Department of Molecular Therapy, National Institute of Neuroscience, National Center of Neurology and Psychiatry, Tokyo, Japan
Department of Molecular Therapy, National Institute of Neuroscience, National Center of Neurology and Psychiatry, 4-1-1 Ogawa-higashi, Kodaira, Tokyo 187-8502, Japan. Telephone: +81-42-346-1720; Fax: +81-42-346-1750
Skeletal muscle satellite cells play key roles in postnatal muscle growth and regeneration. To study molecular regulation of satellite cells, we directly prepared satellite cells from 8- to 12-week-old C57BL/6 mice and performed genome-wide gene expression analysis. Compared with activated/cycling satellite cells, 507 genes were highly upregulated in quiescent satellite cells. These included negative regulators of cell cycle and myogenic inhibitors. Gene set enrichment analysis revealed that quiescent satellite cells preferentially express the genes involved in cell-cell adhesion, regulation of cell growth, formation of extracellular matrix, copper and iron homeostasis, and lipid transportation. Furthermore, reverse transcription-polymerase chain reaction on differentially expressed genes confirmed that calcitonin receptor (CTR) was exclusively expressed in dormant satellite cells but not in activated satellite cells. In addition, CTR mRNA is hardly detected in nonmyogenic cells. Therefore, we next examined the expression of CTR in vivo. CTR was specifically expressed on quiescent satellite cells, but the expression was not found on activated/proliferating satellite cells during muscle regeneration. CTR-positive cells reappeared at the rim of regenerating myofibers in later stages of muscle regeneration. Calcitonin stimulation delayed the activation of quiescent satellite cells. Our data provide roles of CTR in quiescent satellite cells and a solid scaffold to further dissect molecular regulation of satellite cells.
Disclosure of potential conflicts of interest is found at the end of this article.
Muscle satellite cells, which account for 2%–5% of the total nuclei in adult skeletal muscle, play a major role in muscle regeneration [1, 2]. Under normal conditions, satellite cells are found external to the myofiber plasma membrane and beneath the muscle basal lamina  and are mitotically quiescent in the adult skeletal muscle [4, 5]. When activated by muscle damage, they proliferate, differentiate, fuse with each other or injured fibers, and eventually regenerate mature myofibers under the influence of innervation . Recently, it was clearly demonstrated that the proliferation capacity of satellite cells in vivo is robust and that the contribution of interstitial cells or bone marrow-derived cells to muscle fiber regeneration is limited . Importantly, a small fraction of activated satellite cells exit the cell cycle and return to the quiescent satellite state during muscle regeneration to maintain their numbers and the regenerative capacity of muscle.
Besides muscle fiber repair, satellite cells are also responsible for postnatal growth  and hypertrophy of skeletal muscle , and impairment of their functions is related to several pathological conditions, for example, muscular dystrophies and aging-related muscle atrophy . Moreover, several studies showed that satellite cells differentiate into adipogenic cells or osteocytes in vitro [11, –13], implying that they contribute to the fatty infiltration seen in Duchenne muscular dystrophy. Thus, normal functioning of satellite cells is indispensable for the integrity of skeletal muscle, and the cells themselves are an important source of cells for cell therapy of muscle diseases, making it valuable to clarify the molecular regulation of maintenance, activation/proliferation, and differentiation in satellite cells.
Like hematopoietic stem cells, most satellite cells are in a quiescent and undifferentiated state in the adult. Although quiescence is important to retain the proliferative and differentiative potential of satellite cells throughout the lifetime, the molecular regulation of quiescence remains poorly defined. Recent studies suggested that myostatin, a skeletal muscle-specific transforming growth factor-β superfamily member, suppresses the activation of satellite cells . Myostatin has been shown to induce a potent cyclin-dependent kinase inhibitor, p21(Cdkn1a), in vitro . Other in vitro studies suggested that the decrease of MyoD protein and induction of another cyclin-dependent kinase inhibitor, p27(Cdkn1b) , and a Rb-related pocket protein, p130 [16, 17], are involved in the attainment of quiescence by proliferating myoblasts.
