Molecular responses of human muscle to eccentric exercise

Yi-Wen Chen, Monica J. Hubal, Eric P. Hoffman, Paul D. Thompson, Priscilla M. Clarkson


We examined the effect of eccentric exercise on the transcriptome of skeletal muscle in three male human volunteers who performed 300 concentric contractions with one leg and 300 eccentric contractions with the opposite leg. Vastus lateralis muscle biopsies were taken from both legs at 4–8 h after exercise, and expression was profiled by using 12,000 gene Affymetrix U95Av2 microarrays. We found a high concordance of expression responses to eccentric contractions between our human and rat data from a previous study (Chen YW, Nader GA, Baar KR, Fedele MJ, Hoffman EP, and Esser KA. J Physiol 545: 27–41, 2002) (∼50% of gene expression changes shared between species). Potential human-specific changes included greater inflammatory responses [chemokine (C-C motif) ligand 2, C/EBP delta, and IL-1 receptor] and vascular remodeling (tenascin C and lipocortin II). Induction of c-fos and lipocortin II were confirmed at the protein level, with c-fos localized to myofiber nuclei and lipocortin II to intramuscular capillaries. We also confirmed the eccentric-induced expression of six transcripts by quantitative RT-PCR (cardiac ankyrin-repeated protein, chemokine ligand 2, CCAAT/enhancer binding protein delta, IL-1 receptor, tenascin C, and cysteine-rich angiogenic inducer 61). These data provide the first characterization of the transcriptional response of skeletal muscle to eccentric exercise in humans and represent a preliminary step in understanding the molecular processes underlying muscle remodeling (including a new focus on rapid changes in the capillary bed) and inflammatory responses after damaging lengthening contractions.

  • muscle damage
  • c-fos
  • lipocortin II
  • immediate early genes
  • expression profiling

eccentric (muscle lengthening) contractions have been found to cause skeletal muscle damage in both human and animal studies. Manifestations of this damage, such as morphological disturbances to muscle cells, muscle soreness, prolonged strength loss, and increases in muscle proteins in the blood have been clearly described (4, 15, 41). This muscle damage is repairable and results in adaptation such that the muscle is more resistant to subsequent damaging exercise (13). The adaptation process also provides a potent stimulus for muscle growth, greater than that seen after concentric or isometric exercise (20, 22). However, little is known regarding the molecular mechanisms that mediate these processes. Furthermore, possible species differences in response to exercise-induced muscle damage could confound the issue, given that both human and animal (typically rodent) models are used to study this phenomenon.

Several known molecular pathways and/or known transcription factors have been examined after muscle stretch or active eccentric exercise, yielding some preliminary information about what molecular processes are taking place in the tissue. In animal models, Michel et al. (40), Osbaldeston et al. (45), Dawes et al. (14), and Goldspink et al. (18) all detected increases in the transcription factor c-fos after electrical stimulation of rabbit muscle. Goldspink et al. (18) found a biphasic induction of c-fos and c-jun mRNAs after a combination of electrical stimulation and stretch in rabbit extensor digitorum longus (EDL), with the first peak after ∼1 h and the second peak between 4 and 6 h. In humans, Boppart et al. (7) found increases in the transcription factor c-jun in muscle after eccentric exercise. Also, Puntschart et al. (48) revealed significant increases in c-fos and c-jun proteins at 4 and 30 min after exercise before returning to baseline by 3 h after exercise. These data suggest some similarity in the human and animal response to eccentric exercise; however, there has been no systematic comparison between species.

Previous studies have yielded limited information concerning molecular responses to muscle-damaging exercise because they have used what is described as a “vertical approach,” examining single or a few molecules or genes in series. This approach provides only a “snapshot” of the possible cellular events in the damage and repair process. Concerns with the traditional molecular biology approach are that the genes or molecules chosen for study are generally the “obvious” ones, and only limited information is provided given that many pathways are likely activated in response to eccentric contraction stress. Emergent gene microarray technology now provides a sensitive and powerful tool to simultaneously evaluate changes in gene expression in relatively small amounts of tissues or cells. Using this unbiased “horizontal” or survey method, one can readily determine what a cell is “thinking” in molecular, genetic terms when confronted with an environmental change, such as damage-inducing muscular contractions. This method also affords researchers the ability to identify novel candidate genes that respond to eccentric exercise.

