1 Department of Human Genetics, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania 15261
2 Department of Kinesiology, College of Health and Human Performance, University of Maryland, College Park 20742
3 National Institute on Aging, National Institutes of Health, Baltimore, Maryland 21224
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
aging; exercise; gender; transcription; transcriptome
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Recent evidence has indicated that changes in gene expression with advancing age may contribute to a deterioration in muscle function (11, 16, 19, 21, 25). Similarly, limited evidence supports the importance of differences in gene expression in explaining muscle phenotype differences between men and women (7, 31, 37). Changes in gene expression have been well-characterized for a number of muscle-specific genes in response to acute resistance-type exercise (3). A general limitation to previous gene expression research is the reliance on genes previously characterized in muscle or known a priori to be important for muscle structure and function. For this reason, several investigators have recently begun to use " exploratory" gene expression analyses to identify genes not previously known to be important to muscle function (15, 16, 4345). For example, Lee et al. (16) reported that of 6,347 genes expressed in mouse soleus muscle, 58 genes displayed a greater than twofold increase in expression in old compared with young animals, whereas 55 genes exhibited greater than a twofold decrease in expression in old compared with young adult animals, although only a fraction of these corresponded with changes in human muscle (44, 45). In the comparison study of human muscle, Welle et al. (44) compared global gene expression patterns among young and older men and found that 89 of 702 genes were differentially expressed in relation to age, many of which were related to metabolic enzymes. Jozsi et al. (15) performed a medium-density (500 genes) cDNA microarray experiment in young and older men at baseline and in response to acute resistance exercise (24 h after exercise) and observed differential expression in several stress-related genes in the older men compared with the young men at baseline and differential responses in gene expression in some of these same genes in response to resistance exercise in the two groups.
To expand the limited work completed to date, the purpose of the present experiments was to identify genes whose expression in skeletal muscle is influenced by age and sex differences, and/or responses to ST using microarray and quantitative PCR (qPCR) techniques. To achieve this goal, we used a previously validated high-density cDNA microarray (22, 36, 38) representing over 4,000 human genes in a design that included 12 total microarray hybridizations providing 32,000 experimental data points. This report outlines our analysis of these data, including general characteristics of gene expression patterns, as well as specific candidate genes indicated as being important for skeletal muscle function in relation to age, sex, and/or ST.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
ST program.
The details of the study design and methods, as well as additional phenotype data (strength, mass, etc.) have been described previously (14, 17). Here, we briefly describe the ST intervention performed by the subjects (32). The intervention consisted of unilateral ST of the dominant legs knee extensors with the nondominant leg serving as a control. The training program consisted of four sets of high-volume, heavy-resistance knee extension exercise performed 3 days per week. Subjects completed a set of five repetitions of knee extension at the 5RM resistance (following a warm-up). The resistance of subsequent sets was initially set at the 5RM, with the resistance incrementally decreased by the subject in a way that elicited near maximal effort on every repetition to perform a total of 10, 15, and 20 repetitions per set, respectively. Specified rest periods were allowed between sets. An exercise specialist directly supervised all sessions to verify compliance. Progressive increases in resistance occurred throughout the 9-wk ST program.
Biopsies.
Biopsies were obtained from the vastus lateralis muscle of the trained leg before ST (2 wk prior to ST familiarization and baseline testing) and after ST (4872 h following last training session) for each subject as outlined previously (32). Tissue samples were immediately dissected of fat, blood, and connective tissue, enclosed in cryovials, snap-frozen in liquid nitrogen, and stored at -80°C until analysis.
Microarray molecular biology.
