Nitric oxide-dependence of satellite stem cell activation and quiescence on normal skeletal muscle fibers


  • Ashley C. Wozniak,

    1. Department of Human Anatomy and Cell Science, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada
    Search for more papers by this author
  • Judy E. Anderson

    Corresponding author
    1. Department of Human Anatomy and Cell Science, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada
    • Department of Human Anatomy & Cell Science, Faculty of Medicine, University of Manitoba, 730 William Avenue, Winnipeg, Manitoba, R3E 0W3, Canada
    Search for more papers by this author


Satellite cells (quiescent precursors in normal adult skeletal muscle) are activated for growth and regeneration. Signaling by nitric oxide (NO) and hepatocyte growth factor (HGF) during activation has not been examined in a model that can distinguish quiescent from activated satellite cells. We tested the hypothesis that NO and HGF are required to regulate activation using the single-fiber culture model. In normal fibers, HGF and inhibition of NO synthase (NOS) each increased activation without stretching, and NOS inhibition reduced stretch-activation. Activation in unstretched mdx and NOS-I(−/−) fibers was three- to fourfold higher than normal, and was reduced by stretching. Distinctions were not due to different pax7-expressing populations on normal and mdx fibers. The population of c-met–expressing satellite cells on normal fibers was increased by stretch, demonstrating functional heterogeneity among normal satellite cells. Cycloheximide did not prevent the stretch-related increase in c-met expression, suggesting c-met may be an immediate–early gene in satellite cell activation. Results have important implications for designing new therapies that target the role of exercise in health, aging, and disease. Developmental Dynamics 236:240–250, 2007. © 2006 Wiley-Liss, Inc.


Satellite stem cells are precursor cells essential for the growth of normal skeletal muscle and regeneration of injured and diseased muscle (Hawke and Garry,2001; Collins et al.,2005). However, the mechanism of satellite cell activation from quiescence, entry into the cell cycle, and the regulation of stretch-induced activation are not well-established (reviewed by Wozniak et al.,2005).

The role of nitric oxide (NO) and hepatocyte growth factor (HGF) as positive regulators of satellite cell activation from quiescence has been demonstrated by distinctive approaches using cultured muscle cells (satellite-derived cells or myoblasts), cultured single fibers that retain their resident satellite cells in quiescence, and in vivo. NO activates satellite cells before HGF release from the extracellular matrix in stretched in myoblast cultures and unstretched single fibers (Anderson and Pilipowicz,2002; Tatsumi and Allen,2004; Tatsumi et al.,2006a). Downstream from activation and proliferation in myogenesis, NO also has a major role in muscle cell fusion, regeneration, differentiation, and overload hypertrophy (Lee et al.,1997; Anderson,2000; Smith et al.,2002; Pisconti et al.,2006; Sellman et al.,2006; Soltow et al.,2006). Studies examining skeletal muscle in vivo, in myoblast cultures, and using the single fiber cultures show that HGF binding to the satellite cell c-met receptor initiates satellite cell activation and that both quiescent and activated satellite cells express c-met (Cornelison and Wold,1997; Tatsumi et al.,1998,2002; Anderson and Pilipowicz,2002). However, satellite cell responses to NO and HGF have not been tested in a model system that allows cell-by-cell study of expression and activation responses in satellite cells that are resident on single muscle fibers.

NO is produced by NO synthase (NOS; Alderton et al.,2001). In skeletal muscle, NOS-I is localized to the fiber sarcolemma by association with the normal dystrophin-glycoprotein complex (DGC) and down-regulation by loss of dystrophin (Brenman et al.,1995,1996; Silvagno et al.,1996). NO is normally produced in low-level pulses by muscle under conditions where satellite cells are quiescent (Tidball et al.,1998). Through a complex role as a mechanical transducer, the expression and activity of NOS-I in muscle are up-regulated by shear forces, stretch, exercise, loading, and injury, which result in a bolus release of NO that activates satellite cells (Balon and Nadler,1997; Fujii et al.,1998; Tidball et al.,1998,1999; Roberts et al.,1999; Anderson,2000; Tidball,2005). NO partly alleviates dystrophy progression and improves muscle repair and function in mdx dystrophic mice, the mouse homologue of Duchenne muscular dystrophy (Wehling et al.,2001; Anderson and Vargas,2003; Segalat et al.,2005; Archer et al.,2006). However, the impact of NOS-I down-regulation on satellite cell activation is not known.

