Replicative aging down-regulates the myogenic regulatory factors in human myoblasts

Authors


To whom correspondence should be addressed (email mouly@ext.jussieu.fr).

Abstract

Background information. Aging of human skeletal muscle results in a decline in muscle mass and force, and excessive turnover of muscle fibres, such as in muscular dystrophies, further increases this decline. Although it has been shown in rodents, by cross-age transplantation of whole muscles, that the environment plays an important role in this process, the implication of proliferating aging of the muscle progenitors has been poorly investigated, particularly in humans, since the regulation of cell proliferation differs between rodents and humans. The myogenic differentiation of human myoblasts is regulated by the muscle-specific regulatory factors. Cross-talk between the muscle-specific regulatory factors and the cell cycle regulators is essential for differentiation. The aim of the present study was to determine the effects of replicative senescence on the myogenic programme of human myoblasts.

Results. We showed that senescent myoblasts, which could not re-enter the cell cycle, are still able to differentiate and form multinucleated myotubes. However, these myotubes are significantly smaller. The expression of muscle-specific regulatory factors and cell cycle regulators was analysed in proliferating myoblasts and compared with senescent cells. We have observed a delay and a decrease in the muscle-specific regulatory factors and the cyclin-dependent kinase inhibitor p57 during the early step of differentiation in senescent myoblasts, as well as an increase in the fibroblastic markers.

Conclusions. Our results demonstrate that replicative senescence alters the expression of the factors triggering muscle differentiation in human myoblasts and could play a role in the regenerative defects observed in muscular diseases and during normal skeletal-muscle aging.

Abbreviations used:
CDK

cyclin-dependent kinase

CKI

CDK inhibitor

CTGF

connective tissue growth factor

DMD

Duchenne muscular dystrophy

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

HRP

horseradish peroxidase

IGF-1

insulin-like growth factor 1

MRF

myogenic regulatory factor

Rb

retinoblastoma

RT

reverse transcription

Introduction

Skeletal muscle has a remarkable capacity to regenerate in response to injury. However, during aging, there is a gradual decline in the regenerative properties of the skeletal muscles, as well as a decrease in muscle mass and force. Although the maintenance and repair of skeletal muscle rely on a small population of quiescent progenitors, which have been called satellite cells due to their location underneath the basal lamina of muscle fibres, several studies have underlined the importance of the environment for muscle regeneration. Although an old muscle regenerates less efficiently than a young muscle, its ability to regenerate when grafted into a young rat host is significantly enhanced (Carlson and Faulkner, 1989). In addition, the regenerative capacity of the muscle of old mice can be increased by heterochronic parabiose that exposed the old mice to factors present in young serum (Conboy et al., 2005). The general decrease in the level of circulating growth factors and hormones such as IGF-1 (insulin-like growth factor 1) (Benbassat et al., 1997) with aging may explain, at least in part, the slower regeneration of skeletal muscle after injury in elderly subjects. Changes in the number of motor units (Doherty et al., 1993) or in the microvascularization (Scelsi et al., 2002) may also contribute to the reduced regenerative capacity of the skeletal muscle with aging.

The satellite cells, once activated in response to injury or damage, will proliferate as myoblasts and then withdraw from the cell cycle, differentiate and fuse to form newly regenerated muscle fibres or be incorporated into growing post-mitotic myofibres. The regenerative capacity of skeletal muscle is linked not only to the number of satellite cells, which declines with age in humans (Renault et al., 2002), but also to their activation, proliferation and differentiation capacities. The eventual modulation of these complex processes with proliferative aging still needs to be investigated, particularly in human cells.

Myoblast proliferation and differentiation are tightly regulated programmes. The proliferative capacity of human myoblasts, like most other human somatic cells, is limited by a terminal growth arrest called replicative senescence. Replicative senescence is an important barrier to tumour progression (Sedivy, 2007), which at the same time limits the regenerative capacity of skeletal muscle and must be taken into consideration in muscular disease, cell-mediated therapy and aging. Progressive telomere shortening at each cellular division acts as a mitotic clock that regulates the mechanism of replicative senescence (Bodnar et al., 1998). Additional pathways such as the p16INK4a stress pathway could also interfere with the proliferation capacity of human myoblasts by inducing premature senescence (Zhu et al., 2007).