Total RNA was extracted using the SV RNA Isolation Kit (Promega) according to manufacturers instructions (which included DNase I treatment) and quantitated by determining absorbance at 260 nm in triplicate, with the values averaged. For each microarray experiment, a total of 1 µg of total RNA was used for each hybridization, thus 200 ng of total RNA was taken from each sample and pooled for each group. Arrays were hybridized according to the manufacturers instructions, once for each experimental condition (baseline, ST) within a single group. Thus four total microarrays, one for each of the four groups, were hybridized twice each (baseline and after ST). Total RNA was hybridized to GF211 Human Named Genes GeneFilters (Release I; ResGen, Invitrogen) according to the manufacturers instructions. Briefly, 1 µg of total pooled RNA was reverse transcribed using 12- to 18-mer oligo dT primers and SuperScript II (GIBCO-BRL). RNA was labeled with an [ -33P]dCTP probe during reverse transcription. Labeled cDNA was hybridized to the array for 18 h, washed, and exposed to a phosphor storage screen (Molecular Dynamics) for 92 h. The phosphor image was obtained using a Storm 860 PhosphorImaging system (Molecular Dynamics) at a 50 µm resolution. Image files were then imported into Pathways v3.0 microarray software (ResGen, Invitrogen) for analysis. Following image acquisition, arrays were stripped according to the manufacturers instructions. Stripping efficiency was determined by assessing radioactive emissions, as well as by imaging stripped arrays with the PhosphorImaging system. In all cases, the baseline sample was hybridized first, followed by the after-ST sample. In addition to the experimental samples, these same microarrays were hybridized (final hybridization) as outlined above with cDNA generated from a commercially available skeletal muscle total RNA source (Ambion). Thus all four arrays were hybridized with an identical "control" sample, allowing for an analysis of interarray variation (see RESULTS).
Walker and Rigley (38) and others (22, 36) have performed validation studies of these same microarrays (GF211) using quantitative PCR and Northern blot techniques for several differentially expressed genes. Based on this work, Walker and Rigley (38) recommended criteria of greater than 10-fold above background with at least 1.5-fold differential expression. In the present study, we relied on the more stringent criterion of >1.7-fold for defining differential expression to further decrease the chance of false-positive detection.
Quantitative PCR.
qPCR gene expression studies were performed for caldesmon, SWI/SNF (BAF60b), and four-and-a-half LIM domains 1 (FHL1) as validation of the microarray results. Different subjects from those studied in the microarray experiments were selected for the qPCR experiments to provide an independent sample for the validation studies (see Subjects, above). These subjects (within their respective groups) did not differ significantly with regard to age, physical characteristics, or ST response from the subjects analyzed for the microarray experiments.
Total RNA was extracted from the before- and after-ST muscle samples using a standard phenol-based extraction method (Ambion), quantified by determining absorbance at 260 nm in triplicate, with the values averaged. Total RNA was treated with DNase I (Ambion) and reverse transcribed using the Reverse Transcription Reagents kit (Applied Biosystems) according to the manufacturers instructions using random hexamers, with 350 ng of total RNA reverse transcribed in a 20-µl reaction. qPCR was performed using the ABI 7700 DNA Sequence Detection System (TaqMan; Applied Biosystems) using standard fluorescent chemistries and thermal cycling conditions. Primer and probe sequences were designed for each experimental genes mRNA sequence using Primer Express software (Applied Biosystems) as shown in Table 1. 18S rRNA was used as an internal expression control and was amplified using the Ribosomal RNA Control Reagents kit (Applied Biosystems). Primer and probe concentrations were optimized using the TaqMan Universal PCR Master Mix (Applied Biosystems). For each reaction, 14 ng of cDNA was added to optimized primer and probe concentrations, with 25 µl of PCR Master Mix. A corresponding well contained 7 ng of cDNA with reaction reagents for the qPCR of 18S rRNA. All reactions were performed in duplicate. Thermal cycling conditions were as specified by the manufacturer: 50°C for 2 min, 95°C for 10 min, and 40 cycles as follows: 95°C for 15 s, ramp to 60°C for 1 min. Known concentration standards were developed using cDNA (produced as outlined above) from a commercially available skeletal muscle total RNA source (Ambion). Standard curves were generated from five concentrations of total RNA (2, 4, 8, 16, and 32 ng) performed in duplicate for each experimental gene and 18S rRNA. Optimization reactions were performed to ensure that all experimental samples fell within the range of the standard curves for each gene.
|
Data analysis.