We previously reported a method to purify quiescent satellite cells from adult skeletal muscle using the fluorescence-activated cell sorting (FACS) technique and a novel antibody named SM/C-2.6 . In this study, to clarify the molecular regulation of quiescent satellite cells, we performed genome-wide gene expression profiling of quiescent satellite cells isolated from C57BL/6 mice. Expression analysis of individual genes identified several candidate genes that regulate dormancy of satellite cells. Gene set enrichment analysis (GSEA) revealed that the gene sets involved in cell-cell adhesion, cell growth, copper and iron ion homeostasis, lipid transport, and formation of extracellular matrix were coordinately upregulated in quiescent satellite cells. Furthermore, we demonstrate that calcitonin receptor (CTR) is expressed specifically on quiescent satellite cells in vivo and that calcitonin significantly attenuates the activation of satellite cells. Our study is the first report of in-depth gene expression analysis of quiescent satellite cells and will greatly facilitate the investigation of molecular regulation of satellite cells in both physiological and pathological conditions.
Materials and Methods
All procedures using experimental animals were approved by the Experimental Animal Care and Use Committee at the National Institute of Neuroscience. C57BL/6 mice were purchased from Nihon CLEA (Tokyo, http://www.clea-japan.com).
Preparation of Satellite Cells and Nonmyogenic Cells from Mouse Limb Muscles
Mononuclear cells were prepared from fore- and hindlimb muscles of 8- to 12-week-old female C57BL/6 mice as described  and incubated on ice for 30 minutes in the presence of a 1:200 dilution of phycoerythrin-conjugated anti-CD45 (clone: 30-F11) and biotinylated SM/C-2.6 . Cells were then incubated with streptavidin-labeled allophycocyanin on ice for 30 minutes and resuspended in phosphate-buffered saline (PBS) containing 2% fetal bovine serum (FBS) and 2 μg/ml propidium iodide (PI). Cell sorting was performed on a FACSVantage SE flow cytometer (BD Biosciences, San Diego, http://www.bdbiosciences.com). Dead cells were excluded by PI gating. All antibodies and reagents for FACS analysis were purchased from BD Pharmingen (San Diego, http://www.bdbiosciences.com/index_us.shtml).
Satellite cells were cultured in growth medium consisting of high-glucose Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Carlsbad, CA, http://www.invitrogen.com) containing 20% fetal calf serum (FCS; Trace Biosciences, New South Wales, Australia), 2.5 ng/ml basic fibroblast growth factor (Invitrogen), and penicillin (100 U/ml)-streptomycin (100 μg/ml) (Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com) on culture dishes coated with Matrigel (BD Biosciences). Single living myofibers were prepared as described  and transferred to Matrigel-coated 24-well culture dishes (one fiber per well). After a 2-day culture in growth medium with or without elcatonin, satellite cells that had detached from muscle fibers were counted.
FACS-sorted cells were collected on glass slides by Cytospin 3 (Thermo Shandon Inc., Pittsburgh, http://www.thermo.com) and immunostained as described . Cultured cells were fixed on 8-well Lab-Tek Chamber Slides (Nunc, Rochester, NY, http://www.nuncbrand.com) and stained as described [19, 21] with mouse anti-Pax7 (1:100; clone: Pax7; Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.uiowa.edu/∼dshbwww), mouse anti-MyoD (1:200; clone: 5.8A; NeoMarkers; Lab Vision, Fremont, CA, http://www.labvision.com), mouse anti-myogenin (1:100; clone: F5D; Developmental Studies Hybridoma Bank), rabbit anti-Ki67 (1:2; Ylem, Rome), or rabbit anti-p57 antibodies (1:50; GeneTex, San Antonio, http://www.genetex.com) at 4°C overnight and then reacted with secondary antibodies conjugated with Alexa 488 or Alexa 568 (Molecular Probes, Eugene, OR, http://www.probes.invitrogen.com). Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI). Images were photographed using a phase-contrast and fluorescence microscope IX70 (Olympus, Tokyo, http://www.olympus-global.com) equipped with a Quantix air-cooled CCD camera (Photometrics, Kew, VIC, Australia, http://www.photometrix.com.au) and IP Lab software (Scanalytics, Rockville, MD, http://www.scanalytics.com).
Immunostaining of muscle cryosections was performed as previously described  using rat anti-laminin α2 (1:200; clone 4H8–2; Alexis Biochemical, Lausen, Switzerland, http://www.axxora.com), rabbit anti-M-cadherin , rabbit anti-human CTR (1:200; Serotec Ltd., Oxford, U.K., http://www.serotec.com), goat anti-Notch 3 (1:100; R&D Systems Inc., Minneapolis, http://www.rndsystems.com), or mouse anti-Pax7. Rabbit anti-mouse HeyL polyclonal antibody was produced in our laboratory. In brief, the DNA fragment corresponding to amino acids 220–287 of mouse HeyL (GenBank: NM_013905) was fused to glutathione S-transferase in the pGEX-1 Lambda T vector (GE Healthcare, Uppsala, Sweden, http://www.gehealthcare.com). The purified fusion protein was used to immunize New Zealand White rabbits. The obtained serum was affinity-purified. For Pax7 staining, an M.O.M. kit (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) was used to block endogenous mouse IgG. For CTR staining, horseradish peroxidase-conjugated anti-rabbit IgG donkey secondary antibody (1:100; GE Healthcare) and Alexa 568-conjugated Tyramid (Molecular Probes) were used to amplify the signal. Nuclei were counterstained with TOTO-3 (1:5,000; Molecular Probes) or DAPI. The images were recorded using a confocal laser scanning microscope system TCSSP (Leica, Heerbrugg, Switzerland, http://www.leica.com) or Axiophot microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com).