Stretching normal fibers in culture rapidly induces satellite cell activation (Wozniak et al.,2003). Unstretched fibers cultured with bromodeoxyuridine (BrdU) showed a very low baseline level of BrdU incorporation into satellite cell DNA over 24 hr. By observing cultures after 24 hr exposure to BrdU, the stability of quiescence could, therefore, serve as an effective control for comparison with responses to putative activators such as mechanical stretching. Studies of cultured myoblasts prepared from muscles of old animals showed these quiescent myoblasts were activated by stretch-induced NO release, which led to HGF release and HGF binding to the c-met receptor (Tatsumi et al.,2002,2006a). Fewer reports have investigated the dynamics of satellite cell activation while the cells are resident on fibers, where the satellite cell population can be examined cell by cell to detect immunopositive markers for satellite cells (e.g., pax7) or activated satellite cells (muscle regulatory genes). Fewer yet combine studies of activation where control fibers are shown to retain satellite cells in quiescence or report comprehensively on changes in the expression of satellite cell populations (for example, Wozniak et al.,2003; Zammit et al.,2004; Shefer et al.,2006). Such reports demonstrate that very few satellite cells are typically activated in a fresh preparation of fibers, as demonstrated by nucleotide incorporation into DNA or immunodetection of muscle regulatory genes, and a wide range of pax7 immunopositive cells is reported (3.5–5.3 ± 0.7 and 9 ± 4 pax7+cells/fiber in mouse extensor digitorum longus muscle). However, because satellite cells are resident within the matrix around muscle fibers, and both quiescence and activation are influenced by fibers, detailed studies of activation signaling are best investigated in models that retain the normal fiber environment, such as the single fiber culture model, and also retain satellite cells in quiescence (Anderson and Wozniak,2004; Shefer and Yablonka-Reuveni,2005; Anderson,2006).

We hypothesized that stretch activation of satellite cells on normal fibers depends on both NO and HGF signaling and that NOS-I down-regulation (as in mdx fibers) alters the response to stretch. Fibers from NOS-I(−/−) mice served as a genetic-negative control. Fibers were isolated to avoid stretching, because the tissue level of c-met expression increases rapidly after stretch (Anderson and Wozniak,2004). BrdU incorporation into satellite cell DNA and c-met expression by satellite cells were used to assay the response to stretch. Results demonstrated an important role for NO in stretch-activation and the regulation of satellite cell quiescence. Functional heterogeneity among satellite cells with respect to c-met expression defined a novel role for c-met receptor as an immediate–early gene in satellite cell activation.


L-NAME NG-nitro-L-arginine methyl ester HGF hepatocyte growth factor NO nitric oxide NOS nitric oxide synthase DGC dystrophin–glycoprotein complex


Relative Roles of NO and HGF

Satellite cells on normal fibers showed a low level of activation under control conditions (without stretch or treatment; Fig. 1). Activation was measured as the number of BrdU+ satellite cells per fiber in each culture. Stretching for either 0.5 hr or 2 hr increased the number of BrdU+ satellite cells per fiber (P < 0.05), as expected from previous experiments (Wozniak et al.,2003). Additionally, NG-nitro-L-arginine methyl ester (L-NAME) treatment during stretch prevented stretch-induced activation and was not different from the level observed for unstretched control (i.e., untreated) fibers.

Figure 1.

Relative roles of nitric oxide (NO) and hepatocyte growth factor (HGF) in stretch activation. Stretch-related satellite cell activation (bromodeoxyuridine-positive [BrdU+] cells/fiber, mean ± SEM) was measured in unstretched fibers, and after 0.5 hr or 2 hr exposure to stretching. Fiber cultures treated with (1) no additional treatment in controls fibers; (2) NG-nitro-L-arginine methyl ester (L-NAME), a NOS inhibitor (0.2 μg/ml); (3) HGF (20 ng/ml); or (4) a combination of L-NAME and HGF. Fibers were fixed 24 hr after the start of stretch. Data are plotted as mean ± standard error (SEM). Activation in control conditions was increased by stretching. Treatment of unstretched cultures with L-NAME or HGF also increased activation, while the same treatments in stretched cultures decreased activation. Combined treatment (L-NAME + HGF) of cultures stretched for 2 hr induced a large increase in activation; N = 72–1,121 fibers per group (three to six experiments). The asterisk indicates a difference from unstretched controls; The pound sign indicates a difference from 0.5 hr stretch control; the circle indicates a difference from 2 hr stretch control; triangle indicates a difference between two treatment groups connected by a line.

HGF increased satellite cell activation in unstretched fibers (P < 0.05). Surprisingly, L-NAME increased activation to a level 2–3 times greater than in control unstretched fibers (P < 0.05), and HGF treatment of stretched fibers reduced activation. Treatment with L-NAME and HGF together prevented activation above control levels for unstretched or 0.5-hr stretched fibers. Stretching for 2 hr under combined L-NAME and HGF treatment resulted in high levels of satellite cell activation.

