Resistance training using lengthening (eccentric) contractions induces greater increases in muscle size than shortening (concentric) contractions, but the underlying molecular mechanisms are not clear. Using temporal expression profiling, we compared changes in gene expression within 24 h of an acute bout of each type of contractions conducted simultaneously in the quadriceps of different legs. Five healthy young men performed shortening contractions with one leg while the contralateral leg performed lengthening contractions. Biopsies were taken from both legs before exercise and 3, 6, and 24 h afterwards, in the fed state. Expression profiling (n = 3) was performed using a custom-made Affymetrix MuscleChip containing probe sets of ∼3,300 known genes and expressed sequence tags expressed in skeletal muscle. We identified 51 transcripts differentially regulated between the two exercise modes. Using unsupervised hierarchical clustering, we identified four distinct clusters, three of which corresponded to unique functional categories (protein synthesis, stress response/early growth, and sarcolemmal structure). Using quantitative RT-PCR (n = 5), we verified expression changes (lengthening/shortening) in SIX1 (3 h, −1.9-fold, P < 0.001), CSRP3 (6 h, 2.9-fold, P < 0.05), and MUSTN1 (24 h, 4.3-fold, P < 0.05). We examined whether FBXO32/atrogin-1/MAFbx, a known regulator of protein breakdown and of muscle atrophy was differentially expressed: the gene was downregulated after lengthening contractions (3 h, 2.7-fold, P < 0.05; 6 h, 3.3-fold, P < 0.05; 24 h, 2.3-fold, P < 0.05). The results suggested that lengthening and shortening contractions activated distinct molecular pathways as early as 3 h postexercise. The molecular differences might contribute to mechanisms underlying the physiological adaptations seen with training using the two modes of exercise.
- temporal profiling
it is well known that repeated high-intensity contractions results in muscle hypertrophy with greater increases in mass and strength reported for training regimens including lengthening contractions (so called eccentric muscle contractions), although at least one study suggests there is few or no difference in the responses to the two types of contractions (23, 31, 34, 36, 38, 42, 55, 58, 61). Responses to lengthening contractions may result in indices of muscle damage including Z-band streaming, loss in force generating capacity, soreness, swelling, and leaking of sarcoplasmic proteins into the blood (25, 47, 48, 53, 60). Nevertheless, some reports in the literature suggest that the changes listed above may be more related to rapid adaptation than damage (66–68). Proposed mechanisms to explain the long-term differences after training with lengthening and shortening contractions include differences in rates of and types of neural activation, protein turnover, and intracellular signaling. There is conflicting evidence on possible differential neural activation (35, 36, 57), whereas no marked differences have been shown for muscle protein synthetic rates measured up to 48 h after lengthening or shortening contractions even when the mechanical work done is equalized between the two modes (51).
Previously, Cuthbertson et al. (19) carried out a study to determine if PKB and p70(S6k) signaling after exercise might be different between the modes of exercise. However, there were no differences in phosphorylation of PKB and p70(S6k) or in myofibrillar or sarcoplasmic protein synthesis between the two types of contractions at the time points (3, 6, and 24 h) studied. Thus it seems possible that although changes in protein synthesis resulting in net anabolism do eventually occur, such changes are not initiated via alterations in translation of preexisting mRNA but may be the result of the altered expression of muscle genes i.e., an increase of mRNAs for a variety of proteins involved in hypertrophy and remodeling.
In an attempt to elucidate the molecular mechanisms underlying the observed phenotypic differences resulting from the two modes of contractions, gene expression profiling has been performed in rats (15). In that study, gene expression was examined after an acute bout of high-intensity lengthening contractions at 1 and 6 h postcontraction. Results from that work identified genes associated with cell growth-related pathways that were activated primarily at 6 h with very few gene expression changes at 1 h postcontraction. Afterwards, molecular responses to lengthening vs. isometric contractions in mouse muscle at 48 h postexercise were defined (5). Using high-density microarrays the investigators identified 36 genes (including five muscle-specific genes) that were upregulated to a greater extent after lengthening contractions than after isometric contractions. In a study of exercise by human subjects in whom a biopsy was taken at a single time point 6 h postexercise (14), there was concordance between the observed changes and previous results from rodents.
The aim of the present study was to define temporal patterns of gene expression changes in human muscle following lengthening vs. shortening contractions. The strength of this design is that the temporal expression pattern of each gene at 3, 6, and 24 h after the two modes of exercises (concentric/shortening or eccentric/lengthening contractions) was determined in muscle biopsies from the same individuals. The design in this study allowed the expression profiling data obtained from each subject to be normalized to their own baseline, thus avoiding potential problems arising from the considerable interindividual variability often observed.