Cell Cycle Analysis
Muscle-derived mononucleated cells or cultured SM/C-2.6 positive cells were suspended at 106 cells per milliliter in DMEM (Invitrogen) containing 2% FBS (Trace Biosciences), 10 mM Hepes, and 10 μM Hoechst 33342 (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) and incubated for 45 minutes at 37°C. An additional incubation was performed in the presence of 10 μg/ml Pyronin Y (Sigma-Aldrich) for 45 minutes at 37°C. Cells were then washed with PBS containing 2% FCS. Muscle-derived mononucleated cells were stained with SM/C-2.6 antibody and analyzed by FACSVantage SE flow cytometer.
Cell Proliferation Assay
After cell sorting, quiescent satellite cells were plated on 96-well culture plates at a density of 3,000–8,000 in the absence or presence of elcatonin (0.01–0.1 U/ml) (Asahi Kasei Pharma Corporation, Tokyo, http://www.asahi-kasei.co.jp/asahi/en) and cultured for 1–2 days. Then 5-bromo-2′-deoxyuridine (BrdU) (10 μM) was added to the culture. To examine the effects of elcatonin on activated satellite cells, satellite cells were cultured for 3 days and then elcatonin was added to the culture 24 hours before addition of BrdU. Twenty-four hours later, BrdU uptake was quantified by cell proliferation enzyme-linked immunosorbent assay, BrdU Kit (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com), and lumi-Image F1 (Roche). In Figure 6B, cells were exposed to elcatonin for 30 minutes and washed twice with PBS and then plated on culture dishes.
Detection of Apoptotic Cells
Cells were cultured on 8-well Lab-Tek chamber slides with or without elcatonin. Apoptotic cells were detected by rhodamine fluorescence using an ApopTag Red In Situ Apoptosis Detection Kit (Chemicon, Temecula, CA, http://www.chemicon.com).
RNA Extraction and Reverse Transcription-Polymerase Chain Reaction
Total RNA was extracted from sorted or cultured cells with a Qiagen RNeasy Mini kit according to the manufacturer's instructions (Qiagen, Hilden, Germany, http://www1.qiagen.com) and then reverse-transcribed into cDNA by using TaqMan Reverse Transcription Reagents (Roche). The polymerase chain reaction (PCR) was performed with cDNA products under the following cycling conditions: 94°C for 3 minutes followed by 30–40 cycles of amplification, annealing, and extension (94°C for 15 seconds, 58°C for 30 seconds, and 72°C for 30 seconds) with a final incubation at 72°C for 5 minutes. Specific primer sequences used for PCR are described in supplemental online Materials and Methods.
Target Synthesis, Gene Chip Hybridization, and Data Acquisition
To label antisense RNA (aRNA) with biotin for microarray hybridization, we followed the protocol supplied by the manufacturer (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). Because the starting amount of total RNA was 100 ng for the sorted SM/C-2.6+ cell fraction, we used a two-cycle biotin aRNA synthesis kit (Affymetrix). Labeled aRNA was fragmented according to Affymetrix GeneChip protocol and then hybridized to Affymetrix MOE430A GeneChip arrays for 16 hours. After washing, the gene chips were stained according to the instrument's standard Eukaryotic GE WS2v4 protocol using antibody-mediated signal amplification. The signal was determined, using the Microarray Suite (MAS) 5.0 absolute analysis algorithm, as the average fluorescence intensity among the intensities obtained from the probe set. The signal of a probe set was calculated as the one-step biweight estimate of combined differences of all the probe pairs (perfectly matched and mismatched) in the probe set. A one-sided Wilcoxon's signed rank test was used to calculate a p value that reflects the significance of differences between perfectly matched and mismatched probe pairs. The p value was used to make the absolute call for probe sets. A “Present” call was assigned to transcripts for p values between 0 and .04, a “Marginal” call was assigned to transcripts for p values between .04 and .06, and an “Absent” call was assigned to transcripts for p values between .06 and 1.0.
