|Home | About | Journals | Submit | Contact Us | Français|
Polymyositis and dermatomyositis are orphan, chronic skeletal muscle disorders characterized by weakness, infiltrations by mononuclear inflammatory cells, and fibrosis. Until recently, patients were advised to refrain from physical activity because of fears of exacerbation of muscle inflammation. However, recent studies have shown that moderate exercise training in combination with immunosuppressive drugs can improve muscle performance. Despite the positive effects of exercise training, the molecular mechanisms underlying the exercise-associated clinical improvements remain poorly understood. The present study was designed to define, at the molecular level, the effects of resistance exercise training on muscle performance and disease progression in myositis patients. We evaluated changes in muscle strength, histology and genome-wide mRNA profiles to determine the beneficial effects of exercise and determine the possible molecular changes associated with improved muscle performance. A total of 8 myositis patients underwent a 7-wk resistance exercise training program that resulted in improved muscle strength and increased maximal oxygen uptake (VO2max). Training also resulted in marked reductions in gene expression, reflecting reductions in proinflammatory and profibrotic gene networks, changes that were also accompanied by a reduction in tissue fibrosis. Consistent with the exercise-associated increase in VO2max, a subset of transcripts was associated with a shift toward oxidative metabolism. The changes in gene expression reported in the present study are in agreement with the performance improvements induced by exercise and suggest that resistance exercise training can induce a reduction in inflammation and fibrosis in skeletal muscle.
Polymyositis and dermatomyositis are chronic, autoimmune skeletal muscle disorders characterized by proximal weakness and infiltration of mononuclear inflammatory cells. Current pharmacological treatment is based on high doses of glucocorticoids in combination with other immunosuppressive drugs. Most patients respond with improved muscle performance, but many are left with impaired muscle function and reduced health-related quality of life (1). Several factors could contribute to the sustained muscle impairment despite immunosuppressive treatments. Longitudinal studies of patients with persisting muscle weakness have demonstrated phenotypical changes of muscle tissue, including persisting major histocompatibility complex (MHC) class I expression in muscle fibers and activation markers in endothelial cells of microvessels (2). In some cases, muscle fibrosis develops, indicating repeated cycles of damage and repair. In addition, metabolic impairment occurs, leading to an acquired metabolic myopathy characterized by low levels of adenosine triphosphate (ATP) and phosphocreatine and decreased fatigue resistance (3). All these muscle features are shared by the two subsets of the disease (polymyositis and dermatomyositis). Until recently, patients were advised to refrain from physical activity because of fears of exacerbation of muscle inflammation and disease progression. However, recent studies have shown that moderate exercise in combination with immunosuppressive drugs can improve muscle performance without signs of increased muscle inflammation, suggesting that exercise represents a viable therapeutic intervention for autoimmune myositis patients (4,5). Therefore, understanding the molecular mechanisms underlying the exercise-induced performance improvements could yield important information for the development of novel interventions for autoimmune inflammatory myopathy patients.
The predominating molecules in muscle tissue of polymyositis and dermatomyositis patients with muscle weakness are proinflammatory cytokines and chemokines, as well as profibrotic transforming growth factor (TGF)-β. Both subsets have a similar molecular expression profile. The most consistently expressed cytokines in different phases of both polymyositis and dermatomyositis are interleukin (IL)-1 and the alarmin high-mobility group box chromosomal protein (HMGB)-1 (6–8). These cytokines have been detected in muscle tissue with a higher expression than in healthy individuals in both the early and late chronic phase of the disease, even without detectable inflammatory cell infiltrates. This occurrence suggests a potential role in muscle function impairment, similar to the negative effect of tumor necrosis factor (TNF) on muscle fiber contractility (9). Another mechanism that could lead to muscle weakness in chronic muscle inflammation is infiltration of muscle tissue by fibrosis. When present, muscle fibrosis is characterized by excessive accumulation of collagen and other extracellular matrix (ECM) components. This dynamic process is controlled by a host of processing factors responsible for enzymatic cleavage, assembly, cross-linking, elasticity and turnover of collagen. Fibrosis development involves extensive structural disorganization and remodeling of the ECM, in part owing to the altered release of fibrogenic cytokines such as TGF-β1 (10).
The aim of the present study was to define, in molecular terms, the potential mechanisms underlying the beneficial effects of resistance exercise in autoimmune inflammatory myopathy patients. Consistent with our previous findings, we show that 7 weeks of resistance exercise resulted in increased performance along with the modulations of proinflammatory and profibrotic genes. In addition, several genes associated with enhanced metabolism were also positively modulated in line with the increase in performance.
Eight autoimmune inflammatory myopathy patients (five patients with dermatomyositis and three with polymyositis) (11) participated in a resistance exercise program at the Karolinska University Hospital, Stockholm, Sweden (12). Median age was 51 years (range, 44–61 years), and median disease duration was 4.5 years (range, 2.7–29.0 years). More detailed clinical characteristics have previously been published (12). All patients had been treated with glucocorticoids and other immunosuppressive therapies for >12 months, with some improvement of muscle function but with persisting muscle function impairment. The patients were on stable medication before and during the study and had undetected disease activity for 3 months before the beginning of the resistance exercise training program. Maximal oxygen uptake per kilogram at peak exercise (VO2max) was measured according to standard procedures and was determined as the highest value recorded during the last minutes of the exercise test (13). Disease activity was measured by the clinical outcome measure myositis intention-to-treat activity index (MITAX) and by serum creatinine kinase levels (14). Serum complement 1q (C1q) was measured by kinetic nephelometry and compared with a reference curve from a pool of sera from the blood of healthy individuals. The study was approved by the local ethics committee at the Karolinska University Hospital (Solna, Stockholm, Sweden) and the Children’s National Medical Center Institutional Review Board (Washington, DC, USA). All patients gave informed consent to participate.
The resistance exercise training protocol was described in detail previously (12). Briefly, patients underwent a supervised resistance exercise training regimen with an intensity of 10 voluntary repetition maximum (VRM) 3 d/wk for 7 wks. The training involved five muscles groups (deltoid, quadriceps, latissimus dorsi/biceps, gastrocnemius and trunk muscles) 3 d/wk for 7 wks. The clinical and laboratory assessment as well as functional tests were defined previously (12).
Three to five muscle biopsy samples per patient were obtained from the vastus lateralis muscle using the percutaneous conchotome technique (15). The prebiopsy was taken 1 wk before the initiation of the training period, and the postbiopsy was taken 1 wk after the last exercise training session. The pre-and postbiopsies were taken from opposite limbs to avoid effects of successive biopsies (16) frozen in liquid nitrogen–cooled isopentane and stored at −70°C until further analysis.
Expression profiling was performed for each individual sample using Affymetrix microarrays following standard operating procedures and quality controls as reported (17). Briefly, total RNA was extracted from frozen muscle biopsies using TRIzol (Invitrogen, Carlsbad, CA, USA), cleaned and concentrated using the RNeasy Minielute Kit (Qiagen, Santa Clara, CA, USA) and quantified spectrophotometrically at 260 nm. RNA integrity was verified by agarose gel electrophoresis, and 2 μg was used for cDNA synthesis. The double-stranded cDNA was subjected to a 16-h in vitro transcription at 37°C in the presence of biotin-labeled nucleotides using a one-round amplification strategy. The resulting biotin-labeled cRNA was cleaned, fragmented and hybridized to whole genome Affymetrix U133 plus 2.0 arrays. After overnight hybridizations at 45°C, microarrays were washed and stained on an Affymetrix Fluidics Station 450 and scanned on an Affymetrix GeneChip Scanner 3000 (Qiagen).
All array images were visually inspected for errors and probe set grid alignment. Once the arrays passed, all stringent quality-control measures, including scaling factor <5, present calls >35% and 3′/5′ GAPDH ratios <3, were subjected to probe set absolute-intensity calculations in MAS 5 (Qiagen). In addition, all arrays were subjected to a cross-array analysis to identify any outliers using DChip (18). Statistical analysis was done using GeneSpring GX (Silicon Genetics, Redwood City, CA, USA). Additional stringent quality-control measures were performed by filtering out probe set intensities close to zero and using only probe sets flagged as “present” in half of the samples (50% P calls). Sample normalizations included a “per chip” normalization (50th percentile) and a paired sample normalization in which each subject’s pretraining biopsy served as the control for the posttraining biopsy. Statistical significance was determined using a paired t test with an α level of 0.05 and a 1.5-fold difference in gene expression between pre-and posttraining. All data corresponding to the microarray experiments, including chip files, are publicly available via http://www.cmm.ki.se/gustavo.
Quantitative real-time polymerase chain reaction (qRT-PCR) was performed to validate the expression of selected genes obtained with the microarray platform. Briefly, cDNA was synthesized from total RNA (100 ng) from the same patient population using the cDNA archive kit (Applied Biosystems, Foster City, CA, USA). Labeled primers (FAM) were added to the samples together with 12.5 μL TaqMan Universal PCR Master Mix. Human GAPDH (VIC labeled; ABI, Seattle, WA, USA) was used as an endogenous control. Reactions were performed on the ABI PRISM 7900HT Sequence Detection System, as recommended by manufacturers. PCR amplifications were performed in triplicate, multiplexed with the endogenous control primers. Differences in gene expression were calculated using the ΔΔCt method, which is based on the difference in threshold cycles (ΔCt) between the target gene and endogenous control. The following primers were used: C1S (Hs00156159_m1), COMP (Hs00164359_m1), CCL-14 (Hs00234981_m1), CCR1 (Hs00174298_m1), COL1α (Hs00164004_m1), CTGF (Hs00170014_m1), FGF1 (Hs00265254_m1), IL10Rβ (Hs00175123_m1), IL2Rγ (Hs00173950_m1), IRAK3 (Hs00200502_m1), LTBP1 (Hs00386448_m1), RFX-2 (Hs00172177_m1), and VAV1 (Hs00232108_m1).
To determine functional relationships among the identified genes, we analyzed our dataset with the Ingenuity Pathways Analysis (Ingenuity Systems®, www.ingenuity.com). From the experimental dataset, genes of interest were selected and a functional analysis was carried out. The resulting associations were computed and displayed in a series of interacting networks. The significance, which is a measure for how likely it is that genes from the dataset file participate in that function, is expressed as a P value. This result is calculated using the right-tailed Fisher exact test by comparing the number of user-specified genes of interest (i.e., functional analysis genes) that participate in a given function or pathway, relative to the total number of occurrences of these genes in all functional/pathway annotations stored in the Ingenuity Pathways Knowledge database. Significance in the ingenuity analyses refers to the −log (P value). For example, if the P value for a specific high-level function in one analysis is 1 × 10−10, its significance is 10. The lower the P value for a function, the more significant it is.
Immunohistochemistry was performed as described by Grundtman et al. (19) in seven of eight patients because there was an insufficient amount of tissue in one patient. Primary antibodies used were monoclonal antihuman CD3 (SK7), CD4 (SK3) (both from Becton-Dickinson, San Jose, CA, USA), CD163 (Ber-MAC3, Dako Denmark A/S, Glostrup, Denmark), FOXP3 (236A/E7, eBioscience, San Diego, CA, USA), MHC class I [W6/32(5) Dako Denmark A/S], MHC class II (L243 Becton-Dickinson), IL-1α (1277-89-7) and IL-1β (2D8) (both from Immunokontakt/AMS Biotechnology [Europe], Bioggio [Lugano], Switzerland), and IL-1Ra (10309), IL-1RI (35730) and IL-1RII (832437) (all from R&D Systems, Abingdon, UK). For HMGB-1, a rabbit antibody was used (BD, Pharmingen, San Diego, CA, USA). As secondary antibodies, we used biotinylated F(ab)2 fragmented goat antimouse IgG (Caltag Laboratories, Burlingame, CA, USA) and biotinylated goat antirabbit IgG (Vector Laboratories, Burlingame, CA, USA). As negative controls, we used mouse irrelevant IgG1 (DAK-G01), IgG2 (DAK-G059) and negative control rabbit immunoglobulin fraction, all from Dako Denmark A/S. Coded tissue sections were evaluated by conventional microscopic assessment on whole tissue sections by two independent observers with concordant results. For quantification, we measured specific immunostaining on whole tissue sections by computed image analysis for CD3, CD163, IL-1α and HMGB-1. The results are presented as a percentage of total positively stained tissue area. For MHC class I and class II expression, the percentage of positive fibers was estimated (2). Because of variations between individual tissue samples, a significant change was defined as at least doubling or reduction to half expression between the two biopsies (20).
Histochemical analysis was carried out on coded frozen skeletal muscle tissue sections stained by Gomori’s trichrome procedure (Sigma, St. Louis, MO, USA). Digital images were obtained using light microscopy (objective 40×) and computerized image capture (Olympus C.A.S.T. Stereology System, Olympus America, Center Valley, PA, USA). Samples were analyzed randomly to avoid any bias when determining fibrosis. Connective tissue area was expressed as a ratio of the number of counted dots divided by the total number of dots on the overlay (21). In addition, collagen I content was assessed by Western blotting using an anticollagen I antibody (1:1,000, Abcam). Proteins were homogenized in radioimmunoprecipitation assay (RIPA) buffer containing 1% NP-40 and 10× of each protease and phosphatase inhibitor (Pierce, Rockford, IL, USA), separated in 10% resolving gels and transferred to nitrocellulose membranes. Protein bands were visualized by enhanced chemiluminescence (Amersham, Buckinghamshire, UK) and quantified densitometrically with Quest software (Bio-Rad, Hercules, CA, USA). The band intensities corresponding to collagen I (~95 kDa) were normalized to the intensity of the corresponding lanes in the membranes visualized after Indian ink staining (loading control).
StatView (Abacus Concepts, Berkeley, CA, USA) was used for statistical analyses of clinical, immunohistochemistry and serum data. Because of non-normal distribution of immunological markers in muscle biopsies, nonparametric methods were used. The Wilcoxon signed-rank procedure tested for within-group changes. A parametric method was applied (Student t test for paired observations) to establish the level of statistical significance in VO2max changes. Changes in collagen I content were calculated from the Western blot band intensities using a paired t test with significance established at P < 0.05. The Pearson moment correlation was calculated in Microsoft Excel.
Muscle strength increased, serum levels of creatinine kinase decreased and clinical disease activity score (MITAX) improved as previously reported (12). Resistance exercise also resulted in decreased median serum levels of C1q from 141 mg/L (range, 80–330 mg/L) to 69 mg/L (range, 42–147 mg/L) (P = 0.01) (reference value 70–300 mg/L). A significant increase was demonstrated in maximal absolute and relative oxygen uptake (VO2max) from 1.88 ± 0.3 to 2.20 ± 0.3 L × min−1 (P < 0.001) and from 26 ± 3 to 31 ± 3 mL × min−1 × kg−1 (P < 0.001), respectively. VO2max values are means ± SD.
The 7-wk training program resulted in the modulation of 265 transcripts. The predominant enriched ontologies associated with these transcripts were inflammation, fibrosis and metabolism. When these genes were cross-validated against a disease database during the ingenuity pathways analysis, we found that the majority are involved in diseases characterized by inflammation and fibrosis. Ontological analyses of selected genes revealed that 41 transcripts (15.5%) were associated with inflammatory processes, 25 transcripts (9.4%) with fibrotic processes and 7 transcripts (2.6%) with metabolic regulation. We focused on these enriched categories because they best represent the pathogenesis of autoimmune myositis (Table 1). As an initial validation of our microarray results, key representative target genes were analyzed by qRT-PCR. There was a good correlation (R2 = 0.62) between the changes in gene expression detected by microarray and qRT-PCR with the changes showing both qualitative and quantitative similarity (Figure 1).
In total, 41 genes involved in inflammation changed significantly, and 34 of these were downregulated (Table 1). The genes with reduced expression could be categorized into those that are involved in T-cell activation and regulation, such as chemokine (C-X3C motif) receptor 1 (CCR1), CD44 antigen, IL-2Rγ, MHC–F and dedicator of cytokinesis 2 (DOCK2), or those involved in macrophage/monocyte activation, such as colony stimulating factor 2 receptor β (CSF2RB), VAV1 oncogene (VAV1), interleukin receptor-associated kinase 3 (IRAK3) and lipopolysaccharide-induced TNF factor (LITAF). The gene for the proinflammatory cytokine HMGB-1 was significantly downregulated (P < 0.03), but the change did not pass the 1.5 cutoff fold change (−1.27). Other genes involved in inflammation that were down-regulated were clusterin (CLU), neurofilament light polypeptide (NEFL), CCR1, and prostaglandin-endoperoxide synthase 1 (PTGS1). Additionally, a few antiinflammatory genes were upregulated, as was FOXP3, a marker of regulatory T-cells. Additionally, our bioinformatic analysis revealed that several genes regulated by TNFα were negatively modulated, suggesting that they may play a role in the antiinflammatory effects of resistance exercise in polymyositis/dermatomyositis (Figure 2A). A hypothetical model of this regulatory gene network that may mediate the effects of exercise on local tissue inflammation is presented in Figure 2B.
Twenty-five genes related to collagen synthesis changed significantly (Table 1). The majority of these profibrotic genes displayed decreased expression (22 of 25). Increased expression was found in the latent TGF-β1 binding protein 1 (LTBP-1), which is a negative regulator of TGF-β (for example, an antifibrotic gene). Consistent with the decrease in tissue fibrosis genes, collagen gene expression was also reduced (COL1A1, COL14A1, COL5A2) (Table 1). The functionally relevant TGF-β receptors 1 and 2 (TGFBR 1 and 2) did not pass the fold-change cutoff (1.36-and 1.15-fold) but did reach statistical significance (P < 0.02 and P < 0.04, respectively). Again, our bioinformatic analysis indicated that several genes regulated by TGF-β were downregulated, and a negative modulator of TGF-β (LTBP-1) was induced by resistance exercise (Figure 3A). A hypothetical model of this regulatory gene network leading to decreased fibrosis is displayed in Figure 3B. Consistent with this model, a marked decrease in connective tissue (nonmuscle) area was observed (Figure 4A, B), which reflected the significant 24.7 ± 8.2% decrease in collagen I content (Figure 4C, D).
Resistance exercise resulted in the modulation of genes involved in energy metabolism (Table 1). This is reflected by a reduction in genes involved in lipid biosynthesis, for example, acetyl-coenzyme A carboxylase α (ACACA) and an increased expression of the mitochondrial genes, for example, ATPase Na+/K+ transporter (ATP1B2).
A low degree of inflammation with scattered T-cells and macrophages was found in muscle biopsies in six of seven patients, and this was unchanged after the resistance exercise period. One patient who had several inflammatory cell infiltrates displayed a >100% reduction in expression of CD3+ T-cells and CD163+ macrophages by computerized image analysis and also a decreased extranuclear HMGB-1 expression (data not shown). Proinflammatory cytokines and receptors, previously detected in muscle tissue of myositis patients (IL-1α, IL-1β, IL-1 receptor [R] I and IL-1-RII), were present in most biopsies with a low expression and were unchanged after training. Generally, <5% of the fibers expressed MHC class I and II, and there was no statistically significant change after the resistance exercise period.
Exercise training has beneficial effects on systemic inflammation, as determined by a reduction in selective serum markers such as IL-6, C-reactive protein and TNF in healthy individuals (22). However, little is known about the effects of exercise training on inflammation in patients suffering from chronic inflammatory disorders. In the present study, we demonstrate that 7 weeks of resistance exercise modulated the expression of genes involved in inflammation, fibrosis and metabolism in muscle from polymyositis/dermatomyositis patients. These postexercise changes in gene expression may explain the improved muscle performance observed in our patients, and seem to agree with the conclusion that resistance exercise can improve muscle performance in polymyositis/dermatomyositis patients without disease exacerbation by positively modulating genes involved in the disease process.
The modified expression pattern of genes involved in inflammation suggests that resistance exercise induced a coordinated reduction of proinflammatory transcripts and an increase in antiinflammatory transcripts. For example, PTGS1, previously found to be chronically upregulated in myositis and refractory to conventional immunosuppressive treatment (23), was reduced after training, and the upregulation of FOXP3 was consistent with the down-regulation of SMAD7-dependent signaling, suggesting a coordinated modulation of T-cell activity (24). Similarly, the upregulation of NPY2R, a negative effector of monocytes (25), further supports a resistance exercise–mediated antiinflammatory effect. These findings suggest that resistance exercise can improve muscle function in myositis patients by modulating inflammation-associated gene expression in skeletal muscle and seem to agree with previous reports from chronic heart failure and frail elderly men and women, where exercise training resulted in a significant reduction in the local expression of inflammatory molecules such as TNFα (26,27).
Although a reduction in TNFα mRNA was not detectable in our patients, the interpretation of a TNF-related gene network as the target for the local immunosuppressive effects of resistance exercise is consistent with a role in the overall decrease in disease activity after training (Figure 2B). In the present study, serum C1q levels were reduced after resistance exercise and likely modulated by the reduced expression of C1S mRNA. Altogether, our findings support the interpretation that resistance exercise can reduce the expression of proinflammatory genes in autoimmune myositis patients. The lack of a significant decrease at the protein level of inflammatory molecules may be explained by the low expression before exercise and the low sensitivity of immunohistochemistry for quantitative measures. Another possibility is that the observed changes in mRNA levels do not reflect changes in protein synthesis.
The second largest group of genes that was changed with resistance exercise was genes involved in fibrosis. Muscle fibrosis may over time expand and increase relative to the contractile tissue resulting in weakness and decreased performance. In the present study, 7 weeks of resistance exercise was sufficient to produce a significant reduction of tissue fibrosis, as determined by gross morphological examination and the specific decrease of collagen I levels. Consistent with this finding, posttraining analysis of gene expression revealed that resistance exercise also suppressed the expression of a profibrogenic molecular signature consisting of genes involved in collagen synthesis, processing and extracellular matrix assembly and deposition (network depicted in Figure 3). Our bioinformatics analysis identified that a subset of transcripts modulated by resistance exercise is associated with TGF-β signaling (Figure 3A). Furthermore, the increased expression of LTBP-1, a negative modulator of TGF-β (28), supports the interpretation that down-regulation of a TGF-β–associated gene network may mediate the effects of resistance exercise by decreasing muscle fibrosis, not only by suppressing the synthesis of collagen and other ECM components, but also by increasing collagenolysis and elastolysis by matrix metalloproteinases, such as matrix metallopeptidase 16 (MMP16), and other collagen-cleaving enzymes, such as prolyl endopeptidase (PREP) (Figure 3B).
Noticeably, resistance exercise exerted a modulatory effect on both networks, but the precise mechanisms by which this occurred, or whether they can modulate one another in this particular condition, remain to be elucidated. For example, depletion of T lymphocytes in mdx mice resulted in a reduction in muscle fibrosis, suggesting a direct link between muscle inflammation and the aberrant accumulation of ECM (29).
Finally, we found that resistance exercise improved the metabolic profile of myositis muscle. Our training protocol resulted in a shift in the muscle’s metabolic gene profile toward a more oxidative state. Both AK3 and ATP1B2 were reduced after training, suggesting a decreased reliance on immediate energy production mechanisms (30). Conversely, the increased expression of transcripts coding for respiratory enzymes UCP3 (a mitochondrial proton carrier) and UQCRC2 and the decreased ACACA expression suggest that training may have stimulated the activity of oxidative metabolism and decreased lipid biosynthesis, respectively. The increase in oxidative gene expression is in line with the observed increase in VO2max. A similar effect on VO2max was recently reported after strength training in elderly men (31).
A potential limitation of the study is the absence of a nontrained control group. However, we feel that the repeated muscle biopsy design, where each individual served as its own control, compensated for the absence of a separate control group and was better suited for the microarray analysis, since this design reduces the effects of genetic heterogeneity on gene expression (32). Repeat muscle biopsies also compensated for the low subject number and, together with the high stringency microarray analysis, contributed to a final reliable dataset from which some selected genes could be confirmed by qRT-PCR and could functionally begin to explain the effects of resistance exercise on both inflammation and fibrosis outcomes. Furthermore, we did not find any differences in gene expression profiles between polymyositis and dermatomyositis patients, which is in agreement with previously published data on muscle cytokines and chemokines despite differences in inflammatory cell infiltrates. This finding could also suggest that, in the chronic phase of disease with a low degree of inflammation, shared disease mechanisms could contribute to the impaired performance between the myositis subsets.
In conclusion, resistance exercise may restore muscle function in autoimmune myositis patients by reducing inflammation and tissue fibrosis and by improving metabolic homeostasis, events that were reflected by the corresponding molecular signatures represented by the reduced expression of proinflammatory and profibrotic gene networks and by the increased expression of oxidative metabolism genes. In addition, our results demonstrate that high-throughput analysis of skeletal muscle gene expression can provide useful information for the identification of novel disease bio-markers and potential targets for pharmacological interventions of autoimmune myositis patients.
We would like to thank Eva Lindroos and Karin Lundin for support with immunohistochemical analyses and Christina Ottosson for assistance with the muscle biopsies.
This work was supported by the Swedish Research Council (03642) K2005-74X-14045-05AK, the Myositis Association, the Swedish Rheumatism Association, King Gustaf V 80-year foundation, Karolinska Institutet Foundation, Stockholm County Council, through the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet, AutoCure, the EU FP6 project (contract number LSHB-CT-2006-018661), National Institutes of Health grants NS-029525 and NIH RO1-AR050478 (to K Nagaraju) and National Center for Medical Rehabilitation Research (NCMRR) (grant 5R24HD050846), the Mental Retardation Developmental Disabilities Research Center (MRDDRC) at Children’s National Medical Center (grant 1U54HD053177), and NIH grant 3R01-NS29525-13 to EP Hoffman. The Swedish Center for Sports Research (CIF) supported GA Nader.
The funding organizations had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
The authors declare that they have no competing interests as defined by Molecular Medicine, or other interests that might be perceived to influence the results and discussion reported in this paper.
Online address: http://www.molmed.org