1 Biomedical Sciences and Dalton Cardiovascular Research Center, University of Missouri-Columbia, Missouri 65211
2 Research Center for Genetic Medicine, Childrens National Medical Center, George Washington University, Washington, District of Columbia 20010
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
microgravity; hindlimb suspension; exercise; aging; non-weight-bearing activity
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In contrast, the genes that are involved in initiating the early responses eventually leading to these phenotypic changes remain unknown. The goal of the present study was to utilize an unbiased, discovery-oriented approach to identify a compendium of genes whose expression is most responsive to brief periods of inactivity/unloading and of low levels of muscular activity. These genes may be involved in the initial steps of muscle adaptations to HU or to repeated episodes of inactivity common in sedentary lifestyles.
In the current study, we aimed to identify the changes in gene expression that are responsive to short periods (12 and 4 h) of altered muscle usage to determine the sensitivity of transcriptionally responsive events involved in skeletal muscle plasticity. In prior work, we and others have used oligonucleotide microarrays to document the more chronic differential gene expression profiles with skeletal muscle fiber types (10), aging (36), and also with early hypertrophy (11). The brief duration of our treatments (HU and reloading) and the fact that the weight-bearing period in control and reloaded groups consisted of only low-intensity ambulatory activity allowed us to identify those transcripts that are most sensitive to changes in contractile activity and loading. Genes whose expression changes during such short periods of muscle unloading are more likely to be part of the primary causes initiating the problems associated with muscle inactivity than secondary responses that follow the adaptive processes seen after days and weeks of inactivity/unloading (8, 25, 48, 51). This expression profile may be important both in the etiology of muscle alterations and diseases for which physical inactivity is a risk factor and in the development of muscle atrophy that generally occurs during prolonged periods of bed rest or microgravity. Thus these genes will be new potential targets in the development of credible rationales for countermeasures to microgravity or physical inactivity. The soleus is a predominantly slow-twitch muscle that is well-recruited during weight bearing (1, 26), and it has been the most frequently studied muscle during HU (8, 47, 48, 51). In summary, the present study was designed to establish the most sensitive changes in gene expression in posturally activated muscle during relatively short and intermittent periods of inactivity and activity.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
RNA processing.
We used 2029 animals per treatment group so that outliers and individual variability had minimal effect on reported results. Muscles (10 mg) were pooled into groups of 810 animals. Muscle was homogenized on ice in TRIzol Reagent (GIBCO-BRL), and total RNA was isolated using the manufacturers recommended protocol. Eighteen micrograms of total RNA was converted into double-stranded cDNA using the SuperScript Choice system (GIBCO-BRL) with an oligo-dT24 primer containing the T7 RNA polymerase promoter (Genset). Double-stranded cDNA was purified by phenol/chloroform extraction and precipitated. In vitro transcription was subsequently performed using an Enzo BioArray RNA transcript labeling kit (Enzo). Biotin-labeled cRNA was purified by an RNeasy kit (Qiagen) and fragmented. Biotin-labeled cRNA samples were also prepared from total RNA of four pools (20 animals/pool) of heart and brain to determine technical variability between duplicates as part of the determination of meaningful fold changes.
Total RNA was also isolated from the soleus muscle of subsequent independent groups of control and suspended rats (n = 56) to measure changes in gene expression by quantitative real-time PCR procedure.
Microarray processing.
Ten micrograms of fragmented cRNA was hybridized for 16 h on the Affymetrix rat U34A microarray. Each microarray was washed and stained in the Affymetrix Fluidics Station 400 using the manufacturers instructions and reagents. This involved removal of nonhybridized material followed by incubation with streptavidin-phycoerythrin (SAPE) to detect hybridized cRNA. The signal intensity was amplified by a second staining with biotin-labeled anti-streptavidin antibody followed by SAPE staining. Fluorescent images were read before and after amplification using a Hewlett-Packard G2500A gene array scanner.
Data analysis.
Microarray images were analyzed using statistically based Affymetrix Microarray Suite 5.0 software (37). Each gene is represented by 20 probe pairs of 25-mer oligonucleotides that span different sequences of the coding region. Each probe pair consists of a perfect match sequence (PM) that is complementary to the cRNA target and a mismatch sequence (MM) that includes a change of a single base critical for hybridization. Comparison of the hybridization signals from the PM and MM probes was performed for the purpose of removing any nonspecific hybridization from the data analysis. Bacterial sequences were also included on the arrays as external controls for hybridization. Complete transcription and hybridization were validated using the Affymetrix recommended criteria based on the bacterial controls and several housekeeping genes.