Results demonstrated that activation of satellite cells on normal fibers by stretch was attenuated by NOS inhibition and by HGF, whereas each treatment stimulated activation on normal fibers in the absence of stretch. These results indicate that the mechanism of activation by NO and HGF is modified by stretch.

Response to HGF

The interaction of stretch and HGF in satellite cell activation was examined further through dose–response experiments. Stretched and unstretched fiber cultures were exposed to HGF at concentrations of 0–30 ng/ml, and the level of satellite cell activation was compared between groups (Fig. 2). Without stretch, activation was increased by 20–30 ng/ml HGF (P < 0.05), as reported by others for myoblasts and in vivo (Tatsumi et al.,2001,2006a). While activation was increased by stretch in the presence of low-level HGF (0–10 ng/ml), at 20–30 ng/ml HGF activation was decreased by stretch in comparison to unstretched cultures at the same concentrations.

Figure 2.

Hepatocyte growth factor (HGF) dose–response with and without stretch. This experiment tested the effect of HGF (0–30 ng/ml) on satellite cell activation on fibers in stretched and unstretched conditions (0.5 hr of stretching). Activation was measured as no. of bromodeoxyuridine-positive (BrdU+) satellite cells per fiber (mean ± SEM, n = 8–78 fibers/group). Without stretching, satellite cell activation was increased by 20–30 ng/ml HGF over untreated fibers. Stretching increased activation when the level of exogenous HGF was low (0–10 ng/ml). Activation was reduced by exogenous HGF (20–30 ng/ml) in stretched cultures compared with unstretched cultures at the same level of HGF. The asterisk indicates a difference from the unstretched fibers at 0 ng/ml HGF; the triangle indicates a difference from unstretched cultures at the same [HGF].

Satellite Cell Response to Stretch on mdx and NOS-I(−/−) Fibers

The role of fiber-derived NO signaling in satellite cell activation was investigated by studying the time course of activation in cultures of NOS-I–deficient fibers from muscle of mdx and NOS-I(−/−) mice. Respectively, these served as natural and transgenic comparisons to NOS inhibition of normal fibers.

Fibers from mdx mice showed high levels of satellite cell activation compared with normal fibers (Fig. 3A vs. controls in Fig. 1). There were three- to fourfold more BrdU+ satellite cells per mdx fiber compared with normal (P < 0.01; Fig. 3A). Stretching for 0.5 hr reduced activation on mdx fibers compared with unstretched mdx fibers (P < 0.05) to a level that was higher than in normal stretched fibers (P < 0.05). After 1 hr of stretch, activation was not different from the level in unstretched mdx fibers.

Figure 3.

Time course of activation in stretched mdx and nitric oxide synthase (NOS)-I(−/−) fiber cultures. Single fibers from mdx (A) and NOS-I(−/−) (B) flexor digitorum brevis (FDB) muscles were stretched for 0–3 hr, and activation was measured (number of BrdU+ satellite cells per fiber, mean ± SEM). Data are plotted to show activation at each time point. The level of satellite cell activation on mdx and NOS-I(−/−) fibers in unstretched conditions was higher than in normal unstretched fibers (Fig. 1). Stretching for 0.5 hr reduced activation in both mdx and NOS-I(−/−) fibers compared with respective unstretched fibers of the same strain (P < 0.05). The level of activation at 0.5 hr in mdx stretched fibers was also higher than in normal stretched fibers at the same time (P < 0.01), whereas activation in 0.5 hr-stretched NOS-I(−/−) fibers did not differ from the level in stretched or unstretched normal fibers at 0.5 hr. The stretch-related decline in activation was transient in mdx cultures and more persistent in NOS-I(−/−) fiber cultures (mdx: N = 30–432 fibers per time point, ≥ 8 cultures per condition and time point; NOS-I(−/−): N = 8–156 fibers per group, ≥ 6 cultures per condition and time point).

The time course of satellite cell activation on NOS-I(−/−) fibers was strikingly similar to that in mdx fibers. There were more activated satellite cells per fiber at 0 hr in unstretched NOS-I(−/−) fibers than in normal unstretched fibers (P < 0.01; Fig. 3B). Stretching for 0.5 hr decreased activation on NOS-I(−/−) fibers (P < 0.05); at this time, activation was not different from that in normal unstretched fibers. For NOS-I(−/−) fibers, the stretch-related decrease in activation persisted for 1 hr, returned to the level in unstretched NOS-I(−/−) fibers, and decreased again by 3 hr.