The primary purpose of the present study was to use microarray technology to study transcriptional changes associated with exercise-induced muscle damage in humans in the hours after eccentric exercise. To date, most of the research on the molecular responses of skeletal muscle has been done with animal models. Although muscle is clearly a highly conserved tissue, considerable evidence exists that there are species-specific responses to muscle damage (particularly in the muscular dystrophies). Our laboratory has recently published an extensive characterization of the response of rat muscle to eccentric contractions, where both transcriptional and translational regulation were studied at 1 and 6 h after aggressive eccentric exercise (10). In this present report, we were thus able to compare the human and rat responses and identify responses that appear preferentially activated in humans.

Our study of human volunteers has a number of additional advantages over previous studies. Specifi-cally, the genomic resources for humans are relatively mature, with many more genes characterized in human than in other vertebrates. Also, we are able to use a longitudinal design in human volunteers, using one leg as the “control” and one leg as the “experimental” (eccentric), with paired t-tests allowing us to eliminate molecular noise from genetically heterogeneous humans.


Subjects. Three young adult men (mean age = 21 ± 2 yr) completed the study. Subjects signed an informed consent document approved by the University of Massachusetts School of Public Health and Health Sciences and completed a medical screening form. Subjects had not resistance trained in the past 6 mo and did not have a prior history of musculoskeletal injury to the lower limbs. Subjects also refrained from taking anti-inflammatory or analgesic medications for the duration of the study and were instructed to maintain their habitual diet.

The testing protocol consisted of baseline measures, including knee extension maximal force and muscle soreness evaluation of the vastus lateralis. Baseline measures were collected three times with 48 h between testing bouts. After the third baseline measure, subjects performed an eccentric/concentric lower body exercise. At 4, 6, or 8 h, subjects had a muscle biopsy taken from both legs. The reason that a range of time points was used is so that we could, in this initial study, identify major changes that occurred over a given time frame rather than focus on genes for which only a brief transient response would be found. At 24, 72, and 144 h postexercise, subjects returned for reassessment of baseline measures.

Knee extension strength testing. Maximal voluntary contraction force was assessed at 90° of knee flexion on an isokinetic dynamometer (Biodex System 3, Shirley, NY). Subjects maximally internally rotated each foot during each contraction to assure sufficient activation of the vastus lateralis. Each subject performed five 3-s contractions with a 1-min rest between contractions for each leg. The order of testing (i.e., which limb was tested first) was randomized across trials. Data were sampled at 100 Hz, and a three-point moving average was used to smooth the data for analysis. Maximal voluntary contraction was defined as the highest three-point average within each trial.

Exercise. Subjects performed a series of movements in which the left leg underwent 300 concentric contractions (rising from a chair) and the right leg underwent 300 concentric and 300 eccentric contractions (rising from and controlled lowering to a seated position). Thus both legs were used to stand up, but only the right leg was used to lower the body to the seated position. Each contraction lasted 1 s with 1 s between contractions. Subjects exercised to the beat of a metronome and to the verbal instructions of the investigator. This exercise protocol has been shown to result in damage (assessed via magnetic resonance imaging analysis) to the vastus lateralis muscle (55). Furthermore, this protocol allowed us to attribute the postexercise changes to a response to eccentric contractions and not concentric contractions known to produce little or no damage.

Muscle soreness. Muscle soreness was measured by using a visual analog scale, with the left end labeled “no soreness” and the right end labeled “very, very sore.” Subjects marked the scale after two hip/knee flexion and extensions by completing the sentence “Currently, I would describe the soreness (pain) during knee and hip flexion as....” The distance from the left end of the scale to the mark was taken as the level of soreness.

Biopsy. A percutaneous needle muscle biopsy was obtained 4–8 h postexercise from the left and right vastus lateralis muscles by using a Bergstrom 5-mm biopsy needle. All biopsy procedures were done at Hartford Hospital in Hartford, CT. The skin was first lightly anesthetized with lidocaine, a small incision was made through the skin and fascia, the biopsy needle was inserted, and ∼100 mg of tissue was removed and rapidly frozen in liquid nitrogen. The tissue was then packed in dry ice and sent to Children's National Medical Center for expression profiling and mRNA and protein studies.