MATERIALS AND METHODS
The subjects (n = 5) for this study came from a group of eight subjects described in a previous paper (19), they were 24 ± 2 yr, 1.74 ± 0.02 m, 73.2 ± 3 kg, and body mass index 24.2 ± 0.4 kg/ m2. The three subjects for microarray analysis were 21 ± 0.58 yr, 1.75 ± 0.03 m, 73.7 ± 4 kg, and body mass index 24.1 ± 0.7 kg/ m2. These subjects were habitually active at a recreational level. They were instructed to adhere to their usual diet and to refrain from strenuous physical activity for 2 days prior to the study. The subjects were informed of the experimental protocol both verbally and in writing before giving their informed consent. The study protocol was approved by the Tayside Ethics Committee and was carried out according to the Helsinki Declaration. Three of the eight subjects were randomly selected for microarray analysis. For quantitative (q) RT-PCR assays, two additional individuals were selected randomly to make a total five individuals.
This experimental protocol was previously reported in detail (19). In brief, the subjects attended the laboratory having fasted from 2000 on the previous evening. Muscle biopsies were taken from the quadriceps muscle preceding exercise from either leg and 3, 6, and 24 h after exercise from both legs (21). All biopsies were taken via separate incisions and made from distal to proximal areas of the quadriceps. After 4 h at rest, the subjects began the exercise regiment that involved stepping up with one leg onto a box (height = 0.55 m) and stepping down with the contralateral leg at ∼0.5 Hz for 6 min, carrying a pack with 25% of their body mass. The “up/concentric” and “down/eccentric” legs were randomly assigned by the investigators. After 2 min rest, they performed two additional 3-min bouts so the protocol lasted for a total of 16 min, 12 min of which was during exercise. On completion of the exercise bout, all subjects reported being fatigued, were unable to continue further, and had to be helped to a seat. After exercise, in the 2 h preceding each biopsy, subjects were given a 500-ml drink containing 45 g of essential amino acids and 135 g of sucrose. The amino acid/sucrose solution was given an initial aliquot of ∼140 ml and then six subsequent aliquots of 70 ml every 20 min as previously described (19).
An Affymetrix custom microarray chip containing 4,601 probe sets representing 1,150 known genes expressed in muscle and 2,075 expressed sequence tag (EST) clusters from a muscle sequencing project was used for this study (9). The expression profiling procedures and quality control are as previously reported (2, 16, 29). All microarrays were required to fulfill the quality control metrics to be analyzed. Briefly, total RNA was isolated from the muscle biopsies using TRIzol reagent (Invitrogen) and processed for production of biotinylated cRNAs and hybridization to the microarrays. Florescent images were obtained using Affymetrix GeneChip Scanner for data analysis.
Hybridization signal of each probe set was determined using MAS 5.0 (Affymetrix) and dCHIP (43). After the absolute analysis using MAS 5.0 and dCHIP, the gene expression levels were imported into GeneSpring software. Data filtering was done by retaining only those probe sets that showed at least two MAS 5.0 “present calls” across all 24 profiles. This resulted in retention of 2,382 (51%) probe sets on the MuscleChip microarrays. All the profiles from each subject were normalized to their own baseline (0 h time point) prior to additional statistical or clustering analysis. Normalized data were then compared for differential gene expression analysis at given times between muscles that underwent shortening and lengthening contractions. Results are reported as fold change of lengthening/shortening contractions. To reduce false positives, only genes that were statistically significant (Welch t-test, P < 0.05) using expression levels determined by both MAS 5.0 and dChip algorithms were retained for hierarchical clustering analysis, whereas genes that were statistically significant using expression values determined by only one or the other algorithm were not included in our analysis. It has been shown that using MAS 5.0 and dChip algorithms in combination, results in a very high level of stringency in reducing false positives, though it also increases false negatives (20). Due to the small sample size (n = 3), proper variance comparisons would be speculative; therefore, the Welch t-test (P < 0.05) was used to calculate the probabilities of significant gene expression changes of the 2,382 probe sets with each time point analyzed separately. Since the study used a longitudinal (in time) design and the temporal data were analyzed to categorize genes into functional groups (as described in the following paragraph), we did not apply multiple testing corrections to the data to avoid overstringency.
