sequences;

Muscle damage assessment

The muscle damage was assessed before and 24 h after both maximal eccentric tests by three indirect markers that have been used in previous studies (5,33). Blood samples (5 mL in heparin tubes) were drawn from an antecubital vein and centrifuged to obtain plasma. Plasma CK activity was determined by routine spectrophotometric techniques used in the Department of Clinical Chemistry. To assess muscle stiffness, the examiner attempted to flex the knee of the exercised leg while the volunteers were lying in prone position. When the subject perceived painful sensations, the examiner stopped the knee flexion movement and measured the distance from the heel to the buttock. The subjective presence and intensity of pain perceived during daily activities were evaluated using a visual analog scale (VAS) graded from 0 (without pain) to 10 (intolerable pain) arbitrary units (a.u.). Although changes in muscle function are reported to be the best indirect marker of muscle damage, muscle strength assessment 1 d after exercise was not performed because it would have biased the muscle proteome results.

Muscle biopsy

Volunteers were submitted to needle biopsies in the rectus femoris muscle of the right leg in three successive conditions. The control biopsies were collected 4 wk before the first eccentric protocol. The second and the third biopsies were collected 24 h after each eccentric test. This time point was chosen to analyze the muscle proteome when indirect markers of muscle damage were observed, before the early phase of the recovery process.

Muscle samples were obtained using an automatic biopsy system (Bard Magnum Biopsy System, Tempe, AZ) and biopsy needle (14-gauge) according to the procedure described by Magistris et al. (24). Skin was first anesthetized with xylocaine (2%), and a small incision was made in the skin and fascia to insert the biopsy needle. Two muscle samples were collected in each condition. One sample was immediately stored at −80°C for proteomic analysis, and the other was incubated in a stabilization solution to stop RNAse activity (RNAlater; Ambion, Foster City, CA) for 1 h and then stored at −80°C until use for genomic analysis (Barrey et al., unpublished observations).

Statistical analysis of physiological parameters

Values in the text and figures are expressed as means ± SEM. For each parameter, differences between the preexercise and postexercise measures and between the first and the second eccentric bouts were evaluated using a paired Student’s t-test. The level of statistical significance was set to P < 0.05.

Proteomic Analysis

Preparation of total skeletal muscle extracts

The muscle samples were suspended in DIGE-compatible lysis buffer containing 30 mM Tris, pH 8.5 (GE Healthcare, Diegem, Belgium), 7 M urea (GE Healthcare), 2 M thiourea (GE Healthcare), and 2% ASB14 (Sigma, Bornem, Belgium). The suspensions were solubilized by sonication and centrifuged at 20,000g to remove any insoluble material. The protein pellets were resuspended in the DIGE-compatible buffer, the pH was adjusted to 8.5–9, and protein concentrations were measured using the RC/DC Protein Assay (Bio-Rad, Eke-Nazareth, Belgium).

Protein CyDye labeling

For the purpose of comparative proteomic profiling, 25 μg of muscle samples proteins was labeled with 200 pmol of CyDye (GE Healthcare), either Cy3 or Cy5. An internal standard constituted by an equimolar amount of all the biological replicates was labeled with Cy2. The protein/CyDye mixture was vortexed, centrifuged briefly, and incubated for 30 min in the dark. The labeling reactions were stopped by adding 10 mM lysine, vortexing, and incubating for 10 min in the dark. Then, experimental samples were mixed in pairs, together with 25 μg of internal standard before loading on the first dimension gel of electrophoresis.

Two-dimensional electrophoresis

Rehydration and isoelectric focusing was carried out in the first dimension using pH 3–11 (24 cm) immobilized ph gradient (IPG) strips and Ettan IPGphor focusing system (GE Healthcare). The following electrophoresis steps were successively applied: 8 h of rehydration, 1 h at 500 V, 1 h at 1000 V, 3 h at 8000 V, and 8 h 15 min at 8000 V.

Before initiating the second dimension step, the IPG strips were reduced for 15 min in an equilibration buffer (6 M urea, 50 mM Tris–HCl, pH 8.8, 2% sodium dodecyl sulfate (SDS), 30% glycerol) containing 1% dithiothreitol and were then alkylated in the same equilibration buffer supplemented with 2.5% iodoacetamide. Protein IPG strips were briefly washed in SDS running buffer and layered on top of a 12.5% SDS–polyacrylamide gel. IPG strips were held in place using gel-sealing agarose poured on top of the strip and left to solidify. Electrophoretic migration was conducted in an Ettan Dalt apparatus (GE Healthcare) at 1 wt per gel for 1 h and then, overnight at a constant voltage of 40 V until the dye front had just migrated off the lower end of the gels.