Microarray Data Analysis
Scanned output files were analyzed by the probe level analysis package MAS 5.0 (Affymetrix). The Present/Absent call provided by the Affymetrix programs was used for the first selection. The MAS 5.0-generated raw data were uploaded to GeneSpring software version 7.0 (Silicongenetics, Redwood, CA, http://www.chem.agilent.com/scripts/PHome.asp). Data normalization was achieved by one of two methods: (a) each signal was divided by the 50th percentile of all signals in a specific hybridization experiment or (b) each signal was divided by the median of its values in all samples. A more reliable list of “5-fold changing” genes was obtained by applying the filtering options of GeneSpring. Present calls in all (four) quiescent or activated satellite cell probes were selected and a restriction, which passed genes with raw data above 100, was applied. Then, using all the quiescent and activated satellite cells as data, we performed a one-way analysis of variance test between the quiescent satellite cell group and the activated satellite cell group. In particular, a parametric test, with variances assumed equal (Student's t test, p value cut-off .05; multiple testing correction: Benjamini and Hochberg false discovery rate), was applied. The genes passing all these filters and tests were selected as “5-fold changing genes.” Nonmyogenic cells (SM/C-2.6−/CD45− cells) were also prepared four times.
Gene Set Enrichment Analysis
GSEA  is a statistical analysis of sets of gene expression profiles, separated by phenotypic labels. Using GSEA, we can test hypotheses concerned with predefined sets of genes; the rank orderings of the genes in the whole gene set calculated with a given ranking method are random with regard to a given classification of samples. As a result of the analysis, nominal p values, family-wise error rate p values, and false discovery rate (FDR) q values for test hypotheses (thus for gene sets) were obtained.
In our analysis, we used the GSEA-P software package , which is available from the Broad Institute (Cambridge, MA, http://www.broad.mit.edu). We prepared, as input to the GSEA-P, the MAS 5.0-generated raw signal data and gene sets derived independently. We chose genes on the chip that were detected (the Present call was assigned) in at least one sample (17,150 of 22,626). The raw signals of the chosen genes were normalized so that their total sum was 1. Because the total amount of mRNA in a quiescent satellite cell (QSC) is much less than that in an activated satellite cell (ASC), the normalized signal should be understood as a relative signal among the chosen genes. To compile the gene sets, we assigned each probe to a gene ontology (GO) category  using annotations of the MOE430A chip (September 22, 2005) provided by Affymetrix. Therefore, these gene sets reflect the structure of the GO categories and subcategories of molecular function (MF), biological process (BP), and cellular component (CC). The 17,150 genes chosen comprised 1,674, 1,698, and 412 gene sets in the MF, BP, and CC subcategories, respectively, and were reduced to 162, 218, and 85 after filtering out gene sets with sizes smaller than 20 or larger than 1,000. We ran the GSEA-P with the signal-to-noise option for its ranking metric, with permutation over phenotype labels of QSC and ASC samples, and repeated it 2,000 times with the “weighted” option for its scoring scheme.
Results and Discussion
Isolation of Quiescent Satellite Cells from Mouse Skeletal Muscle
First, to obtain RNA samples for microarray analysis, we prepared mononuclear cells from 8- to 12-week-old C57BL/6 mouse muscle, and the SM/C-2.6+ fraction was collected as the satellite cell fraction by FACS  (Fig. 1A). Consistent with our previous report, more than 97% of fresh SM/C-2.6+ cells expressed Pax7 (Fig. 1B) but were mostly negative for both MyoD (Fig. 1Ca, 1Cb) and Ki67 (Fig. 1Ce, 1Cf). After 4–5 days of culture, more than 98% of SM/C-2.6+ cells expressed MyoD (Fig. 1Cg, 1Ch) and Ki67 (Fig. 1Ck, 1Cl). Both freshly isolated, uncultured SM/C-2.6+ cells and SM/C-2.6+ cells cultured in growth medium were negative for myogenin expression (Fig. 1Cc, 1Cd, 1Ci, 1Cj), but these cells started to express myogenin and differentiated well into multinucleated myotubes after mitogen withdrawal (data not shown). In contrast, more than 99% of freshly isolated SM/C-2.6−/CD45− cells were negative for Pax7 expression (Fig. 1Bc, 1Bd), and cultured SM/C-2.6−/CD45− cells did not express MyoD (data not shown), again indicating that myogenic cells are highly enriched in the SM/C-2.6+ fraction.