The myogenic differentiation of human myoblasts as described for all mammals is regulated by a family of muscle-specific regulatory factors [MRFs (myogenic regulatory factors)], which includes MyoD, Myf5, Myogenin and MRF4. These transcription factors have a bHLH (basic helix—loop–helix) central domain, which is involved both in protein interactions to form heterodimers and in DNA binding to transactivate muscle-specific genes. MyoD and Myf5 have been described to be involved in the specification and activation of satellite cells, Myogenin in their differentiation, whereas MRF4 was thought to be involved in the maturation (Charge and Rudnicki, 2004), until findings demonstrated that MRF4 might be involved also in the determination of muscle progenitors (Kassar-Duchossoy et al., 2004). Cross-talk between the MRFs and the cell cycle regulators is necessary for myogenic differentiation (Kitzmann and Fernandez, 2001). In order to undergo differentiation, myoblasts have to exit irreversibly from the cell cycle through the G1 checkpoint. Cell cycle exit is under the control of the CKIs [CDK (cyclin-dependent kinase) inhibitors)], which are divided into two families: the Cip—Kip (p21, p27 and p57) and the Ink4 family (p15, p16, p18 and p19) that strictly regulate cyclin—CDK complexes and, subsequently, the Rb (retinoblastoma) phosphorylation status and progression through S-phase.

Most of the data in the literature have been generated either in avian myoblasts or in murine muscle cell lines, which do not share with human cells the same pathways to regulate proliferation and therefore do not allow one to investigate the effect of proliferative aging on human myogenic differentiation. Human satellite cells can be isolated in vitro and, following an initial phase of exponential growth, the proliferative rate of myoblasts derived from these satellite cells progressively slows down with proliferative aging until they stop dividing, thus indicating that the establishment of senescence is a progressive process (Mouly et al., 2005). Since satellite cells in vivo are not a synchronously dividing population of cells, aging may result in focal senescence, a process that would become more generalized in degenerative diseases such as muscular dystrophies. Senescent myoblasts are cells that have irreversibly withdrawn from the cell cycle and can no longer re-enter it; however, they are still able to differentiate (Renault et al., 2000). Myotubes derived from human satellite cells of a young donor but aged in vitro have been shown to exhibit a delay in the excitation—contraction coupling mechanism similar to that observed in myotubes derived from satellite cells of old donors, suggesting that replicative aging of the myoblasts can modify their intrinsic properties both in vivo and in vitro (Lorenzon et al., 2004).

The aim of the present study was to determine the effects of replicative senescence on the myogenic programme of human myoblasts. Human satellite cells have been aged in vitro in order to dissociate the effects of the proliferative aging from those due to the environment. We demonstrate that senescent myoblasts are still able to differentiate and form multinucleated myotubes. However, these myotubes are significantly smaller. The expression of MRFs and CKIs, the key factors involved in the early steps of differentiation, were analysed in proliferating myoblasts and compared with senescent cells. A delay and a dramatic reduction in the expression of the MRFs and CKI p57 were observed during the early steps of differentiation of the senescent myoblasts, whereas fibroblastic markers were increased in senescent human cells as compared with young cells. Our results demonstrate that replicative senescence deregulates the myogenic programme, resulting in impaired myogenesis, which could play a role in the regenerative defects observed in muscular diseases and during normal skeletal-muscle aging.

Results

Senescent cells are still able to differentiate

Human myoblasts were cultivated in proliferation medium until they reached an irreversible cell cycle arrest in senescence. Myoblasts were considered as young during the first third of their life span and as senescent at the end of the life span when they failed to divide during 3 weeks of repeated feeding. Young and senescent myoblasts were induced to fuse into myotubes during 6 days of differentiation. As observed in Figure 1(A), the senescent cells are still able to form myotubes but they are smaller and less branched than those formed by the young cells. These experiments were performed on myoblasts isolated from a 5-day-old infant. To confirm these results, two other cultures were tested, and we observed the same modification in fusion with smaller myotubes in the senescent cultures isolated from a foetal and 17-year-old individual. To establish whether there was any difference in the terminal differentiation, both the fusion index [the ratio of the number of nuclei per myotube (>2 nuclei) to the total number of nuclei] and the number of nuclei per myotube were measured. Although the fusion index of the senescent cell cultures remained relatively elevated (63%), a significant decrease was detected when compared with the 81% attained in young cells (Figure 1B). In order to determine whether this decrease was due in part to a defect in the growth of the myotubes to form large multinucleated structures, the number of nuclei per myotube was counted. Young cells formed very large branched myotubes and we calculated that 90% of the myonuclei were in myotubes that had more than 50 nuclei. Senescent cells formed much smaller myotubes: 63% of the myonuclei were in myotubes that had less than 25 nuclei, 35% in myotubes containing between 25 and 50 nuclei and only 2% in myotubes containing more than 50 nuclei (Figure 1C). The size of the myotubes in culture is largely determined by the probability of finding a pre-existing myotube with which to fuse once a cell makes the decision to differentiate. However, this decrease in fusion properties was not due simply to a decrease in the motility of senescent cells, since similar results were obtained when differentiation was induced at very high density (twice the number of cells at confluence; results not shown), indicating a defect in differentiation and/or fusion properties of the senescent myoblasts.