Image acquisition, gel analysis, and statistical analysis of protein expression

The CyDye-labeled proteins were visualized using the Typhoon 9400 imager (GE Healthcare), and the DeCyder 2D 7.0 software package for DIGE (GE Healthcare) was used for gel maps analysis. Spot detection and spot quantification relative to the corresponding spot in the internal standard were performed by the DIA module (differential in-gel analysis). Intergel matching and statistical analysis of spot quantifications were obtained by running DeCyder BVA (biological variance analysis). We created distinct BVA workspaces for the “untrained group” and “trained group” analyses because proteomic analysis of these groups was performed with a different internal standard. In the experimental design window of the BVA modules, spot maps were assigned to the three conditions: 1) control (biopsies collected before any test), 2) Test1 (biopsies made 24 h after the first eccentric test), and 3) Test2 (biopsies obtained 24 h after the second bout of maximal eccentric exercise). Within all groups, each biopsy condition was compared with another condition using a t-test (P < 0.05). Spots exhibiting a statistically significant variation in relative abundance of at least twofold and present in two-thirds of all the protein maps were selected for picking. Then, the two BVA workspaces were imported in the DeCyder EDA module (extended data analysis). After applying a normalization to correct for nonbiological variation between the BVA workspaces, a differential expression analysis was performed. Protein spots that showed a statistically significant Student’s t-test (P < 0.05) for at least a twofold increase or decrease in normalized abundance were accepted as being differentially expressed between two biopsy conditions or between the trained and untrained groups.

Protein identification by MS

For protein identification, preparative gels were loaded with unlabeled samples (300 μg per strip) and matched with the master image of the analytic gels. Spots of interest were picked out from these preparative gels with a Spot Picker robot (GE Healthcare). Excised gel spots were digested with trypsin, and their identification was revealed by matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry (MALDI-TOF MS/MS; Ultraflex II; Bruker Daltonics, Billerica, MA) as described by Poirrier et al. (35). The variable and fixed modifications were methionine oxidation and cysteine carbamylation, respectively, with a maximum number of missed cleavages of 1. Mass precision tolerance error was set to 100 ppm. The four peaks with the highest intensities or all peaks obtained in TOF/MS mode were next analyzed by ``LIFT’’ mode of tandem mass spectrometry. Proteins were identified with the Biotools 3.1 software (Bruker Daltonics) using the Mascot (Matrix Science, Boston, MA) search engine. The probability score (MOWSE score) calculated by the software was used as a criterion for correct protein identification.

Western Blot Analysis

Samples containing 25 μg of protein were separated by electrophoresis on NuPage gels (Invitrogen, Merelbeke, Belgium) at 200 V and transferred to polyvinylidene fluoride membranes (Roche, Brussels, Belgium) using NuPage transfer buffer. Nonspecific binding was prevented by preincubation of the polyvinylidene fluoride membranes in I-block (0.1%) (Tropix, Bedford, MA) for 1 h. The membranes were then incubated overnight at 4°C with primary antibodies. Antibodies against aconitase 2 (85 kDa, raised in rabbit, 1:250), VDAC1 (31 kDa, raised in rabbit, 1:500), and actin (42 kDa, raised in mouse, 1:400) were purchased from Abcam (Cambridge, UK), and the antibody against aldolase (40 kDa, raised in rabbit, 1:1000) was purchased from Cell Signaling (Beverly, MA). Bound primary antibody was detected using a horseradish peroxidase–conjugated secondary antibody (raised in mouse or rabbit, 1:3000; Sigma) and SuperSignal West Pico Chemiluminescent Substrate Kit (Thermo Scientific) as a chemiluminescent substrate. The blots were stripped in Restore Western Blot Stripping Buffer (Thermo Scientific, Geel, Belgium) for 15 min and probed again with antiactin antibodies to account for variability in protein loading and transfer in each lane. Scanned images were quantified by ImageMaster 1D Prime Software (GE Healthcare). The signal intensity of the metabolic protein bands was normalized with the corresponding actin band.


Eccentric Exercises and Muscle Damage Assessment

Eccentric tests

The total work developed by the quadriceps muscles was significantly higher in Test2 compared with Test1 for the subjects who performed the five training sessions between the two bouts of eccentric exercises (P < 0.05) but not for the untrained group (Fig. 1A). All subjects assigned to the trained group showed an increase of total work after the training, but the magnitude of the benefit varied according to the participants (from +3.97% to +27.62%).

Eccentric tests and muscle damage assessment (mean ± SEM,n = 6 per group). A, Eccentric work performed by quadriceps muscles during each set of the Test1 and Test2; total work of the two tests. *Significant difference between Test1 and Test2 values for the trained group (P<0.05). B, Mean plasma CK response, expressed by the log(CK post/CK pre). *Significant difference between Test1 and Test2 for the trained group (P < 0.05). C, Mean rating scores on a VAS of the pain and discomfort in the subject’s quadriceps assessed 24 h after the Test1 and the Test2. a.u., arbitrary units. Note that this measure is 0 before Test1 and Test2 in both trained and untrained groups. **Significant difference between Test1 and Test2 in the trained group (P < 0.01). D, Mean heel–buttock distance (cm) obtained by passive flexion of the knee 24 h after the Test1 and Test2. *Significant difference between Test1 and Test2 values for both groups (P < 0.05).

Plasma CK activity

Mean values of plasma CK activity increased significantly (P < 0.01) in both groups 1 d after the first test. CK activity increased from 108 ± 17 to 3908 ± 1859 IU·L−1 in the untrained group and from 266 ± 174 to 11,485 ± 7020 IU·L−1 in the trained group (Table 1). No significant difference between groups was evident for the changes in plasma levels of this enzyme after Test1. The plasma CK activity after Test2 was, on average, 33% of the activity after Test1 for the untrained group but only 5% for the trained group. Thus, compared with Test1, the CK response was significantly (P < 0.05) attenuated after Test2 for the trained group but not for the untrained group (Fig. 1B).

Plasma CK activity (IU·L1) measured before (Pre) and 24 h after exercise (Post) in both untrained and trained groups.

