Relationships between circadian rhythms and modulation of gene expression by glucocorticoids in skeletal muscle

Richard R. Almon, Eric Yang, William Lai, Ioannis P. Androulakis, Svetlana Ghimbovschi, Eric P. Hoffman, William J. Jusko, Debra C. DuBois


The existence and maintenance of biological rhythms linked to the 24-h light-dark cycle are essential to the health and functioning of an organism. Although much is known concerning central clock mechanisms, much less is known about control in peripheral tissues. In this study, circadian regulation of gene expression was examined in rat skeletal muscle. A rich time series involving 54 animals euthanized at 18 distinct time points within the 24-h cycle was performed, and mRNA expression in gastrocnemius muscles was examined using Affymetrix gene arrays. Data mining identified 109 genes that were expressed rhythmically, which could be grouped into eight distinct temporal clusters within the 24-h cycle. These genes were placed into 11 functional categories, which were examined within the context of temporal expression. Transcription factors involved in the regulation of central rhythms were examined, and eight were found to be rhythmically expressed in muscle. Because endogenous glucocorticoids are a major effector of circadian rhythms, genes identified here were compared with those identified in previous studies as glucocorticoid regulated. Of the 109 genes identified here as circadian rhythm regulated, only 55 were also glucocorticoid regulated. Examination of transcription factors involved in circadian control suggests that corticosterone may be the initiator of their rhythmic expression patterns in skeletal muscle.

  • corticosteroids
  • clock genes
  • gene arrays
  • expression profiling
  • transcriptome

virtually all organisms have biological rhythms associated with the light-dark cycle (4, 24, 25, 34). Although a good deal is known about the central control of such rhythms by the brain, less is known about how that control is translated to peripheral tissues. The central core of the mammalian circadian clock is located in the suprachiasmatic nucleus (SCN) in the anterior part of the hypothalamus, which receives direct input by way of the retinohypothalamic tract. Outputs from the SCN are directed to other parts of the hypothalamus that control both anterior and posterior pituitary hormones, as well as to the area of the hypothalamus and medulla that control the autonomic nervous system. It is these hormonal and autonomic outputs that to a large degree convey light-dark rhythmicity to the rest of the body. Besides receiving inputs from the retina, the SCN also receives inputs from forebrain areas and from the locus ceruleus that modulate the influence of the SCN. The nonretinal inputs to the SCN provide the flexibility for the existence of diurnal and nocturnal animals, as well as for animals being able, when necessary, to shift wake-sleep periods from night to day and vice versa (5, 6).

Simplistically, the central clock mechanism involves an autoregulatory negative feedback loop with a periodicity of approximately 24 h (5, 6, 34). The elements of this basic feedback loop are several transcription factors, including CLOCK and BMAL1, which heterodimerize and enhance the expression of Period (Per) and CRYPTOCHROME (CRY). These two transcription factors heterodimerize and repress the expression of CLOCK and BMAL1. The core system is entrained to the light-dark cycle with CLOCK:BMAL being high during the light period and Per:CRY being high during the dark period. However, as many as 17 transcription factors have been defined as also being involved in the central clock mechanism (10). These include within the basic mechanism several isoforms of Per and CRY, as well as neuronal PAS domain protein 2 (Npas2, Mop4), which has been reported as an alternative to CLOCK as a heterodimerizing partner with BMAL (7). The precise role of the other transcription factors in the central clock is still a subject of investigation.

The input from the SCN to the regulation of both pituitary hormones and the autonomic nervous system impart rhythmicity to peripheral tissues. However, rhythmicity in peripheral tissues is further complicated by more diffuse behavior-related factors such as sleep-wake patterns, eating, and physical activity, which alter systemic energy demands. Many of the transcription factors involved in regulating the central clock are also expressed in peripheral tissues (10). However, their regulation is complicated by variations in ancillary factors, such as hormonal patterns, autonomic output, and behavior that orchestrate peripheral rhythmicity (23). The existence of both diurnal and nocturnal mammals and the phenomena of phase shifting by food restriction illustrate both the complexity and flexibility in peripheral rhythmicity. Perhaps one of the best illustrations of the flexibility of rhythmic behavior outside the SCN is the observation that rhythmic behavior with a periodicity of ∼24 h can be induced in a variety of cells in culture (34).

The central clock anticipates the change in photoperiod and prepares the animal for the upcoming period of activity and feeding, regardless of whether that period is in the light or dark. The hypothalamus-pituitary-adrenal (HPA) axis is of particular importance to the active feeding period, as illustrated by the fact that the effector hormones, glucocorticoids, are high during the light period in diurnal animals and high during the dark period in nocturnal animals (6). The generally accepted mechanism for most glucocorticoid effects involves binding of free steroid to a cytoplasmically localized receptor, translocation of ligand-bound receptor into the nucleus, binding of a ligand receptor dimer to specific DNA sites (GREs, glucocorticoid response elements), and modulation of the amount of selective mRNA (13). Although some effects on mRNA stability have been noted, the common mechanism involves increasing or decreasing the rate of transcription of particular genes. Even though more recent evidence suggests that this mechanistic view may be overly simplistic, it is clear that changes in the amount of specific mRNAs are the basis of the vast majority of glucocorticoid actions (2, 3). By virtue of their circadian rhythmicity, glucocorticoids themselves are effectors of many but not all circadian changes in gene expression.