Figure 1.

Differentiation in vitro of young and senescent myoblasts

(A) Immunofluorescence was achieved with an antibody directed against Desmin after 6 days of differentiation. Specific antibody labelling was revealed using Alexa Fluor® 488 green fluorochrome. Nuclei were visualized with Hoechst (blue). (B) Fusion index, i.e. the number of nuclei incorporated into myotubes as a percentage of the total number of nuclei, was calculated in young and senescent cultures after 6 days of differentiation. (C) Representation of the relative sizes of young and senescent myotubes. The number of nuclei per myotube was counted and results were graded into three groups: less than 25 nuclei, between 25 and 50 nuclei and more than 50 nuclei per myotube. The results are presented as the percentage of each group within each culture.

Factors involved in the regulation of the early steps of myogenic differentiation

Very few studies have been carried out concerning the consequences of proliferative aging on the differentiation of human myoblasts. We investigated the expression of cell cycle regulators and MRFs involved in the initial process of myogenic differentiation. Young human myoblasts were induced to differentiate for 6, 12, 24, 48 and 72 h, and mRNA levels of factors involved in the control of cell cycle such as Rb, p53 and members of the Ink4 and Cip—Kip families were determined using an RNase protection assay (Figure 2A). At the RNA level, p57, a member of the Cip—Kip family, was the only factor that showed significant regulation during this early period of differentiation. A 6-fold increase in the amount of p57 mRNA was observed after 48 h of differentiation, while the expression of the other factors (e.g. p21, p53 or Rb) remained unchanged.

Figure 2.

Kinetics of factors triggering the early steps of myogenic differentiation

(A) The mRNA levels of Rb, p53, the Ink4 and Cip—Kip families were analysed by RNase protection assays. The left panel shows a sample of the probe, yeast tRNA negative control and HeLa-positive mRNA control. Right panel: young human myoblasts were induced to differentiate for 6, 12, 24, 48 and 72 h. The results were normalized to GAPDH mRNA and the ratios were expressed as a comparison with those obtained on undifferentiated samples (sample 0), to which an arbitrary value of 1 was attributed. (B) MyoD, Myogenin, Myf5 and MRF4 mRNA levels were analysed by Northern blot during the first hours of differentiation in young cultures. The results were normalized to the expression of the 18S rRNA. (C) Western-blot analyses of MyoD, Myogenin and p57 protein expressions before differentiation (sample 0) and after 10, 15, 24, 30, 36, 48 and 60 h of differentiation in young cultures. The results were normalized to the expression of the nuclear protein Emerin.

The expression of the MRFs during the early steps of differentiation was determined during the same time course by Northern-blot analysis and signals were normalized to 18S mRNA (Figure 2B). MyoD and Myf5 mRNAs were up-regulated 17- and 3-fold respectively after 10 h of differentiation, and then their mRNA steady-state level decreased concomitantly with an increase in Myogenin mRNA. MRF4 mRNA was not detected during this period of time. To confirm at the protein level the up-regulation of p57, MyoD and Myogenin mRNAs during early myogenic differentiation, Western-blot analyses were carried out. The levels of the protein signals were compared with that of Emerin, a ubiquitous nuclear envelope protein. As seen in Figure 2(C), the MyoD protein was up-regulated after 10 h of differentiation and then decreased concomitantly with an increase in the myogenic factor Myogenin and the cell cycle regulator p57. This result confirms that up-regulation of MyoD is one of the earliest events in the decision to undergo differentiation in human myoblasts, even earlier than the expression of p57 involved in the exit from the cell cycle.