Pax7+ Satellite Cell Population Size

Because the large number of activated satellite cells on fibers from mdx (and NOS-I(−/−)) muscle could be explained by strain-dependent differences in the size of the satellite cell population, the number of pax7+ cells per fiber was determined using in situ hybridization on fibers. Observations confirmed the previous report of pax7 expression by satellite cells on normal and mdx fibers (Fig. 4A). One third of mdx fibers showed pax7-expressing central nuclei (Fig. 4A: xi), although there were fibers with central nuclei that did not display myonuclear pax7 expression. In two fibers isolated from normal muscle, pax7+ cells in the satellite position showed a condensed pattern of staining similar to apoptotic figures (Fig. 4A, v). The number of pax7+ cells was not different between unstretched normal and mdx fibers (control, 3.45 ± 0.38; mdx, 2.82 ± 0.26). A graphical representation of the populations shows an identical pattern of pax7+ cells on each type of fiber, and pax7-expressing nuclei were observed in the satellite cell position of fibers in sections and in cultures (Fig. 4B). The pax7 data demonstrate that the differences in activation at time = 0 hr between normal and mdx fibers that were observed in Figure 3 could not be explained by differences in total satellite cell population.

Figure 4.

Pax7 expression. A: Pax7+ cells on fibers were observed by oil immersion and counted as the number per fiber in triplicate fiber cultures prepared from young (5- to 6-week-old) normal and mdx mice. B: Counts were plotted as the proportion of fibers displaying from 0–11 pax7+ cells (180 control fibers, 98 mdx fibers). Pax7+ satellite cells are shown on control fibers (A, i–vi), mdx fibers (A, vii–x), in a section of normal muscle (A, xii), and in proximity to the central nuclei (A, xi) observed in one third of mdx fibers. Note that (A, vi) is a negative control for in situ hybridization on fibers. Not all central nuclei in a single mdx fiber demonstrated the pax7+ in situ stain, and not all fibers containing central nuclei showed pax7 expression internally in the fiber. In addition, the pax7 in situ hybridization signal was observed in condensed nuclear regions of apparently apoptotic satellite cells on two fibers from normal muscle. Original magnification, ×540. B: There was no difference in the mean number of pax7-expressing satellite cells per fiber and no difference in the pattern of pax7+ satellite cell populations observed on fibers from normal and mdx mice. Scale bar = 25 μm in B.

Stretch Effects on c-met Expression in Normal and mdx Satellite Cells

Heterogeneity in satellite cell c-met expression was explored by testing the influence of 0.5-hr stretching on the pattern of c-met+ cell populations that is resident on normal and mdx fibers. Statistically, this pattern is the frequency distribution of the number of fibers displaying 0–6 c-met+ cells per fiber (plotted in Fig. 5 as proportion of total fibers). Stretching normal fibers increased the number of c-met+ satellite cells per fiber, as indicated by a shift toward higher numbers of c-met+ cells (Fig. 5A; P < 0.005). Notably, the pattern of c-met+ satellite cells per fiber on mdx fibers did not change with stretching (Fig. 5B). There was no difference in the pattern of c-met+ satellite cells on unstretched normal and mdx fibers. Because there was no difference in the overall population of pax7+ satellite cells in normal and mdx fibers, the changes in c-met expression suggest there are functional differences in the response to stretch among two or more subpopulations of satellite cells on normal fibers.

Figure 5.

The pattern of c-met–expressing satellite cells with and without stretch. The numbers of c-met+ cells per fiber were determined for stretched and unstretched cultures using in situ hybridization; data are expressed as the proportion of fibers per group that displayed from zero to six c-met+ cells per fiber. A,B: Activation-related changes in normal (A) and mdx (B) single fibers (350–450 fibers per group) were compared using χ2 statistics. In unstretched cultures, the number of c-met+ cells did not differ between normal and mdx fibers. Stretching increased the number of c-met+ cells in normal fiber cultures (P < 0.005). There was no change in the numbers of c-met+ satellite cells after stretching mdx fibers. The distribution of c-met+ cells on stretched fibers was higher in normal than on mdx cultures (P < 0.05).

C-met Response to Cycloheximide

The role of c-met in activation was examined further in experiments designed to test whether c-met acts as an immediate–early gene. Whole-muscle cultures were treated with the protein synthesis inhibitor cycloheximide, and c-met expression was determined (Fig. 6A). An increase in c-met expression after stretch was not prevented by cycloheximide.

Figure 6.