Hierarchical clustering analysis was performed using the Gene Tree algorithm in GeneSpring. Expression values generated using MAS 5.0 were used. The genes were clustered based on the ratio between the lengthening and shortening contractions using standard correlation as the similarity measure. Hierarchical clustering of temporal data has previously been used in profiling studies and is considered one of the most useful tools for analyzing time series expression profiling data (4, 69). In this study, four major clusters were identified visually.
All profiles are publicly accessible via National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) (accession no. GSE7286) (http://www.ncbi.nlm.nih.gov/geo/) and the Children's National Medical Center PEPR (Public Expression Profiling Resource) (http://pepr.cnmcresearch.org) (13). Online data queries are also publicly accessible, both through a single gene query via PEPR (http://pepr.cnmcresearch.org/home.do) and via NCBI GEO DataSets.
Reverse transcription and TaqMan quantitative PCR analysis.
Reverse transcription and qRT-PCR was performed as previously described (14). Briefly, total RNA (3 μg) was reverse transcribed to cDNA using oligo dT primer (0.05 μg/μl) and Superscript II Reverse Transcriptase (Invitrogen), cDNA was amplified in triplicate in SYBR Green PCR Master Mix (Applied Biosystems). The thermal cycling conditions include 94°C for 5 min, followed by 40 cycles of amplification at 94°C for 30 s, followed by 60°C for 1 min. Each reaction contained 10 ng of template cDNA with 75 nM primer. TaqMan PCR primers were designed using a Primer Express program v. 1.01 (Applied Biosystems). Primer sequences used for human F-box only protein 32 (FBXO32) were (forward) 5′-TTTCCTGGAAGGGCACTGAC-3′, (reverse) 5′-TATTCCCAGCTCTCCAGTCAGCA-3; sine oculis homeobox 1 (SIX1) (forward) 5′-GGAGGCCAAGGAAAGGGAG-3′, (reverse) 5′-GTCTGGACTTTGGGGAGGTGA; cysteine and glycine-rich protein 3 (CSRP3) (forward) 5′-CAAGCCTTGGCACAAGACCT, (reverse) 5′-AGGCCTCCAAACCCAATACC; and musculoskeletal embryonic nuclear protein 1 (MUSTN1) (forward) 5′-ATGCGAGAGTGTGAGCAAGCT, (reverse) 5′-TTCTCAGCCGAAGACACTCTTGT-3. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as internal control. The primers of GAPDH were purchased from Applied Biosystems. All primers were tested for nonspecific amplicons and primer dimers by visualizing PCR products on 2% agarose gels prior to qRT-PCR, as well as by dissociation curve analysis following the RT-PCR assays. Additionally, no template controls were checked for each reaction by melting curve, agarose gel, and ΔRn. The 2^ΔΔCT method was used to determine fold differences and a paired t-test was used (P < 0.05) to test statistical significance. As no probe set was available for FBXO32 and no other data existed for a comparison directly to the zero time point as was available for all genes represented on the Muscle Chip, a paired t-test was used to analyze FBXO32 expression level at each time point compared with the zero time point.
Using our strategy of combined algorithms, we identified 51 transcripts represented by 54 probe sets that were differentially regulated (35 transcripts upregulated and 16 transcripts downregulated) when comparing muscle that had been involved in lengthening or shortening exercise. Among the 51 transcripts, two transcripts were ESTs resulting in 49 known genes identified. Among the 51 transcripts, 8 were differentially expressed at 3 h postexercise, 11 transcripts at 6 h postexercise, and 35 transcripts at 24 h postexercise. Three genes, SIX1, crystallin alpha B (CRYAB), and pyruvate kinase 2 (PKM2) were significantly upregulated at two time points (Tables 1, 2, 4). The change in expression of only eight genes at the 3-h time point is consistent with previous reports that few genes change expression early after exercise (15, 39). There was no clear pattern of increased or decreased expression at any given time point, and only the 24-h time point revealed a particular gene category (sarcomere/structural) that dominated the changes in expression seen at that time. For comparisons with the zero time point, Supplemental Table S1 and Supplemental Fig. S1, A–D, are available as online material.1
Using unsupervised clustering to gain an overall picture of temporal changes, genes that shared similar temporal expression patterns being grouped together, we identified four distinct clusters (Fig. 1). The genes belonging to three of the four clusters belong to a specific functional categorization, based on gene ontology classifications and known function previously reported. The order of genes in Fig. 1 is representative of the gene order in Tables 1–4. Cluster A consists primarily of genes associated with the translation of proteins (6/11), many of them encoding ribosomal proteins, including ribosomal proteins S17, S20, S27, S28, L15, and SA (Fig. 1, Table 1). Cluster B contains genes downregulated after lengthening contractions compared with shortening at each time point (10/11) but did not display any clear functional categorization (Fig. 1, Table 2). Cluster C contains genes involved in early growth and also heat shock proteins (8/10) (Fig. 1, Table 3), whereas cluster D was noteworthy in containing a large number of muscle structural genes (6/14) with one EST (Fig. 1, Table 4). Four transcripts, tetraspanin 8, metallothionein 2A, synaptogyrin 2, and an EST (GenBank number F37840), could not be simply categorized in any clearly defined cluster or functional group (Table 3).