The GF211 microarray contains probe spots for over 4,000 known human genes. After each array image was imported into the Pathways v3.0 software, comparison images were normalized using the data point method according to the manufacturers instructions. The data point method normalizes the data for each particular probe to the background intensity of the entire array, thereby controlling for hybridization variability. Following normalization, microarray comparisons were performed between age, sex, and ST conditions. In general, we performed two microarray comparisons (e.g., A vs. B and C vs. D) and then identified those genes differentially expressed in both comparisons to more specifically target genes related to a particular condition (e.g., age, sex, or ST). To decrease the prevalence of false-positive results, only genes expressed at 10-fold above background with a minimum of 1.7-fold differential expression between arrays were considered truly differentially expressed. To further decrease the chance of false-positive identification, comparisons were performed among the control microarrays to determine "differentially expressed" genes, which are not predicted. Any genes identified as differentially expressed on the control arrays were excluded in the experimental analysis. For brevity, some tables include only those genes with >2.0-fold differential expression. All Supplementary Tables1
(labeled Tables S1S5) are available online at the Physiological Genomics web site and at http://www.inform.umd.edu/knes/research/genomics/data.
For the analysis of the qPCR results, averaged normalized data for each experimental gene was compared between and among groups using either independent or paired-samples t-tests, or one-way or repeated-measures ANOVA with LSD post hoc. Data are presented as means with standard deviation (SD) or means ± standard error (SE) as indicated, with P < 0.05 accepted as statistically significant.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Influence of sex.
To identify genes differentially expressed in relation to sex, we first compared men vs. women both before and after ST to identify genes differentially expressed regardless of the influence of ST. A total of 210 genes were identified, with 175 of these with higher expression in men (see Supplementary Table S1 for all genes differentially expressed at >2.0-fold levels). We then determined sex differences regardless of the influence of age by comparing before-ST samples both between older men and women and between young men and women. That analysis revealed 179 differentially expressed genes (>1.7-fold), with 136 genes expressed higher in men (Supplementary Table S2). Of the 179 genes shown in Table S2 (sex differences regardless of age), only 28 of the genes (16%) are also represented in Table S1 (sex differences regardless of ST), with all but 5 of these being more highly expressed in men. Table 3 shows the majority of those 28 genes that were identified in both comparisons (i.e., present in both Supplementary Tables S1 and S2).
|
|
|
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
We performed the microarray experiments in the present study using pooled samples of five individuals per group. Our intent with this strategy was to define general patterns of gene expression for each group by diminishing the influence of interindividual variation in expression, which has been noted previously (2, 15). This strategy has been used successfully by other groups (15, 44) and has recently been recommended for the analysis of skeletal muscle gene expression using microarrays (2). Specifically, Bakay et al. (2) reported that high interindividual variability is common in skeletal muscle array experiments, which can obscure general patterns of expression. The authors further showed that pooling of patient samples was an effective way to normalize much of the intersubject "noise" while still identifying the large majority of differentially expressed genes. Bakay et al. (2) concluded that stringent yet robust data can be generated by pooling individuals (n = 5), with the caveat that follow-up experiments will be required to define interindividual variability.
The present results represent the first large-scale skeletal muscle gene expression study to address sex-related differences and indicate greater differential expression between the sexes than any other comparison, with 200 genes differentially expressed between men and women regardless of either age or ST. Within these identified genes, which represent several functional classes, no specific patterns of genes emerged as being consistently differentially expressed between men and women, and only a fraction of the genes in either Supplementary Table S1 or S2 has been studied previously in the context of skeletal muscle. The basis for these extensive sex-related differences is unclear. Whether common transcription factor binding sites (e.g., estrogen and androgen response elements) within these genes might influence the sex-related differences in gene expression, as a result of hormonal differences between the sexes, is uncertain, although a remarkable 75% of the identified genes were more highly expressed in men than women. Another possibility is that these differences are related to the significant differences in relative body fat proportions between men and women, although we expect that hormonal differences are the predominant factor.
