Adapted physical exercise enhances activation and differentiation potential of satellite cells in the skeletal muscle of old mice

Abstract

During ageing, a progressive loss of skeletal muscle mass and a decrease in muscle strength and endurance take place, in the condition termed sarcopenia. The mechanisms of sarcopenia are complex and still unclear; however, it is known that muscle atrophy is associated with a decline in the number and/or efficiency of satellite cells, the main contributors to muscle regeneration. Physical exercise proved beneficial in sarcopenia; however, knowledge of the effect of adapted physical exercise on the myogenic properties of satellite cells in aged muscles is limited. In this study the amount and activation state of satellite cells as well as their proliferation and differentiation potential were assessed in situ by morphology, morphometry and immunocytochemistry at light and transmission electron microscopy on 28‐month‐old mice submitted to adapted aerobic physical exercise on a treadmill. Sedentary age‐matched mice served as controls, and sedentary adult mice were used as a reference for an unperturbed control at an age when the capability of muscle regeneration is still high. The effect of physical exercise in aged muscles was further analysed by comparing the myogenic potential of satellite cells isolated from old running and old sedentary mice using an in vitro system that allows observation of the differentiation process under controlled experimental conditions. The results of this ex vivo and in vitro study demonstrated that adapted physical exercise increases the number and activation of satellite cells as well as their capability to differentiate into structurally and functionally correct myotubes (even though the age‐related impairment in myotube formation is not fully reversed): this evidence further supports adapted physical exercise as a powerful, non‐pharmacological approach to counteract sarcopenia and the age‐related deterioration of satellite cell capabilities even at very advanced age.

Introduction

During ageing, a progressive loss of skeletal muscle mass and a parallel decrease in muscle strength and endurance take place. This condition, termed sarcopenia, has important healthcare and socio‐economic implications for humans as it contributes to frailty, functional loss, disability, high healthcare costs, and premature death (Cruz‐Jentoft et al. 2010).

The mechanisms of sarcopenia are complex and still remain to be completely elucidated; however, it has long been suggested that muscle atrophy is associated with a remarkable decline in the efficiency of tissue regeneration (recent reviews in Sayer et al. 2013; Schiaffino et al. 2013; Alway et al. 2014).

The main contributor to muscle maintenance, growth, and repair is the satellite cell (SC), the undifferentiated mononuclear myogenic precursor located between the plasma membrane and the basal lamina of the adult myofibre (Mauro, 1961; Campion, 1984; Anderson & Wozniak, 2004). In adult muscles, SCs are normally quiescent, but in response to physiological or pathological stimuli (e.g. stretching, exercise, mechanical injury, denervation or muscle dystrophy), they activate and proliferate to generate myoblasts. These myoblasts can either repair damaged segments of existing fibres or fuse together to generate entirely new multinucleated syncytia; some daughter cells return to quiescence, thereby maintaining the basal pool of resident SCs (Relaix & Zammit, 2012). Parise et al. (2008) showed that muscle satellite cells are the primary responders to exercise‐induced stress, and Lepper et al. (2011) demonstrated that SCs are essential for acute injury‐induced muscle regeneration. Fry et al. (2014), using an SC‐depleted mouse model, were able to prove the role of SCs in the maintenance of muscle mass following hypertrophy, although evidence has also been provided (McCarthy et al. 2011) that mechanically induced hypertrophy does not depend directly on SCs.

Actually, decreasing numbers of SCs and/or alteration in SC potential to activate, proliferate and/or differentiate has been shown in the ageing muscle (recent reviews in García‐Prat et al. 2013; Alway et al. 2014). It is, however, worth recalling that in male sedentary mice which had been experimentally depleted of SCs in young adulthood, sarcopenia was neither accelerated nor worsened, thus suggesting that SCs may not contribute directly to the maintenance of fibre size during ageing (Fry et al. 2015).

Physical exercise proved beneficial in aged skeletal muscle, without apparent histological or cytological damage (e.g. Yarasheski, 2002; Marcell, 2003; Zancanaro et al. 2007; Bautmans et al. 2009). Resistance training has been shown to impact favourably on SCs (Snijders et al. 2009, 2014). However, knowledge on the effect of adapted physical exercise on the myogenic properties of SCs in aged muscles is limited (Shefer et al. 2010); in particular, the effect on SCs of aerobic exercise, a commonly used training modality in the elderly (Pillard et al. 2011), has not yet been explored.

In a previous study (Malatesta et al. 2011) we demonstrated that sarcopenia is mitigated in 28‐month‐old mice by adapted aerobic physical exercise on a treadmill. Here, we deepen the analysis on the beneficial effect of this physical exercise focusing on SCs of old muscles, in the same experimental model. The amount and activation state of SCs as well as their proliferation and differentiation potential were assessed in situ by morphology, morphometry and immunocytochemistry at light and transmission electron microscopy. Considering that SCs are a scanty and possibly heterogeneous cell population, which can be unequivocally recognized on the base of topological location and/or molecular markers, this approach is especially suitable for studying their structural and functional features (Pellicciari, 2013). The effect of physical exercise in aged muscles was further analysed by comparing the myogenic potential of SCs isolated from old running and old sedentary mice using an in vitro system, which allows observation of the differentiation process under controlled experimental conditions. Sedentary adult mice were used as a reference for an unperturbed control at an age when the capability of muscle regeneration is still high.