Expression profiling. Expression profiling was conducted with Affymetrix Human Genome U95Av2 microarrays, containing ∼12,000 full-length genes and expressed sequence tags (EST). Biotinylated cRNAs prepared from total RNAs isolated from muscle having undergone concentric contractions (control) and those from the muscle having undergone eccentric and concentric contractions (eccentric) were hybridized to U95Av2 GeneChips. Procedures of cRNA preparation and microarray processing were performed as previously described (11).

Data analysis. After microarray images were obtained, we followed quality control criteria developed at Children's National Medical Center Microarray Center for each array. Quality-control measures included greater than fourfold cRNA amplification (from total RNA/cDNA), scaling factors of <2 to reach a whole-chip normalization of 800, and visual observation of hybridization patterns for chip defects (see

Absolute and comparison analyses of Affymetrix image data were done by using Affymetrix Microarray Suite 5.0 as previously described (35). Briefly, each gene shown was queried with 16 “perfect match” 25-bp oligonucleotides and with paired “mismatched” oligonucleotides designed with a single mismatch in the center position. Comparison of the hybridization signals from the perfect match and mismatched probes allowed a specific measure of signal intensity, and elimination of most nonspecific cross hybridization from the data analysis. Values of intensity difference as well as ratios of each probe pair were used for determination of whether a gene was called “present” or “absent.”

For comparison analyses (e.g., muscle that underwent eccentric and concentric contractions vs. muscle that underwent only concentric contractions), each probe set in an experimental GeneChip was compared with the control chip (other leg) from the same individual. One such pairwise comparison was performed for each of the three individuals. The difference calls that showed consistent results (same direction) in all three pairwise comparisons were retained for further analysis. To increase the stringency of the difference analysis, we limited our focus to transcripts with average changes of greater than fivefold (see Table 2). The changes of each individual were calculated by dividing the expression level of eccentric exercise by that of concentric exercise. Because each paired sample was from the same individual, average changes were calculated by averaging changes obtained from the three individuals. We calculated paired t-test statistics and corrected false positives due to multiple testing by using Benjamini and Hochberg false discovery rate methods (6, 49); however, the small sample size generally provided inadequate power for adequate measurements of significance. For this reason, we chose to confirm all discussed changes by either quantitative mRNA or protein studies. The P values in Tables 2 and 3 were generated by paired t-test without correcting multiple testing.

View this table:
Table 2.

Transcriptional changes of genes after eccentric exercise

View this table:
Table 3.

Quantitative RT-PCR confirmation of transcriptional changes of 6 genes after eccentric exercise

Immunoflorescent staining. Serial 4-μm-thick frozen muscle sections were cut with a Microme cryostat, mounted to Superfrost Plus Slides (Fisher Scientific), and fixed in cold anhydrous acetone. Sections were then blocked for 30 min in 10% horse serum and 1× PBS, and incubated with primary antibody for 3 h at room temperature. Polyclonal antibodies against c-fos and lipocortin II (annexin II) (Santa Cruz Biotechnology, Santa Cruz, CA) were applied with 1:20 dilution. Washes were done with 10% horse serum and 1× PBS, and sections were then incubated for 1 h with a secondary antibody Cy3-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories) diluted 1:500.

Immunoblotting. Frozen cryosections were solubilized in Laemli buffer, and 50 μg were loaded on 4–12% Tris · glycine SDS-PAGE gels (Zaxis). Gel-separated proteins were transferred by electrophoresis on 0.45-μm nitrocellulose membranes (Amersham Pharmacia). After transfer, membranes were blocked for 1 h at room temperature with 5% powdered milk in Tris-buffered saline + 0.1% Tween (TBST). After being washed three times for 15-min intervals in TBST, membranes were incubated for 1 h with the lipocortin II primary antibody (Sata Cruz Biotechnology) (1:100 dilution). Membranes were washed three times for 15-min intervals in TBST and then were probed with horseradish peroxidaseconjugated secondary antibody (Santa Cruz Biotechnology) for 1 h (1:500 dilution). Blots were developed by using enhanced chemiluminescence (Amersham Pharmacia) after another series of 3 × 15-min washes. Tubulin α1 (Santa Cruz Biotechnology) (1:100 dilution) was used as an internal loading control.