Deregulation of the factors triggering differentiation in senescent cells

In order to determine whether replicative senescence modulates the expression of the MRFs and the p57 protein, we investigated the kinetics of expression of these factors during the early steps of differentiation in young and senescent human myoblasts. As shown in Figure 3(A), induction of the p57 protein expression was delayed by 15 h in senescent myoblasts and it reached its highest level at 45 h, as in young cells, but the signal remained 2.4-fold lower in the senescent myoblasts. A similar pattern of expression was observed for Myogenin, with a delay of 15 h and a 7-fold decrease in the senescent cells. Concerning MyoD, an increase in the amount of protein was detected after 15 h of differentiation in both young and senescent myoblasts, but the level remained 3.3-fold lower in senescent cells and did not change over the next 45 h. To verify that the deregulation of the factors triggering the differentiation is a mechanism specific to senescent myoblasts and not just to this particular human cell strain, the kinetics of the expression of these factors were tested in two other human cultures. Identical results were obtained, the kinetics of the expression of these factors were delayed and their level of expression was decreased (results not shown).

Figure 3.

Deregulation of key factors during the differentiation of senescent human myoblasts and increase in fibroblastic markers in senescent cells

(A) Western-blot analyses of the level of expression of MyoD, Myogenin and p57 during the first hours of differentiation in senescent and young cultures. The results have been normalized to the expression of the nuclear protein Emerin. (B) DNA-binding activity of MyoD and Myf5 during the first 48 h of differentiation in young (white column) and senescent (grey column) cultures. (C) Expression of CTGF and collagen IV mRNAs, two fibroblastic markers analysed by RT—PCR and normalized to GAPDH. –, Negative control; +, fibroblast mRNA; Y, young myoblast mRNA; S, senescent myoblast clone mRNA.

Since the activity of the MRFs relies not only on the quantity of protein present but also on its capacity to bind DNA, we measured the DNA-binding activity of MyoD and Myf5. Figure 3(B) shows that there was no difference in the DNA-binding activity of MyoD measured in the proliferating myoblasts (time 0) or after 8 h of differentiation in either young or senescent cells. In contrast, at 16 h of differentiation the activity was dramatically up-regulated, 7.7-fold, in young cells, whereas only a 4.8-fold increase was detected in senescent cells. In these senescent cells, a progressive increase in the MyoD DNA-binding activity was measured during the first 48 h of differentiation, even if it remained lower than the value observed in young cells. Figure 3(B) shows that Myf5 activity increases progressively to reach a peak at 16 h of differentiation (Figure 3B) in young myoblasts, whereas in senescent cells, such an increase was not observed and the activity remained very low. Therefore the binding activities of both MyoD and Myf5, which have been defined as key regulators for myogenesis, are both delayed and decreased in senescent cells when compared with young cells.

In 2002, Taylor-Jones et al. suggested that a balance exists between myogenic and adipogenic programmes in myoblasts, which is controlled by the relative level of myogenic and adipogenic factors, and that this balance shifts with age (Taylor-Jones et al., 2002). More recently, Brack et al. (2007) demonstrated that myoblasts from aged mice tend to convert from a myogenic lineage into a fibrogenic lineage. In order to determine whether the down-regulation of myogenic markers was associated with an increase in adipogenic or fibrogenic factors, we analysed the expression of some of these factors by RT—PCR (reverse transcription—PCR). The mRNAs coding for the adipogenic markers PPARγ (peroxisome-proliferator-activated receptor γ) or LPL (lipoprotein lipase) were not detected either in young myoblasts or in senescent myoblasts, whereas a high level of expression was detected in fat tissue (results not shown). However, fibroblastic markers were increased in senescent myoblasts as compared with young myoblasts. To ascertain that this increase was only due to a deregulation in myoblasts, a clonal study was performed and the expression of these markers was investigated in senescent clones. Figure 3(C) shows that the expression of two fibroblastic markers, CTGF (connective tissue growth factor) and collagen IV, was increased in senescent myoblast clones. The ratio of the mRNA coding for collagen IV to the mRNA coding for the housekeeping gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase), and the ratio CTGF/GAPDH, were increased 2.4- and 3.7-fold respectively in senescent cells as compared with young cells.

Discussion

As stated above, one of the hallmarks of skeletal-muscle aging is a decline in regenerative capacity. This is now known to be a multifactorial process involving modification in circulating trophic and environmental factors and a decrease in satellite cell number (Renault et al., 2002). However, the influence of proliferative aging and senescence on the myogenic programme of the satellite cells, which are the human muscle progenitors, has been poorly investigated. In the present study, we wanted to determine the effects of replicative senescence (i.e. the irreversible growth arrest reached by cells that have exhausted their proliferative capacity) on the differentiation programme of human satellite cells. The human muscle progenitor cells were aged in vitro and used as a model system to dissociate the intrinsic modifications due to proliferative aging from those due to the environment.