Materials and methods

Animals and physical exercise

Eight adult (12‐month‐old) and 16 old (28‐month‐old) male mice from the INRCA breed (Ancona, Italy) were used in this study. The INRCA breed is a 40‐year established Balb‐c mice strain which has been widely used for studies on physiological ageing: these mice have a long life (mean life span 25 months; maximal life span 34 months; Mocchegiani et al. 2007) and a relatively low incidence of pathologies (Staats, 1980; Bronson & Lipman, 1993) in comparison with the usual Balb‐c strains, which generally have a life span of about 18 months (Storer, 1966).

Animals were bred as a close colony, maintained under standard conditions (24 ± 1 °C ambient temperature, 60 ± 15% relative humidity, and 12 h light/dark cycle), and fed ad libitum with a standard commercial chow diet. Eight old mice were trained on a Harvard Instruments treadmill (Crisel Instruments, Rome, Italy) for 45 min a day at 9 m min⁻¹ belt speed, 5 days a week for 1 month (old running group, OR). Current treadmill protocols for adult mice consistently use 1 h running a day at belt speed > 10 m min⁻¹. In this work, physical training was adapted to optimize old mice compliance to training (Fabene et al. 2008). Eight old mice (old sedentary group, OS) and eight adult animals (adult sedentary group, AS) had only spontaneous free‐moving activity in the cage. AS animals were used as a reference group characterized by a high muscle regeneration capability.

To avoid possible interference of acute with chronic effects of physical exercise, the animals were killed 3 days after the last treadmill session.

The experimental protocols have been approved by the Ethical Committee for Animal Experimentation (CESA) of the University of Pavia and authorized by the Italian Ministry of Health.

Ex vivo analyses of SCs

Four mice from each group (OS, OR, AS) were used: they were deeply anaesthetized with diethyl ether and then killed by cervical dislocation. Quadriceps femoris was chosen as the most suitable muscle for this study because of the prevalence (about 90%) therein of type II fast‐twitch myofibres, which are especially affected by sarcopenia (Larsson et al. 1978; Lexell, 1995). The muscles were quickly removed and fixed in 4% paraformaldehyde in 0.1 m phosphate buffer, pH 7.4 for 24 h at 4 °C, extensively washed in tap water for 24 h, dehydrated with ethanol, and embedded in paraffin wax.

To identify SCs, 5‐μm‐thick cross‐sections of muscle samples were rehydrated, pre‐treated with H₂O₂ to block endogenous peroxidases and then incubated overnight at 4 °C with a rabbit polyclonal antibody (Abcam, Cambridge, MA, USA) recognizing the SC‐specific paired box protein 7 (Pax7) transcription factor (Seale et al. 2000), diluted 1 : 400 in PBT [phosphate‐buffered saline (PBS) containing 0.5% Tween‐20 and 0.01% bovine serum albumin (BSA)]. The antigen–antibody complex was revealed as a brown reaction product by using a biotinylated antibody against rabbit IgG (Abcam), followed by the ABComplex (Vectastain, DBA Italia S.r.l., Milan, Italy) and, finally, 3,3'‐diaminobenzidine (DAB). To identify activated SCs, other muscle sections were incubated with a mouse monoclonal antibody (Abcam) recognizing the myogenic differentiation transcription factor D (MyoD; Legerlotz & Smith, 2008) diluted 1 : 400 in PBT and then revealed as above, with the addition of 0.5% NH₄Cl to obtain a blue precipitate. However, as MyoD also occurs in myonuclei and can be expressed also by other stem cells (Legerlotz & Smith, 2008), a simultaneous staining of laminin was performed to distinguish SCs unequivocally: briefly, after MyoD labelling the sections were incubated with a rabbit polyclonal anti‐laminin antibody (Abcam) diluted 1 : 400 in PBT, then reacted with a biotinylated antibody against rabbit IgG (Abcam) followed by the ABComplex (Vectastain) and finally DAB, thus visualizing as a brown staining the antigen–antibody complex detecting laminin.

Sections were observed in an Olympus BX51 light microscope using a 100× objective lens. In muscle tissue labelled for Pax7, Pax7⁺ SCs were counted around individual myofibres over 400 myofibres per animal; the percentage of myofibres showing 0–6 Pax7⁺ SCs was also calculated. In sections labelled for MyoD and laminin, MyoD⁺ SCs as well as the total number of SCs (labelled and unlabelled SCs surrounded by basal lamina) were counted around each myofibre, and the percentage of MyoD⁺ SCs per myofibre was calculated over 400 myofibres per animal; the percentage of myofibres showing 0–6 MyoD⁺ SCs was also calculated.