Reverse transcription and TaqMan quantitative PCR analysis. All reagents for reverse transcription were from Invitrogen unless otherwise noted. Total RNA was reverse transcribed by using oligo (dT) primer (0.5 μg/μl). Samples were heated to 70°C for 10 min to denature the primer and RNA. After heating, master mix consisting of 5× cDNA synthesis buffer, 0.1 M DTT, 10 mM dNTP, and superscript II RT (200 U/μl) was added to the samples. Samples were transferred to a thermocycler and incubated at 42°C for 60 min. TaqMan PCR primers were designed by using a Primer Express program version 1.01 (Applied Biosystems). GAPDH primers were purchased from Applied Biosystem. Specifically, primer sequences used for human cystein-rich angiogenic inducer 61 (CYR61) were (forward) 5′-AGTGTACAGCAGCCTGAAAAAGG-3′ and (reverse) 5′-GTGCGTCCTTGGCGTCATG-3; for cardiac ankyrin-repeated protein (CARP) were (forward) 5′-CTCAACATCAAGAACTGTGCTGG-3′ and (reverse) 5′-ACGGAGTCTTACATCGATACGC-3′; for chemokine ligand 2 (CCL2) were (forward) 5′-TCCCAAAGAAGCTGTGATCTTCA-3′ and (reverse) 5′-GTCTAAGAACCCAACACCTCACT-3′; for CCAAT/enhancer binding protein delta (CEBPD) were (forward) 5′-CTACAGCCTGGACTTACCACCACT-3′ and (reverse) 5′-TTACATGGAATCGACGTAGTTGTCC-3′; for IL-1 receptor were (forward) 5′-TAAGGAGGGACAAGAATCAATGGA-3′ and (reverse) 5′-AAGTGCCCCTTGATCCTTACAC-3′; for tenascin C were (forward) 5′-CAATAACCACAGTCAGGGCGTT-3′ and (reverse) 5′-CGGACGGAAGTTCTAAAGACTT-3′; and for human GAPDH were (forward) 5′-GAAGGTGAAGGTCGGAGTC-3′ and (reverse) 5′-GAAGATGGTGATGGGATTTC-3′. TaqMan ABI 7700 (Applied Biosystem) was used for mRNA quantification from muscles that underwent concentric and concentric plus eccentric exercises. Briefly, cDNA was added to SYBR Green PCR Master Mix (Applied Biosystems). Samples were amplified in triplicate by using the following thermal cycling conditions: 94°C for 5 min, followed by 40 cycles of amplification at 94°C for 30 s, followed by 60°C for 1 min to allow for denaturing and annealing extension. cDNA quantification was directly related to fluorescence of SYBR Green I dye during 40 cycles of amplification.

Estimation of amplified gene products were normalized to GAPDH (housekeeping gene), compensating for variations in quantity as well as for differences in RT efficiency. We used GAPDH gene as our internal control because expression level of the gene was constant in our samples according to the array data (P = 0.7). All primers were tested for nonspecific amplicons and primer dimers by visualizing PCR products on 2% agarose gels before performing qualitative RT-PCR (qRTPCR). Only primers that did not generate nonspecific products and primer dimers were used for the qRT-PCR assay. The final primer concentration used in the amplification reaction was 500 nM. Gene expression was determined from measurements of the increase in fluorescence, corresponding with amplification and incorporation of SYBR Green I dye during the PCR reaction. The cycle in which fluorescence exceeds the manually set threshold level of detection is termed the “threshold cycle” (Ct). Greater amounts of initial specific template added to the reaction would result in threshold-exceeding fluorescence during earlier cycles (lower Ct value). Thus the Ct value is inversely proportional to the log of the input target-specific cDNA. The Ct value of the internal control gene was used to calculate normalized target gene expression, referred to as delta Ct, to correct for differences between samples. Calculation of relative gene expression was then performed, which allowed for comparison of the gene expression in one sample relative to another. The delta-delta Ct was calculated by subtracting the eccentric delta Ct from the concentric delta Ct. Paired t-test was performed to obtain P values.


Exercise and muscle function. Three male subjects performed timed exercise consisting of chair standing/sitting, where the left leg underwent 300 concentric contractions (rising from a chair) and the right leg underwent 300 concentric and 300 eccentric contractions (rising from and controlled lowering to a seated position) (see methods). Both concentric only (left leg) and concentric plus eccentric (right leg) exercise produced strength losses immediately after exercise (-16 vs. -33%, respectively) (Fig. 1). At 72 h postexercise, the right leg maintained a 9% drop in strength, which was reduced to 3% by 144 h postexercise. The left leg showed no strength decrements between 72 and 144 h postexercise.