Previous work characterizing large-scale patterns of gene expression in human skeletal muscle is limited. In a smaller-scale gene expression study, Welle et al. (41) reported no age-related differences in muscle mRNA content of several muscle-specific contractile protein genes. Comparison with the present study is difficult, since -actin is the only gene reported by Welle et al. (41) that is also represented on the GF211 array used here. No age-related differences in
-actin were observed in the present study, which supports their conclusion (41). More recently, Welle et al. (43, 44) used serial analysis of gene expression (SAGE) to generate transcript libraries of skeletal muscle from young and older men, the results of which indicated that
12% of genes were differentially expressed as a result of age. In the present study, only
5% of the
1,000 genes expressed above background were differentially expressed in relation to age; however, our results were distilled from a conservative multiple comparison process. Of the genes identified by Welle et al. (44), seven were available for direct comparison on the GF211 microarray used in the present study (GAPDH, glycogen phosphorylase, triosephosphate isomerase, pyruvate kinase, glycogen synthase, phosphoglycerate kinase, and cytochrome c oxidase). Of those seven genes, all but one (phosphoglycerate kinase) similarly demonstrated higher expression in the young compared with older males in the present study, although similar differences were not observed in women, thus explaining their exclusion from the final lists of genes (Tables 4 and 5). This highlights the possibility that the molecular mechanisms of muscle aging may differ between men and women, consistent with the finding of broad sex-related differences in the present study. Finally, Jozsi et al. (15) performed a medium-density microarray investigation of skeletal muscle gene expression in young and older men both at baseline and 24 h following an acute bout of resistance exercise. They reported elevated expression of stress-related genes (e.g., HSP27 and XRCC1) in the older men at baseline, with increased expression of several stress-related genes (e.g., HSP27, MAP kinase kinase 3, and VEGF) in response to exercise, with differential expression indicated for some genes between young and older men. Of the genes identified in Jozsi et al. (15), five were available for comparison with the present study (XRCC1, IL-1ß, RANTES, VEGF, and EGR-1). In relation to aging, only XRCC1 could be compared with the present study, and we observed higher XRCC1 levels in young compared with older men, consistent with Jozsi et al. (15), but this relationship was not found in women. Comparisons between acute exercise (24 h post) and chronic exercise training (4872 h post; present study) require caution, as activity-induced transcription is likely transient (29). In that context, we did not observe changes in VEGF, XRCC1, RANTES, or EGR-1 levels in response to ST. We did observe small increases in IL-1ß with ST in young men and women, but not in older men and women, which is consistent with the acute exercise response reported by Jozsi et al. (15). In summary, our data compare favorably to existing reports for comparable genes and extend the existing literature by providing the first investigation to examine skeletal muscle gene expression in relation to sex and chronic ST.
Several genes that might be hypothesized to be differentially expressed as a result of age (11, 16, 19, 21, 25) could not be verified in the present study because they were not represented on the GF211 microarray (e.g., muscle transcription factors). Few of the genes shown in Supplementary Tables S3 and S4, genes differentially expressed in relation to age, are immediately recognizable as previously studied muscle-related genes, with GAPDH, carbonic anhydrase III, and tropomyosin-ß being the most studied in skeletal muscle. The genes most differentially expressed with age (Tables 4 and 5) represent several classes, including structural, metabolic, and regulatory genes. Similar to the issue raised with regard to sex-related differences in gene expression, whether shared transcription factor binding sites within these genes (with corresponding age-related changes in transcription factor levels) explain their differential expression or whether these genes are independently regulated remains to be seen.
With regard to ST-induced changes in gene expression, only a moderate number of genes (n = 69) were found to be differentially expressed as a result of ST across all groups (Table S5). Our aim with the present exploratory investigation was to characterize "trained" skeletal muscle, thus we chose our after-ST biopsy time point for 4872 h after the last training session to minimize the residual effects of the last bout of acute exercise (29). Although translation of existing mRNA has been shown to be important to myofibrillar protein synthesis after resistive exercise (42), the hours following an acute ST bout are likely associated with substantial changes in transcription for several genes, which then return to near baseline by 4872 h after exercise (4, 29). Twenty of the 69 genes in Supplementary Table S5, genes differentially expressed in response to ST, are immediately recognizable as previously studied muscle-related genes, and include several structural and metabolic genes such as carbonic anhydrase III, GAPDH, nebulin, and troponin I. The specific roles of the other identified genes in skeletal muscle adaptation remain to be determined. Multiple myosin light chains (MLC; no myosin heavy chains are represented on the GF211 array) were also identified as differentially expressed in response to ST. Although only limited work has been done in relation to MLC expression and ST (46), the decreases in expression of the MLCs are not predicted; however, these results must be viewed with caution. An inherent limitation in cDNA microarrays is cross-hybridization of genes within the same gene family (8). Moreover, the GF211 was not designed to distinguish between MLC isoforms, information that is critical for determining functional significance. Although Jozsi et al. (15) reported altered expression of stress-related genes between young and older men in relation to ST, signal transduction, stress, repair, and growth factor genes were specifically represented on the commercial array used in their investigation; the GF211 microarray used here does not represent specific classes of genes, but rather contains a broad spectrum of known genes selected without regard to functional class. Thus the GF211 microarray is well-suited for identification of novel genes (22, 36, 38), which was the purpose of our work. For example, caldesmon was highly expressed on the arrays in all groups (i.e., highly expressed in muscle) but was more highly expressed in women compared with men in all analyses. Caldesmon has been most studied with regard to the regulation of smooth muscle contraction, as it inhibits myosin ATPase activity by binding to actin and tropomyosin (5, 27, 40), especially when dephosphorylated (10). To our knowledge, its role in skeletal muscle contraction is unclear, although Heubach et al. (12) have determined that caldesmon does inhibit skeletal muscle force production in vitro.