In vitro analyses of SC‐derived myoblasts and myotubes

Four mice for each group were used. Samples of quadriceps femoris were trimmed out of blood vessels, fat and connective tissue, and rinsed in PBS. SCs were isolated by enzymatic dissociation of skeletal muscle (slightly modified from Musarò & Barberi, 2010). Briefly, the muscle tissue was minced with small surgical scissors, incubated in collagenase type II (Sigma‐Aldrich, Buchs, Switzerland; 0.1 mg mL⁻¹ PBS) for 20 min at 37 °C on a rocker, and then centrifuged. The pellet was resuspended in the collagenase solution and incubated again for 30 min at 37 °C. The tissue was dispersed by pipetting, filtered using appropriate nylon mesh cell strainers (Falcon 2360, 2340 and 2350; Euroclone, Milan, Italy), centrifuged and resuspended in culture medium (see below). After pre‐plating in Petri dishes in an incubator (5% CO₂, 37 °C) for 1 h to remove fibroblasts, the solution enriched in myoblasts was seeded in 12‐multiwell dishes (Euroclone) in a culture medium containing 50% DMEM (Dulbecco Minimal Essential Medium, Euroclone) and 50% F10 HAM (Euroclone), supplemented with 15% fetal bovine serum (Euroclone), 15% horse serum (Euroclone), 0.5 mg mL⁻¹ BSA (Sigma), 0.5 mg mL⁻¹ fetuin (Sigma), 0.39 μg mL⁻¹ dexamethasone (Sigma), 10 ng mL⁻¹ epidermal growth factor (Sigma), 0.05 mg mL⁻¹ insulin (Sigma), 3 mg mL⁻¹ glucose (Sigma), 0.5% chick embryo extract, 4 mm L‐glutamine (Euroclone), 100 U mL⁻¹ penicillin (Euroclone) and 100 μg mL⁻¹ streptomycin (Euroclone). The cell cultures were grown at 37 °C in a humidified 95% air/5% CO₂ atmosphere for three passages.

To evaluate the percentage of myogenic cells, aliquots of isolated cell suspensions were fixed with 2% formaldehyde in PBS for 15 min, washed in PBS, and incubated for 60 min with a mixture of the rabbit polyclonal antibody recognizing Pax7 and the mouse monoclonal antibody recognizing MyoD, diluted as described above. Primary antibodies were then revealed with a mixture of an Alexa Fluor 488‐conjugated anti‐rabbit IgG antibody and an Alexa Fluor 594‐conjugated anti‐mouse IgG antibody (both antibodies were diluted 1 : 200 in PBS). A drop of the immunolabelled samples was placed onto a glass slide and covered with a coverslip to be observed in an Olympus BX51 fluorescence microscope equipped with a 100 W mercury lamp and the appropriate filtre sets for the fluorochromes used. Using a 40× objective lens, 10 fields per sample were randomly selected and, for each field, the total number of cells, the number of Pax7⁺ (i.e. satellite) cells and of Pax7/MyoD double‐positive (i.e. activated satellite) cells as well as their percentages were estimated; at least 500 cells per sample were scored.

To monitor adhesion of isolated SCs to the growing plastic substrate of the 12‐multiwell dishes, the cultures were observed daily in phase contrast under an Olympus CX41 inverted microscope equipped with a 40× phase‐contrast objective lens: the percentage of adhering cells was estimated in 10 randomly selected fields per culture.

Transmission electron microscopy

The myoblasts obtained by the above‐described procedure were allowed to grow until they were approximately 80% confluent and then split 1 : 3 for propagation. At the third passage, myoblasts were planted on gelatin‐coated glass coverslips and, when 80% confluent, some samples were processed for transmission electron microscopy as described below, while others were placed in the differentiation medium, consisting of DMEM supplemented with 5% horse serum, 3% chick embryo extract, 4 mm l‐glutamine, 100 U mL⁻¹ penicillin and 100 μg mL⁻¹ streptomycin, and allowed to fuse and differentiate into myotubes for 7 days.

For ultrastructural analysis, both myoblasts and myotubes were fixed with 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 m Sörensen phosphate buffer at 4 °C for 1 h, washed, post‐fixed with 1% OsO₄ at 4 °C for 30 min, dehydrated with acetone and embedded in Epon, as described in Malatesta et al. (2013).

Ultrathin sections were stained with 4.7% aqueous solution of uranyl acetate and lead citrate for ultrastructural observation.

Morphometrical analyses were performed on randomly selected myoblast nuclei of Epon‐embedded samples (20 per sample) to compute an index of nuclear shape. Areas and perimeters of nuclei were measured (× 11 000) and the index was expressed as the ratio between the perimeter and the circumference of the equivalent circle (index = P/2πr, where P is the observed perimeter, r is the radius of equivalent circle having the same area A; thus r = √A/π).

Ultrastructural immunocytochemistry

For immunoelectron microscopy, ultrathin sections of myoblast samples were treated with the following probes: mouse monoclonal antibodies directed against the active phosphorylated form of RNA polymerase II (Research Diagnostic Inc., Flanders, NJ, USA), or the (Sm)snRNP (small nuclear RNP) core protein (Abcam) or MyoD (Abcam), or a rabbit polyclonal antibody directed against Pax7 (Abcam). Before starting the immunocytochemical procedure, the samples were submitted to an etching procedure with a 0.2 m aqueous solution of sodium metaperiodate for 60 min at room temperature to improve antibody binding (Bendayan & Zollinger, 1983). Sections were briefly floated on normal goat serum diluted 1 : 100 in PBS and then incubated for 17 h at 4 °C with the primary antibodies diluted with PBS containing 0.1% BSA (Fluka) and 0.05% Tween 20. After rinsing, sections were floated on normal goat serum and then allowed to react for 30 min at room temperature with the appropriate 6‐nm or 12‐nm gold‐conjugated secondary antibodies (Jackson ImmunoResearch Laboratories Inc.) diluted 1 : 10 in PBS. Finally, the sections were rinsed and air‐dried. As controls, some grids were incubated in buffer without the primary antibody and then processed as described above. The sections were weakly contrasted with 2.5% uranyl acetate.

All samples were observed in a Philips Morgagni TEM operating at 80 kV and equipped with a Megaview II camera for digital image acquisition.

Quantitative assessment of immunolabelling was carried out by estimating the gold particle density over the nucleoplasm in sections treated in the same run. The surface area of nucleoplasm was measured in 20 randomly selected electron micrographs (× 22 000) of myoblasts from each animal using a computerized image analysis system (AnalySIS Image processing, Soft Imaging System GmbH). Background evaluation was carried out on resin (in the areas devoid of tissue) of the immunolabelled samples as well as on tissue of control samples. Gold particles present over the nucleoplasm were counted and the labelling density was expressed as number of gold particles per μm².

Statistical analysis

For each analysed variable, the Kolmogorov–Smirnov two‐sample test was performed to verify the hypothesis of identical distribution among animals of each experimental group. Data for each variable were then pooled according to the three experimental groups, OS, OR and AS, and the mean ± standard deviation (SD) was calculated. Statistical analysis of the results was performed by the Kruskal–Wallis test, and the two‐tailed Mann–Whitney U‐test was used for pairwise comparisons. Statistical significance was set at P ≤ 0.05.

Results

Ex vivo analyses of SCs

The immunolabelling for Pax7 allowed the visualization of SCs in skeletal muscle sections: in all mice they were distributed at the periphery of myofibres (Fig. 1a–c). Quantitative analysis of Pax7‐labelled cells demonstrated lower SC number in OS mice and similar values in OR and AS mice (Fig. 1d, Table 1).

Figure 1.

SCs immunostained for Pax7. (a‐c) Immunohistochemical detection of Pax7 on quadriceps femoris sections from AS (a), OS (b), and OR (c) mice. Scale bar: 50 μm. (d) Mean percentage ± SD of myofibres with 0 to 6 associated Pax7⁺ SCs in the three experimental groups: it is worth noting that in OS mice, the fraction of myofibres without associated SCs is significantly higher than in AS (P = 0.04) or OR (P = 0.04) animals.

Table 1.

Total number of Pax7⁺ SCs and percentage of MyoD⁺ cells therefrom in myofibres from individual OS (n = 4), OR (n = 4) and AS (n = 4) mice. The mean value per group is also presented together with P values for the Kruskal–‐Wallis and the Mann–Whitney U‐tests. Data are mean ± SD

	Pax7+ cells		MyoD+ cells
	Individual value (cells per myofibre)	Mean*	Individual value (%)	Mean**
OS	1.01 ± 1.01	1.00 ± 0.11°, ^	25.34 ± 27.11	27.53 ± 2.08§, °
	1.01 ± 0.96		28.07 ± 31.37
	1.13 ± 1.08		26.54 ± 29.49
	0.86 ± 0.85		30.16 ± 35.04
OR	1.25 ± 0.97	1.56 ± 0.25	57.41 ± 32.10	59.30 ± 4.75¶
	1.85 ± 1.16		53.67 ± 33.84
	1.54 ± 0.98		61.60 ± 33.55
	1.58 ± 1.12		64.52 ± 32.72
AS	1.89 ± 1.16	1.65 ± 0.21	82.50 ± 28.01	80.49 ± 4.48
	1.74 ± 1.18		85.82 ± 24.03
	1.46 ± 0.98		77.45 ± 31.35
	1.49 ± 1.09		76.21 ± 32.08

*P = 0.006; **P = 0.001.

^° P = 0.007 vs. OR; ^{^} P = 0.008 vs. AS.

^§ P = 0.012 vs. OR; ^° P = 0.007 vs. AS; ^¶ P = 0.024 vs. AS.

Simultaneous labelling of MyoD and laminin allowed for the unambiguous detection of activated SCs: in fact, MyoD⁺ SCs were clearly surrounded by the basal lamina (Fig. 2a). Quantitative analysis demonstrated that OS mice showed the lowest percentage of MyoD⁺ SCs per myofibre; in OR mice this figure was higher than in OS, while remaining lower than in AS (Fig. 2b, Table 1).

Figure 2.

SCs immunostained for MyoD. (a) Representative micrograph showing the dual immunohistochemical detection of MyoD (blue signal) and laminin (brown) on a quadriceps femoris section from an OR mice: the black arrow points to an activated (MyoD⁺) SC, and the red and the green arrows respectively indicate a quiescent (MyoD⁻) SC surrounded by the basal lamina and a MyoD⁺ myonucleus. Scale bar: 50 μm. (b) Mean percentage ± SD of myofibres with 0–6 associated MyoD⁺ SCs in the three experimental groups: it is worth noting that in OS mice, the fraction of myofibres with 0 associated SCs is significantly higher than in AS (P = 0.03) or OR (P = 0.04) animals.