Fig. 1.

Loss of strength is greater in concentric plus eccentric exercise compared with concentric exercise alone. The left leg performed 300 concentric knee flexions. The right leg performed both 300 concentric and 300 eccentric knee flexion/extensions. Data are group means during maximal isometric voluntary contraction testing.

Soreness. Muscle soreness measured across time is shown in Table 1. Soreness was not different between legs before exercise (4 mm for the left leg vs. 3 mm for the right leg). Twenty-four hours after exercise, muscle soreness ratings were 62% higher in the damaged (right) leg. This difference was exacerbated at 72 h postexercise (382% difference) and at 144 h postexercise (145% difference).

View this table:
Table 1.

Muscle soreness measures

Gene expression profiling. Gene expression profiling analyses were conducted by using Affymetrix human genome U95Av2 microarrays containing probe sets representing ∼12,000 known genes and ESTs. To reduce the influence of genetic background (interindividual variation), the expression profile of each muscle that underwent eccentric plus concentric contractions was compared with the control (concentric) muscle from the same subject by using Affymetrix Microarray Suite 5.0. The percentage of present calls was found to be 28 ± 5%. Three pairwise comparisons were generated (1 for each subject). This led to the identification of potentially differentially regulated genes (subject 1: 625 “difference” calls; subject 2: 1,125 difference calls; subject 3: 1,096 difference calls). The complete gene list of each subject can be viewed at the website Genes that showed consistent direction of regulation in the eccentric plus concentric leg, relative to the concentric exercise leg, and also showed greater than a fivefold average fold change, were retained for further analysis (Table 2). Our experimental design and small sample size precluded additional statistical corrections for false discovery rates, and determination of paired t-tests generally showed inadequate power. For these reasons, we set very high thresholds for retaining genes for further analysis and focused on those that were confirmed at the mRNA or protein level.

Twenty-eight genes (31 probe sets) were identified as being upregulated specifically in the eccentric plus concentric exercised muscles. No gene was identified as consistently downregulated. The upregulated genes include those involved in cell growth regulation, DNA damage response, stress response, energy metabolism, inflammation, extracellular matrix structure and function, muscle differentiation and signaling, and two genes with unknown functions (Table 2). The genes that showed changes in all three individuals but less than fivefold can be viewed at

c-fos and lipocortin II protein studies. The most dramatic upregulation that we observed was for the immediate early gene and transcription factor c-fos. The strong upregulation of c-fos in human (23.0-fold increase; average of 3 probe sets) was also found in rat muscle exposed to eccentric exercise (19.8-fold increase) (10). We had previously localized the c-fos protein to rat muscle nuclei after eccentric exercise (10).

We also found a 2.8-fold increase in lipocortin II (also called annexin II), which is not reported elsewhere in the exercise-induced muscle damage literature. Lipocortin II was upregulated 1.9-fold in the previously reported rat studies but did not reach the 2-fold change criteria in that publication (10). We opted to confirm the increase of lipocortin II at the protein level because this protein is thought to be associated with vascular remodeling and fibrinolysis. Note that this transcript did not meet the fivefold increase necessary for Table 2.

To investigate whether these particular gene changes identified with microarrays correlated to protein changes, immunostaining of c-fos and lipocortin II was performed. The muscle biopsies obtained after eccentric contractions showed that many more myonuclei stained positive for the c-fos protein than those obtained from muscle from the same subjects having undergone concentric contractions only (Fig. 2). Immunostaining for lipocortin II stained capillaries between muscle fibers after both eccentric and concentric contractions; however, the protein expression difference seen subjectively by immunostaining was less dramatic than c-fos (data not shown). Because lipocortin II was expressed in both the leg that underwent eccentric plus concentric contractions and the leg that underwent concentric contractions, and the change was relatively minor (2.8-fold), immunoblotting was performed to quantify the protein amount. Figure 3 demonstrates that lipocortin II was upregulated 1.7-fold at the protein level at 8 h after eccentric contractions, whereas the other two subjects showed less dramatic protein changes. The average change of the protein in all three subjects is 1.5 ± 0.8-fold (P = 0.2).