Subjective muscle soreness

No muscle soreness was described before tests (0 a.u.) by any volunteer in the trained or untrained groups. After the first eccentric test, volunteers reported significant (P < 0.001) muscle soreness evaluated at 5.8 ± 0.5 a.u. (for the trained group) and 4.3 ± 0.7 a.u. (for the untrained group) on a 10-cm VAS (Fig. 1C). The level of perceived soreness was significantly smaller (P < 0.01) after the second test compared with the result obtained after the first test for the volunteers assigned to the trained group (2.4 ± 0.4 a.u.) but not for the untrained group (3.1 ± 0.3 a.u.).

Muscle stiffness

The first bout of eccentric exercise induced a significant (P < 0.01) increase in quadriceps muscle stiffness that was responsible for a change in the mean heel–buttock distance of 10.5 ± 1.1 and 11.7 ± 2.4 cm in the trained and untrained groups, respectively. The change in the quadriceps muscles extensibility was significantly (P < 0.05) smaller after the second test in both groups (trained group = 6.2 ± 1.1 cm, untrained group = 9.4 ± 2.3 cm). This decrease in stiffness amounted to 41% in the trained group but only to 20% in the untrained group (Fig. 1D).

Proteomic Analysis

The fresh mass of muscle sample biopsies was 16.7 ± 0.7 mg, and the mean protein concentration after extraction was 4 ± 0.4 mg·mL−1. After coupling, samples with CyDyes, their separation by two-dimensional electrophoresis, and the scan of the analytical gels generated three images per gel, an example of which is shown in Figures 2A–D, with the color of the laser used to detect each CyDye.

A–H, Gel images: internal standard (A; Cy2), Test2 sample (B; Cy3), control sample (C; Cy5), and superposition of B and C (D). In this superposition,green spots correspond to proteins more abundant in Test2 biopsies, while red spots correspond to proteins more abundant in control biopsies and yellow spots correspond to proteins equally abundant in control and Test2 biopsies. E and F, Gel maps illustrating the position of the spots that exhibit a differential expression between control/Test1 (E) and control/Test2 (F) conditions for the untrained group and identification of proteins in these spots by MALDI-MS/MS. G and H, Gel maps illustrating the position of the spots that exhibit a differential expression between control/Test1 (G) and control/Test2 (H) conditions for the trained group and identification of proteins in these spots by MALDI-MS/MS.

For both groups, the spot images were classified in three conditions: control, which corresponds to the muscle biopsy obtained before any eccentric exercise; Test1 and Test2, which correspond to muscle biopsies obtained after the first and the second maximal eccentric tests, respectively. After statistical analysis with the DeCyder BVA software, 191 spots were selected for picking in the trained group analysis and 118 spots in the untrained group analysis. From these, the MS/MS analysis identified 61 unique proteins. After normalization in the DeCyder EDA module, 52 unique proteins present in 128 spots showed a differential expression between conditions or between groups. These protein spots are listed in Table 2 (A–I) with their ID numbers, t-test values, average ratio, identification score, queries matched, and coverage. Figures 2E–H show the position of the identified proteins that exhibit a differential expression between control/Test1 and control/Test2 conditions for the trained and untrained groups. Some proteins such as myosin heavy chains were sometimes identified at a lower molecular mass than expected and thus could have been subjected to a limited proteolysis.

Human muscle proteins whose abundance differs between the biopsy conditions in the untrained (A-C) and trained groups (D-F) or between the trained and untrained groups (G-I).

Untrained Group

Effect of an intense eccentric test on human muscle proteome (Test1/control)

The proteomic comparison between Test1 and control biopsies revealed a decreased expression of two spots containing myosin 1 (MyHC-IIx/d) and myosin 2 (MyHC-IIa) after Test1.

Effect of a second eccentric test performed 6 wk after the initial bout (Test2/control)

The second test induced a differential expression of 19 unique proteins compared with the control biopsies. Protein spots containing myosin 1 (MyHC-IIx/d) and myosin 2 (MyHC-IIa), which already decreased after the first test (mean ratio = −3.44), decreased at an even lower level after the second eccentric test (mean ratio = −7.48). Spots identified as myosin 1 and α2-actin also decreased after the second test. The expression of desmin, a protein of the myofibrillar network, was higher after Test2. Increases in expression were also found for two structural proteins involved in the plasticity of actin microfilaments (gelsolin, moesin) and for a protein playing a role in protein assembly or remodeling processes (protein disulfide–isomerase A3). Collagen α1, being involved in basement membrane anchorage, and one isoform of keratin were also upregulated, indicating that structural changes occur at multiple levels in the muscle fiber reaction after a second eccentric bout. The expression of HSP71, a chaperone protein, was also higher after Test2. The glycolytic enzyme, L-lactate dehydrogenase M chain and the mitochondrial enzyme, ATP synthase subunit beta, were upregulated after the second test. Differential expression was also observed for myoglobin, annexin A2, and several plasma proteins (serotransferrin, serum albumin, Ig κ chain C region, Ig γ1 chain C region, and fibrinogen β chain). Except for myoglobin, the proteomic analysis revealed an increased expression of all these proteins.

Proteomic comparison between Test1 and Test2

One stress protein, HSPB1, showed a higher expression in Test2 condition compared with Test1, whereas the expression of myoglobin was decreased after Test2.