Response of c-met expression to cycloheximide treatment. A: Whole normal muscle cultures (one to three repeat experiments) were exposed (or not) to cycloheximide treatment for 0.5 hr, 4 hr, or 24 hr, with some cultures also subjected to brief stretching for 0.5 hr. The c-met expression was measured using RNAse protection assays (sample blot is inset for stretched (S) and unstretched (US) cultures, treated (+) or not (−) with cycloheximide; control RNA at 760 bp migrated to the level of the arrow). Blots were scanned and data (optical density/μg RNA, mean ± SD, duplicate blots) are plotted relative to respective untreated cultures. A: Without treatment, stretching for 0.5 hr induced a large rise in c-met expression, and c-met expression was greater after cycloheximide treatment in muscles stretched for 0.5 and 24 hr compared with untreated unstretched muscles. B: Satellite cell cultures from normal mice were treated with cycloheximide for 0–90 min and RNA extracts of each culture were probed for gene expression as above (one experiment, two to three cultures per treatment and time point). Duplicate blots were scanned for c-met and c-fos expression from each culture and data (OD/μg RNA loaded, mean ± SD) are plotted relative to untreated cultures. The c-fos expression increased rapidly over 30 min and then declined. The c-met expression was increased at 90 min. The asterisk indicates significant difference from untreated cultures.

The timing of c-met expression was examined using normal myoblast cultures treated with cycloheximide. Expression of a positive-control immediate–early gene, c-fos, peaked at 30 min after cycloheximide and then declined (Fig. 6B). Cycloheximide also did not prevent the increase in c-met expression at 90 min compared with controls.


These experiments demonstrated for the first time that NO and HGF have dual roles in satellite cell activation and quiescence, and that c-met is an immediate gene in activation. The single fiber model and a baseline of stable quiescence enabled the study of cell-by-cell responses that are essential to dissecting the details of activation.

NOS-inhibition experiments on normal fibers and time-course studies on mdx and NOS-I(−/−) fibers showed that both quiescence and activation are NO-dependent. NOS inhibition of normal fibers prevented activation by stretching and indicated a role for NO in activation. In the absence of NOS activity, the level of satellite cell activation was increased, as demonstrated in unstretched fibers from mdx and NOS-I(−/−) mice, and in unstretched normal fibers treated with L-NAME. These observations are definitive, given that identical sizes of pax7-expressing satellite cell populations were detected on normal and mdx single fibers, in agreement with an earlier report (Reimann et al.,2000). Studying pax7 expression in satellite cells on fibers also suggests that many pax7+ cells in regenerating muscles reported by Seale et al. (2000) are likely interstitial myoblasts or stem cells and not in the satellite position on fibers. Together these findings are interpreted to support the novel idea that NO is required to maintain satellite cell quiescence. Because the level of activation was decreased with stretch in both mdx and NOS-I(−/−) fibers, results implicate either alternate production of NO, possibly by NOS-III within mdx fibers or from NOS-I activity in satellite cells on mdx fibers, was responsible for some restoration of a quiescence-level NO concentration upon stretching (Fujii et al.,1998; Anderson and Vargas,2003; Tatsumi et al.,2006a). NO-dependent stretch activation and the modulation of activation responses by deficiencies in dystrophin and NOS-I may, therefore, explain the accelerated differentiation by mdx satellite cells that was recently reported (Yablonka-Reuveni and Anderson,2006).

The role of HGF in stretch-related satellite cell activation was examined by HGF dose–response experiments. Without stretch, exogenous HGF increased activation, confirming earlier work on fiber cultures (Anderson and Pilipowicz,2002). Surprisingly, higher levels of HGF consistently reduced satellite cell activation in stretched fibers. We propose this finding may be due to c-met receptor desensitization, as in stretched myoblast cultures exposed to high concentrations of HGF (Tatsumi, personal communications), although each activating stimulus would likely up-regulate c-met expression and increase stretch-induced release of HGF from the fiber matrix (Tatsumi et al.,2006a). Alternatively, it is possible that the matrix pH or the molecular conformation of the multifunctional HGF-binding site in the c-met receptor, receptor tyrosine-kinase activity of c-met or PS1 domain that helps in positioning the binding site may be influenced by mechanical stretching (Ponzetto et al.,1994; Kozlov et al.,2004; Tatsumi et al.,2006b). Therefore, in both myoblasts and fibers, HGF is involved in stretch-activation, and high HGF reduces c-met activity in satellite cells during stretch.

To investigate the interdependence of NO and HGF signals, activation was examined while manipulating the two signals together. While activation increased in unstretched fibers treated with L-NAME or HGF alone, during the combined treatment, satellite cell activation was increased in fibers stretched for 2 hr. These observations suggest that NOS inhibition dominated the short-term outcome of activating stimuli (stretch and HGF), while HGF and stretching outweighed NOS inhibition in the longer term. These results are consistent with the notion of normal satellite cell diversity with the possibility that one subpopulation may be more responsive to NO and brief stretching, while another responds more to HGF and more persistent stretch stimulation (Parry,2001; Collins et al.,2005). These data may also be the basis of previous observations of two peaks in the time course of satellite cell activation after stretching (Wozniak et al.,2003).