Differences in expression of protein synthesis associated genes with mode of exercise were modest.
It is notable that cluster A (protein synthesis) contains only one gene (of 11) that exhibited a change of >1.5-fold, whereas the other clusters contain few genes with a <1.5-fold change (Tables 1–4). The data suggest only modest differences due to transcriptionally driven alterations in protein synthesis between the two modes of contraction up to 24 h postexercise.
Alterations in both protein synthesis and degradation can influence the overall muscle net protein balance. FBXO32, also known as atrogin-1/MAFbx, is an E3 ubiquitin ligase that exhibits a very muscle-specific pattern of expression. Atrogin-1 has been suggested to play a critical role in maintenance of muscle mass through regulating targeting of muscle proteins for degradation through ubiquitinylation (8, 27); thus we decided to examine whether FBXO32 is differentially expressed in muscle after the two modes of contractions. As there was no probe set for FBXO32 on the MuscleChip, we performed real-time qRT-PCR to determine the expression level of the gene. We found that FBXO32 was downregulated by 3.4-fold (P < 0.05) at 6 h postexercise in the lengthening-exercised muscle in relation to its expression in the shortening-exercised muscle (Fig. 2D, Table 5). Additionally, FBXO32 decreased expression in the lengthening-exercised muscle at 3 h (2.7-fold, P < 0.05), 6 h (3.3-fold, P < 0.05), and 24 h (2.3-fold, P < 0.05) in relation to the zero time point, whereas no significant change occurred at any time point for the shortening-exercised muscles. This result is similar to findings in rat tibialis anterior muscle following eccentric contraction, mouse plantaris muscle in response to mechanical overload (John McCarthy and Karyn Esser, unpublished data) and is the first report to our knowledge of the downregulation of FBXO32 after lengthening contractions in human muscle.
SIX1 expression was downregulated after lengthening exercise at 3 and 6 h.
Two of the genes at the 3-h time point are known transcription factors, SIX1 and zinc fingers/homeoboxes 2 (ZHX2). When clusters were considered, SIX1 belonged to the cluster with genes downregulated in the muscles after lengthening rather than shortening contractions (cluster B). SIX1 is a homeodomain-containing transcription factor, which is expressed in skeletal muscle during development and plays an important role in myogenesis as well as in fiber-type switching (28, 44, 59). The array data showed a decreased expression of SIX1 at both 3 and 6 h (Fig. 2A). When checked by qRT-PCR, it was found that the gene was 2.1-fold downregulated (P < 0.05) in the muscles that underwent lengthening compared with those that underwent shortening contractions at 3 h postexercise (Table 5).
Stress and early growth response genes were upregulated by lengthening exercise.
Consistent with our previous human and animal studies, one cluster (cluster C) contained several genes involved in early growth response. Heat shock proteins are normally expressed during developmental periods and in response to environmental stress (mechanical, thermal, chemical). In the present study, heat shock 27-kDa protein 1 (HSPB1) and heat shock 22-kDa protein 8 (HSPB8) showed an increase at 6 and 24 h (Table 3). Additionally, CRYAB showed an increase (cluster D) at 6 and 24 h (Table 4). Increases in CRYAB and other heat shock proteins have previously been reported in response to exercise (3, 52).
An enzyme of major importance for cAMP hydrolysis, phosphodiesterase 4D (PDE4D), was increased twofold for muscle that had undergone lengthening rather than shortening contractions. Additionally, two genes involved in cell growth, four and a half LIM domains 1 (FHL1) and nexilin (NEXN), were upregulated in the muscles that underwent lengthening contractions. Both these genes, which modulate cell shape, have been shown to increase expression during muscle cell hypertrophy in animals and during muscle and nonmuscle cell migration (46, 49, 64). Groups of genes that modulate cell shape such as these (together with CSRP3) are believed to be important in repairing muscle and likely stimulating growth, though the exact function in response to exercise is currently not known (56).