Skeletal muscle represents about 40% of the body mass of most mammals. From a systems point of view, the musculature is central to not only movement and posture, but also to energy metabolism. The musculature uses a very large amount of energy, most of which is in the form of lipid fuels, to maintain posture against gravity and body temperature relative to the environment. The musculature is also responsible for about 75% of the insulin-directed glucose disposal, which supports the phasic mechanical functions of type 2 fibers (15). In addition, under the influence of glucocorticoids, the musculature is shifted into net degradation of proteins, which provide amino acid carbon for gluconeogenesis in the liver and kidney (2). Energy expenditure by muscle should vary as a function of circadian rhythms with the highest expenditure during the animal's active feeding period.

Previously, we profiled the time-dependent response of skeletal muscle from adrenalectamized rats to two different dosing regimens of the synthetic glucocorticoid, methylprednisolone (MPL) (3). By using such animals, we removed not only the endogenous source of glucocorticoids, but also major aspects of the systemic influence of the sympathetic nervous system. One time series involved giving a single bolus dose of MPL (50 mg/kg) and euthanizing animals at 16 time points over a 72-h period. The second time series involved delivering MPL (0.3 mg·kg−1·h−1) via Alzet pump and euthanizing animals at 10 times over a 168-h period. In the present report, we describe the use of Affymetrix arrays to analyze skeletal muscles from intact rats maintained on a strict light-dark regimen consisting of 12:12-h light-dark cycle with three animals killed at 18 time points during the 24-h period. This rich time series allowed us to identify genes whose expression was rhythmically modulated and to group those genes into eight relatively discrete temporal circadian clusters. We grouped the genes with circadian rhythms into 11 functional categories to examine cellular activity as a function of temporal cluster. In addition, comparisons with our previous work allowed us to evaluate the response profiles of genes with circadian rhythms to the two dosing regimens of exogenous corticosteroid, providing insight into the role of glucocorticoids in regulating rhythms of gene expression in skeletal muscle.



Fifty-four normal (150–175 g) male Wistar rats were purchased in two separate batches of 27 from Harlan Sprague Dawley (Indianapolis, IN), and experiments were initiated at body weights between 225 and 275 g. Animals were housed and allowed to acclimatize in a constant-temperature environment (22°C) equipped with a 12:12-h light-dark cycle. Twenty-seven rats (group 1) were acclimatized for 2 wk before study to a normal light-dark cycle, in which lights went on at 8 AM and off at 8 PM. The onset of light cycle was considered as time 0. The other 27 rats (group 2) were acclimatized for 2 wk before study to a reversed light-dark cycle, where lights went on at 8 PM and off at 8 AM. Two weeks of acclimation was sufficient time to reestablish circadian cycles, as indicated by blood corticosterone concentrations in the two groups. Rats in group 1 were killed on three successive days at 0.25, 1, 2, 4, 6, 8, 10, 11, and 11.75 h after lights on to capture the light period. Rats in group 2 were killed on three successive days at 12.25, 13, 14, 16, 18, 20, 22, 23, and 23.75 h after lights on to capture the dark period. Animals killed at the same time on successive days were treated as triplicate measurements. Because normal rats were used, minimal animal handling with the least possible environmental disturbances was used to minimize stress. Night vision goggles were used to carry out animal procedures conducted in the dark period. At death, rats were weighed, anesthetized by ketamine:xylazine (80:10 mg/kg), and killed by aortic exsanguination. Blood was drawn from the abdominal aortic artery into syringes using EDTA (4 mM final concentration) as anticoagulant. Plasma was harvested from blood by centrifugation (2,000 g, 15 min, 4°C) and frozen at minus 80°C until analyzed for corticosterone. Gastrocnemius muscles were excised and frozen in liquid nitrogen immediately after death and stored at minus 80°C until RNA preparation. Both acute and chronic MPL dosing experiments have been previously published (26, 32). In brief, populations of adrenalectomized male Wistar rats were given doses of the synthetic glucocorticoid, MPL. In the acute experiment, the animals were given a single bolus dose (50 mg/kg) of MPL and were killed at 16 times over a 72-h period following dosing. In the chronic experiment, the animals were administered 0.3 mg·kg−1·h−1 MPL via Alzet osmotic pumps and were killed at 10 time points over a 168-h period. All rats had free access to rat chow and 0.9% saline drinking water. Our research protocol adheres to the “Principles of Laboratory Animal Care” (NIH publication 85–23, revised in 1985) and was approved by the University at Buffalo Institutional Animal Care and Use Committee.