We observed that the differentiation of human myoblasts into multinucleated myotubes is impaired by replicative senescence, the terminal stage of proliferative aging. Even though the senescent myoblasts were still able to fuse, a significant reduction in the number of nuclei per myotube and in the fusion index was observed when compared with young cells. Under our conditions, 90% of the myonuclei in young cultures were found in large and branched myotubes that had more than 50 nuclei, whereas this number decreased to less than 3% in senescent cultures. Most of the myonuclei were located in small myotubes that had less than 25 nuclei. Reduction in the motility of senescent cells did not explain this difference since the decrease in the size of the myotubes was observed even when differentiation was induced at densities high enough (twice the number of cells at confluence) so that myoblasts were already in contact with each other.

This differentiation and/or fusion defect may result from modifications in satellite cell functions. Therefore we analysed the regulation of key factors involved in myogenic differentiation, the MRFs and the cell cycle regulators, in both young and senescent human myoblasts. During early differentiation, the level of the MyoD protein and its DNA-binding activity were both significantly reduced and delayed in senescent when compared with young satellite cells. Although the decrease in MyoD protein level could be explained in part by the general decrease in transcription and a change in protein stability that is known to occur in senescent cells, the changes in its DNA-binding activity represent a strong argument in favour of a MyoD loss of function in senescent myoblasts. Interestingly, MyoD−/− satellite cells derived from adult mouse muscle also showed a reduced differentiation capacity and a decrease in fusion index (Sabourin et al., 1999). It may be noted that MyoD−/− cells displayed a flattened morphology with an enlarged cytoplasm and extended cytosolic processes, a morphology reminiscent of that observed when satellite cells reach a senescent state. Interestingly, our results suggest that MyoD does not down-regulate the expression of Myf5 in the context of human myoblast differentiation as observed in the MyoD−/− mouse myoblasts. As for MyoD, Myf5 DNA-binding activity was down-regulated in senescent satellite cells as compared with young cells.

As illustrated in Figure 4, the expression of the MRF Myogenin and the cell cycle inhibitor p57 was also both delayed and decreased during the early hours of differentiation in senescent cells as compared with young cells. Both Myogenin and p57 expression was subsequent to the up-regulation of MyoD in young cells, and the decreased expression of MyoD in senescent cells was accompanied by a decrease in the expression of both p57 and Myogenin. Since Myogenin expression is suggested to be controlled by MyoD (Hollenberg et al., 1993; Cao et al., 2006) and since MyoD is able to induce p57 protein expression in p21−/− mouse fibroblasts (Vaccarello et al., 2006), we propose that the delayed expression of both Myogenin and p57 may be the consequence of a deregulation of MyoD in the senescent human myoblasts. The increased activity of MyoD at the onset of differentiation may have a dual and essential role: first, by inducing p57 expression to trigger cell cycle exit and, secondly, to induce Myogenin expression to trigger the transcription of the cascade of differentiation-specific muscle genes. Reynaud et al. have suggested that p57 stabilizes the MyoD protein and this could play an important role in the accumulation of MyoD at the onset of myoblast differentiation (Reynaud et al., 1999, 2000). These results suggest the existence of a regulatory loop involving both MyoD and p57. The presence of this cross-talk between these two factors is reinforced by our results demonstrating that both are down-regulated in senescent cells: p57 expression is delayed during the differentiation of senescent cells and its level remained lower than in young cells. However, an increase in p57 is still observed in totally senescent cultures when cells are irreversibly arrested. This result supports also the idea that p57, known to be an inhibitor of the cell cycle, could play a role independent of the cell cycle exit, such as stabilizing MyoD.

Figure 4.

Schematic representation of the major modifications of key factors during the first steps of differentiation of young and senescent myoblasts

In young myoblasts, MyoD and Myf5 were quickly up-regulated during the first hours of differentiation. These increases were followed by the expression of the cell cycle regulator p57 and the MRF Myogenin. In senescent myoblasts, both a delay and a down-regulation of MyoD were observed, and senescent myoblasts failed to increase their DNA-binding activity for Myf5. p57 and Myogenin were also delayed and did not reach the level observed in young myoblasts.

As suggested by other studies, the down-regulation of the myogenic programme could be associated with an adipocyte-like or fibroblastic-like conversion (Guan et al., 2002; Taylor-Jones et al., 2002; Brack et al., 2007). Our results show that myoblasts did not express adipogenic markers either in young myoblasts or in senescent myoblasts. However, we observed an increase in the expression of two fibroblastic factors, CTGF and collagen IV, with proliferative aging. Interestingly, Brack et al. demonstrated that the fibroblastic conversion of mouse myoblasts is mediated by factors present in the systemic environment (Brack et al., 2007). We show in the present study that replicative aging could also play a role in this process. This pathway triggering a fibroblastic-like conversion in senescent myoblasts still needs further investigation.