C-met receptor activity was pivotal during satellite cell activation, as previously demonstrated during muscle development (Dietrich et al.,1999). In nonmuscle tissues, c-met signaling promotes migration, cell survival, and proliferation and a high level of expression is associated with metastasis, tumor aggressiveness, and a poor prognosis (Abounader et al.,2001; Maulik et al.,2002a,b; Derksen et al.,2003; Corso et al.,2005). A wide range of c-met+ satellite cell population sizes (n = 0 to 6) occupied single normal flexor digitorum brevis (FDB) fibers under quiescent conditions (Wozniak et al.,2003). Because the tissue level of c-met expression increased after stretching muscles in culture (Anderson and Wozniak,2004), we tested whether 0.5 hr of stretching increases the number of c-met+ satellite cells. In situ hybridization experiments demonstrated that stretching increased the number of c-met+ satellite cells on normal fibers, an observation that extends the idea of satellite cell heterogeneity. For example in normal muscle, there may be an “early activating” satellite cell population (i.e., activated after 0.5 hr of stretch) that expresses c-met in the “basal” or unstimulated, quiescent state. These cells may be prepared in advance for activation by means of the HGF signal, as evidenced in earlier reports of HGF-c-met colocalization 10 min after injury (Tatsumi et al.,1998; Anderson,2000). Other later-activating satellite cells (e.g., those activated here only after 2 hr of stretch) may up-regulate c-met expression after an initial period of stretch, and only then would be responsive to the HGF that is released by stretch-induced NO release. It is interesting to speculate that the different intensities of the in situ signal for pax7 expression that were observed in satellite cells on fibers in a native state (in the absence of any activating stimuli) may also represent functionally distinctive groups of satellite cells in adult muscle, possibly self-renewing and differentiative subpopulations as proposed by others (Zammit et al.,2004,2006). The potential for treatments to be directed at either population in a particular disease or condition strongly encourages further investigation of this heterogeneity.

Notably with reference to the broader field of muscular dystrophy research, only normal fibers showed stretch-related changes in satellite cells expressing c-met receptor. The pattern of c-met+ cell populations on fibers did not differ between normal and mdx fibers in the present experiments. Because stretch did not increase the number of c-met+ satellite cells on mdx fibers, it seems possible that the satellite cell population on mdx fibers, although of similar size to that on normal fibers, is more homogeneous with respect to c-met expression. By this reasoning, the abundance of c-met expression in each satellite cell increased after stretch, because the tissue level of c-met was increased by stretching mdx muscle cultures. These findings are in strong agreement with the idea that one population of satellite cells is exhausted in older mdx muscle, as reported in vivo (Heslop et al.,2000). It remains to be determined whether an early- or another later-activating subpopulation (discussed earlier) is exhausted, or whether ongoing cycles of damage and regeneration or the NOS-I deficiency per se could account for the population change. The pattern of population loss in dystrophic muscle may also impact disease progression or feedback to affect regenerative capacity. At the cellular level, the immediate impact is that satellite cells on mdx fibers do not depend on NO for activation, although their activation can be down-regulated by NO and stretch. Because both mdx and NOS-I muscle display altered regulation of satellite cell activation and retain the capacity for muscle regeneration, NO-independent HGF/c-met signaling and possibly other signaling pathways control activation in muscle deficient in dystrophin- and NOS-I.

Rapid changes in c-met expression suggested c-met may be an immediate–early gene, similar to c-fos and c-jun, that are expressed rapidly in response to growth factors and without requirements for protein synthesis (Cohen and Curran,1988). Immediate–early genes are “super-induced” by cycloheximide (Muller et al.,1984). Here, c-met expression increased in normal muscle in stretched muscles during exposure to cycloheximide. This finding had a similar time course to the rise in c-fos expression in myoblast cultures and was more rapid than de novo assembly of the 8kb c-met transcript (Moghul et al.,1994). These experiments showed that c-met acts as an immediate–early gene in satellite cell activation (Greenberg and Ziff,1984). A similar role for c-met as an early–immediate gene was reported for stimulated hepatocytes and epithelial cells (Boccaccio et al.,1994; Desiderio et al.,1998).