Plasma steroid assays.

Plasma corticosterone concentrations were determined by a sensitive normal-phase HPLC method, as previously described (11). The limit of quantitation was 10 ng/ml. The interday and intraday coefficients of variation (CV) were less than 10%.


Muscle samples from each animal were ground into a fine powder in a mortar cooled by liquid nitrogen, and 100 mg was added to 1 ml of prechilled TriZOL Reagent (Invitrogen, Carlsbad CA). Total RNA extractions were carried out according to manufacturer's directions and were further purified by passage through RNeasy mini-columns (Qiagen, Valencia, CA) according to manufacturer's protocols for RNA cleanup. Final RNA preparations were suspended in RNase-free water and stored at minus 80°C. The RNAs were quantified spectrophotometrically, and purity and integrity were assessed by agarose gel electrophoresis. All samples exhibited 260/280 absorbance ratios of ∼2.0, and all showed intact ribosomal 28S and 18S RNA bands in an approximate ratio of 2:1, as visualized by ethidium bromide staining. Isolated RNA from each muscle sample was used to prepare target according to manufacturer's protocols. The biotinylated cRNAs were hybridized to 54 individual Affymetrix GeneChips Rat Genome 230A (Affymetrix, Santa Clara, CA), which contained 15,967 probe sets. The 230A chip was used in the chronic infusion experiment as well, allowing direct comparison between the two experiments. The 230A gene chips contain over 7,000 more probe sets more than the ones used (U34A) in our earlier muscle bolus dose MPL study. The high reproducibility of in situ synthesis of oligonucleotide chips allows accurate comparison of signals generated by samples hybridized to separate arrays. This data set has been submitted to GEO (GSE8989).

Quantitative real-time RT-PCR.

The quantity of muscle glutamine synthetase along with gene-specific in vitro transcribed cRNA standards were determined by quantitative real-time RT-PCR using TaqMan probes. Briefly, primer and probe sequences were designed using PrimerExpress software (Applied Biosystems, Foster City, CA) and custom synthesized by Biosearch Technologies, (Novato, CA) The RT-PCR was performed using Brilliant QRT-PCR Core Reagent Kit, 1-Step (Stratagene, La Jolla, CA) in a Stratagene MX3005P thermocycler, according to manufacturer instructions. A standard curve was generated using the in vitro transcribed sense cRNA standards. Primer and probe sequences are as follows: forward primer (5′-CGCCCGCCGTCTGA-3′), reverse primer (5′-TCTCCTGGCCGACAATCC-3′), and probe (5-FAM-TCCACGAAACCTCCAACATCAAACGACTTT-BHQ-3′). Intraday and interday CVs were 18% and 10%, respectively.

Data set construction.

As discussed above, animals were killed at precise times on three successive days to obtain data points for the light period and three successive days to obtain data points for the dark period. Animals killed at the same time on different days were treated as three replicates for that time to construct a 24-h light-dark cycle. To obtain a clear picture of an entire cycle, two 24-h periods were concatenated to obtain a 48-h period, which allowed visualization of rhythms that spanned the dark-light and light-dark transitions.

Data mining.

A nonlinear curve fit using MATLAB was conducted, which fitted a sinusoid function [A·sin(Bt + c)] to the data, including the replicates. Genes that could be curve fitted with a R2 correlation of greater than 0.8 were kept. This curve-fitting approach enabled use of replicate information instead of depending on the ensemble average necessary with Fourier transforms or Lombs-Scargle methods. This approach is viable due to our relatively large number of time samples. This data set was then loaded into a data mining program, GeneSpring 7.0 (Silicon Genetics, Redwood City, CA), and we normalized the value of each probe set on each chip to the average of that probe set on all chips. To identify genes with similar patterns of oscillation within the daily cycle, we applied Quality Threshold Clustering (QT) in GeneSpring using Pearson's correlation as the similarity measurement.


Data mining.