Trained Group

Effect of an intense eccentric test on human muscle proteome (Test1/control)

Compared with control biopsies, the first test induced a decreased expression of six spots identified as myosin 1 (MyHC-IIx/d) and one spot containing myosin 1 and myosin 2 (MyHC-IIa). At the same time, the expression of serum albumin decreased after Test1. The lowered occurrence of the latter protein may probably result from fluid shifts into the interstitial compartments associated with localized edema.

Effect of a specific training performed before an intense eccentric test (Test2/control)

Thirty proteins displayed differences in their relative expression level between control and Test2 biopsies. Most of these proteins possess contractile, metabolic, or structural functions. Spots of myosin 1 or containing myosin 1 and myosin 2, which already decreased after the first test (mean ratio = −2.30 and −2.04, respectively), decreased at an even lower level after the second eccentric test (mean ratio = −4.69 and −6.69, respectively). Myosin 7 (MyHC-beta), myosin light chain 1/3, fast isoforms of troponin T and troponin I, tropomyosin 1 α chain isoform 1, α1-actin, and α2-actin were also downregulated after Test2 compared with control. Four proteins associated with the z-disk (myozenin 1, LIM domain binding 3, α-actinin 3, and PDZ and LIM domain 3 protein) were markedly decreased after the second test (mean ratios = −3.91, −3.94, −7.43, and −3.76, respectively). The two-dimensional difference gel electrophoresis analysis of the trained group shows a differential expression of metabolic proteins after Test2. A large set of glycogenolytic and glycolytic enzymes (glycogen phosphorylase, aldolase, triosephosphate isomerase, glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase, phosphoglycerate mutase, enolase, pyruvate kinase, muscle L-lactate dehydrogenase M chain) were markedly downregulated after the eccentric test that was preceded by five sessions of eccentric training (Fig. 3). Muscle CK, an enzyme essential for rapid replenishment of cellular ATP stocks, was also differentially expressed with a lower expression after Test2. Other enzymes whose expressions were also downregulated after Test2 included citrate synthase, myoglobin, and an antioxidant enzyme, peroxiredoxin 6. The only protein that showed an increased expression after the second test compared with control was identified as the 14-3-3zeta protein. This protein is reported to be involved in the regulation of the kinase mammalian target of rapamycin signaling, which is recognized as a key regulator of skeletal muscle size (27).

The glycolytic pathway in the trained subjects. The glycolytic pathway incorporating the names of the identified enzymes and their respective ratios of abundance in the indicated comparisons for the trained group are mentioned. All glycolytic enzymes from aldolase to pyruvate kinase are significantly downregulated after Test2 compared with control or Test1 (the extreme in ratio of down-regulation observed in individual subjects for each glycolytic enzymes are indicated).

Proteomic comparison between Test1 and Test2

Eighteen proteins displayed differences in their relative expression level between Test1 and Test2 conditions, all showing a lower expression in Test2 compared with Test1. Except for cytosolic malate dehydrogenase, the proteins that decreased after Test2 were previously identified in the Test2/control comparison.

Trained versus Untrained Groups

The comparison of “control” condition between the trained and untrained groups revealed a differential expression of six different proteins (Table 2G). These included two different isoforms of actin, CK, mitochondrial glutamate dehydrogenase, serum albumin, and serotransferrin. Only one spot identified as keratin was differentially expressed between groups after Test1 (Table 2H), whereas 15 unique proteins showed a differential expression after Test2 (Table 2I). The fast isoforms of two contractile proteins (myosin 1 and troponin T) and eight metabolic enzymes including CK and seven enzymes involved in the glycolytic pathway were less expressed in the trained group. In contrast, the expression level of α-actin 2, 14-3-3 protein γ, and apolipoprotein 1 were higher in the trained group. Concerning the structural proteins, desmin and moesin were expressed at a higher level in the untrained group.

Western Blot Analysis of the Trained Group

Our proteomic analysis of the trained group showed a decrease in the abundance of most enzymes of the glycolytic pathway after Test2 but failed to document an increase in Krebs cycle enzymes that would compensate for the lost ATP-producing glycolytic machinery. This could have resulted from a lower relative abundance of mitochondrial enzymes in muscle biopsies that would render them undetectable in the proteomic analysis. On the other hand, the lack of oxidative enzymes identification in the DIGE analysis could also be explained by the absence of change in their expression. To clarify the changes in metabolic enzymes expression observed in the trained group, we applied a more sensitive method, a Western blot analysis using specific antibodies, to measure the change in abundance of one glycolytic enzyme, aldolase, and of two mitochondrial proteins, anion-channel VDAC1 and aconitase 2. Figures 4A, B show the immunodetection of aldolase in biopsies of trained subjects taken at control, Test1, and Test2 conditions. As observed in the proteomic analysis, the expression level of this glycolytic enzyme was lower in Test2 conditions compared with the other two conditions. Nevertheless, the decreased abundance of aldolase after Test2 did not reach the level of significance, probably because of the small number of replicates in this Western blot quantification. The anion-channel VDAC1, a mitochondrial membrane protein complex, was significantly more abundant in Test2 compared with Test1 (Figs. 4C, D). For aconitase 2, a mitochondrial matrix enzyme that catalyzes the conversion of citrate into isocitrate in the Krebs cycle, normalization of the results with actin was not possible because of a problem with membrane stripping. We observed, however, a significantly stronger signal for this oxidative protein in Test2 compared with Test1 (see Figure, Supplemental Digital Content 2, which shows the immunoblots of three trained subjects for aconitase 2; These Western blot results suggest that the glycolytic enzymes expression tend to decrease in the trained subjects in Test2 with respect to Test1 and control conditions and support an increase of oxidative proteins in the test after the training sessions compared with the first test.