In vitro analyses of SC‐derived myoblasts and myotubes

In the cultures from all experimental groups the myogenic index based on the immunopositivity for Pax7 as well as the percentage of activated SCs was always fairly high (Table 2).

Table 2.

Cells isolated from the quadriceps femoris muscle of OS (n = 4), OR (n = 4), and AS (n = 4) mice, counted in 10 fields per sample. The total number of Pax7⁺ SCs, Pax7⁺ and MyoD⁺ SCs, and their respective percentage per sample and group are reported. Data are mean ± SD

	Total cells	Pax7+ cells			Pax7+/MyoD+ cells
	Sample	Sample	%	% mean	Sample	%	% mean
OS	533	404	75.63 ± 7.57	74.67 ± 10.39	358	67.10 ± 8.04	65.77 ± 10.17
	587	425	72.35 ± 9.78		377	63.31 ± 8.36
	611	445	70.26 ± 15.03		414	64.43 ± 14.90
	501	373	74.67 ± 8.39		342	68.25 ± 8.58
OR	578	433	74.58 ± 8.34	74.93 ± 11.74	336	59.57 ± 12.83	63.09 ± 13.67
	633	489	78.37 ± 10.71		412	66.70 ± 15.68
	592	438	73.30 ± 14.52		391	64.63 ± 13.80
	544	398	73.49 ± 13.57		336	61.45 ± 13.25
AS	625	485	78.91 ± 9.84	77.55 ± 10.58	410	65.98 ± 6.33	68.04 ± 11.34
	576	439	76.62 ± 9.98		358	63.50 ± 12.99
	600	461	77.41 ± 12.33		418	70.18 ± 13.16
	631	489	77.24 ± 11.54		459	72.48 ± 10.99

The SCs from all the experimental groups were able to adhere to the culture support and become proliferating myoblasts, although the adhering process was much less efficient in OS than in OR myoblasts, with AS showing the better performance (flattened and firmly adhering SCs reached 50% in about 24, 72, and 120 h for cells isolated from AS, OR, and OS mice, respectively) (Fig. 3a–c).

Figure 3.

Light microscopy micrographs of cultured SC‐derived myoblasts and myotubes. Primary cultures of myoblasts from AS (a), OS (b), and OR (c) mice 12 h after seeding of freshly isolated SCs: in (a), nearly all myoblasts are flattened and firmly adhering (arrows), whereas adhering cells (arrows) are much less numerous in (b) and (c). Examples of multinucleated myoblast‐derived myotubes from AS (d), OS (e,e'), and OR (f) mice: myotubes from AS mice (d) exhibit regularly aligned nuclei (arrowheads), whereas those from OS mice (e) are often roundish in shape with clustered nuclei (arrowheads) and even at the beginning of myoblasts' fusion (e'), extensive vacuolization is observed in the cytoplasm (thin arrows). The myotubes from OR mice (f) are more similar to those from the adults, although the nuclei (arrowheads) are less regularly arranged. Scale bars: 30 μm (a‐c), 50 μm (d‐f).

Group difference was also observed in the processes of myoblast fusion and myotube organization. Myoblasts from AS mice SCs easily fused, forming long myotubes with regularly aligned nuclei (Fig. 3d) showing spontaneous contraction (a cue for the correct assembly of myofibrils). In contrast, OS myoblasts rarely formed myotubes and, when they did, these were morphologically abnormal (i.e. roundish), with clustered nuclei (Fig. 3e) and extensive cytoplasmic vacuolization (Fig. 3e'), and were never found to contract. OR myotubes were similar to the AS ones, although they were shorter and with less regularly arranged nuclei (Fig. 3f), and were never observed to contract.

Transmission electron microscopy

Transmission electron microscopy revealed fine structural differences among myoblasts from OS, OR, and AS mice. AS myoblasts (Fig. 4a‐c) had roundish nuclei (mean shape index ± SD: 1.13 ± 0.09) with one or two reticular nucleoli; the cytoplasm was rich in free ribosomes, rough endoplasmic reticulum (RER), and Golgi apparatus (GA); some residual bodies were also present. Bundles of cytoskeletal filaments showing incipient sarcomere‐like arrangements were frequent, and elongated mitochondria were often lined in between the bundles. OS myoblasts (Fig. 4d,e) showed roundish nuclei (mean shape index ± SD: 1.16 ± 0.12; OS vs. AS P = 0.386) with one to two compact nucleoli; the cytoplasm contained large amounts of heterogeneous residual bodies and vacuoles, whereas free ribosomes, RER, GAs, and mitochondria were morphologically similar those in AS myoblasts, though less abundant; cytoskeletal filaments were irregularly arranged (not shown). OR myoblasts (Fig. 4f,g) showed irregularly shaped nuclei (mean shape index ± SD: 1.28 ± 0.12; OR vs. AS: P < 0.001; OR vs. OS: P = 0.005) with one or two compact nucleoli; free ribosomes were abundant, and RER, GAs, and mitochondria were numerous and well developed; the cytoplasm also contained some residual bodies and bundles of cytoskeletal filaments with sarcomere‐like arrangements.