The forward and side scatter profiles of freshly isolated SM/C-2.6+ cells showed that they are small and uniform in granularity (data not shown). In fact, as shown in Figure 1D, the cell size of fresh SM/C-2.6+ cells was estimated to be approximately one-half that of cultured SM/C-2.6+ cells based on the forward scatter profile, indicating that the freshly isolated SM/C-2.6+ cells were not activated yet. Pyronin Y staining showed the small amount of RNA content in freshly isolated SM/C-2.6+ cells (Fig. 1D). In general, a Pyroninlow and Hoechst 33342low fraction is considered to be G0 cells . Pyronin Y and Hoechst double staining shows that approximately 90% of fresh SM/C-2.6+ cells were in the G0 phase of the cell cycle. In contrast, 90% of cultured SM/C-2.6+ cells were cycling (Fig. 1E).
Thus, our procedure, which takes 5–6 hours in total to isolate 1–2 × 105 SM/C-2.6+ cells from one C57BL/6 mouse, enables us to isolate satellite cells still in a quiescent and undifferentiated state. The yield corresponds to 10%–15% of the total mononucleated cells obtained from mouse hind limb muscles by enzymatic digestion. Therefore, in this report, we call freshly isolated SM/C-2.6+ cells “quiescent satellite cells” and cultured, proliferating SM/C-2.6+ cells “activated satellite cells.” Our procedure was also applicable to dystrophin-deficient mdx muscle with modifications, although 30%–40% of mdx satellite cells are Ki67-positive (M. Ikemoto et al., submitted manuscript). Unfortunately, SM/C-2.6 did not react with satellite cells from dystrophin-deficient dystrophic dogs (data not shown).
Single Gene Analysis of Quiescent and Activated/Proliferating Satellite Cells
We prepared RNA samples from quiescent satellite cells and activated satellite cells and performed microarray analysis using Affymetrix GeneChips. Hybridization and data collection were performed four times using independent preparations of cells and RNA samples for each cell fraction. Raw data are available at http://www.ncbi.nlm.nih.gov/geo. The Gene Expression Omnibus accession number is GSE3483.
First, we compared the expression levels of individual genes in quiescent and activated states using GeneSpring software. We found that 507 genes (665 probes) were expressed in quiescent satellite cells at more than fivefold higher levels than in activated satellite cells (Fig. 2A). We roughly categorized these 507 genes into 11 gene groups: cell adhesion (15 genes), cell cycle regulation (26), proteolysis (21), cytoskeleton (13), cell surface (41), extracellular (61), immunoresponse (22), signal transduction (81), transcription (67), transport and metabolism (82), and unknown (78) based on Gene Ontology and listed all of them in supplemental online Table 1. On the other hand, 659 genes (814 probes) were upregulated (>fivefold) in the activated state (supplemental online Table 2). We also examined the gene expression of proliferating satellite cells/myoblasts in vivo that were directly isolated from regenerating muscle 2 days after cardiotoxin injection. The activated and proliferating satellite cells in vivo showed an expression profile quite similar to satellite cells cultured in vitro (data not shown).
Upregulation of Cell Cycle Regulators in Quiescent Satellite Cells
Under normal conditions, most satellite cells are in the G0 phase of the cell cycle, possibly preventing their premature exhaustion. It is of note that nine genes encoding negative regulators of the cell cycle were highly upregulated in the quiescent stage: Rgs2 (regulator of G-protein signaling 2) (×69, ×23), Rgs5 (×37, ×21), Pmp22 (peripheral myelin protein 22)/Gas3 (growth arrest specific 3) (×25), Cdkn1c (cyclin-dependent kinase inhibitor 1C)/p57 (×14), Spry1 (sprouty homolog 1) (×11), Gas1 (×7, ×6), Reck (reversion-inducing-cysteine-rich protein with kazal motifs) (×6), Ddit3 (DNA-damage inducible transcript 3) (×6), and Trp63 (transformation-related protein 63) (×5) (supplemental online Table 1). Reverse transcription (RT)-PCR confirmed that Rgs2, Pmp22, p57, and Spry1 are highly expressed in quiescent satellite cells and downregulated in activated satellite cells (Fig. 2Ba).