Evidence of sarcomerogenesis induced by 24 h after the lengthening contractions.
Cluster D is the largest cluster and consists primarily of structural components of muscle, such as synaptopodin 2 (SYNPO2), PDZ and LIM domain 3 (PDLIM3), myomesin 2 (MYOM2), LIM domain containing preferred translocation partner in lipoma (LPP), CSRP3, and myosin binding protein H (MYBPH). The pattern of this cluster (Fig. 1, cluster D) displays a small increased expression as early as 3 h with a continued increase and highest expression at 24 h. This pattern suggests that transcription of these genes may be activated as early as 3 h postexercise, although the differences between muscles experiencing different modes of contractions was most obviously significant at 24 h (Table 4). Using real-time qRT-PCR, we verified the results from the array and these findings define changes at both 6 and 24 h postexercise for two genes, CSRP3 and MUSTN1 (Table 5).
CSRP3 is part of the Z-disk of sarcomeres, where it is bound to telethonin and is thought to be involved in the regulation of the mechanotransduction/mechanosensory machinery of skeletal muscle by influencing other myogenic factors including myogenic differentiation factor (MYOD), myogenin, and muscle regulatory factor (MRF) 4 (26, 40). MUSTN1 was previously identified in a study examining bone remodeling in mice after fracture (45). The gene is highly expressed during embryogenesis and inactivated in most adult tissues with the exception of skeletal muscle, bone, and tendon. The function of MUSTN1 in muscle is unknown. By using real-time qRT-PCR, we verified that CSRP3 was 2.9-fold (P < 0.05) higher in the lengthening muscles at 6 h and 4.2-fold (P < 0.01) higher at 24 h. MUSTN1 was found 2.2-fold upregulated at 6 h (P > 0.05) and 4.3-fold (P < 0.05) upregulated in muscles previously contracted during lengthening at 24 h. Although the changes reported by microarray analysis in CSRP3 and MUSTN1 genes reached significance at one time point only (Fig. 2, B and C), results from qRT-PCR showed that both genes were significantly upregulated at both 6 and 24 h time points.
We designed our study to elucidate the temporal changes in the molecular responses to lengthening or shortening contractions. The aim of this project was to begin to identify contraction-specific molecular signatures that underlie phenotypic alterations with acute exercise. To our knowledge, this is the first study to use multiple samples from the same individual for expression profiling after exercise. The results of this analysis identify very distinct gene expression differences in the responses in the muscles undergoing the two modes of contraction.
The results of our temporal clustering analysis for each contraction type revealed differential responses in four major clusters of genes, sharing a similar pattern of expression. Three of the clusters were representative of specific functional categories (protein synthesis, early growth and stress response, and structural components of muscle), suggesting that genes belonging to each functional group shared common or coordinated patterns of change according to the previous mode of exercise. The fourth cluster exhibited a consistent decrease in gene expression in response to lengthening exercise but could not be functionally classified using current knowledge. The temporal changes observed in the clustering analysis should provide insights of the molecular mechanisms of the physiological effects of the two types of exercise. Additionally, we reported (and verified by real-time qRT-PCR) marked differences of expression of individual genes (FBXO32, SIX1, CSRP3, and MUSTN1) between the muscles that had undergone lengthening vs. shortening contraction.
Genes likely to regulate the translation of proteins showed only a slight change after either lengthening or shortening contractions (Fig. 1, cluster A). All of the ribosomal genes showed changes <1.5-fold when lengthening contractions are compared with shortening (Table 1), an increase in contrast to that observed in the rates of myofibrillar and sarcoplasmic protein synthesis rates in these same human subjects reported previously (19) where there was upregulation of synthesis rate without differences between legs. These findings indicate that although protein synthesis may not be differentially altered in response to either acute lengthening or shortening contractions, there may be an increase in translational efficiency in response to chronic training utilizing lengthening contractions due to the increase in protein synthetic machinery. Increases in translation of the ribosomal proteins may also be occurring, which would be invisible using the currently applied methods.
Our data suggest that slight increases at the molecular level of protein synthesis may explain some of the small differences seen between lengthening and shortening contractions at 4–8 h after exercise (51). Whether the differences of the protein synthesis genes between the two types of exercise will widen at a later time point or whether, though subtle, a persisted upregulation of these genes will contribute to the differences in hypertrophy needs to be further investigated.