Multiple criteria were applied to determine differential expression between treatments. The Affymetrix Microarray Suite 5.0 was utilized. Results are reported only for transcripts that passed each of the analytical filters (described below) and therefore displayed the most robust responses to unloading and reloading. We focus more on results for the fully annotated genes (see Table 2 and Fig. 3) because of less confidence in the identity of expressed sequence tags (ESTs) (see Supplemental Table 3). (Supplemental Table 3 is available online, published at the Physiological Genomics web site.)1
Published statistical algorithms (37) were used to determine whether a transcript was first detectable in a given sample based upon the 1620 independent probe pairs for each transcript. A one-sided Wilcoxons signed rank test (WSR) was applied to the PM and MM intensities of each probe set to determine which genes were expressed above background (37). This nonparametric test was chosen because it has been shown to be robust, insensitive to outliers, and does not assume a normal data distribution (37, 54). Genes were reported as significantly expressed above background at P 0.04. Using this criterion, we found 2,949 (34% of the array) transcripts were expressed in all HU groups or all control groups and then considered for further statistical analysis. Statistical algorithms based on a WSR test (37) were then used to determine significant differential expression in comparative analyses between treatment groups. Each unloaded and reloaded group was compared with each control group, resulting in nine comparisons for unloading (3 x 3) and six for reloading (3 x 2). A WSR test was applied to 3240 probes of each possible comparison. The average P value of all possible comparisons was then adjusted for multiple hypotheses testing using the false discovery rate (FDR) method by Benjamini et al. (6). In this method, all P values were ranked and tested against different thresholds. The lowest P value was tested against the Bonferroni threshold (0.05/2,949 present genes = 1.7 x 10-5), and subsequent P values were adjusted by ranking the most significant transcripts (0.05 x rank/2,949 = FDR P value) until this threshold became lower than the P value of each possible comparison tested. This last FDR threshold (cutoff for declaring significance) was found at P = 0.0044 for HU vs. control comparisons and 0.0059 for reload vs. control comparisons. This improves upon the strict Bonferroni validation test of single events, with a more tolerable version validating a group of events while minimizing both type I and II errors. This FDR method has been previously used to analyze microarray data (19, 40, 50). The magnitude of change in expression for 264 transcripts that meet our FDR criterion (for 9 of 9 pairwise comparisons) was then considered as a third criterion. Only differences in gene expression
1.5-fold for all possible comparisons (9 of 9 for HU) were considered to be most reliable and reported. This 1.5-fold change filter was based upon an error rate analysis that we performed. We found a reasonable amount of confidence in changes at this level, because there were 0 false calls between pooled samples in the same treatment; e.g., there were no genes called differently expressed between different pools of animals in the same treatment group. This partly reflects low variability in each group (Fig. 1B) and also our findings that 98 ± 1% of the technical error between duplicate cRNA was with <1.5 fold difference (Table 1). Because of the statistical criteria, the difference in gene expression between control and unloaded was
1.9-fold for 88% of the genes (Fig. 1A). This magnitude was reported as the average of fold changes for each detected probe in all possible pairwise comparisons. This ratio approach using individual probes has been shown to significantly reduce errors and increase precision of the fold changes compared with a fold change calculated from the average signal intensities (53). Technical and biological variability was accessed and results were used to minimize false calls. Pooling tissue for multiple samples has been proposed to reduce errors related to biological variability and increase confidence for many of the messages in the genome (3, 40). We tested the variability (coefficient of variation = mean/standard deviation x 100) in the average signal intensities of 2,233 commonly expressed transcripts between pooled and single samples. Approximately two times as many transcripts had a coefficient of variation (CV) < 15% (<1.2 fold error) in pooled samples compared with single samples, and 11 times fewer genes had a CV > 30% (>1.5 fold) with pooling (Fig. 1C). From these criteria to filter results, the final list of genes responsive to HU had a high rate of reproducibility between groups (Fig. 1B). About 75% of the genes had a CV that was <15% of the mean (Fig. 1B) for the differentially expressed genes discussed in RESULTS.
|
|
|
|
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
The second global pattern was for transcripts that increased during unloading and rarely reversed completely during reloading. Among the 38 genes for which transcript concentrations increased during unloading, only one (uncoupling protein 3, UCP3) decreased during reloading below control level, and most remained significantly elevated (Fig. 3 and Table 2). Previous studies have also shown that UCP3 mRNA concentration decreases in response to more vigorous exercise (9), but the present global view reveals this transcript has an unique responsiveness to altered contractile activity. Among the functional classes responsive to 12-h HU, the expression for genes coding for transcription factors and glucose metabolism proteins returned at least to baseline with reloading (Table 2), suggestive of shorter decay times for the concentration of these mRNA classes in muscle. In contrast, many of the changes in gene expression related to protein synthesis and protein degradation did not respond to short-term reloading.
ESTs are defined as transcripts with potential homologies with known genes or still unknown identity. Although there is less confidence in their identity, ESTs generally supported the global response of fully sequenced genes. In this study there was identification of 58 ESTs whose expression is changed with unloading (Supplemental Table 3). Sequence homologies seem to indicate that some of those ESTs code for similar proteins or can be classified in similar functional classes as the 63 previously described genes.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Several genes found as differentially expressed after 14 (12) to 35 days HU (55) were already influenced after 12-h HU, with examples such as HSP70, DNAJ, and cathepsin L. Yet, most of the genes in this investigation have not been studied before in skeletal muscle. This opens the door to future studies of these highly "inactivity responsive genes" in cell signaling, metabolism, and other putative processes that linked physical inactivity/unloading to the earliest onset of processes related to metabolic disease and muscle alterations.
Data reproducibility.
This study reports the responses of a small set of transcripts (4% of expressed genes on the array) that were most reproducibly influenced by short periods of altered muscle activity/loading. The microarray approach that we used included several levels of replication. First, each mRNA was analyzed with 1620 distinct probe pairs (where one-half controlled for nonspecific hybridization). We only considered transcripts that were significantly expressed (as determined by statistics of 1620 probes). The multiple probe sequences for each mRNA were also used to gain confidence in differential expression not being biased by any specific probe sequences. Second, we included 2029 animals per treatment group to increase the likelihood that sample means approached population means. Third, we used two to three replicate microarrays for each treatment group. Fourth, to make a conclusion about differential expression, statistical criteria were used to filter the data. Statistical differences between group treatments were tested by a nonparametric test avoiding a priori assumptions about the distribution of gene expression profiles. These significant differences were adjusted using a modified Bonferroni threshold to locate the promising genes for further examination with minimal spurious events. The fold changes 1.5 (generally >2.0) were reported when all possible comparisons between control and HU/reloaded were different at this level. As a result, the variability between groups in the same treatment was minimal as revealed by
81% of the genes having a CV
20% of the mean (Fig. 1B). There was also low variance due to treatment effects (see SE in Table 2 and Supplemental Table 3). Therefore, we report for the first time with reasonable confidence a global gene expression profile that is responsive to transient periods of altered activity. This may add insight to understanding gene expression during sedentary lifestyles containing long periods of skeletal muscle inactivity followed by occasional ambulatory activity.
Gene expression during the early muscle adaptations to HU.
We examined gene expression profiles during brief unloading to identify genes that are most sensitive to reduction of normal muscle function. Importantly, several studies show the rat soleus muscle is not significantly atrophied in this model for 3 days or more (8, 30), and fiber-type changes also take longer. Thus the genes reported here may be involved in the early stages of regulatory processes that eventually lead to the well-established muscle alterations observed with longer term HU.
Muscle tissue, especially postural muscles, frequently undergoes extremely wide swings in energy expenditure and mechanical stress during the course of routine physical activity in an active animal. A variety of biochemical and physiological measures have shown that the soleus is a muscle type especially dependent upon the contractile activity and loading and thus highly responsive to HU (5, 20, 33, 34, 44). This is largely because it is a muscle rich in the type of fibers that are most frequently recruited by normal ambulatory activity for several hours per day (26). HU and spinal isolation cause a slow to fast fiber type transition (30, 38, 48). Some of the identified candidate genes may be tied to initiating the fiber type transition, but it is important to note that it takes several weeks of disuse for the transition to become evident at the level of structural proteins like myosin. Some muscles contain sections rich in fibers rarely activated in normal life, especially the type IIB fibers that are activated for only a few minutes per day (26). An earlier microarray study in control mice (10) contrasted a pure type IIB section of the white vastus lateralis to the soleus (a mixture of both type IIA and I fibers in the mouse) and noted many fiber type differences. Taken together, it is now clear that dozens of transcriptional events are very sensitive to physical inactivity in posturally activated muscles like the soleus, and small amounts of daily ambulatory activity are necessary for optimal health of the tissue and the whole animal.