This report provides novel insights into very early events in muscle satellite cell activation and the satellite cell populations on normal and dystrophic fibers. Together these results show that NO concentration regulates a balance between quiescence and activation on fibers, as illustrated in a proposed model of activation (Fig. 7). One “normal” concentration of NO may maintain quiescence, such that higher or lower NO levels induce activation. (Real-time measurement of NO would corroborate the relationship with activation.) The NO-activation curve may be shifted higher by stretching through increased assembly of c-met transcripts. Exogenous HGF may shift the curve toward lower activation during stretch due to a desensitized c-met receptor. This explanation would account for the observed heterogeneity in activation, such that later-activating satellite cells respond to NO at a different set point. The fascinating complexity of the activation response is highlighted by observations of more than one satellite cell population on normal and NOS-I(−/−) fibers, distinctly modified by dystrophin-deficiency. Results suggest that satellite cell activation by means of NOS-I and c-met activity may be useful targets for new therapeutic compounds that would promote normal muscle growth and regeneration and restore regulation in muscular dystrophy.

Figure 7.

A model of activation as a function of nitric oxide (NO). The model proposes that activation can be represented as a function of the NO concentration acting on satellite cells that are resident on fibers. At some concentration, X, of NO in normal muscle and conditions, satellite cells are maintained in a quiescent state through NO synthase (NOS)-I activity localized to the subsarcolemmal cytoskeleton. For normal muscle, stretch induces NO release, which activates satellite cells by means of hepatocyte growth factor (HGF)/c-met signaling, while a lower NO concentration (from NOS inhibition in normal muscle or NOS-I down-regulation in mdx and NOS-I(−/−) muscle), also allows activation according to the left arm of the curvilinear function. Stretch may stimulate NOS activity from other NOS isoforms expressed in muscle fibers and possibly also from NOS-I activity in mdx satellite cells, and shift satellite cells toward the right along the curve, due to higher NO levels. In satellite cell cultures, this fiber-dependent NO signaling would be absent.



Vitrogen 100 was supplied by Cohesion Technologies (Palo Alto, CA). Dulbecco's modified Eagle's medium (DMEM), antibiotic/antimycotic, chick embryo extract, fetal bovine serum (FBS), trypsin, and gentamycin were supplied by Invitrogen (Carlsbad, CA). BrdU, Serum Replacement-2 (S-9388; different from the serum replacement used in earlier studies by Yablonka-Reuveni and colleagues, and which is currently not available), L-NAME, HGF, diaminobenzidine, anti-BrdU antibody, and HRP-linked anti-mouse antibodies were obtained from Sigma Chemical Co. (Oakville, Canada). FlexCell plates were from FlexCell International (Hillsborough, NC).

Fiber Isolation

Fiber cultures were prepared from FDB muscles of normal, mdx, and NOS-I(−/−) mice (5–8 weeks old; Wozniak et al.,2003; Wozniak and Anderson,2005). Briefly, muscles were gently isolated, cleaned in proliferation medium (1× DMEM with 10% fetal bovine serum, 2% chick embryo extract, 1% antibiotic/antimycotic, and 0.1% gentamycin), and incubated in 0.2% collagenase at 37°C and 5% CO2 for 2.5 hr. Muscles were transferred to fresh proliferation medium without collagenase and gently triturated to release single fibers. Fibers were further cleaned by gravity sedimentation and plated in six-well FlexCell plates, which had been precooled and coated with 80 μl of Vitrogen100 (manufactured as 98% collagen type I, 2% collagen type III). Fibers adhered during incubation (20 min at 37°C, 5% CO2) and then were covered with basal growth medium (DMEM, 20% Serum Replacement-2, 1% FBS, 1% antibiotic antimycotic, and 0.1% gentamycin). Fibers were allowed to attach overnight before stretching (defined as time = 0 hr). Preparations that contained tissue debris or broken, hypercontracted fibers were discarded. At the start of experiments (0 hr), BrdU (0.002% weight/volume) was added to medium and cultures were maintained for 24 hr before fixation.

Roles of NO and HGF in Stretch-Activation

Normal fiber cultures were treated at time = 0 hr with one of the following: a nonspecific NOS inhibitor, L-NAME (0.2 μg/ml), HGF (20 ng/ml, the concentration known to activate quiescent satellite cells on single fibers; Anderson and Pilipowicz,2002), or L-NAME and HGF together. HGF-response experiments used 0–30 ng/ml HGF added to the medium (time 0 hr).

Plates were placed in a FlexCell system and stretched (Wozniak et al.,2003) at 4 cycles per min for 0 to 3 hr. One cycle was 8 sec in a stretched position and 7 sec in a neutral or unstretched position. A 10% stretch, similar to the physiologic range of length changes during a contraction, was applied along the radius of each culture as provided by a vacuum 20 kPa applied to the bottom of culture plates (Anderson et al.,1993). After stretch, the vacuum was disconnected and cultures were maintained (without stretch) until time = 24 hr and fixed in acid–alcohol.