Cyclin-dependent kinase inhibitors (CKIs) play a key role in controlling the cell cycle in many cell types. p21 (CIP1) triggers the cell cycle exit of proliferating myoblasts to initiate myoblast terminal differentiation in response to differentiation signals . p57 (KIP2) is induced in myoblasts upon differentiation. Gene targeting experiments showed that these two CKIs redundantly control cell cycle exit during myogenesis . Compared with irreversible cell cycle arrest upon differentiation, however, attainment of a reversible G0 state by satellite cells is poorly understood. In vitro studies suggested that Rb family members p130 and p27 are involved in the reversible cell cycle exit of proliferating myoblasts to return satellite cells to quiescence . In our experiments, p21 (×0.5), p27 (×1.5), and p130 (×2–3) were not significantly upregulated in quiescent satellite cells. Reflecting the levels of p57 mRNA, p57 protein was found in more than 90% of freshly isolated SM/C-2.6+ cells (Fig. 2Ca). Whether p57 is required for acquisition and maintenance of quiescence of satellite cells remains to be determined in a future study.
Upregulation of Myogenic Inhibitors in Quiescent Satellite Cells
Quiescent satellite cells barely express myogenic basic helix-loop-helix (bHLH) factors. Activity of the Myf-5 locus was revealed through a reporter gene, but Myf-5 protein is hardly detected in dormant satellite cells. On activation, satellite cells upregulate Myf5 and start to express MyoD  (Fig. 1). Our microarray analyses revealed that several myogenic inhibitory molecules were upregulated in quiescent satellite cells: Bmp6 (bone morphogenetic protein 6) (×214), Bmp4 (×66), Bmp2 (×82), Heyl (hairy/enhancer-of-split related with YRPW motif-like)/Herp3/Hrt3/hesr3 (×101, ×33, ×32), Musculin/MyoR (×83), Notch3 (×9). Upregulation of Bmp4, Bmp6, Msc/MyoR, and Heyl in quiescent satellite cells was confirmed by RT-PCR (Fig. 2Bb). BMP4 is reported to negatively regulate MyoD expression in somite myogenesis  and differentiation of satellite cells, where BMP4-induced inhibition of myogenic differentiation requires Notch signaling . Notch signaling is reported to inhibit the differentiation of myoblasts by repression of MyoD expression . In postnatal muscle, Notch signaling controls satellite cell activation and their cell fate , and insufficient upregulation of the Notch ligand Delta is casually related to impaired regeneration of aged muscle . Among several molecules in the Notch signaling pathway, our microarray analysis showed that Notch3 and one of the Notch-effector genes, Heyl, are highly expressed in quiescent satellite cells. When cross-sections of normal mouse tibialis anterior (TA) muscle were stained with specific antibodies, HeyL was found in nearly all Pax7-positive nuclei, and Notch3 was expressed on the surface of mononuclear cells beneath the basal lamina (Fig. 2Cb, 2Cc). These results suggest that Notch3 and HeyL play roles in Notch signaling to inhibit muscle differentiation of satellite cells. Musculin/MyoR is a bHLH transcription factor originally cloned as a repressor of MyoD . Musculin-null mice do not exhibit any skeletal muscle defect, but musculin is likely to negatively regulate MyoD in muscle regeneration .
In addition to negative regulators, two positive regulators of myogenesis, Gli2 (GLI-Kruppel family member GLI2) (×29, ×13) and Meox2 (mesenchyme homeobox 2) (×17), are preferentially expressed in quiescent satellite cells. Gli2 directly upregulates Myf5 , and Meox1 and 2 regulate Pax3 and Pax7 expressions . These observations suggest that Gli2 and Meox2 maintain lineage identity in quiescent satellite cells.
Identification of Quiescent Satellite Cell-Specific Genes
To identify quiescent satellite cell-specific genes from 507 genes (Fig. 2A), we next prepared RNA samples from nonmyogenic cells (SM/C-2.6−/CD45− in Fig. 1A) and performed microarray analysis using Affymetrix GeneChips. Statistical analysis validated that 63 genes out of 507 genes were preferentially expressed (>fivefold) in quiescent satellite cells compared with nonmyogenic cells or activated satellite cells (genes in bold letters in supplemental online Table 1).
To confirm the microarray results, we next performed RT-PCR on 14 genes of interest. In addition to microarray samples, the results for TA muscle and a myogenic cell line, C2C12 cells, are also shown (Fig. 3). Two well-established satellite cell markers (Pax7 and M-cadherin) were expressed not only in quiescent satellite cells but also in activated satellite cells and/or C2C12 cells. In contrast, two cell surface molecules, Odz4, a mouse homolog of the Drosophila pair-rule gene Odd Oz , and CTR, a signaling molecule Tribbles1, and two extracellular molecules, endothelin3 and chordin-like2, were all confirmed to be expressed exclusively in quiescent satellite cells.