An increased rate of protein degradation with HU is well established (8, 35, 47, 51). Thus identifying changes in gene expression contributing to protein degradation is important in understanding either proteins that have a high rate of turnover or in chronic atrophy conditions. A group of gene products mediating protein degradation increased in parallel during 12 h of unloading and surprisingly also persisted to stay elevated during the reloading period. Cathepsin L is a protease in the lysosomal-dependent proteolytic pathway (42), and C3 and C8 proteins are subunits of the 20S proteasome in the ubiquitin-dependent proteolytic pathway (18, 49). With 12-h HU, we found that mRNA concentrations for cathepsin L, C3, and C8 increased by 4-, 4-, and 2-fold, respectively (Table 2). Goldberg and colleagues (32) have reported this rise in cathepsin L independent of 11 other lysosomal enzymes, and one study identified this message with skeletal muscle wasting due to sepsis, dexamethasone treatment, and tumor cachexia (16). Additionally, mRNA concentration for ingensin, a neutral protease in rat and pig skeletal muscle (13, 31), was 3-fold higher in 12-h unloaded soleus muscle compared with control. Previous studies have implicated a role for interleukin-6 (IL-6) and its receptor (IL-6R) in the stimulation of processes mediating protein degradation and the development of muscle atrophy (23, 52). IL-6R mRNA concentration was already increased
2-fold after only 12-h HU (Table 2), implying that this potential trigger for protein metabolism development is being regulated quickly at the mRNA concentration.
Previous studies have also indicated that protein synthesis is inhibited at the initiation (29) and elongation (29, 33) phases of translation in the soleus muscle with acute unloading. Several genes coding for proteins involved in these phases were regulated at the mRNA level during HU (Table 2). For example, a decrease in HSP70 protein mass has been implicated in slowing the elongation of nascent polypeptides during acute unloading (34). Consistent with this finding, we found that HSP70 mRNA concentrations decreased quickly by 4-fold in 12-h unloaded muscle and rapidly rebounded with 4 h of loading. In addition, mRNA concentration for calreticulin, another endoplasmic reticulum chaperone involved in the folding and maturation of newly synthesized proteins (39), was also decreased by 2-fold during 12-h HU (Table 2). Interestingly, two recent studies of atrophied muscle from immobilized (46) or unloaded (55) animal models for several weeks have both reported an increase in expression for many genes involved in translational machinery, including eukaryotic initiation factor-2
(eIF-2
), elongation factor-2 (EF2), and 10 ribosomal proteins. In the present study, we also found large increases of mRNA concentrations for eIF-4E and a large group of ribosomal proteins (S6, L26, L23, S4, and S27) during 12-h HU (Table 2), and Supplemental Table 3 reports ESTs of similar interest. As for most of the genes related to protein metabolism, the time required to decrease this class of mRNAs is apparently longer (>4 h) compared with most of the other mRNAs (e.g., transcription factors), possibly because of a difference in mRNA decay processes.
Recent work demonstrates that chronically inactive humans (tetraplegic) (28) and rats (unloading) (14) have high UCP3 mRNA concentration in their vastus lateralis or soleus muscles. Here we have shown that UCP3 mRNA concentration increases in rat soleus muscle even during acute HU (12 h) (Table 2) and thus are more rapidly induced during unloading than previously known. The increase of UCP3 gene expression during HU may be involved in the overall regulation of lipid oxidation concomitant to the prevention of lipid-induced disturbances (17).
Prolonged HU has been associated previously with a greater reliance on carbohydrate utilization (5). A rate-limiting enzyme in glycolysis, phosphofructo-2-kinase/fructose-2,6-biphosphatase (PFK-2/FBPase-2), was increased 3.4-fold. Also, these array results suggest that changes in Rad and SNAP-23 (synaptasome-associated membrane protein, 23 kDa) transcripts may be involved in the complex orchestration of insulin-mediated glucose uptake during the initial stages of HU. We found that Rad mRNA level was dramatically reduced by eightfold after 12-h HU, whereas SNAP-23 mRNA was increased twofold (Table 2). Overexpression of Ras-related protein associated with diabetes (Rad) inhibited the insulin-mediated glucose uptake in myocytes (41). In contrast SNAP-23 promoted insulin-dependent glucose uptake (21). The present study alludes to pretranslational control of these and other important metabolic genes during unloading, even after several hours.
The molecular events that lead to cell growth arrest and apoptosis involve the activation of c-Jun N-terminal kinases (JNK) (22). Both overexpression of MUK (27) and repression of mitogen-activated kinase phosphatase-1 (MKP-1) (15) resulted in the activation of JNKs. Here we show that both MUK and MKP-1 are sensitive genes to HU, and 12-h HU resulted in a surprising 20-fold increase of MUK and 2-fold decrease of MPK-1 mRNA concentrations in the soleus muscle (Table 2). These data suggest that mRNA expression for these important kinases is changing during the early period of unloading preceding the resultant cellular events of apoptosis which have been described before in remodeled muscle (2).