Fig. 2.

c-fos is strongly induced in myofiber nuclei after eccentric contractions. Shown are immunostaining of c-fos on muscle biopsies obtained 8 h after concentric (right) and eccentric plus concentric (left) contractions from a single individual. Significantly more c-fos-positive nuclei with higher intensity were observed in the eccentric exercise muscle.

Fig. 3.

Top: lipocortin II protein is strongly induced by eccentric contractions. Shown is immunoblot analysis of lipocortin II in muscle biopsies from a normal human volunteer 8 h after the exercise stimulus. Control is vastus lateralis subjected to 300 concentric contractions (Con), and experimental is 300 concentric plus 300 eccentric contractions (Ecc). Bottom: tubulin-α1 was used as control on the same immunoblot. These data show strong upregulation of the lipocortin II protein in the eccentric exercised leg, consistent with the 2.8-fold upregulation seen at mRNA level by microarray analyses.

Real-time qRT-PCR confirmation. We selected three immune genes, (C-C motif) CCL2, CEBPD, and IL-1 receptor type I (IL-1R1) for real-time qRT-PCR confirmation. Tenascin C, which is known to be involved in tissue, including blood vessel, remodeling and CYR61, an angiogenic factor, were also selected. Because CARP is known to play important roles in both cardiac and skeletal muscle hypertrophy, it was selected for real-time qRT-PCR confirmation. Each gene was tested in all three human volunteers, and the average change was calculated and shown in parentheses in Table 2, adjacent the Affymetrix GeneChip change. Thus five of six genes tested were confirmed by qRT-PCR by using a P < 0.05 cutoff, with the changes similar to those found by Affymetrix microarrays (Table 3). CYR61 did not reach the statistically significant level (P < 0.05) due to a much larger change in one of the three subjects (52-fold vs. 6- and 4-fold). The array data showed the same large change in the same subject (22-fold vs. 5- and 3-fold). This is very often seen when the expression level of the gene is close to the background (no expression in the control leg) and can cause an extremely large change when it is upregulated.

Comparison between human and rat. Chen et al. (10) used a rat model to study the effects of eccentric exercise by expression profiling. In that study, rat tibialis anterior muscles were expression profiled at 1 and 6 h after eccentric contractions by using Affymetrix rat genome U34A microarrays containing ∼8,000 genes and ESTs. By comparing our results to those from the rat exercise model, 9 of 18 genes represented on both human and rat arrays showed similar changes in both the human and rat model, including heparin-binding EGF-like growth factor, c-myc, heat shock proteins 40 and 70, GADD45A, CARP, interferon-related developmental regulator 1 (also known as PC4), activating transcription factor 3 (ATF3), and c-fos (Table 2). By comparing human data to those previously identified in the rat model, we identified more genes involved in immune responses and vasculature remodeling that appeared more specific for humans eccentric contractions (Table 2).


Exercises biased toward eccentric, or muscle lengthening, contractions cause repairable muscle damage, whereas exercises using primarily concentric contractions cause little damage. We recently reported expression profiling of rat muscle at 1 and 6 h after eccentric exercise to determine what transcriptional and translational changes take place early in exercise-induced muscle damage (10). In this present study, we used a more physiological model in human volunteers with a longitudinal design where expression changes were defined only within an individual and then only consistent changes that survived all three individual comparisons were selected. An additional advantage of this human study is the more mature nature of the human genome resources compared with those of the rat.

The present study used an exercise model where one leg performed eccentric plus concentric contractions, and the contralateral leg performed concentric contractions of the knee extensor muscles to examine the differences in molecular signaling between eccentric and concentric exercise modes. Subsequent gene microarray analysis using intrasubject comparisons between legs was performed to determine which changes in gene expression were specific to damage-inducing exercise stress. Although the leg that performed both concentric and eccentric contractions experienced twice the number of contractions as the leg that performed concentric contractions only, we do not believe that the differences seen in transcription are attributable to this difference. Concentric contractions are associated with metabolic muscle fatigue but not with muscle damage, as indicated by ultrastructural damage (42) or pain/tenderness (43). Although we are unaware of any studies examining a large number of genes with relation to exercise mode and number of contractions, a few studies have indicated that concentric contractions have little effect on some well-known intracellular pathways in muscle. Wretman et al. (60) found that neither concentric contractions nor mild static stretch caused increases in p38 phosphorylation, indicating that passive mechanical strain or contraction-induced metabolic changes do not upregulate this pathway. Additionally, Martineau and Gardiner (39) demonstrated that peak tension during contractions was a better predictor of the c-Jun NH2-terminal kinase pathway activity after exercise than time-tension integrals in rat muscle, indicating that the greater force levels experienced during eccentric contractions would have more of an effect on pathway activity than the time during which muscle is active. In the present study, we found increases in several transcription factors known to be downstream of these pathways such as c-fos. Therefore, although it is possible that some of the changes found in this study could be sensitive to the number of contractions performed, the majority of the changes are likely due to differences between contraction modes (i.e., strain, peak tension).