Fibers were immunostained to detect the incorporation of BrdU using anti-BrdU antibody (1:1,000; Sigma Chemical Co.); secondary HRP-linked anti-mouse antibody (1:300; Sigma Chemical Co.) was visualized with diaminobenzidine (25 mg/ml) (Wozniak et al.,2003). Fiber cultures were coded and viewed at ×200 magnification. The total number of satellite cells cannot be counted in these experiments because quiescent satellite cell nuclei (and myonuclei) do not incorporate BrdU. The number of BrdU+ satellite cells per fiber was counted without knowledge of source, by systematically scanning the entire area under each coverslip. Data on normal fibers were compiled from eight independent experiments. Experiments with mdx and NOS-I(−/−) fibers used identical protocols, and data were obtained from three repeat experiments for each strain.

In Situ Hybridization

Normal and mdx fibers were stretched (or not) for 0.5 hr and immediately processed for in situ hybridization to detect c-met message (Wozniak et al.,2003). Antisense, digoxigenin-labeled riboprobes were synthesized following Boehringer Mannheim protocols (Anderson and Vargas,2003). Labeled probes were run on formaldehyde–agarose gels, transferred to nylon membrane, and visualized using anti-digoxigenin antibodies and alkaline phosphatase color detection to confirm a probe size of 1.27 kb. The target mRNA was 9 kb. Hybridized transcripts were localized in satellite cells resident on fibers.

To examine the total satellite cell population, normal and mdx fibers were processed for in situ hybridization to detect pax7 message (Seale et al.,2000). The riboprobe for pax7 was made as follows. Templates were obtained by transforming bacteria with a pBluescript KS plasmid (Stratagene) containing bases 1–872 of the pax7 cDNA, a generous gift from Dr. P. Gruss (Jostes et al.,1990). An antisense, digoxigenin-labeled riboprobe was synthesized and characterized as above to confirm probe size (0.812 kb). The target mRNA was 4.9 kb. Hybridized pax7 transcripts were localized in sections of normal muscle to confirm the identification of satellite cells on fibers that were fixed in acid–alcohol immediately after plating, as reported (Wozniak et al.,2003). In other experiments, sense riboprobes showed no signal, similar to procedural controls that omitted probe, anti–digoxigenin antibody, or color detection steps. In each case, the number of message+ cells per fiber was counted for all fibers in a well. Data are reported as mean ± SEM or shown (in a frequency distribution plot) as the proportion of fibers displaying a range of satellite cells per fiber.

C-met as an Immediate–Early Gene

Whole-muscle cultures and myoblast cultures were treated with cycloheximide to determine whether c-met acts as an immediate early gene. Whole extensor digitorum longus (EDL) muscles were isolated and pinned at resting length, as reported (Wozniak et al.,2005). Normal primary myoblasts (derived from satellite cells isolated from pooled limb muscles of normal mice) were cultured in proliferation medium until 75% confluence. EDL cultures were treated (or not) with cycloheximide (10 μg/ml) for 4 hr in unstretched conditions and snap frozen in liquid nitrogen, or stretched for 0.5 hr and either frozen or maintained until 24 hr before snap freezing (Boccaccio et al.,1994). Myoblast cultures were treated with cycloheximide (10 μg/ml) for 0 to 1.5 hr. RNA was isolated from muscles and myoblasts by phenol–chloroform extraction (Chomczynski and Sacchi,1987). The c-met and c-fos mRNAs were detected using RNAse protection assays.

Statistical Analysis

Data were compared by analysis of variance (ANOVA) with repeated measures, as appropriate for multiple groups in each experiment, using parametric (n ≥ 6) or nonparametric (n = 2–3) tests according to sample size (Statpak; NorthWest Analytical Inc.). Data are plotted as mean (± standard error of the mean, SEM, except as indicated in Fig. 5). Pair-wise differences between groups were assessed using least significant difference tests when there were significant effects determined by ANOVA. The χ2 statistics were applied in testing changes in the pattern (i.e., the frequency distribution) of c-met+ cells on fibers; raw frequency data were used for χ2 comparisons, and data are plotted as proportions. Significance was determined at the P < 0.05 level; only significant changes are reported.


The plasmid containing the pax7 sequence was a generous gift from Dr. Peter Gruss in the Department of Cell Biology at the Max Planck Institute of Biophysical Chemistry. The authors thank Ms. C. Vargas for technical assistance. This work was funded by a grant from the Manitoba Institute of Child Health, program support from the Muscular Dystrophy Association and a Canada Graduate Scholarship (Doctoral) from the Natural Sciences and Engineering Research Council (A.W.) supported these studies.