The assumption used to process this data set is that genes whose expression levels are part of the circadian rhythm will show one full oscillation every 24 h. To exploit this assumption, a nonlinear curve fit was conducted, which fitted a sinusoid to the data, including the replicates. We identified 122 probe sets which fit the model [A·sin(Bt + c)] with a R2 correlation greater than 0.8. Because of probe set redundancy (more than one probe set for a single gene), the 122 probe sets represent 109 genes. Using GeneSpring, we normalized the value of each probe set on each chip to the average of that probe set on all chips, such that the expression pattern of all probe sets oscillated approximately around 1. There appear to be two major patterns, as illustrated in Figure 1, which presents the expression pattern of all 122 probe sets. Some probe sets (represented by individual lines in Fig. 1) show expression patterns greater than one during the light period but less than one during the dark period, reflecting maximum expression during the animal's light-inactive period. Conversely, other probe sets reach a maximum (values greater than one) during the dark/active period. However, because of the richness of the data set, we were able to identify more discrete relationships between oscillations in expression and the light-dark periods. To group genes with similar patterns within the daily cycle, we applied Quality Threshold Clustering (QT clustering), yielding eight clusters. Figure 2 shows these eight clusters with the centroid (average of all of the genes in that cluster) highlighted by a white line. Fifty-seven percent of the genes (62 of 109) are in clusters 6, 7, or 8 with maximum expression during the dark/active period. Thirty-one percent of the genes (34 of 109) are in clusters 1, 2, 3, or 4, with maximum expression during the light/inactive period, and the remaining 12% (13 genes) are in cluster 5 with a maximum at hour 12, the transition between light and dark. Corticosterone reaches its maximum plasma concentration at hour 13.3 (Fig. 3). Table 1 provides a detailed list of all genes in each cluster, including probe set ID, accession number, Pearson's correlation coefficient with the centroid, gene symbol, gene name, and gene function.

Fig. 1.

A nonlinear curve fit using MATLAB was conducted which fitted a sinusoid function [A·sin(Bt + c)] to the data, including the replicates. Genes that could be curve fitted with a R2 correlation of greater than 0.8 were kept.

Fig. 2.

QT clustering of genes in muscle measured over 24 h. Each probe set has greater than a 0.75 Pearson's correlation with the centroid of the cluster.

Fig. 3.

Plasma corticosterone (CST) as a function of circadian time as measured by HPLC. Unshaded areas indicate light period and shaded areas indicate dark period.

View this table:
Table 1.

Circadian-regulated genes arranged by cluster


We examined the chip for probe sets for genes previously identified as involved in regulation of circadian patterns (34). The 230A chip contained probe sets for 17 of these transcription factors. However, only eight (Per1, Per2, BMAL1b, Bhlhb3, DBP, Nfil3, Nr1d1, and Nr1d2) showed distinct circadian oscillation. Figure 4 provides profiles for these eight transcription factors with circadian rhythms and their respective clusters. The chip did contain probe sets for Per3, Bhlhb2, RORB, RORC, CRY2, and CLOCK. On visual inspection, the signals for Per 3, RORB, and CRY2 were very low, which may prevent identification of rhythmicity. However, the signals for Bhlhb2, RORC, and CLOCK were reasonably intense but did not show any indication of oscillation.

Fig. 4.

Expression patterns of 8 clock-related transcription factors in skeletal muscle as a function of circadian time. Unshaded areas indicate light periods, and shaded areas indicate dark periods.

Previously, we conducted two time series experiments in which cohorts of adrenalectamized rats were given MPL either as a single bolus dose or chronic infusion, and the muscles were analyzed by gene arrays. Because a major regulator of circadian rhythms is the HPA axis, these arrays were examined for clock genes. Figure 5 shows the acute and chronic profiles for Per2, Bhlhb3, DBP, Nr1d1, and Nr1d2. In the acute profile, all five genes respond to the single-dose with a transient oscillation. Per2 began the oscillation with increased expression while DBP, Nr1d1, Bhlhb3, and Nr1d2 begin the oscillation with down-regulation. With chronic infusion, Per2 appears to oscillate throughout the 168-h infusion period beginning with enhanced expression. Consistent with the response to acute dosing DBP, Bhlhb3, Nr1d1, and Nr1d2 all show initial downregulation in response to infusion. DBP recovers baseline by 36 h and starts a slow decline throughout the remainder of the infusion. Nr1d1 recovers baseline by 24 h, shows a period of dampened oscillation for an additional 36 to 48 h, then begins a period of slow decline similar to DBP. Bhlhb3 looks almost like an inverse of Per2, in that it recovers baseline by about 48 h, after which it goes down again, and then shallowly starts back up again. Nr1d2 almost recovers baseline by 36 h, then maintains a new expression level slightly lower than baseline throughout the remainder of the infusion period. Because of their fluctuating behavior, for the most part, these five genes did not meet our criteria for corticosteroid regulation when the acute and chronic treatment data sets were initially mined (3). With acute dosing, BMAL1b oscillated with a pattern similar to Per2. However, with chronic infusion BMAL1b showed a strong peak of enhanced expression followed by a lower level of enhanced expression (Fig. 6). This figure also shows just the chronic profiles of Per1 and Nfil3 as the U34A chip did not contain probe sets for these genes. Both showed enhanced expression throughout the infusion period.

Fig. 5.

Expression patterns of 5 clock-related transcription factors in skeletal muscle as a function of time after methylprednisolone (MPL) administration to adrenalectomized animals. Left: data from acute (bolus 50 mg/kg) MPL dosing. Right: data from chronic (0.3 mg·kg−1·h−1) MPL infusion.