Gene Set Enrichment Analysis Revealed Gene Groups Upregulated in Quiescent Satellite Cells
Single-gene analysis permitted us to identify candidate genes that regulate quiescence and undifferentiated state of satellite cells in vivo. To complement the analysis at the single gene level, we performed gene set enrichment analysis . GSEA is an analytical method that identifies small but coordinated changes of predefined gene sets but not up- or downregulation of individual genes, which therefore would help us to identify important signaling pathways or regulatory mechanisms for satellite cells. We used GO annotations  to group all genes on GeneChips and tried to extract gene sets that are upregulated as a whole in quiescent satellite cells compared with activated and proliferating satellite cells (Fig. 4). When all genes were categorized into 1,674 gene sets according to their biological process ontology, only three gene sets were judged to be coordinately upregulated in quiescent satellite cells (FDR < 0.25): cell-cell adhesion, regulation of cell growth, and transmembrane receptor protein tyrosine phosphatase signaling pathway (Table 1). When all genes were grouped into 1,698 gene sets according to cellular component ontology, three gene sets, insoluble fraction, extracellular region, and collagens, were found to be coordinately upregulated in quiescent satellite cells compared with activated/proliferating satellite cells (Table 1). When grouped into 412 gene sets based on their predicted molecular functions, three gene sets, extracellular matrix structural constituent conferring tensile strength, copper ion binding, and lipid transporter activity, were found to be coordinately upregulated in quiescent satellite cells (Table 1). Seven genes listed in Table 1 (Tek, Socs3, Igfbp7, Ptprz1, End3, Wnt4, and Col3a1) were confirmed to be upregulated in quiescent satellite cells by RT-PCR (Fig. 2Bc). A more detailed discussion on GSEA results is in the supplemental online Discussion.
Table Table 1.. Gene sets upregulated in quiescent satellite cells and genes with high enrichment scores
Gene Sets That Are Coordinately Upregulated upon Activation
Many gene sets were found to be coordinately upregulated in activated/proliferating satellite cells compared with quiescent satellite cells (Fig. 4). These are involved in active synthesis of DNA, RNA and protein, progression of cell cycle (Cdc2a, Cdc20, Cdc25c, Ccnb1, Ccna2, etc.), mitochondrial activities, and so on. The gene sets are all listed in supplemental online Table 3. The results well reflect active cell cycling and high metabolic activity of satellite cells.
Expression of Cell-Cell Adhesion Molecules on Satellite Cells
Both single gene analysis and GSEA suggest that cell-cell adhesion is one of the key elements in the regulation of satellite cells. Preferential expression of the following genes in quiescent satellite cells was confirmed by RT-PCR and quantitative PCR (supplemental online Fig. 1A, 1B): VE-cadherin (cadherin 5), Vcam1, Icam1, Cldn5 (claudin 5), Esam (endothelial cell-specific adhesion molecule), and Pcdhb9 (protocadherin beta 9). To date, several cell surface markers for satellite cells have been identified, including M-cadherin, syndecan3, syndecan4, c-met, Vcam-1, NCAM-1, and CD34 [5, 38, , , , –43]. Vascular endothelial (VE)-cadherin, Icam1, claudin5, Esam, and Pcdhb9 should be added to the list. Because Esam is upregulated in long-term hematopoietic stem cells and mammary gland side population cells [44, 45], the expression of Esam in quiescent satellite cells is quite intriguing. When transverse sections of adult skeletal muscle were stained with specific antibodies, M-cadherin was found at the site of contact between satellite cells and myofibers (supplemental online Fig. 1C) . Vcam-1 and VE-cadherin proteins are also detected at the boundary of satellite cells and myofibers. Although their roles in regulation of satellite cells remain to be determined, our observations suggest that cell-cell adhesion molecules have critical roles in keeping satellite cells in an undifferentiated and quiescent state and in protecting satellite cells from cell death. We also confirmed that FACS with Vcam-1 anti-body efficiently enriches quiescent satellite cells as SM/C-2.6 does (supplemental online Fig. 2).