Figure 4.

Electron micrographs of SC‐derived myoblasts from muscles of AS (a‐c), OS (d,e), and OR (f,g) mice. In (a) the myoblast shows a roundish nucleus (N) with a large reticular nucleolus (Nu); the cytoplasm (b) is rich in RER (arrows), GAs (G), and mitochondria (M); some residual bodies (R) are also present. Moreover, bundles of cytoskeletal filaments show incipient sarcomere‐like arrangements (c). In (d,e), the myoblast shows an irregularly shaped nucleus (N) with a compact nucleolus (Nu), whereas the cytoplasm contains large amounts of heterogeneous residual bodies (R); RER (arrows), GAs (G), and mitochondria (M) are less abundant. In (g,f), the myoblast shows an irregularly shaped nucleus (N) with a compact nucleolus (Nu); RER (arrow), GAs (G), and mitochondria (M) are well developed similar to myoblasts from AS mice; moreover, these filament bundles show sarcomere‐like figures (inset in g). Scale bars: 2 μm (a,d,f), 1 μm (b,c,e,g, inset).

Transmission electron microscopy also demonstrated structural differences among myotubes differentiated from OS, OR, and AS myoblasts. AS myotubes (Fig. 5) were elongated in shape, with longitudinally arranged nuclei showing scarce heterochromatin and one or two reticular nucleoli. In the cytoplasm, several bundles of longitudinally arranged myofibrils showing sarcomere‐like structures occurred. Numerous elongated mitochondria were either orderly arranged in between myofibril bundles or clustered around the nuclei. Large amounts of free ribosomes, RER, smooth endoplasmic reticulum and well developed GAs were present. Heterogeneous residual bodies were also observed. OS myotubes (Fig. 6a) were roundish; their nuclei exhibited numerous indentations, and contained clumps of heterochromatin and one or two compact nucleoli. In the cytoplasm, residual bodies and electron‐lucent vacuoles were prominent. Mitochondria, RER, and GAs were well developed, though being less numerous than in AS myotubes. The myofibrils were scarce and irregularly arranged, and occasionally formed thin bundles that were always devoid of sarcomere‐like patterns. OR myotubes (Fig. 6b) had irregular shapes, and their nuclei contained some heterochromatin and one or two compact nucleoli. In the cytoplasm, the myofibril bundles were less numerous and thinner than in AS myotubes, but they often showed sarcomere‐like arrangements. Numerous elongated mitochondria, free ribosomes, and well developed RER and GAs were present, as well as large numbers of residual bodies and vacuoles.

Figure 5.

Electron micrographs of myotubes derived from myoblasts of AS mice. The myotubes show elongated shapes, with longitudinally arranged nuclei (a). Mitochondria (M), RER (arrows), and GAs (G) are numerous and well developed (b‐c). (d‐e) Several bundles of longitudinally arranged myofibrils (F) show sarcomere‐like arrangements (arrows), and elongated mitochondria (M) are often lined in the small cytoplasm cords (asterisks) between the bundles. Scale bars: 2 μm (a), 1 μm (b–e).

Figure 6.

Electron micrographs of myotubes derived from myoblasts of OS (a) and OR (b) mice. OS myotubes (a) are roundish with irregularly shaped nuclei (N); residual bodies (asterisks) and vacuoles (v) are prominent. Mitochondria (M) and GAs (G) are scarce (a'). Myofibrils are rare and irregularly arranged (a''). OR myotubes (b) exhibit irregular shapes with centrally located nuclei (N); residual bodies (asterisks) and vacuoles (v) are numerous. Mitochondria (M) are abundant, and RER (arrow) and GAs (G) are well developed (b'). Myofibril bundles are thin and often show a sarcomere‐like pattern (arrows) (b''). Scale bars: 2 μm (a,b), 1 μm (a',a'',b',b'').

Ultrastructural immunocytochemistry

Immunocytochemical analyses revealed that, in addition to these cytological modifications, OS, OR, and AS myoblasts showed significant differences in the immunolabelling density for the transcription factors RNA polymerase II, Pax7, and MyoD as well as for the early splicing factor (Sm)snRNP, although the intranuclear distribution of these nuclear factors was similar for all the experimental groups. The transcription factors were always associated to the perichromatin fibrils (i.e. the site of pre‐mRNA transcription, splicing and 3′‐end processing; Biggiogera & Fakan, 2008) and the splicing factors to both perichromatin fibrils and interchromatin granules (i.e. the storage, assembly, and phosphorylation sites for transcription and splicing factors; Puvion & Puvion‐Dutilleul, 1996; Bogolyubov et al. 2009) (Fig. 7a‐c). The labelling for RNA polymerase II, (Sm)snRNPs, Pax7, and MyoD markedly increased in OR in comparison with OS mice (Fig. 7d).

Figure 7.