The qPCR results confirmed a reduction in the expression of FHL1 following ST, as well as higher expression in women than men. FHL1 is a LIM protein family member, with three primary isoforms, FHL1, 2, and 3 (23, 24, 26). FHL1 and FHL3 are highly expressed in skeletal muscle, whereas FHL2 is expressed predominantly in cardiac muscle (24). Expression of FHL1 was associated with muscle cell differentiation in C2C12 cell culture studies (24), but no further work has been reported as of this writing. FHL1 differs from a related protein, muscle LIM protein (1).
Finally, SWI/SNF, or BAF60b, expression was higher in women than men. BAF60b is one of a series of proteins that are involved in the remodeling of chromatin during development (30) and which appear to be necessary for transcriptional activation of many genes, as well as cell cycle control (6, 39). BAF60b is one of three family members, a, b, and c, of which BAF60b and BAF60c are highly expressed in skeletal muscle (39). Recently, de la Serna et al. (6) reported that during myoD-mediated induction of muscle differentiation, the myogenic phenotype was completely absent when mutated forms of BAF60b and BAF60c genes were expressed. Moreover, the BAF60 enzymes appear to promote myoD-mediated differentiation by altering chromatin structure in promoter regions of endogenous, muscle-specific loci, including myogenin and myosin heavy chain (6), demonstrating the importance of this protein to muscle.
The three genes outlined above were chosen for "real-time" or qPCR validation experiments based on their patterns of expression, available background literature, and other unpublished microarray data. All three genes demonstrated similar expression patterns in the qPCR experiments compared with the microarray results, despite the fact that the subjects used in the qPCR experiments were different from those of the microarray experiments, thus providing an independent validation.
Perspective.
The current investigation was an exploratory analysis designed to identify candidate genes related to differences in skeletal muscle gene expression in relation to age and sex, as well as changes in muscle in response to ST. Despite our validation experiments and the extensive validation of these particular GF211 microarrays by others (22, 36, 38), these results should be viewed as preliminary and not as a definitive view of the skeletal muscle "transcriptome" in these conditions. While the use of microarray technology is within the grasp of most laboratories, the analysis of the resulting data is complex and often specific to a particular type of array, with few established techniques for in-depth statistical analysis of the accompanying large datasets. Moreover, in practical terms it is impossible to verify each identified gene using qPCR or similar single gene techniques. Thus we refrain from in-depth speculation about the "interpretive meaning" of the results and rather present the data as an important initial step in understanding the underlying biology of skeletal muscle. From these data, specific genes can be targeted for individual study (e.g., caldesmon, FHL1, and BAF60b). The results of this and other large-scale gene expression studies will need to be combined and compared, with the ultimate goal of developing a model that describes the global patterns of gene expression in skeletal muscle.
![]() |
ACKNOWLEDGMENTS |
---|
This work was supported by National Institutes of Health Grants AG-42148 and DK-46204 and by the Competitive Medical Research Fund of the Univ. Pittsburgh Medical Center Health System. S. M. Roth was supported by National Research Service Award Grant AG-05893.
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: S. M. Roth, A300 Crabtree Hall-GSPH, Dept. Human Genetics, Univ. of Pittsburgh, Pittsburgh, PA 15261 (E-mail: sroth{at}hgen.pitt.edu).
10.1152/physiolgenomics.00028.2002.
1 Supplementary materials (Tables S1S5) to this article are available online at http://physiolgenomics.physiology.org/cgi/content/full/10/3/181/DC1.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|