Calcitonin Receptor Is Sharply Downregulated on Activated Satellite Cells and Reappeared on Renewed Satellite Cells During Muscle Regeneration
RT-PCR verified that CTR is exclusively expressed in quiescent satellite cells but not in activated satellite cells or in nonmyogenic cells (Fig. 3). In addition, we confirmed that calcitonin mRNA is expressed in satellite cells (data not shown). Therefore, we examined the expression of CTR protein in vivo using immunohistochemistry. As shown in Figure 5A, CTR protein was observed in Pax7-positive mononuclear cells beneath the basal lamina in uninjured muscle. We next stained cross-sections of regenerating muscle with anti-CTR antibody. Three days after cardiotoxin injection, many activated satellite cells were stained with anti-M-cadherin antibodies, but CTR expression was not detected on activated satellite cells on the serial sections (Fig. 5B). Furthermore, there were no Pax7+/CTR+ cells on muscle sections until 7 days after injury (cardiotoxin [CTX]-7d), when Pax7+/CTR+ cells were again found at the periphery of centrally nucleated, relatively large myofibers but not of small regenerating fibers (Fig. 5C, 5D). The number of Pax7+/CTR+ cells gradually increased thereafter and reached the level of uninjured muscle by CTX-14d (Fig. 5D). Interestingly, approximately 20% of Pax7+/CTR+ cells on CTX-7d were found outside the basal lamina (Fig. 5E). This atypical position of satellite cells was transient, and the ratio of satellite cells residing beneath the basal lamina increased during myofiber maturation (data not shown). Taken together, the results suggest that the expression of CTR is found not only on quiescent satellite cells but also on newly formed satellite cells that are closely associated with maturing myofibers.
Calcitonin Inhibits Activation of Quiescent Satellite Cells
To investigate the roles of CTR in the regulation of satellite cells, eel calcitonin, elcatonin, was added to the culture of quiescent satellite cells in vitro before or after activation. Addition of calcitonin before activation significantly inhibited BrdU uptake by quiescent satellite cells (Fig. 6A) but not by already activated satellite cells (Fig. 6A). Interestingly, a short exposure (0.5 hours) to calcitonin was enough to suppress the activation of quiescent satellite cells (Fig. 6B).
MyoD staining of satellite cells revealed that calcitonin/CTR signaling delays the induction of MyoD in quiescent satellite cells (Fig. 6C). The lower percentage of Ki67-positive cells in calcitonin-treated satellite cells also indicated that calcitonin delays the entry of quiescent satellite cells into the cell cycle (Fig. 6C). Calcitonin-treated cells were considerably smaller than control cells on the second day of culture (Fig. 6D), again indicating delayed activation of satellite cells in the presence of calcitonin. A terminal deoxynucleotidyl transferase dUTP nick-end labeling assay excludes the possibility that calcitonin induced apoptosis in satellite cells (Fig. 6E).
To further investigate the effects of calcitonin on activation of quiescent satellite cells, we prepared living single muscle fibers from mouse extensor digitorum longus muscles by using the collagenase digestion method  and plated them onto Matrigel-coated 24-well plates at a density of one fiber per well in the presence or absence of calcitonin. In control wells, many satellite cells had detached and migrated from the myofibers 2 days after plating (Fig. 6F). Calcitonin significantly reduced the numbers of satellite cells that had detached from myofibers (Fig. 6F). It was reported that calcitonin signaling was mediated via cAMP . An analog of cAMP, dibutyryl cAMP, and an activator of adenylate cyclase, forskolin, also attenuated the activation of satellite cells in vitro (data not shown). Collectively, our results suggest that calcitonin/CTR signaling inhibits activation of satellite cells but not their proliferation or survival. The downstream target molecules of calcitonin/CTR remain to be determined.
Single gene-level analysis revealed several candidate genes that negatively regulate cell cycling of satellite cells. Furthermore, our results suggested that satellite cells express both myogenic and antimyogenic molecules to maintain their delicate state.
GSEA showed that dormant satellite cells coordinately express gene groups involved in cell-cell adhesion, cell-extracellular matrix interaction, copper and iron homeostasis, lipid transport, and regulation of cell growth. Although the result shows one aspect of regulation of quiescent satellite cells, more elaborate gene grouping might be needed to further understand the molecular regulation of quiescent satellite cells.
Finally, we showed that calcitonin receptor is specifically expressed on quiescent satellite cells and transmits signals that attenuate the entry of quiescent satellite cells into the cell cycle. Our results would greatly facilitate the investigation of molecular regulation of satellite cells in both physiological and pathological conditions.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
This work was supported by Grants for Research on Nervous and Mental Disorders (16B-2), Health Science Research Grants for Research on the Human Genome and Gene Therapy (H16-genome-003), for Research on Brain Science (H15-Brain-021) from the Japanese Ministry of Health, Labor and Welfare, Grants-in-Aids for Scientific Research (14657158, 153,90281, and 165,90333) from the Japanese Ministry of Education, Culture, Sports, Science and Technology, and “Ground-Based Research Program for Space Utilization” promoted by Japan Space Forum.