Immunoelectron microscopy of myoblast nuclei. Representative high magnification details of myoblast nuclei immunolabelled for (a) activated RNA polymerase II, (b) (Sm)snRNPs, and (c) Pax7 (12‐nm gold particles) and MyoD (6‐nm gold particles). The labelling distribution was similar in all animal groups: activated RNA polymerase II, Pax7, and MyoD were specifically associated to perichromatin fibrils (arrows), whereas (Sm)snRNPs occurred on perichromatin fibrils (arrows) and interchromatin granules (IG). Scale bars: 0.5 μm. (d) Labelling density (gold particles per μm²) of the four factors in the nucleoplasm is shown (means ± SD); all signals markedly decrease in OS mice and increase in OR animals. In each histogram, asterisks indicate values that are significantly different from each other (OR vs. OS P < 0.001 for all probes; OS vs. AS P < 0.001 for all probes; OR vs. AS P < 0.001 for MyoD and (Sm)snRNPs, P = 0.243 for activated polymerase II and P = 0.186 for Pax7).

Background levels were always negligible (not shown).

Discussion

Sarcopenia has often been associated with a decline in SC number and/or activation potential both in rodents and humans (recent reviews in García‐Prat et al. 2013; Alway et al. 2014). Accordingly, we found that in the quadriceps femoris of OS vs. AS mice the total number as well as the activated fraction of SCs is reduced; in addition, SCs isolated from OS mice adhere more slowly to the culture support, and exhibit lower amounts of transcription and splicing factors with reduced ability to differentiate into well developed myotubes. Based on this in vitro observation, the lower regeneration potential in muscle from old sarcopenic mice might be due not only to the smaller number of SCs and/or their slower activation, but also to their impaired differentiation capability. This is consistent with findings by Pietrangelo et al. (2009) that showed limited capability of myoblasts from old people to execute a complete differentiation program, as well as Corbu et al. (2010), indicating that SCs isolated from ageing human muscle biopsies were able to proceed through the myogenic program in culture and form myotubes, although taking longer than SCs isolated from young controls. Corbu et al. (2010) also observed by RT‐PCR analysis and electron microscopy that the myogenic potential of SCs in vitro is compromised in the ageing human muscle. Studies on muscle precursor cells isolated from skeletal muscle of 32‐month‐old rats also revealed that the onset of proliferation and the exit from the cell cycle were delayed compared with the same cells from 1‐month‐old rats, suggesting that their myogenic potential and the ability to undergo differentiation could be impaired in ageing (Zwetsloot et al. 2013). It is, however, worth mentioning that other reports in the literature (Shefer et al. 2006; Alsharidah et al. 2013) indicate that muscle precursor cells from young and elderly people may have a similar myogenic behaviour in culture, so that the reduced muscle fibre maintenance in ageing should be accounted for by the decreased number of SCs rather than by their reduced myogenic potential (see also Renault et al. 2002). As suggested by Barberi et al. (2013), the limited ability of muscle repair in aged subjects may also depend on the age‐related changes in the systemic muscle environment, which becomes less effective at maintaining the regenerative activity of muscle stem cells while facilitating their conversion to fibrogenic differentiation.

Physical exercise proved to stimulate SCs from severely atrophic muscles, thus eliciting a hypertrophic response in the skeletal muscles of old individuals albeit to a reduced extent compared with their young counterparts (e.g. Welle et al. 1996; Grounds, 1998; Lowe & Alway, 1999; Blough & Linderman, 2000). Accordingly, previous work on severely sarcopenic 28‐month‐old mice demonstrated that adapted physical exercise by treadmill running induces a slight size increase in type II myofibres, while stimulating transcriptional and post‐transcriptional activities in both myonuclei and SC nuclei (Malatesta et al. 2011) without evident cell or tissue damage (Zancanaro et al. 2007; Malatesta et al. 2011). In the present work, we expand previous data by showing that such physical exercise contrasts the age‐related SC decline by increasing both the total number and the activated fraction of SCs per myofibre. Interestingly, these data are supported by findings in old rats in which a 4‐week endurance training of moderate intensity increased the amount of myofibres and SCs as well as the fraction of myogenic clones (Shefer et al. 2010), consistent with the observation that SCs increase in number after chronic endurance exercise during the rat maturational growth (Smith & Merry, 2012). Altogether, these findings suggest that physical exercise may exert positive effects on SCs irrespective of the exercise intensity. In addition, the absence of ultrastructural fibre damage under our experimental conditions proves that a non‐injurious adapted physical exercise is sufficient to activate SCs in old animals.

We also found that adapted physical exercise improves structural and functional features of SC‐derived myoblasts from old mice. In particular, cytoplasmic organelles involved in proteosynthesis, and the cytoskeletal components which are markedly altered in OS mice, partially reverted towards AS morphology in OR mice; moreover, the transcription factor MyoD (one of the earliest markers of the myogenic commitment) and the splicing factor (Sm)snRNP (involved in the early steps of pre‐mRNA maturation) significantly increased in nuclei of OR vs. OS myoblasts, while the level of expression of the transcription factors RNA polymerase II (main responsible for pre‐mRNA transcription) and Pax7 (expressed in proliferating myoblasts) reverted to the AS values. Furthermore, SC‐derived myoblasts from old animals submitted to adapted physical exercise are able to differentiate into myotubes that appear much better developed than those isolated from sedentary old mice, with special regard to the cytoskeletal apparatus. However, the administered exercise seems to be unable to fully restore the SC myogenic potential typical of adulthood; in fact, bearing in mind the features of the AS‐derived myotubes, it is evident that some structural alterations persist in OR mice‐derived myotubes, both in the cytoplasm (i.e. thinner myofibril bundles, vacuolization and accumulation of residual bodies) and the nucleus (i.e. higher amounts of heterochromatin and compact nucleoli, suggestive of a reduced nuclear activity). Beccafico et al. (2011) observed impaired differentiation of human myoblasts from old subjects when cultured in differentiation medium, and suggested that intrinsic factors exist which may account for the defective SC function in aged skeletal muscles.