Fig. 6.

Expression patterns of 3 clock-related transcription factors in skeletal muscle as a function of time after MPL administration to adrenalectomized animals. Left: data from acute (bolus 50 mg/kg) MPL dosing. Right: data from chronic (0.3 mg·kg−1·h−1) MPL infusion.

The Affymetrix 230A chip was used for chronic infusion experiments, while the acute dosing experiment used the older U34A chip, which contained fewer probe sets. Because both this circadian experiment and the chronic infusion experiment used the same chip, we were able to make a direct comparison between corticosteroid-responsive genes and those with circadian rhythms. Data mining identified over 2,000 genes in the chronic infusion data set that were responsive to the infusion of 0.3 mg·kg−1·h−1 MPL via Alzet osmotic pumps (3). Of the 109 genes with circadian rhythms, only 55 were responsive to MPL infusion, while 67 were not. The circadian rhythm-regulated probe sets that are also responsive to MPL infusion are highlighted in bold in Table 1. Just as the 109 genes are not equally distributed among the eight clusters, the percentage of genes in each cluster that respond to MPL is not the same. For example, only 25% of the genes in cluster 1 are responsive, while 71% of the genes in cluster 6 are responsive. Cluster 6 is the first cluster to reach a maximum after plasma corticosterone reaches its maximum. The reverse is also true; not all genes that are responsive to MPL infusion have circadian rhythms. For example, Fig. 7 shows the response profiles of two genes to MPL infusion (bottom) together with their response profiles over the 24-h circadian time series (top). Both of these genes are likely involved in the atrophy caused by MPL infusion. The first is glutamine synthetase (GS), which is involved in the export of amino acid carbon from the musculature (22). The second is eukaryotic translation initiation factor 4e binding protein (Eif4ebp), which binds to the initiation factor and prevents the initiation of translation (33). Both genes are responsive to MPL. GS has a strong circadian rhythm, which has also been verified by quantitative RT PCR (Fig. 8). In contrast, Eif4ebp does not exhibit a circadian rhythm.

Fig. 7.

Expression patterns of GS (left) and Eif4ebp1 (right) following chronic infusion of MPL (0.3 mg·kg−1·h−1) over 168 h (top), and within the 24 h circadian cycle (bottom).

Fig. 8.

Circadian pattern of expression of glutamine synthetase measured by gene arrays (•) and by kinetic-based quantitative RT PCR (○).

Functional groupings.

The accession numbers for all 122 probe sets were analyzed using the National Center for Biotechnology Information Basic Local Alignment Search Tool (BLAST) to both definitively identify the gene and to obtain alternative names and symbols. Using this information, we extensively researched each gene to identify and categorize its function. The genes were then separated into the following 11 functional groupings: carbohydrate metabolism, cell cycle, cytoskeleton/extracellular, immune, lipid metabolism, mitochondrial, protein degradation, signaling, small molecule metabolism, transcription, and transport. In several cases, we needed to decide on the most appropriate category because some genes could fit into more than one group. For example, the transcription factor, sterol regulatory element-binding transcription factor 1 (SREBP1), was placed in the category of lipid metabolism because it is critical to this function, while alternative placement in transcription would also have been valid (8). Similarly, two genes, WD repeat and suppressor of cytokine signaling (SOCS) box-containing 1 and ankyrin repeat and SOCS box-containing protein 2, were placed in the immune functional category rather than signaling. Both of these genes are adaptor/regulatory modules that contain a SOCS box that is essential for the physiological inhibition of interferon signaling (28). Table 2 provides the grouping of genes into functional categories.

View this table:
Table 2.

Functional categorization of circadian regulated genes


This report describes a gene array analysis of circadian oscillation of mRNA expression in the skeletal muscle of adult male Wistar rats. Animals were killed at nine time points during a 12-h light period and nine corresponding time points during a 12-h dark period (n = 3 per time point). Muscle RNAs from each of the 54 animals were applied to individual Affymetrix GeneChips Rat Genome 230A, which contained 15,967 probe sets. Analysis yielded 122 probe sets (representing 109 genes) with a 24-h periodicity. Because of the richness of this data set, we were able to apply QT clustering and identified eight clusters with maxima at different times during the cycle. Four clusters peaked during the light period and contained 31% of the genes, three clusters peaked during the dark period and contained 57% of the genes, and one peaked very close to the light to dark transition (12% of the genes). In 13 cases, there were two probe sets on the chip for a gene. In 12 of these cases, both probe sets were in the same cluster with similar correlations to the centroid. However, in one case, dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 2 (Dyrk2), one probe set was in cluster 1 with a maximum at 2 h, and the other was in cluster 2 with a maximum at 4 h. Both probe sets BLAST to the consensus sequence for this gene (NM_001108100.1). Both have similar scores around 1,000, indicating a very good match with the consensus sequence. Visual inspection of the expression patterns confirms that they are different in both time and intensity.