In addition to protein synthesis, protein degradation plays a major role in regulating protein turnover in tissues. Specific skeletal muscle proteins are targeted for degradation by FBXO32 also known as atrogin-1/MAFbx. FBXO32 was identified in muscle wasting and atrophy as unique to striated muscle and was concurrently identified upregulated in various models of skeletal muscle atrophy, including denervation, immobilization, and hindlimb suspension (8, 27). FBXO32 contains an F-box domain that links its specific protein substrates to be ubiquitinylated with the apparatus of the E3 and ubiquitinylation machinery and might thereby promote protein degradation (12). Although it is recognized that acute strenuous resistance exercise promotes muscle protein breakdown (7), we found (using qRT-PCR) that FBXO32 was downregulated at the 3-, 6-, and 24-h time points in the lengthening-exercised muscle in relation to the zero time point and at 6 h in relation to the shortening muscle. There may not be a one-to-one relationship between FBXO32 expression and bulk protein breakdown, which may explain the results. The fact that both muscles of both legs were exposed to the same nutritional environment (in fed subjects, see Ref. 19) cannot explain the differential responses in the muscles exercised in different modes. The exact role of FBXO32 in this regard remains to be determined but will be an intriguing area of research.
Recently it was reported that although isometric resistance training offset the increased expression of FBXO32 seen in mouse model of muscle atrophy (30), it did not abolish the loss of muscle mass; it may be that additional stimulus, such as lengthening contraction, is required to reverse the expression in FBXO32 and abolish the atrophy. In the only extant report concerning human muscle, of an acute bout of resistance training in trained endurance or strength-trained athletes, there were no significant changes in FBXO32 expression by 3 h after exercise. (18). This may possibly be a reflection of training status: in our subjects unaccustomed to high levels of training, we detected significant decreases in FBXO32 expression at 3-, 6-, and 24-h time points.
Though most genes increased expression in this study, it is notable that cluster B actually decreased in response to lengthening contractions. These genes are of diverse functions with little known about them except for SIX1. The SIX1 gene was initially identified in Drosophila as an essential gene in eye development (17). SIX1 is part of a family of proteins characterized by a protein-protein interaction (Six) domain and a DNA binding homeodomain, which for SIX1 binds to myocyte enhancer factor (MEF) 3 binding sites. MEF3 sites have been proposed to be linked to regulation of fast twitch-specific genes in skeletal muscle (59). Indeed, SIX1 has been implicated in fiber-type switching from the slow to fast phenotype (28). SIX1 is expressed in skeletal muscle during development and results from SIX1 knockout mice show multiple organ defects including substantial impairments in muscle development (44). This same study demonstrated the requirement of SIX1 for myoblast proliferation in vitro and differentiation of muscle cells during development (44). SIX1 is known to act as a repressor and activator of many genes, including MYC, cyclin A1, myogenin, MYOD, and myosin heavy chain in chicken somites (32, 44). We reported a twofold decreased expression of SIX1 at 3 and 6 h (Fig. 2A). This is one of few genes to show differential expression this early after exercise and is likely an upstream regulator of gene expression changes and possibly fiber-type changes (11, 28). Exact targets were not identified in the current study, thus further studies will be required to determine its role.
In regards to cluster C and cell growth, we found that heat shock proteins, HSPB1 and HSPB8 and CRYAB, were upregulated at 6 and 24 h in response to the lengthening contraction (Tables 3 and 4). Heat shock proteins are normally expressed during developmental periods and in response to environmental stress (mechanical, thermal, chemical). They contribute to protein stabilization during periods of cellular damage, saving proteins from unfolding, and contribute to protein folding during periods of increased protein synthesis, which would likely be a result of the damage incurred due to lengthening exercise and the additional need of new proteins in the damaged muscle.
An important consideration in the interpretation of this study is the comparison of lengthening to shortening contractions. The original intent of the study was to compare the two types of contractions, and the data have been interpreted in this fashion. However, a significant change can be the result of a decreased expression level in response to one type of contraction, while the opposing contraction caused an opposite response. The only instance in which a gene is reported to increase in response to lengthening contractions but is due to a decrease in shortening contractions is PDE4D. To aid in the interpretation of the data, Supplemental Table S1 has been included to show actual expression changes in response to lengthening and shortening contractions. In response to acute contractions, muscle cytosolic cAMP, which is a ubiquitous mediator of intracellular signaling events, increases, pari passu with increased expression of phosphodiesterases, which in turn metabolizes cAMP to 5′-AMP (6, 65). At 3 h, PDE4D displayed a more than twofold increased expression when lengthening contractions were compared with shortening. Yet it was actually a decrease in expression of the shortening leg relative to the zero time point that caused the significant difference between the two legs (Supplemental Table S1). The finding of no change in the lengthening leg is important in that muscle adaptation will ultimately be a balance of increased and decreased expression of genes in certain pathways and, though speculative, could contribute to the reinforcing effects of repeated bouts of exercise in causing hypertrophy, as PDE4D has been linked both to cellular desensitization to cAMP and cardiomyocyte hypertrophy (22, 62). Additionally, FHL1 and NEXN showed increased expression; both genes that have been shown to increase expression during cell hypertrophy, and both modulate cell shape with FHL1 producing hyperelongation of differentiating muscle cells and with NXN mediating cell motility through F-actin binding (46, 49, 64). Similar to PDE4D, the apparent increase in FHL1 is partly due to a decrease in response to shortening contractions (Supplemental Table S1). Groups of genes that modulate cell shape such as these and CSRP3 are believed to be important in repairing muscle damage (56). Thus our results for this cluster of genes extend and confirm previous reports of their importance in the adaptive response to exercise, especially that in the lengthening mode.
The largest cluster (D) identified after hierarchical clustering analysis consists primarily of structural muscle genes, likely involved in sarcomerogenesis, such as SYNPO2, PDLIM3, MYOM2, LPP, CSRP3, and MYBPH. Though their mechanistic significance in relation to the exercise modes used is unknown, some of the genes in this group are known to translocate from their structural position in the cytoplasm into the nucleus and affect gene transcription. These structural genes have been implicated in the mechanotransduction signal of muscle hypertrophy via intracellular signaling (24, 33, 50). Lim protein exhibits dual subcellular localization, both binding to the Z-disk and being found in the nucleus where they are believed to affect gene transcription (24, 33, 50). Though the exact function of this group of genes is currently unknown in response to exercise, our current results are consistent with previous results (5, 14, 15) and suggest the potential importance of this group of genes. It is noteworthy that all of the aforementioned structural genes bind to components of the M-line or Z-disc, which may be relevant in the light of recent evidence that suggests that during myofibrillogenesis, M-bands and Z-disks can assemble independently in spatially segregated regions of the myoplasm as the first structural component of new sarcomeres and before myosin assembles into A-bands (41).
Repair of sarcomeres would coincide with the most obvious morphological evidence of lengthening exercise, which is disruption of the sarcomeres. Additionally, it has long been hypothesized that serial sarcomerogenesis may account for the repeated bout effect observed in response to lengthening exercise, and it has recently been shown that sarcomerogenesis occurs as a unique response to lengthening exercise in rats (10), and some evidence suggests that it occurs in humans (66, 67). Thus, our results may represent the initial stages of structural repair of a damaged muscle fiber to repair damaged sarcomeres and potentially build new sarcomeres in series.
One of the cluster D genes to significantly increase expression at 6 h was CSRP3, which was found 2.8-fold upregulated 6 h postexercise. The increase in CSRP3 is consistent with two previous studies utilizing expression profiling (5, 15). In our rat study, we reported a 4.9- and 4.3-fold increase in muscle after lengthening exercise in polysomal and total RNA respectively at 6 h (15). In a mouse study, a greater than fivefold increase in CSRP3 as determined by expression profiling 48 h after lengthening exercise (5). In a subsequent analysis the group also detected a significant increase at 6, 12, and 24 h by qRT-PCR postlengthening exercise, with peak expression occurring at 12 h (5). In addition to this expression in rodent skeletal muscle, the CSRP3 protein has been reported to be differentially expressed in nuclei of human cardiac muscle undergoing pathogenic hypertrophy (24). In relation to this, a mutation in the CSRP3 gene is believed to be causal in a subset of dilated cardiomyopathy cases, though less is known about the effects of this in skeletal muscle (37, 40).
CSRP3 is expressed at low levels in other tissues than striated muscle, both during development and in adulthood (54). Structurally, it is part of the Z-disk of sarcomeres, bound to telethonin, and is thought to be involved in the mechanotransduction machinery of skeletal muscle whereby it translocates to the nucleus upon stretch or stimulation. Within the nucleus it is a known binding partner of MYOD, myogenin, and MRF4, affecting their ability to bind DNA and activate gene transcription (26, 40). Additionally, CSRP3 has been shown to promote terminal differentiation of skeletal muscle myoblasts. Indeed, in its absence, skeletal muscle cells express myogenin but do not terminally differentiate (1). Furthermore, mice knockout models have produced evidence of involvement of CSRP3 in muscle hypertrophy and regeneration process via intracellular signaling (calcineurin/NFAT) and MYOD expression. While its precise role in response to an exercise stimulus is currently unknown it seems likely to be promoting muscle growth through structural support of the Z-line and sarcomere and assisting MYOD activity (1). Our results are the first reporting its differential expression human skeletal muscle in response to different modes of exercise with a greater increased expression at 6 and 24 h after lengthening contractions.