A–D, Western blot analysis for the trained subjects. A, Aldolase protein expression was detected in control (C), Test1 (T1), and Test2 (T2) conditions using muscle samples of three trained subjects. The intensity of each aldolase signal was normalized with the corresponding actin signal obtained after stripping and reprobing of the membrane. B, The decreased expression of aldolase observed in Test2 does not reach significance. C, VDAC1 protein expression was detected in Test2 (T2) compared with Test1 (T1) condition. The intensity of VDAC1 signals was normalized using actin signal after membrane stripping and reprobing. Experiments were performed with muscle extracts of four participants. D, A significant increase in expression level between Test1 and Test2 values was found (P < 0.05). Values are means ± SEM. *Significant differences between the two conditions.


The present study addressed the question of the effect of two intense eccentric exercises applied 6 wk apart on the expression levels of protein in human muscle. To the best of our knowledge, this is the first time that the eccentric exercise–induced muscle damage has been investigated using a proteomic approach. In addition, this study examined the human muscle proteome when the second eccentric test is preceded by a specific training known to decrease markers of muscle injury (7,8).

In the first part of this study, the occurrence of the muscle damage was evaluated 24 h after each bout of eccentric exercises that consisted of performing three sets of 30 maximal contractions of the quadriceps muscles. As expected, the first isokinetic eccentric test leads to a marked increase of plasma CK activity, subjective muscle soreness and an increase in muscle stiffness. Regarding the CK activity, a qualification as low and high responders for subjects who have low and high plasma CK activity after muscular exercise has been proposed in the literature (3,31). This parameter showed a large interindividual variability in our study, and we can qualify two subjects (subjects 10 and 11) as high responders. Besides this wide variability, no significant difference between groups was evident for the three indirect markers. The low number of subjects could represent a limitation of the present study, but it is justified by the duration and the invasive nature of the experimental protocol. Furthermore, increasing the number of subjects may not necessarily reduce the interindividual variability response to eccentric protocols.

The assessment of the muscle damage manifestations before both eccentric tests indicated that all subjects were in similar conditions before Test1 and Test2 (a low plasmatic CK activity, no muscle soreness, and no difference in the baseline heel–buttock measurements). The evolution of the three indirect markers after the second test is in agreement with numerous studies demonstrating that previous exposure to eccentric exercise reduces the severity of many common markers of muscle damage in subsequent exercises (4,5,25,34). However, the magnitude of the protective effect is higher in the trained subjects but only light to moderate for the untrained group, demonstrating that muscles of subjects who underwent training with submaximal exercises benefited from a more complete adaptation to eccentric exercise. In fact, CK response, muscle stiffness, and subjective soreness were significantly reduced after Test2 when compared with Test1 for the trained group, whereas the untrained group showed a significant reduction only in muscle stiffness. The light adaptation observed in the untrained group could also be explained by the long interval between the two eccentric sessions, as it is known that the magnitude of the protective effect seems to decrease gradually as the time between bouts increases (32).

There are limitations to the analysis of the entire muscle proteome as used in this study. Approximately 50% of a muscle fiber’s protein content is made up of the contractile machinery (11). The high abundance of the contractile proteins impairs the detection and quantification of proteins that are present at lower concentrations in muscle cells such as transcription factors, signaling proteins, inflammatory factors, or mitochondrial proteins. Consequently, most of the identified proteins are contractile, structural, and/or abundant metabolic proteins. To overcome this limitation, subcellular fractionation of the entire skeletal muscle has already been performed successfully (17,38) but is not possible with the very small amounts of tissue in biopsies obtained from human volunteers. Most of the muscle proteome modifications were observed after Test2, when changes in the indirect markers of muscle injury were reduced. This demonstrates that the protein expression changes reflect muscle adaptation rather than necrosis. The comparison of control/Test1 biopsies allowed observing only few significant protein expression modifications. However, it is possible that the first eccentric exercise induced more changes of the muscle proteome because only spots with a differential expression ratio of ±2 were chosen for picking, and all picked spots were not successfully identified.

The comparison between control and Test1 biopsies allowed the observation of lower concentrations of myosin heavy chain isoforms in both groups. Surprisingly, we observed that these isoforms that had already decreased after Test1 decreased further after Test2 in both groups. In addition, the relative amount of a larger set of contractile proteins also decreased after Test2: α1-actin, putative β-actin-like protein 3, myosin 7, myosin light chain 1/3, tropomyosin, fast isoforms of troponin T and troponin I (for the trained group), and α-2 actin (for both groups). Test2 was performed after single or repeated eccentric exercises known to decrease the severity of muscle damage. It was observed that plasma CK activity and the mean rating scores on a VAS of pain and discomfort and the muscle stiffness were decreased after Test2. Because the DOMS was high after Test1 but was reduced after Test2, we can conclude that the occurrence of DOMS is not directly linked to the decreased expression of contractile proteins, which was observed after both eccentric tests. In addition, both groups showed a marked decrease of several spots identified as myosin 1 (MyHC-IIx/d) after Test2, suggesting that this change probably reflects a consequence of the first eccentric bout. Myosin 1 is specific to fast-twitch glycolytic fibers, and these fibers have been shown to be more susceptible to disruption by an unaccustomed eccentric exercise (14,15,21). Fiber-type–specific damage might in part explain why hamstring muscles, which have a relatively high proportion of fast type II fibers, have been shown to be markedly more vulnerable to delayed muscle soreness and muscle damage compared with the quadriceps muscle (9). It should be noted that the present study was applied on quadriceps muscles rather than hamstrings for technical reasons associated with the biopsy procedure. The decreased expression of myosin 1 is thus consistent with the hypothesis that muscle fibers become more resilient and able to withstand a given eccentric exercise because of the removal of stress-susceptible fibers and their replacement by regenerated fibers (26,29). Nevertheless, this hypothesis may not explain completely the repeated bout effect because a moderate protective effect was also observed in the untrained group.

The contractile apparatus disruption is classically associated with a disorganization of cytoskeletal proteins. Several investigators have indeed demonstrated histopathological changes in titin, desmin, and dystrophin staining in eccentric contraction models in rats (1,20). However, because of their high molecular mass, structural proteins such as titin, nebulin, and dystrophin cannot be resolved by two-dimensional electrophoresis. Nevertheless, our proteomic data revealed a differential expression of several structural proteins. The increased expression of desmin observed in the untrained group after the second test is consistent with previous studies (39,40). Feasson et al. (12) reported a dramatic increase in desmin protein level in human muscle biopsies 14 d after cessation of a single bout of eccentric exercise. The authors interpreted this finding as indicative of a remodeling response of the myofibrils, which could reinforce the z-disk through an increased resistance to mechanical constraints. In addition, the increased expression of the chaperone proteins, HSP71 and HSPB1, could be linked to the assembly and maintenance of actin and of the intermediate filament network.

In the trained group, modifications of the cytoskeletal proteins after Test2 seemed to affect predominantly those isoforms that are found in fast-twitch fibers, as already observed for the contractile proteins (troponin T, troponin I, MyHC-IIx/d). Among the structural proteins whose expression is also decreased after Test2 (myozenin 1, LDB3 protein, α-actinin 3, and its interacting PDLIM3 protein, all located at the Z-line), isoform 3 of α-actinin that cross links the thin actin filaments of adjacent sarcomeres is restricted mainly to fast glycolytic or type II fibers. Actinin alpha 2, which did not change in the trained group, is the isoform expressed in all muscular fibers. Similar to α-actinin 3, myozenin 1 (a calcineurin-binding protein also termed “calsarcin 2”) is also exclusively expressed in fast-twitch fibers of the skeletal muscle tissue. Recent data indicate that myozenin 1 and α-actinin 3 could play an important role in the differentiation process toward a fast-twitch, glycolytic profile and that, in their absence, the fiber may tend toward a slower-twitch, more oxidative profile (2,13).

Our results suggest that a potential isoform shift in fiber-type components could be involved in the protective effect observed after repeated eccentric exercises. Indeed, we demonstrate that major contractile and structural proteins that are specific to the fast muscle fibers are indeed diminished after Test2, when the muscle damage indirect markers were reduced. These modifications could be interpreted as a decrease in the amount of mass from type II fibers in relation to type I fibers because of the removal of stress-susceptible fibers consecutive to Test1 (28,37). The large decrement in several proteins may be related to an accelerated proteolytic system that specifically occurs with eccentric exercises (12,28,37). Indeed, Vissing et al. (37) have shown that the proteolytic response is not attenuated 24 h after a second bout of eccentric exercise compared with an identical first bout, which may explain part of the decreased protein expression observed after Test2. However, because decreased expression of fast isoforms (except for myosin heavy chains) was only observed in the trained group, it is more likely that an isoform shift to slower fiber-type components would be specifically induced by the eccentric training sessions.

Potential neural adaptations for the repeated bout effect described in the literature include an increase in motor unit activation and/or a shift to slow-twitch fiber activation, which distributes the contractile stress over a larger number of active fibers (26). Our data are consistent with a decrease in fast glycolytic fiber activation. Thus, we hypothesize that repeated eccentric training sessions could promote a motor unit remodeling, which might induce a more oxidative muscle phenotype. Whether a shift from fast glycolytic fiber type to a slow oxidative fiber type might contribute to prevent the appearance of muscle damage symptoms is an interesting question that should be further investigated. Numerous recent proteomic studies document changes in muscle fiber types after chronic low-frequency stimulation (10), in sarcopenia of old age (16), and in muscular diseases (9), demonstrating the great plasticity of skeletal muscle. In the context of muscle exercise, very few proteomics analysis exist. For instance, Holloway et al. (19) investigated the effects of a 6-wk interval training on the human muscle proteome and reported an increased expression for some contractile and metabolic proteins, contrasting with our proteomic data. As for the interpretation of these divergent findings, one must consider that the exercise protocol (combinations of sprints and jogging on a treadmill) and the time of muscle biopsies (72 h after training) differed strongly from the study of Holloway et al. (19). Because our work specifically addresses eccentric contraction and training, extrapolation of findings of the present study to those using other exercise model may not be possible.

In addition, the comparison between control and Test2 conditions in the trained group revealed a more profound down-expression of a large set of enzymes implicated in the glycolytic pathway. Because such a decrease in the glycolytic pathway enzymatic capacity was observed neither after Test1 nor after Test2 in the untrained group, it is likely that the training sessions are responsible for this modification. The downregulated expression of glycolytic enzymes after training could be interpreted as a reduction of the relative contribution of anaerobic ATP production, implying that ATP used during Test2 would be generated from other sources such as pyruvate oxidation (at the expense of lactic acid production) and/or intramuscular fatty acid oxidation. In addition, the expression of CK is strongly decreased after Test2, providing additional evidence of a reduced anaerobic ATP production. It is well accepted that a program of endurance exercise training can result in a significant increase of muscular mitochondrial density and oxidative enzyme activities (6), but whether anaerobic training, such as the short-duration eccentric exercise on the isokinetic dynamometer, can result in an improvement of either glycolytic or oxidative enzyme activities remains unclear (23). The results of Harmer et al. (18) show attenuated glycogenolysis and glycolysis during intense exercise after sprint training and provide evidence suggesting an enhanced muscle oxidative metabolism. However, our proteomic analysis on total proteins has not revealed an increase in the concentration of enzymes of the oxidative metabolism. This could result from a lower relative abundance of mitochondrial enzymes in muscle biopsies that would place them below the limit of detection in the proteomic analysis (see above). Furthermore, our results concerning metabolic adaptation after training are in agreement with a parallel genomic study (Barrey et al., unpublished observations), which demonstrates an increase of the mitochondriogenesis in the biopsies collected after the test after the training sessions. The metabolic changes are consistent with our results, suggesting that muscles submitted to eccentric training sessions would adopt slower, more oxidative properties.

In conclusion, the present study has shown that a proteomic approach represents a useful technique for investigating the response of the muscle proteome to eccentric exercise in humans. Indeed, our results suggest that the decreased expression of contractile proteins is not responsible for the occurrence of DOMS. They also highlight a metabolic adaptation induced by a short eccentric training, which could be linked to an adaptative decrease of anaerobic catabolism and a possible isoform shift toward a more resilient, slower-contracting phenotype of trained muscle.

This work was supported by the Fonds National de la Recherche Scientifique and the Fonds Spéciaux de l’Université de Liège. S.H. is a research fellow and P.L. is a research associate of the Fonds National de la Recherche Scientifique.

The authors declare no conflicts of interest.

The authors thank P. Gengoux, S. Labrozzo, and P. Piscicelli for their expert technical assistance.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.