Regulation of the central SCN clock involves as many as 17 genes (34). The 230A array contained probe sets for 14 of these. Eight (Per1, Per2, BMAL1b, Bhlhb3, DBP, Nfil3, Nr1d1, and Nr1d2) showed distinct circadian oscillations in skeletal muscle. However, six (Per3, Bhlhb2, RORB, RORC, CRY2, and CLOCK) did not demonstrate oscillations in our study. In the cases of Per3, RORB, and CRY2, this may be due to very low signal intensities. However, the signal intensities of CLOCK, Bhlhb2, and RORC were reasonable. Zambon et al. (40), using a much sparser data set and statistical analysis, did report a circadian rhythm for CRY1 in human skeletal muscle (40). However, visual inspection of their data does not suggest an oscillation with a 24 h periodicity, which was the criterion that we applied here. While the 230A chip used here did not contain a probe set for CRY1, it did contain a probe set for CRY2. McCarthy et al. (20) reported circadian oscillation of CRY2 in mouse skeletal muscle using RTPCR. This suggests that our result may be due to the inability of the CRY2 probe set on the 230A array to measure the signal. In contrast, CLOCK exhibited a strong signal but no circadian pattern in our studies. It has been reported by others that at least in some tissues, CLOCK is expressed at tonic levels and that cycling is due to the rhymicity of its heterodimeric partner BMAL (27). Steeves et al. (30), using Northern blot hybridization, did report the expression of CLOCK in skeletal muscle (30). However, they only measured one time point, so oscillation was not measured. Our result confirms their observation that CLOCK is expressed in skeletal muscle and indicates that it does not have circadian rhythmicity. In addition, in our chronic infusion studies CLOCK does not respond significantly to MPL.

Because skeletal muscle represents about 40% of the mass of a normal healthy mammal, it is a major target of endogenous glucocorticoids. The large protein mass of the musculature provides a major underpinning for systemic glucose homeostasis. Glucocorticoids shift muscle into a net negative nitrogen balance, and much of the resultant amino acid carbon is used to synthesize glutamine. Glucocorticoids also enhance the expression of enzymes in the liver and kidney that use the amino acid carbon released from the musculature for gluconeogenesis. We also used kinetic-based RT PCR to measure the expression of mRNA for glutamine synthesis in the same muscle used for this array experiment. As expected, RT PCR and array data exhibit virtually identical circadian patterns (Fig. 8). The array results presented here put that observation within the broader context of all gene expression changes and demonstrates that this enzyme is in cluster 7, which reaches a maximum about 5 h after the maximum of the corticosterone rhythm. Synthetic glucocorticoids, corticosteroids, are widely used therapeutic agents for modulation of immune/inflammatory responses. Previously, we measured the time course of the responses of skeletal muscle to both a single bolus dose and continuous infusion of the corticosteroid methylprednisolone (3). Because those experiments were conducted with adrenalectamized animals, corticosterone circadian rhythmicity was absent. Although corticosteroids are widely used agents to modulate immune/inflammatory responses, they have a low therapeutic index. The effect of these agents on the musculature contributes substantially to the side effects of corticosteroids. Not only does the musculature atrophy to provide amino acid carbon for increased gluconeogenesis by the liver and kidney but also the musculature becomes insulin resistant. The result is a condition known as steroid diabetes.

The effects of glucocorticoids/corticosteroids on skeletal muscle are quite complex. Some, like the enhanced expression of glutamine synthetase, appear to be due to the interaction of the hormone receptor complex with a GRE 5′ to the coding region. Others are less direct and more complex due to the effects of corticosteroids on, among other things, the expression of other transcription factors. For example, corticosteroids enhance the expression of myostatin, a muscle-specific transcription factor (18). We examined our previously obtained acute and chronic data sets to ascertain the effect of chronic MPL infusion and where possible acute dosing on the eight transcription factors identified in the current data set as having circadian rhythms. In the six cases for which acute profiles were available, the response was that the drug caused the expression to begin oscillation, which dampened with time. The first phase of this oscillation was up for BMAL1b and Per2, while the first phase of the oscillation was down for Bhlhb3, DBP, Nr1d1, and Nr1d2. With chronic infusion, the initial effect was in the same direction as the acute dosing. Per2, Bhlhb3, DBP, and Nr1d1 continued oscillating with different periodicities throughout the infusion period, while Nr1d2 made a significant initial decline and almost recovered baseline. However, three genes BMAL1b, Nfil3, and Per1 showed very significant prolonged changes in expression. BMAL1b showed a very sharp increase in expression that dropped by about 50% by 36 h to a state of continuous enhanced expression over baseline throughout the infusion period. Per1 quickly went up 5- to 6-fold and remained at this level throughout the infusion period. Nfil3 also quickly went up about 5- to 6-fold and slowly began to decline over the infusion period but was still about twice the baseline at the end (168 h). Taken together, the results suggest that the expression of some of these transcription factors, BMAL1b, Nfil3, and Per1, are more closely linked to the corticosterone circadian rhythm than others (Per2, Bhlhb3, DBP, Nr1d1, and Nr1d2). However, the fact that all of the measurable transcripts in our acute bolus dosing study begin to oscillate suggest that the corticosterone rhythm may be the initiator of their rhythmic pattern of gene expression. This is consistent with the observation by McCarthy that some genes in muscle maintain rhythmic expression in CLOCK mutant mice and that diurnal and nocturnal animals can phase shift not only their glucocorticoid rhythm but also the rhythms of gene expression in peripheral tissues (20).