Because our study of three normal human volunteers precluded extensive statistical evaluation of the microarray data, we set stringent thresholds for analysis of our data and focused on verification of eight differentially expressed genes at the mRNA or protein levels. All data (.dat, .cel, and analysis files) are available at our web site, which allows further analyses and comparisons ( and less stringent interpretation.

Increased expression of transcription factor c-fos. c-fos has been extensively studied in a series of animal models of eccentric contractions; however, ours is the first to use data from human muscle. We found c-fos to be markedly upregulated between 4 and 8 h after damaging exercise, which is at later time points than that seen in most rodent studies. Our data concerning c-fos may either represent a more prolonged upregulation than previously described or may represent the second phase of the biphasic c-fos response similar to that seen by Goldspink et al. (18) in actively stretched rabbit muscle.

Increased expression of vascular and extracellular matrix remodeling proteins. Previous studies have identified lipocortin II as being located in the interstitial space and in the capillaries of cardiac tissue and as being upregulated in failing heart muscle (5, 54). The increase of lipocortin II in failing hearts suggests a potential role in the regulation or modulation of proteins responsible for myocardial and vascular remodeling (59). Lipocortin II mRNA was upregulated 2.8- and 1.9-fold in the human and rat studies, respectively. We hypothesized that the observed increase in lipocortin II mRNA could likewise reflect microvasculature damage. To test this, we studied lipocortin II protein localization by immunostaining and immunoblotting. Like c-fos, lipocortin II mRNAs were upregulated at all three time points after exercise. Immunoblot analysis showed that the lipocortin II protein levels were most dramatically changed in one subject (Fig. 3) at 8 h postcontraction and were less clearly differentially regulated in the other two subjects. This could reflect variability in protein levels, or timing of protein expression, and will require further study.

CYR61 is a heparin-binding, extracellular, matrix-associated protein that induces angiogenesis in vivo. It supports cell adhesion, promotes cell migration, and enhances growth factor-stimulated mitogenesis in fi-broblasts and endothelial cells (19, 21, 32). CYR61 was shown to be regulated by mechanical stretch in cultured bladder smooth muscle cells and stimulated revascularization in ischemic limbs (16, 56). Tenascin C has restricted expression in normal skeletal muscles, primarily at the extracellular matrix and vasculature. It is a modular and multifunctional extracellular matrix glycoprotein that is exquisitely regulated during embryonic development and in adult tissue remodeling (25, 28). Recent studies showed that reloading after unloading leads to upregulation of tenascin C in muscle (17, 24). The study of tenascin C knockout mice suggested that tenascin C is also important in the formation and stabilization of the neuromuscular junction (12). We found a 10.1-fold increase of CYR61 (20.9-fold by qRT-PCR) and a 5.3-fold increase of tenascin C (11.2-fold by qRT-PCR) in the muscles that underwent eccentric exercise. The upregulation of lipocortin II, CYR61, and tenascin C were very likely induced by the mechanical myofiber damage and ischemic conditions caused by the eccentric exercise and may play an important role in extracellular matrix remodeling of the myofiber, vasculature, and neuromuscular junction.