1. Barash IA, Peters D, Friden J, Lutz GJ, Lieber RL. Desmin cytoskeletal modifications after a bout of eccentric exercise in the rat. Am J Physiol Regul Integr Comp Physiol. 2002; 283 (4): R958–63.
2. Chan S, Seto JT, MacArthur DG, Yang N, North KN, Head SI. A gene for speed: contractile properties of isolated whole EDL muscle from an alpha-actinin-3 knockout mouse. Am J Physiol Cell Physiol. 2008; 295 (4): C897–904.
3. Clarkson PM, Ebbeling C. Investigation of serum creatine kinase variability after muscle-damaging exercise. Clin Sci (Lond). 1988; 75 (3): 257–61.
4. Clarkson PM, Nosaka K, Braun B. Muscle function after exercise-induced muscle damage and rapid adaptation. Med Sci Sports Exerc. 1992; 24 (5): 512–20.
5. Cleary MA, Kimura IF, Sitler MR, Kendrick ZV. Temporal pattern of the repeated bout effect of eccentric exercise on delayed-onset muscle soreness. J Athl Train. 2002; 37 (1): 32–6.
6. Coffey VG, Hawley JA. The molecular bases of training adaptation. Sports Med. 2007; 37 (9): 737–63.
7. Croisier JL, Camus G, Venneman I, et al.. Effects of training on exercise-induced muscle damage and interleukin 6 production. Muscle Nerve. 1999; 22 (2): 208–12.
8. Croisier JL, Camus G, Crielaard JM. Interest of a specific training to reduce DOMS. Isokinetics Exerc Sci. 2000; 8: 40.
9. De Palma S, Morandi L, Mariani E, et al.. Proteomic investigation of the molecular pathophysiology of dysferlinopathy. Proteomics. 2006; 6 (1): 379–85.
10. Donoghue P, Doran P, Wynne K, Pedersen K, Dunn MJ, Ohlendieck K. Proteomic profiling of chronic low-frequency stimulated fast muscle. Proteomics. 2007; 7 (18): 3417–30.
11. Doran P, Donoghue P, O’Connell K, Gannon J, Ohlendieck K. Proteomic profiling of pathological and aged skeletal muscle fibres by peptide mass fingerprinting. Int J Mol Med. 2007; 19 (4): 547–64. Review.
12. Feasson L, Stockholm D, Freyssenet D, et al.. Molecular adaptations of neuromuscular disease–associated proteins in response to eccentric exercise in human skeletal muscle. J Physiol. 2002; 543 (Pt 1): 297–306.
13. Frey N, Frank D, Lippl S, et al.. Calsarcin-2 deficiency increases exercise capacity in mice through calcineurin/NFAT activation. J Clin Invest. 2008; 118 (11): 3598–608.
14. Friden J, Lieber RL. Segmental muscle fiber lesions after repetitive eccentric contractions. Cell Tissue Res. 1998; 293 (1): 165–71.
15. Friden J, Sjostrom M, Ekblom B. Myofibrillar damage following intense eccentric exercise in man. Int J Sports Med. 1983; 4 (3): 170–6.
16. Gelfi C, Vigano A, Ripamonti M, et al.. The human muscle proteome in aging. J Proteome Res. 2006; 5 (6): 1344–53.
17. Hamelin M, Sayd T, Chambon C, et al.. Differential expression of sarcoplasmic proteins in four heterogeneous ovine skeletal muscles. Proteomics. 2007; 7 (2): 271–80.
18. Harmer AR, McKenna MJ, Sutton JR, et al.. Skeletal muscle metabolic and ionic adaptations during intense exercise following sprint training in humans. J Appl Physiol. 2000; 89 (5): 1793–803.
19. Holloway KV, O’Gorman M, Woods P, et al.. Proteomic investigation of changes in human vastus lateralis muscle in response to interval-exercise training. Proteomics. 2009; 9 (22): 5155–74.
20. Lehti TM, Kalliokoski R, Komulainen J. Repeated bout effect on the cytoskeletal proteins titin, desmin, and dystrophin in rat skeletal muscle. J Muscle Res Cell Motil. 2007; 28 (1): 39–47.
21. Lieber RL, Friden J. Morphologic and mechanical basis of delayed-onset muscle soreness. J Am Acad Orthop Surg. 2002; 10 (1): 67–73.
22. Lieber RL, Thornell LE, Friden J. Muscle cytoskeletal disruption occurs within the first 15 min of cyclic eccentric contraction. J Appl Physiol. 1996; 80 (1): 278–84.
23. MacDougall JD, Hicks AL, MacDonald JR, McKelvie RS, Green HJ, Smith KM. Muscle performance and enzymatic adaptations to sprint interval training. J Appl Physiol. 1998; 84 (6): 2138–42.
24. Magistris MR, Kohler A, Pizzolato G, et al.. Needle muscle biopsy in the investigation of neuromuscular disorders. Muscle Nerve. 1998; 21 (2): 194–200.
25. Mair J, Mayr M, Muller E, et al.. Rapid adaptation to eccentric exercise–induced muscle damage. Int J Sports Med. 1995; 16 (6): 352–6.
26. McHugh MP. Recent advances in the understanding of the repeated bout effect: the protective effect against muscle damage from a single bout of eccentric exercise. Scand J Med Sci Sports. 2003; 13 (2): 88–97.
27. Miyazaki M, Esser KA. REDD2 is enriched in skeletal muscle and inhibits mTOR signaling in response to leucine and stretch. Am J Physiol Cell Physiol. 2009; 296 (3): C583–92.
28. Nedergaard A, Vissing K, Overgaard K, Kjaer M, Schjerling P. Expression patterns of atrogenic and ubiquitin proteasome component genes with exercise: effect of different loading patterns and repeated exercise bouts. J Appl Physiol. 2007; 103 (5): 1513–22.
29. Newham DJ, Jones DA, Clarkson PM. Repeated high-force eccentric exercise: effects on muscle pain and damage. J Appl Physiol. 1987; 63 (4): 1381–6.
30. Newham DJ, Jones DA, Edwards RH. Large delayed plasma creatine kinase changes after stepping exercise. Muscle Nerve. 1983; 6 (5): 380–5.
31. Nosaka K, Clarkson PM. Variability in serum creatine kinase response after eccentric exercise of the elbow flexors. Int J Sports Med. 1996; 17 (2): 120–7.
32. Nosaka K, Newton M, Sacco P. Delayed-onset muscle soreness does not reflect the magnitude of eccentric exercise–induced muscle damage. Scand J Med Sci Sports. 2002; 12 (6): 337–46.
33. Nosaka K, Sakamoto K, Newton M, Sacco P. How long does the protective effect on eccentric exercise–induced muscle damage last? Med Sci Sports Exerc. 2001; 33 (9): 1490–5.
34. Paddon-Jones D, Abernethy PJ. Acute adaptation to low volume eccentric exercise. Med Sci Sports Exerc. 2001; 33 (7): 1213–9.
35. Poirrier JE, Guillonneau F, Renaut J, et al.. Proteomic changes in rat hippocampus and adrenals following short-term sleep deprivation. Proteome Sci. 2008; 6: 14.
36. Proske U, Allen TJ. Damage to skeletal muscle from eccentric exercise. Exerc Sport Sci Rev. 2005; 33 (2): 98–104.
37. Vissing K, Overgaard K, Nedergaard A, Fredsted A, Schjerling P. Effects of concentric and repeated eccentric exercise on muscle damage and calpain–calpastatin gene expression in human skeletal muscle. Eur J Appl Physiol. 2008; 103 (3): 323–32.
38. Vitorino R, Ferreira R, Neuparth M, et al.. Subcellular proteomics of mice gastrocnemius and soleus muscles. Anal Biochem. 2007; 366 (2): 156–69.
39. Yu JG, Carlsson L, Thornell LE. Evidence for myofibril remodeling as opposed to myofibril damage in human muscles with DOMS: an ultrastructural and immunoelectron microscopic study. Histochem Cell Biol. 2004; 121 (3): 219–27.
40. Yu JG, Thornell LE. Desmin and actin alterations in human muscles affected by delayed onset muscle soreness: a high resolution immunocytochemical study. Histochem Cell Biol. 2002; 118 (2): 171–9.


Supplemental Digital Content

©2011The American College of Sports Medicine