We also compared circadian-regulated genes with those identified as regulated by the chronic infusion of MPL. Of the more than 2,000 probe sets identified as responsive to MPL infusion, only 57 probe sets (55 genes) with circadian rhythms were among the glucocorticoid-regulated group. However, these data sets together allowed us to address two basic but related questions. The first is: do all genes that respond to corticosteroids have circadian rhythms? The second is: do all genes with circadian rhythms respond to corticosteroid? The answer to both questions is “No.” Only 55 of the genes identified were both circadian and MPL-responsive. The facts that all genes that respond to MPL are not circadian and that all genes with circadian rhythms do not respond to MPL suggest that there is some diversity in mediating mechanisms. This result is consistent with both the previously described observations comparing our acute and chronic profiles and the current observation that the genes can be separated into eight clusters (3).

If an animal is diurnal, increased mRNA expression near the end of the dark period begins to prepare the animal for activity and feeding. Similarly, downward changes during the end of the light period prepare the diurnal animal for inactivity and rest. For example nocturin, a deadenylase that has peak expression in the early dark period in rodents, has been implicated in the regulation of genes important to feeding and energy metabolism (9). Rats are nocturnal, and cycling of gene expression in peripheral tissues like the muscle is reversed relative to humans who are essentially diurnal. On the basis of extensive literature searching, we placed each of the 109 genes into functional groupings as follows: carbohydrate metabolism, cell cycle, cytoskeleton/extracellular, immune, lipid metabolism, mitochondrial, protein degradation, signaling, small molecule metabolism, transcription, and transport. The objective of this analysis was to ascertain which functions are most influenced by circadian patterns and when they are occurring. The most populated functional group is transcription (Table 2). In addition to the 25 genes that appear on this table, Nfil3 (which we placed in the immune grouping), Klf15 and nr4a1 (in the carbohydrate grouping), and srebpf1 (in the lipid grouping) are also transcription factors. Thus about one-fourth of the genes with circadian rhythms are involved in the regulation of transcription/translation. It is also important to note that genes in this grouping occur in all eight clusters, which illustrate the complex nature of circadian rhythms on gene expression. Almost all of the genes in this grouping are involved in transcription. However, a notable exception is Pumilio homolog 2 (Pum2), which is in cluster 3. Cluster 3 also contains zinc finger homeodomain 4 (Zfhx4). Pum2 is involved in mRNA degradation and has been shown to promote stem cell proliferation (37). Kostich and Sanes (16) observed that Zfhx4 was high during proliferation but decreased with differentiation. These observations, together with the observation that transforming growth factor, beta-2 (tgfb2) in the signaling group, which is an inhibitor of differentiation (29), is also in cluster 3 suggests that muscle stem cell proliferation may be initiated during the middle of the light/inactive period. Reinforcing this conclusion is the presence of nuclear mitotic apparatus protein 1 (numa1), PHD finger protein 17 (Phf17), and protein phosphatase 3, catalytic subunit, alpha isoform (Ppp3ca) from the cell cycle group in cluster 3 as well.

The second most populated group is mitochondrial with 21 genes. In stark contrast to the transcription grouping, 14 of these genes are in cluster 7, which reached its maximum in the middle of the dark/active period of the animal. The gastrocnemius muscle is a mixed-fiber muscle, but because the animals are awake but relatively sedentary, the musculature in general is using primarily lipid fuels that require mitochondria. This grouping also illustrates quite well that genes that work together are expressed together.

The third most populated group is signaling, with 14 genes. In addition to the cluster 3 gene, tgfb2, the signaling group also contains two probe sets for protein phosphatase 2, regulatory subunit B56 alpha (ppp2r5a), a cluster 2 gene that would be consistent with cell proliferation occurring during the light period (21). In contrast, the expression of genes such as Stathmin 1 (35), along with the mitochondrial genes in cluster 7, suggests that the circadian expression is by differentiated cells. All genes in the functional immune group, except Spondin 2 (Spon2), which is involved in innate immune responses (12), are expressed in clusters 6, 7, and 8, which reach maxima during the dark/active period.