The present study also suggests that SCs activation is an essential, but not sufficient, event for myogenesis in vitro. Similar results were obtained by Lees et al. (2006) in cultured muscle precursor cells from 32‐month‐old rats: compared with 3‐month‐old rats, an age‐associated decrease in the expression of muscle‐specific proteins (myosin heavy chain and muscle creatine kinase) was observed in old rats, despite similar levels of MyoD expression.

It should be kept in mind that the procedure of cell isolation entails mechanical disaggregation of the muscles, and this may act as an activating event for SCs, irrespective of the age and exercise condition of the mice (Wozniak et al. 2003).

However, although SCs isolated from AS, OS, and OR mice muscles rapidly activated in comparable amounts, adaptation to the culture condition as shown by surface adhesion was much faster in myoblasts from OR than from OS mice. Moreover, whereas SCs from OS mice gave rise to markedly altered myoblasts and myotubes, both myoblasts and myotubes from OR mice shared similar features with those from AS mice. This suggests that the activating effect of physical exercise is maintained and expressed in culture by SCs from OR mice, confirming the observation in humans by Green et al. (2013), who found that SCs isolated from active individuals retain in vitro some of the metabolic characteristics associated with physical activity.

Ageing implies progressive deterioration of the molecular mechanisms responsible for cell viability and proliferation. It has been suggested (Jameson, 2004) that the nucleus might be the target of ageing at the cellular level, due to a progressive alteration of the chromosome structure, the increasing malfunction of the molecule‐degrading systems, and the age‐related alteration of the mechanisms regulating protein post‐translational modification. Accumulation of both cytoplasmic and nuclear proteins usually occurs in ageing cells (Kim et al. 2001; Hallen, 2002; Malatesta et al. 2009, 2010a,b), mainly due to the lowered metabolic turnover of molecules (Goto et al. 2001; Jameson, 2004), and it is known that in senescing myoblasts the growth factor‐driven mitogenic signals are unable to overcome the cell cycle block due to the accumulation of inhibitory factors preventing cell division (Buricchi et al. 2005; Carlson & Conboy, 2007). Recently, it has been demonstrated that, normally, SCs actively repress genes inducing irreversible cell‐cycle withdrawal, and that this protective mechanism is lost with advanced age, driving SCs from a quiescent to a senescent state (Sousa‐Victor et al. 2014). During and following exercise, the skeletal muscle synthesizes and releases a number of myokines that influence muscle metabolism, regeneration, and/or hypertrophy (Pedersen et al. 2007); in particular, the leukaemia inhibitory factor has been shown to stimulate SC proliferation (Broholm & Pedersen, 2010). Mechanical signals elicited by exercise also induce the production of many growth factors influencing gene expression; among others, the hepatocyte growth factor and the mechano‐growth factor (although being produced at lower levels in aged than young muscle, in response to exercise) promote activation of SCs (Allen et al. 1995; Tatsumi et al. 2002; Goldspink, 2004; Goldspink & Harridge, 2004). In particular, the mechano‐growth factor also delays the onset of senescence of SCs isolated from newborns and young adults, probably by regulating the action of cell cycle‐related proteins (Kandalla et al. 2011).

Concluding remarks

The results of this ex vivo and in vitro study carried out in old mice demonstrated that adapted physical exercise increases the number and activation of SCs as well as their capability to differentiate into structurally and functionally correct myotubes, even though the age‐related impairment in the formation of functional myotubes is not fully reversed. Previous findings showed that a mild physical exercise promotes recovery of certain ultrastructural features and increases in situ transcriptional activity in myofibres and SCs of old mice muscles (Malatesta et al. 2011); consistently, the current data support adapted physical exercise as a powerful, non‐pharmacological approach to counteract age‐related deterioration of SC capabilities and sarcopenia even at a very advanced age.

Work is presently being done to assess whether adjusting the exercise protocol parameters and/or starting physical exercise at an earlier age would augment the benefit to old skeletal muscles.

Conflict of interest

The authors declare no conflict of interest.

Author contributions

BC participated in the study design and data interpretation, and performed the ultrastructural analyses; MG performed in vitro culture of satellite cells and histochemical analyses at light microscopy; MC participated in sample processing and carried out morphometric evaluations; PF was responsible for animal care and training, and participated in sample collection; CZ performed statistical analyses and participated in drafting the manuscript; CP and MM conceived and co‐ordinated the study, and drafted the manuscript.