Effective countermeasures to inactivity and microgravity.
One novel and unanticipated finding of the current study, which has the potential to reshape our understanding of the global regulation of mRNAs during transient periods of physical inactivity, was the clustering patterns during reloading. The genes that were differentially expressed in skeletal muscle during unloading fell into three general clusters when categorized according to their response to a subsequent 4-h period of reloading (Fig. 3). Most (21 of 25) of genes downregulated during HU returned at least to control levels upon reloading (Fig. 3A), and the rest (4 of 25) all partially reversed toward control (Supplemental Table 3). In fact, most of the mRNAs decreased with HU overshot the control level by a remarkably large amount during reloading (Fig. 3A). In contrast, 71% (27 of 38) of the genes upregulated during unloading remained significantly above control during reloading (Fig. 3B), and only one overshot the normal level (Fig. 3C). However, when less stringent criteria were used and trends were considered, 21 of these upregulated mRNAs did partially reverse at least 25% (2580%) toward control in this short 4-h period (Table 2). Therefore, the mRNA concentrations for a majority of those genes increasing with HU remained elevated throughout the 4-h reloading period (Fig. 3B). The diverse functional categories for members of each cluster, combined with the conspicuous absence of a fourth cluster (i.e., genes downregulated during HU and not subsequently upregulated during reloading), suggest that there is a more fundamental reason for the dissimilar response to reloading between mRNAs that increase and those that decrease during HU. The simplest explanation for this disparity is that the cell can synthesize new mRNA in response to reloading more rapidly than it can degrade mRNA that was accumulated earlier during the unloading period. This explanation is plausible because the rate that mRNAs are reduced depends upon mRNA half-lives and associated regulatory mechanisms that may require more than 4 h to reach the new steady state.
The search for effective countermeasures to prevent muscle alterations induced by conditions such as microgravity and bed rest has only recently begun to be based on a rationale including genomic approaches (such as the global gene expression profile). Our knowledge of the molecular mechanisms affected by microgravity and inactivity is still too limited to develop targeted countermeasures that will prevent the myriad of skeletal muscle phenotypic changes induced by microgravity/inactivity or to know the minimal exercise dose necessary. Here, we introduced the most basic countermeasure: returning muscle to contractile activity in a normal 1-g environment for a brief period of time. Reversal of the alterations in gene expression was complete for many of these activity-responsive genes with only 4 h of reloading. However, the challenge will be to design optimal countermeasure paradigms to ensure that brief exposures to activity each day are sufficient to prevent the relatively unresponsive expression changes (e.g., proteasome subunits and ribosomal proteins).
Understanding the molecular events induced in skeletal muscle by physical inactivity/microgravity is a challenge in the treatment of chronic diseases linked to a sedentary lifestyle, rehabilitation of injured muscle, and recovery from space flight adaptations. Thus the identification of changes in gene expression involved in triggering initial responses to physical inactivity/microgravity and the development of effective countermeasures to prevent them are fundamental to these efforts. To identify these triggers, we examined the global gene expression profile in skeletal muscle during acute HU and measured changes in expression for 121 genes and ESTs that were highly sensitive to short periods of inactivity/unloading. Clustering patterns revealed that the mRNAs increasing during these periods were less responsive to 4 h of ambulatory activity than the mRNAs decreasing during the inactivity period, emphasizing the slowness in completely undoing the "inactivity" gene expression profile for some genes. This study identifies genes that are the most sensitive to loading/activity in rat skeletal muscle and provide new targets for understanding the mechanisms initiating muscle alterations during inactivity. These results are also important for discovery of new candidate genes explaining physiological responses favorably influenced by even transient periods of low-intensity ambulatory activity.
![]() |
ACKNOWLEDGMENTS |
---|
This research was supported by a grant from the National Space Biomedical Research Institute and by National Heart, Lung, and Blood Institute Grant HL-57367 (to M. T. Hamilton) and by a grant from the Life Science Mission Enhancement Postdoctoral Fellowship from University of Missouri-Columbia (to L. Bey).
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: M. T. Hamilton, E102 Veterinary Medicine Bldg, 1600 E. Rollins Rd, Univ. of Missouri-Columbia, MO 65211 (E-mail: hamiltonm{at}missouri.edu).
10.1152/physiolgenomics.00001.2002.
1 Supplemental Table 3 is available online at http://physiolgenomics.physiology.org/cgi/content/full/13/2/157/DC1.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|