Other studies in the rat and the mouse have underlined the importance of environment for muscle regeneration (Carlson and Faulkner, 1989; Conboy et al., 2005). However, differences exist between species, and although somatic cells derived from rodents or humans have a proliferative limit, the regulation of this limit is distinct. Whereas in human cells proliferative senescence is triggered by telomere shortening, mouse cells do not have a telomere-driven replicative senescence since their telomeres are too large to impose a barrier to cell division (Kipling, 2001). Furthermore, the laboratory strains of mice frequently maintain telomerase activity in the adult animals (Prowse and Greider, 1995). As an example we demonstrated in a previous study that IGF-1, which has been shown to enhance the proliferative life span of myoblasts in rodents, has little effect on the proliferative life span of human myoblasts (Jacquemin et al., 2004). Involvement of proliferative aging as a factor that modulates the human muscle regenerative process was also proposed by Lorenzon et al. (2004). They showed a delay of the excitation—contraction coupling mechanism in myotubes derived from human satellite cells of old donors compared with young donors, and a similar effect was reproduced in satellite cells from a young donor aged in vitro. These results indicate that the replicative aging of the satellite cells may contribute to the delayed functional maturation observed during aging in vivo.

Clonal studies conducted on human fibroblasts have shown that the life span of individual cells covers a wide range of values (Smith and Hayflick, 1974). In addition, there is a gradual depletion of the replicative potential of individual clones, or a clonal attenuation during aging (Martin et al., 1974). This implies that progenitors will not reach proliferative senescence in a co-ordinated manner, and focal senescence may modify the regenerative capacity of an organ. Although populations of human satellite cells isolated from elderly subjects can still divide, culture conditions, containing many growth factors, will mask the presence of slowly dividing or non-dividing senescent cells in vitro, whereas they will be preserved in vivo. In the present study, we show that senescent cells present a defective differentiation programme that will slow down the regeneration process, as observed during normal muscle aging. Such defects will be even more detrimental to muscle regeneration in diseases that involve repeated cycles of degeneration—regeneration, such as DMD (Duchenne muscular dystrophy). Progenitors isolated from patients suffering from DMD display a defective phenotype (Blau et al., 1983), which can be related to the exhaustion of proliferative capacity (Decary et al., 2000). Defective myoblasts have also been characterized from biopsies of patients suffering from other neuromuscular disorders such as OPMD (oculo-pharyngeal muscular dystrophy) (Perie et al., 2006) and DM1 (myotonic dystrophy type 1) (Furling et al., 2001). Whether these defects are directly linked to a dysregulation of regulatory factors caused by premature proliferative arrest still needs to be confirmed. Our study suggests that senescence is detrimental to muscle regeneration not only by reducing the number of progenitors, but also by inducing modifications in their differentiation properties due to a deregulation of specific key factors.

Materials and methods

Cell culture

Human satellite cells were isolated from muscle biopsies as described previously (Edom et al., 1994). Cells used throughout the present study, unless specified, were isolated from a quadriceps muscle biopsy of a 5-day-old infant. Two other satellite cell cultures were isolated from the muscle of a foetus (Furling et al., 2001) and a 17-year-old healthy subject (Jacquemin et al., 2004). Cells were cultivated at 37°C in a humid atmosphere containing 5% CO2. Myoblasts were grown in F10 medium (Gibco) supplemented with 5 μg/ml gentamicin (Gibco) and 20% (v/v) foetal calf serum (Gibco). Cell populations were trypsinized when they were 80% confluent. For each passage, the number of divisions was calculated as: ln(N/n)/ln 2, where N is the number of cells counted and n is the number of cells initially plated. Myogenic purity of cultures was performed throughout the life span by immunocytochemistry using an antibody specific for Desmin (1:50; D33, Dako). Specific binding was revealed using peroxidase. To determine the percentage of positive cells, 500 cells were counted. The cultures were considered to be senescent when they failed to divide during 3 weeks in high-serum growth medium.

For differentiation studies, cells were trypsinized, counted and 1 million cells were plated on to a 100 mm dish. After 48 h, the proliferation medium was removed and replaced with DMEM (Dulbecco's modified Eagle's medium; Gibco) supplemented with 10 μg/ml insulin (Sigma)/100 μg/ml of Transferrin (Sigma). This condition of plating cells at a high density prior to inducing differentiation was chosen in order to eliminate any bias in the quantification of fusion, which could be decreased by the poor migration of senescent cells.

Differentiation analysis

Immunofluorescence and fusion index

After 6 days of differentiation, cells were fixed in ethanol, washed in PBS and then incubated with an antibody against Desmin (1:50; D33, Dako). Specific antibody binding was revealed using Alexa Fluor® 488 directly coupled with the secondary antibody (Molecular Probes). To visualize nuclei, cells were mounted in Mowiol containing Hoechst (Sigma) at a concentration of 10 μg/ml. All images were digitalized using the MetaView image analysis system.

To calculate the fusion index, the number of nuclei incorporated into the myotubes (>2 nuclei) was counted and the ratio of this number to the total number of nuclei was determined. At least 500 nuclei per dish were counted and three dishes were counted for each experiment.

Number of nuclei per myotube

To determine the size of the myotubes, the number of nuclei per myotube was counted. A total of 500 nuclei were counted per dish and three dishes were counted for each experiment. The myotubes were graded into three groups: less than 25 nuclei, between 25 and 50 nuclei and more than 50 nuclei per myotube and the results are presented as a percentage of the total number of myonuclei.

RT—PCR

Total RNA was extracted from cells by using TRIzol® reagent (Invitrogen) according to the manufacturer's instructions. A 5 μg portion of RNA from each sample was reverse-transcribed into cDNA according to the manufacturer's instructions (Invitrogen). Equal amounts of the RT products were subjected to PCR amplification (ABGene). Amplification was initiated by 5 min of denaturation at 94°C, followed by 20 cycles (GAPDH) or 22 cycles (CTGF and collagen IV) of amplification. Each cycle consisted of 45 s at 94°C, 45 s at 55°C (GAPDH) or 59°C (CTGF and Collagen IV) and 45 s at 72°C. A final step of extension was performed for 10 min at 72°C. After amplification, 15 μl of each PCR reaction product was separated on a 1.5% agarose gel containing ethidium bromide. The mRNA levels of CTGF and collagen IV were normalized with GAPDH mRNA levels. All PCR primers were synthesized by Sigma—Proligo.

Sequences of the primers are: (i) GAPDH sense, 5′-GATGACAAGCTTCCCGTTCTCAGCC-3′; (ii) GAPDH antisense, 5′-TGAAGGTCGGAGTCAACGGATTTGGT-3′; (iii) collagen IV sense, 5′-CGGGGTTACAAGGTGTCATTGG-3′; (iv) collagen IV antisense, 5′-GCCAAGTATCTCACCTGG-3′; (v) CTGF sense, 5′-AACTATGATTAGAGCCAACTGCCTG-3′; (vi) CTGF antisense, 5′-TCATGCCATGTCTCCGTACATCTTC-3′.

RNase protection assay

Total RNA was extracted from the cells using TRIzol® reagent (Invitrogen) according to the manufacturer's instructions. The RNase protection assays were performed in three independent experiments following the instruction for the multi-probes hCC-2 kit (BD Biosciences). Briefly, the antisense RNA probes were synthesized with T7 RNA polymerase with [α-32P]UTP and purified by phenol/chloroform precipitation. A 5 μg portion of mRNA was hybridized with an excess of probes. Free probes and other single-stranded RNAs were digested by RNases. The remaining RNA/probes were purified and resolved on denaturing polyacrylamide gels. Gels were dried and exposed to a Phosphor Screen (Kodak) and the signals were captured using a Personal Molecular Imager FX (Bio-Rad). The signal responses were analysed by a computer-assisted system using NIH Image 1.62. The results were normalized to the GAPDH housekeeping mRNA and the ratios were expressed as compared with those obtained on samples before differentiation (sample 0), to which a value of 1 was attributed.

Northern blot

Total RNA was extracted using TRIzol® reagent (Invitrogen). For each sample, 10 μg of total RNA was separated on a 1% agarose gel containing 0.66 M of formaldehyde and transferred by capillarity on to a nylon membrane (Biodyne-Pall). MyoD, Myf5, Myogenin and MRF4 probes were generated with [α-32P]dCTP by random priming of 25 ng of cDNA. Hybridization was carried out overnight at 68°C in 10% dextran sulfate/2% (w/v) SDS/2×Denhardt's (1×Denhardt's=0.02% Ficoll 400, 0.02% polyvinylpyrrolidone and 0.02% BSA)/1×SSPE (0.15 M NaCl, 10 mM sodium phosphate, pH 7.4, and 1 mM EDTA)/100 μg/ml salmon sperm DNA. An 18S rRNA probe was used to assess the amount and integrity of the total RNA loaded. Membranes were exposed to a Phosphor Screen (Kodak) and the signals were captured using a Personal Molecular Imager FX (Bio-Rad). The signal responses were analysed by a computer-assisted system using NIH Image 1.62 and normalized to the 18S rRNA signal.

Western-blot analysis

A 30 μg portion of total protein extracts was resolved on SDS/10% PAGE. Proteins were transferred on to a nitrocellulose membrane and incubated with antibodies directed against: Myogenin (1:50; F5D hybridome), p57 (1:200; sc-1040; Santa Cruz Biotechnology) and Emerin (1:2000; Novocastra). Binding of a secondary antibody coupled with HRP (horseradish peroxidase) was revealed using an enhanced chemiluminescence kit (Pierce). Concerning MyoD analyses, membranes were blocked with 3% (w/v) BSA/3% fetal calf serum in PBS/0.1% Tween for 6 h at room temperature (22°C). Membranes were then incubated with the MyoD antibody (1:500; clone 5.8A; Dako) in 1% BSA/2% fetal calf serum/PBS/0.1% Tween overnight at 4°C. Secondary antibody coupled with HRP (1:200000) was revealed using the SuperSignal West Femto Maximum Sensitivity Substrate enhanced chemiluminescence kit (Pierce). Signals were detected on a film (Fujifilm) and quantified by densitometry by using NIH Image 1.62 software.

MyoD and Myf5 DNA-binding activity

The DNA-binding activity assays were performed using a TransAM kit (TransAM™ Active Motif) according to the manufacturer's instructions. Briefly, the DNA-binding ELISA assays were based on the use of multiwell plates coated with an oligonucleotide containing an E-box (5′-CACCTG-3′). MyoD or Myf5 contained in the nuclear extract specifically binds to this oligonucleotide. The presence of the DNA-bound transcription factors was detected by specific antibodies directed against MyoD and Myf5, by using the antibodies provided with the kit and following the manufacturer's instructions. The binding of a secondary antibody coupled with peroxidase was revealed by a sensitive colorimetric agent and quantified by spectrophotometry at 450 nm. The results, which represent means for three independent experiments, are expressed after subtraction of the blank values. The blanks were prepared following the same procedure, except that lysis buffer (0.5% Nonidet P40, 20 mM Hepes, 5 mM NaF, 0.1 mM EDTA and 1 mM Na2MoO4) is incubated in the microwells instead of nuclear lysate. Nuclear extracts were prepared as described in the kit instructions. This assay is more sensitive than the electrophoretic mobility-shift assay (Renard et al., 2001) and has already been used in senescence-related studies (de Magalhaes et al., 2004).

Statistical analysis

Statistical analyses were performed using Prism software. Results are given as the means±S.D. for at least three experiments. To determine significance between two groups, comparisons were made using an unpaired Student's t test. Analyses of multiple groups were performed using one-way ANOVA followed by a Newman—Keuls post hoc test. For all tests, the groups were considered statistically different for P value <0.05 (*P<0.05, **P<0.01, ***P<0.001).

Acknowledgements

We acknowledge financial support from: the AFM (Association Française contre les Myopathies), Inserm, University Pierre et Marie Curie, the Duchenne Parent Project-NL and the United Duchenne Parents Project, and particularly Luc Pettavino, the MYORES Network of Excellence (contract 511978) and the MYOAMP STREP (contract 0374479 from the 6th Framework Programme of the European Commission), both from the 6th Framework Programme of the European Commission, and the CPLD (Conseil de Prévention et de Lutte contre le Dopage). We also thank Woodring E. Wright, Elisa Negroni, Ingo Riederer and Virginie François for fruitful discussions, Lidia Dollé, Stéphanie Lorain and Valérie Migeot for technical assistance, and Kamel Mamchaoui (Platform for Human Cells, Institut de Myologie). O.T. and F.D.-C. are respectively Research Associate and Post-doctoral Fellow of the Belgian FNRS (Fonds National de la Recherche Scientifique). We thank the European Commission for ‘Link-Age’ Coordination Action, LSHM-CT-2005-523866, and ‘Proteomage’ Integrated Project, LSHM-CT-2005-518230.

Ancillary