A novel finding of the current study was the differential expression of MUSTN1 in the lengthening exercised muscle increasing over the 24 h (Fig. 2C; significant at 24 h). MUSTN1 encodes an 82-amino acid nuclear protein with no homology to any known protein family. It was identified in a study examining bone remodeling in mice after fracture, in which Lombardo et al. (45) performed suppressive subtractive hybridization between RNA isolated from intact bone to that of callus from postfracture (PF) days 3, 5, 7, and 10 as a means of identifying upregulated genes in the regenerative process. MUSTN1 was identified and was reported to be highly expressed during embryogenesis and inactivated in most adult tissues with the exception of skeletal muscle and tendon. In the bone PF study, MUSTN1 was acutely and differentially expressed during regeneration of bone. MUSTN1 mRNA expression in the callus dramatically rises to a 22-fold increase by PF day 3 and a staggering 54-fold by PF day 5 (45). Taken together, these data support the notion that this novel gene plays a role in the development and regeneration of the mammalian musculoskeletal system as a whole. We have already shown that muscle and tendon show a similar time course of increase of myofibrillar and collagen protein after exercise, and it is tempting to speculate that MUSTN1 is involved in these tissues and may even be involved in the anabolic responses of bone to increased physical activity. However, the function of MUSTN1 is unknown in skeletal muscle and further studies will be needed to confirm its regulation in response to exercise and test its function.
The primary limitation of the current study is the small sample size (n = 3), yet several steps were taken in the design and analysis to discern accurate results. First is the level of stringency undertaken in data analysis (as described in materials and methods). In regards to this, the results of the current paper are in concordance with the few other studies that have used expression profiling in response to an acute bout of lengthening exercise (5, 14, 15). The consistency is seen in both individual genes (CSRP3, ALP, FBXO32, CRYAB) and pathways (protein synthesis, early growth) (5, 14, 15). Furthermore, this study provides a unique design by using only lengthening or shortening contractions and comparing samples within an individual. The technique of within individual comparisons, as opposed to comparing between individuals or using a zero time point has been recently proposed as the ideal design for studying the human gene response to exercise (63). We recognize that this analysis does not describe the relationship to the zero time point, which is an important consideration. Therefore, Supplemental Table S1 and Supplemental Fig. S1, A–C, have been added in order that these comparisons can be examined by the interested reader. Supplemental Figure S1, A–C, exemplifies the point that most changes in the lengthening leg correspond to a similar but less dramatic change by the shortening leg. A final strength and level of stringency of this study are the use of a specific muscle chip for profiling. While it does not contain every gene in the genome, it decreases the likelihood of false positives by using a smaller number of genes and genes that were selected as skeletal muscle specific.
Through further molecular studies examining gene/protein expression, signaling, gene/protein interaction it should be possible to define the course of events that cause muscle hypertrophy and possibly develop specific interventions that can be applied to those with muscle diseases, wasting diseases, the elderly, and those in need of physical therapy to alleviate the physical ailments that coincide with loss in muscle strength and size.
In conclusion, here we report the first human study to utilize expression profiling over 24 h to delineate the differential responses to lengthening or shortening contractions and comparing samples within an individual. We show novel findings in changes of individual genes (SIX1, MUSTN1, CSRP3, ALP, FBXO32, CRYAB) and pathways/processes (protein synthesis, early growth, sarcomerogenesis) in muscle in response to acute bouts of lengthening vs. shortening exercise. These changes may give insight into specific adaptations to different modes of contraction.
Y.-W. Chen and M. Kostek were partially supported by US National Institute of Child Health and Human Development Grant 1R01 HD048051-01. M. J. Rennie is supported by the UK Medical Research Council, UK Biotechnology, and Biological Sciences Research Council (BB/X510697/1 and BB/C516779/1), and The Wellcome Trust and US National Institutes of Health (49869).
The authors thank Dr. Heather Gordish for statistical advice during the project.
↵1 The online version of this article contains supplemental material.
Address for reprint requests and other correspondence: Y.-W. Chen, Research Center for Genetic Medicine, Children's National Medical Center, 111 Michigan Ave. NW, Washington, DC 20010 (e-mail:).
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
- Copyright © 2007 the American Physiological Society