Increased expression of genes involved in the inflammation response. Inflammation is known to play a major role in exercise-induced muscle damage (3638, 46, 53). Inflammation after muscle injury occurs to clear debris from the injured area in preparation for regeneration. The resulting cytokines generated by muscle fibers and immune cells are thought to play important roles in muscle degeneration/regeneration and growth. We found large increases in several gene transcripts involved in the inflammatory response, many that have not been previously described in relation to exercise-induced muscle damage and were not shared with the rat model. Two cytokines previously examined in muscle after eccentric exercise are IL-1 and IL-6 (36, 44, 53, 58). IL-6 is one of the cytokines suggested to be involved in the degeneration/regeneration process of skeletal muscle. This cytokine expresses at low levels in normal skeletal muscle but is dramatically upregulated in both injured myofibers and inflammatory mononuclear cells located at the injury site after muscle damage (8, 29, 30, 33, 34). We did not detect changes of either IL-1 or IL-6 in our study but found strong increases in three other inflammatory proteins, namely CCL2 (8-fold by microarrays; 24-fold by qRT-PCR), CEBPD (7-fold by array; 18-fold by qRT-PCR), and IL-1R1 (6-fold by array; 20-fold by qRT-PCR). CCL2, also known as monocyte chemoattractant protein-1, has been shown to be expressed in both monocular cells and muscle fibers in mdx mice (47). CCL2 is regulated by IL-6 and plays a significant role in the migration of immune cells and the regulation of Th1/Th2 activities (3, 51, 52). Sekine et al. (50) showed that CCL2 is involved in the proinflammatory effect of insulin in the insulin-resistant state. Interestingly, CCL2 has been found to be regulated by CEBPD in vascular smooth muscle cells (VSMC), and CEBPD was also upregulated in our study. CEBPD is a nuclear transcription factor that regulates cellular growth and differentiation. It has been shown to regulate proinflammatory gene expression in VSMC. Targeted overexpression of CEBPD evokes high levels of PDGFαR gene expression, susceptibility to VSMC growth, and proliferation of VSMC to PDGF. The studies suggested that it might play an important role in VSMC growth during the process of vascular remodeling (50, 61).

The IL-1R binds both IL-1α and IL-1β, initiating IL-1 signal transduction. A naturally occurring receptor antagonist (IL-1RA) also binds IL-1R but does not initiate signal transduction. Malm et al. (38) reported increased IL-1β but not IL-1α 6 h after unilateral eccentric cycling exercise. Based on our data, there seems to be a modulation in the IL-1 receptor mRNA from 4 to 8 h after high-force eccentric contractions, suggesting a mechanism by which the muscle increases its receptor availability/sensitivity to IL-1 in the hour after damaging exercise. This, coupled with the protein findings of IL-1β in humans (38), suggests strong activation of the proinflammatory IL-1 signaling cascade after a bout of eccentric contractions.

Genes involved in cell growth regulation and stress response in both human and rat exercise models. Functional clusters previously studied in the rat model (10) included genes involved in promoting cell growth and proliferation, antiproliferation genes, and stress-response genes. Within these groups, most of the transcriptional responses to eccentric contractions were shared between the human and rat studies (Table 2). Shared growth-related genes include heparin-binding EGF-like growth factor, c-myc, PC4, ATF3, and c-fos. c-fos,c-myc, and ATF3 were previously identified in the study of the growth responses of cultured fibroblasts after serum stimulation (23). Novel growth-related genes identified in our human study and showed no change in the rat model include CYR61. Two of the five antiproliferation genes identified in the rat study (10) were also identified in the present study (GADD45 and CARP). GADD45 is a DNA damage response molecule that has also been shown to upregulate during overload-induced hypertrophy (9). It is a well-characterized protein involved in cell cycle arrest at G1/S and/or G2/M through direct inhibition of cdc2a kinase (1, 26, 57). Induction of CARP has been shown to be an early marker of cardiac and skeletal muscle hypertrophy (2, 9, 27). Recent studies have shown that overexpression of CARP acts to decrease proliferation in smooth muscle vascular cells, suggesting that it may function to maintain the differentiated state in heart and skeletal muscle during periods of growth (31).

In summary, our human eccentric exercise data are consistent with the model proposed in recent rat models, where the damaging contractions induce a series of stress response genes, with a simultaneous induction of specific growth-promotion genes (hypertrophy) and antiproliferation genes (retains postmitotic nature of the myofiber) (10). Our human data considerably extend the rat data by showing strong induction of immune modulatory genes, in addition to some vascular and myofiber remodeling genes, most of which were not shared with the rat model, indicating that species differences may exist between rat and human muscle.


Y. W. Chen is supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant 5R21 AR-048318.


We thank Hong Qian for assisting with array data confirmation.


  • 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.


View Abstract