Although there are only seven genes in the lipid group, several are central to the regulation of lipid metabolism in skeletal muscle. The first is SREBP1 which is a master transcription factor that plays a role in the expression of many lipogenic genes (8). SREBP1 is in cluster 1 that reaches a maximum early in the light period. Cluster 2 contains lanosterol 14-demethylase (CYP51), an enzyme important for cholesterol biosynthesis (31). Cluster 3 contains three genes that are central to muscle lipid metabolism. The first is low-density lipoprotein receptor (LDLR). The second is anigopoietin-like 4 (angptl4), which inhibits lipoprotein lipase (39), and the last is insulin-induced gene 1 (insig1), which facilitates the retention of the SREBP cleavage-activated protein/SREBP complex in the endoplasmic reticulum (1). Cluster 6, which has a maximum during the dark period, contains two genes important for the synthesis of triglycerides. The first is acyl-CoA synthetase long-chain family member 6 (acsl6), which catalyzes the ligation of long-chain fatty acids with coenzyme A (19). The second is diacylglycerol O-acyltransferase homolog 2 (Dgat2), which catalyzes the covalent linking of diacylglycerol to long-chain fatty acyl-CoAs (17).

The functional group that we labeled cytoskeleton/extracellular contains eight genes. Only one of these, syndecan 2 (Sdc2) is in a cluster that has a maximum during the light period. For the most part, these genes seem to reflect muscle mechanical activity during the dark period. In contrast, the distribution of genes in the cell cycle group spans many clusters, which illustrates the progression of stem cells through the cell cycle with cell division beginning during the light period.

Surprisingly few genes with circadian rhythms are associated with protein degradation. Enhanced expression of GS, which is glucocorticoid responsive and is associated with several atrophy conditions is found in cluster 7, along with a few other relevant genes. In the case of GS, data suggest that the maximum expression of this gene is directly associated with the maximum of the corticosterone rhythm, which occurs several hours earlier (38). However, no such data are available for other cluster 7 genes associated with protein degradation. What is surprising is that ubiquitin B, C (Ubb, Ubc) is in cluster 1. Enhanced expression of this gene is associated with several muscle atrophy conditions, including atrophy induced by glucocorticoids (14). The genes in the small molecule metabolism group concentrate in cluster 6, which probably reflects the increased activity of the animal during the dark period. Likewise, surprisingly few genes associated with carbohydrate metabolism have circadian rhythms. The fact that 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 1 (Pfkfb1), along with most of the mitochondrial genes are in cluster 7, suggest that this is the period of highest energy demand by the musculature. However, these animals were housed in single cages and were relatively sedentary, which may influence energy demands by muscle fibers.

The last grouping was transport. Three of the four genes in this group were in cluster 6, which is rational, given the activity of the animals. The last, Slc41a3, is a Mg+ transporter found in cluster 2 (36). Very little data are available on the function of this gene in skeletal muscle.

These microarray experiments only reflect changes in the amount of expressed message for a gene. Although we only identified 109 genes whose message had a rhythm with a 24-h periodicity, the nature of some of these genes will magnify the effect of the rhythmicity. For example, rhythmicity in Pum2, which mediates mRNA degradation, will likely change the amount of many other messages. However, these changes may not have a strict 24-h periodicity. Similarly, the fact that many phosphatases and kinases have rhyhmicity will magnify oscillations at the functional level. Similar to secondary changes in the amount of message, rhythmic changes in function may not necessarily have a 24-h periodicity.

Perspectives and Significance

Although much progress has been made in understanding the control of circadian rhythms in the SCN, translation of control to peripheral tissues, which ultimately mediate systemic biological rhythms, is much less well understood. The present study identifies patterns of circadian gene expression changes in one peripheral tissue: skeletal muscle. It also identifies eight transcription factors that have been implicated in central clock mechanisms as having oscillating patterns in muscle as well. Extensions of such studies to other peripheral tissues such as liver and adipose tissue will address the question of what genes are regulated in a circadian pattern in multiple tissues of an organism and will allow a better understanding of how changes in multiple peripheral tissues can be integrated into systemic biological effects, such as rhythmic changes in energy metabolism. This study also demonstrates that although glucocorticoids serve as important signals for translating rhythms from central control centers in the brain to the periphery, factors other than glucocorticoids likely also play a role in this translation.


This work was supported by Grants GM 24211 and GM 67650 from the National Institute of General Medical Sciences, National Institutes of Health (NIH), Bethesda, MD, and by a grant from NASA. This data set was developed under the auspices of a grant from the National Heart, Lung, and Blood Institute/NIH Programs in Genomic Applications HL-66614. E. Yang and I. P. Androulakis also acknowledge support from NSF Grant 0519563 and EPA Grant GAD R 832721-010.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract