Reduced myotube diameter, atrophic signalling and elevated oxidative stress in cultured satellite cells from COPD patients

The mechanisms leading to skeletal limb muscle dysfunction in chronic obstructive pulmonary disease (COPD) have not been fully elucidated. Exhausted muscle regenerative capacity of satellite cells has been evocated, but the capacity of satellite cells to proliferate and differentiate properly remains unknown. Our objectives were to compare the characteristics of satellite cells derived from COPD patients and healthy individuals, in terms of proliferative and differentiation capacities, morphological phenotype and atrophy/hypertrophy signalling, and oxidative stress status. Therefore, we purified and cultivated satellite cells from progressively frozen vastus lateralis biopsies of eight COPD patients and eight healthy individuals. We examined proliferation parameters, differentiation capacities, myotube diameter, expression of atrophy/hypertrophy markers, oxidative stress damages, antioxidant enzyme expression and cell susceptibility to H2O2 in cultured myoblasts and/or myotubes. Proliferation characteristics and commitment to terminal differentiation were similar in COPD patients and healthy individuals, despite impaired fusion capacities of COPD myotubes. Myotube diameter was smaller in COPD patients (P = 0.015), and was associated with a higher expression of myostatin (myoblasts: P = 0.083; myotubes: P = 0.050) and atrogin-1 (myoblasts: P = 0.050), and a decreased phospho-AKT/AKT ratio (myoblasts: P = 0.022). Protein carbonylation (myoblasts: P = 0.028; myotubes: P = 0.002) and lipid peroxidation (myotubes: P = 0.065) were higher in COPD cells, and COPD myoblasts were significantly more susceptible to oxidative stress. Thus, cultured satellite cells from COPD patients display characteristics of morphology, atrophic signalling and oxidative stress similar to those described in in vivo COPD skeletal limb muscles. We have therefore demonstrated that muscle alteration in COPD can be studied by classical in vitro cellular models.


Introduction
Dysfunction and atrophy of the skeletal limb muscles are now recognized as important extrapulmonary manifestations of chronic obstructive pulmonary disease (COPD), contributing to impaired patient exercise tolerance, worsened quality of life and reduced survival [1][2][3]. Furthermore, the altered muscle strength and endurance, and the muscle atrophy, are likely due to a combination of different mechanisms, with oxidative stress being one of the most important [4][5][6][7][8][9].
An impaired capacity for muscle regeneration has also been hypothesized to explain COPD muscle atrophy [10,11]. In addition, skeletal muscle repair mechanisms seem to be altered in COPD patients showing abnormal muscle structure [8]. As satellite cells are the primary contributors to muscle tissue homoeostasis, muscle regeneration during exercise and injury, and muscle repair over the long-term [12], several groups have compared their abundance in the skeletal muscles of COPD patients and healthy individuals, and they consistently found no difference [11,13,14]. However, the number of satellite cells provides no information on proliferation and differentiation capacities or redox status, therefore, the hypothesis of a compromised maintenance of muscle mass and exhausted muscle regenerative capacity of satellite cells [11] has to be assessed. Indeed, the intrinsic capacity of satellite cells to replicate and adopt myogenic development in COPD remains unknown [11].
Primary human satellite cell culture is now a well-developed approach. It has been widely used in studies of myogenesis, muscle regenerative capacity, myotube morphology alterations, signalling pathways, and the role of oxidative stress under physiological and pathological conditions, this last including both pathologies of genetic origin, like muscle dystrophy [15,16], and acquired muscle dysfunction, like type 2 diabetes and insulin resistance [17,18]. Interestingly, the myotubes obtained from satellite cell culture in non-genetic diseases conserve some of the molecular and morphological characteristics seen in vivo in patient muscles, and they may thus be a useful model for studying muscle dysfunction mechanisms [17,18].
The aim of this study was thus to determine whether cultured satellite cells derived from skeletal limb muscles of COPD patients are altered in terms of proliferative and differentiation capacities, morphological phenotype and atrophy/hypertrophy signalling, and redox status in comparison with cells from healthy individuals.

Study population
Sedentary healthy individuals were recruited on the basis of the following criteria: age from 57 to 67.5 years, no disease and less than 150 min. of moderate-to-vigorous physical activity per week. COPD patients were defined on the basis of the following criteria: dyspnea, and/or chronic cough or sputum production, and/or history of exposure to risk factors for the disease, with the diagnosis confirmed by spirometry (post-bronchodilatator FEV 1 /FVC<70%; FEV 1 : forced expiratory volume in 1 sec.; FVC: forced vital capacity) [19]. Exclusion criteria were: other respiratory diagnosis, decompensated co-morbidity, and exacer-bation in the last 2 months. Functional tests are detailed in the Data S1.

Muscle biopsy procedures and conservation
Muscle biopsies were performed in the vastus lateralis of the quadriceps using the usual methodology [20]. One piece of the fresh biopsy was flash frozen in a pre-cooled beaker of isopentane placed in liquid nitrogen, to avoid distortion of the tissue, and lastly conserved at À80°C. Cryosections of this biopsy specimen served to assess muscle fibre cross-sectional area (CSA) by immunohistochemistry, using an anti-dystrophin antibody. Another piece of the fresh biopsy was placed in a cryogenic tube and was then progressively frozen to À80°C for 24 hrs using a Mr. Frosty freezing container (Nalgene Fisher Scientific, Pittsburgh, PA, USA), to preserve cell integrity. The cryogenic tube was then stored in liquid nitrogen until use for myoblast isolation.
Satellite cells were then purified following a 30-min. incubation with an anti-CD56 (NCAM) antibody (BD Biosciences) [21], using an immunomagnetic sorting system (Miltenyi Biotec, Bergisch Gladbach, Germany). Purified myoblasts (passage 1: P1) were then grown in a 100-mm collagen-coated Petri dish in proliferation medium. The purity of the 16 myoblast cultures (eight COPD and eight healthy individuals) was evaluated after immunostaining with an anti-desmin antibody and Hoechst 33258, followed by fluorescence microscopy (see the Data S1). Data analysis of more than 200 cells per culture showed a high and comparable purity of the myoblast cultures derived from healthy individuals and patients [99.8% (97.8-100) versus 99.7% (98.9-100); P = 0.721]. Myoblasts were always used at a passage below P4 for the experiments.
When myoblasts reached 80% confluence, myogenic differentiation was induced by changing the proliferation medium to DMEM/2% FBS (differentiation medium). Myotubes were obtained after 6 days in differentiation medium.
Myoblast and myotube characterization, oxidative stress assessment, antibodies and reagents, quantitative polymerase chain reaction (qPCR) and primers Full details are given in the Data S1.

Statistical analysis
Variables were compared between COPD and control groups using the Student's t-test or the Mann-Whitney test to account for non-parametric data distribution, and data are presented as median (25th percentile-75th percentile), except for the H 2 O 2 -induced oxidative stress experiment (Fig. 8), where data are presented as the means AE standard errors (SEM). Statistical analyses were performed with SigmaStat. Significance is at P ≤ 0.05.

Characteristics of the study groups
The clinical and functional characteristics of the study groups are presented in Table 1. The median predicted FEV 1 value indicated severely impaired lung function and the BODE index [22] indicated moderate-to-severe COPD clinical states. Both 6-minute walking distance (6MWD) and quadriceps muscle voluntary contraction (MVC) values indicated significant exercise limitation and muscle dysfunction in the COPD group. Although the included COPD patients were not selected on a specific phenotype, our patient group reflects a COPD population with a significant impaired clinical state. Fibre CSA tended to be lower in the eight patients compared to the eight healthy individuals (Table 1; P = 0.14). However, our study groups were extracted from larger and gendermatched populations of COPD patients (n = 37, 31 males/6 females) and healthy individuals (n = 14, 12 males/2 females), in which the fibre CSA was significantly lower in COPD patients ver-sus healthy individuals [4588 lm 2 (3022-5731) versus 5463 lm 2 (4630-6453); P = 0.026], and was close to our present working groups [ Table 1; 4091 lm 2 (3090-5178) versus 5671 lm 2 (4789-6618); P = 0.14].

COPD myotubes have a normal commitment to terminal differentiation despite impaired fusion capacities
After determining the myoblast characteristics, we evaluated the differentiation abilities of cultured healthy individual and COPD myoblasts placed in differentiation conditions. Figure Fig. 2E] are significantly reduced in COPD myotubes compared to healthy individual myotubes, suggesting that myotube fusion is impaired in COPD muscle cells. Study of the expression of myogenesis markers reveals that MyoD, Myf5 and myogenin are similarly expressed in myoblasts and myotubes from healthy individuals and COPD patients ( Table 2 and Fig. S1A-G). Furthermore, the expression levels of the two late differentiation markers, myosin heavy chain 1 (MHC1) and myosin heavy chain 2 (MHC2), were assessed in healthy individual and COPD myotubes. As seen in Figure 2F

COPD myotubes have a reduced diameter
The diameter of the troponin T-labelled myotubes was then measured for each culture. Two representative healthy individual and two representative COPD myotube cultures observed by fluorescence microscopy are shown in Figure 3A. Analysis of the healthy individual and COPD myotube cultures (Fig. 3B) revealed that the median myotube diameter was significantly lower for COPD patients than for healthy individuals [21.6 lm (20.7-34.7) versus 41.1 lm (34.9-76.5); P = 0.015], suggesting that in vitro myotubes derived from COPD patients have an altered morphology. Figure 3C showed the significant correlation (r = 0.594; P = 0.024) between the myotube diameter of the in vitro cell cultures and the quadriceps fibre CSA of the healthy individuals and patients included in this study. We also observed a significant correlation (r = 0.649; P = 0.016) between the diameter of the cultured myotubes and the MVC values obtained for all healthy individuals and patients (Fig. 3D). Furthermore, significant correlations were also observed when only the COPD patient group was considered, between the in vitro myotube diameter and fibre CSA (r = 0.855; P = 0.030), as well as MVC (r = 0.899; P = 0.017; Fig. S2A and B, respectively).  Fig. 4D] in COPD myoblasts compared to healthy individual myoblasts. These results suggest that protein synthesis is decreased and protein breakdown is enhanced in COPD muscle cells in culture. The expression of various other markers was studied, but their expression levels did not show any significant variation between COPD and healthy individual myoblasts and myotubes ( Table 2).

Oxidative stress in cultured COPD myoblasts and myotubes
Oxidative stress damage was assessed in the cultured COPD myoblasts and myotubes. Protein carbonylation was significantly more We next studied the expression of four major antioxidant proteins in the cultured COPD myoblasts and myotubes. As seen in We also examined the susceptibility of the cultured COPD myoblasts to an induced oxidative stress by exposing the cells to increases in the concentration of H 2 O 2 . Figure 8 shows that the mortality rate for the COPD myoblasts was significantly higher than for the healthy individual myoblasts at H 2 O 2 concentrations from 100 to 500 lM, with almost 100% mortality at concentrations greater than 600 lM for both study groups. Furthermore, the H 2 O 2 concentration necessary to produce a 50% cell death rate was 392 AE 33 lM for healthy individual myoblasts compared with 148 AE 28 lM for COPD myoblasts (P < 0.001).

Discussion
The major finding of this study is that myoblasts and myotubes obtained from cultured satellite cells derived from the skeletal muscle of COPD patients are altered compared with cells from healthy individuals. Although the COPD myoblasts exhibited growth capacities similar to those of healthy individual cells and the COPD myotubes had a normal commitment to terminal differentiation, we observed that: (i) COPD myotubes had impaired fusion capacities, (ii) the cultured COPD myotubes showed significant reduced diameter compared with healthy individual myotubes, (iii) COPD myoblasts and myotubes showed decreased protein synthesis associated with increased protein breakdown, (iv) protein oxidation and lipid peroxidation were more elevated in myoblasts and myotubes from COPD patients, and (v) the COPD myoblasts were more susceptible to oxidative stress than healthy individual myoblasts. Together, our data indicate that in vitro myoblasts and/or myotubes derived from COPD patients display characteristics of reduced diameter, atrophic signalling and elevated oxidative stress similar to those described in in vivo skeletal limb muscles of COPD patients.
Cultured myotubes derived from human satellite cells have been shown to display morphological and biochemical characteristics similar to those of in vivo human skeletal muscles, under both physiological [23] and pathological conditions like the insulin resistance of type 2 diabetes [17,18]. For this reason, cultured human satellite cells have been successfully used as a cellular model to study muscle regeneration during ageing [24], the muscle biochemical characteristics in type 2 diabetes [25,26], and the susceptibility of muscle to oxidative stress and muscle differentiation in facioscapulohumeral dystrophy [15,16]. We show here that a single progressively frozen muscle biopsy from a COPD patient gave access to millions of purified myoblasts that can be expanded and that retained the capacity to differentiate into myotubes, allowing us to carry out multiple cellular and biochemical studies starting with minimal in vivo samples.
The myoblast and myotube cultures demonstrated that proliferation characteristics and commitment to terminal differentiation were  Table 2). Our in vitro findings are therefore in accordance with some in vivo data showing that no major morphological abnormalities are present in COPD muscle biopsies, in terms of central nuclei, fibre splitting, regenerating fibres and apoptosis, despite the significant atrophy of muscle fibres in these patients [27]. Furthermore, muscle regenerative capacity, as reflected by the number of satellite cells per muscle fibre, is not altered in patients with COPD [13].
Studies using computed tomography have demonstrated that in vivo thigh muscle CSA is reduced in COPD patients [28] and that midthigh muscle CSA is a good predictor of mortality in these patients [2]. In addition, this reduced muscle CSA may explain the reduced quadriceps strength in a population of healthy individuals and COPD patients combined [28]. One of the most interesting findings of our study is the significant reduced myotube diameter observed in cultured cells derived from COPD patients (Fig. 3). Moreover, we observed a correlation between the in vitro myotube diameter and both in vivo quadriceps fibre CSA and in vivo muscle strength (Fig. 3). We also showed that the reduced COPD myotube diameter could result from two mechanisms. First, COPD myotubes have a reduced number of nuclei per myotubes (Fig. 2), suggesting impaired fusion capacities that would result in thinner myotubes. Secondly, we observed an increased expression of the muscle growth inhibitor myostatin and of the muscle-specific ubiquitin E3 ligase atrogin-1 (Fig. 4), showing that atrophic signalling pathways are activated in cultured COPD muscle cells. In parallel, the protein synthesis pathway is repressed in COPD cells as observed by the reduced P-AKT/AKT ratio (Fig. 4). Interestingly, it has been demonstrated that myostatin plays a central role in muscle wasting as it activates myotube atrophy through negative regulation of AKT signalling [29] and positive modulation of the atrogin-1-dependent proteasome pathway [29,30]. The in vitro reduced COPD myotube diameter could therefore result from a combination between impaired myoblast fusion, a mechanism that has not been evocated in the COPD literature yet and that could be a novel pathway to explore, and increased atrophic signalling, a pathway that has been reported in the limb muscles of COPD patients by several authors [31][32][33]. Our in vitro data are therefore in accordance with what is observed in COPD patients, which suggests that the cellular model could be used to study the molecular mechanisms involved in COPD muscle atrophic remodelling.
Elevated oxidative stress, as indicated by increased levels of protein carbonylation and lipid peroxidation, was observed in the cultured myoblasts and myotubes derived from COPD patients (Figs 5 and 6). Under these conditions, constant or higher expression levels of antioxidant enzymes (Fig. 7) suggest that the elevated oxidative stress in cultured COPD myoblasts and myotubes cannot be fully overcome by the antioxidant defence mechanisms present in COPD muscle cells. These data are therefore in accordance with the increased susceptibility to oxidative stress we observed in the cultured myoblasts (Fig. 8).
In various studies, similar high oxidative stress has been demonstrated in human biopsies, as indicated by increased lipid/protein oxidative damage [4][5][6]8] and constant or higher expression levels of antioxidant enzymes [4,8] in the skeletal limb muscles of COPD patients. Furthermore, in this present work, our study groups were extracted from a larger population in which we have observed significant higher levels of protein carbonylation in the quadriceps of COPD Interestingly, in this study, the satellite cells from COPD patients conserved pathological characteristics, such as elevated intrinsic oxidative stress, even when they were taken out of their physiological context and placed in in vitro culture conditions. Different hypotheses can be proposed to explain this mechanism. First, a genetic defect in the satellite cells of COPD patients might be the cause, even though various single nucleotide polymorphism studies have only shown a restricted association with COPD status [34][35][36]. Second, mitochondria from COPD skeletal muscle show significant dysfunction associated with elevated levels of ROS [37], and a decrease in mitochondrial DNA content is observed in the skeletal muscle of COPD patients following exercise [38]. One might thus assume that mitochondrial dysfunction and elevated ROS in in vivo COPD muscles also affect satellite cells, which conserve their pathological characteristics when placed in in vitro conditions. Last, epigenetics can result in inheritable changes in cell phenotype in response to environmental factors through the methylation of DNA, and it has been recently demonstrated that ROS can modulate the expression of several genes by DNA methylation [39,40]. We can therefore speculate that gene expression may be altered in vivo in the muscle satellite cells of COPD patients by ROS-induced DNA methylation and that this epigenetic modulation could be transmitted to in vitro satellite cells. Thus, the in vitro cellular model developed in this study should allow us to study these different hypotheses.
In summary, we demonstrated that cultured satellite cells derived from skeletal limb muscles of COPD patients have a proliferative capacity and a commitment to terminal differentiation similar to those of cells from healthy individuals. We also showed that in vitro myotubes from COPD patients have a reduced diameter associated with an increased atrophic signalling, and that cultured myoblasts and myotu-bes from these patients display elevated oxidative stress. Thus, in vitro myoblasts and myotubes derived from COPD satellite cells exhibit characteristics of morphology, atrophy and oxidative stress similar to those of in vivo quadriceps muscles from COPD patients. We propose that this in vitro model provides a promising basis for research into COPD muscle alteration, which is a key component of muscle dysfunction and atrophy in patients.