Effects of a two-year health-enhancing physical activity program on skeletal muscle protein proles in people with rheumatoid arthritis

Rheumatoid arthritis (RA) induces loss of muscle mass and impairments, which discourage physical activity. Consequently, RA is associated with decreased levels of exercise. Positive effects of exercise on physical and mental health without increasing pain or disease activity have been demonstrated but the inuence of moderate-intensity physical activity on the molecular effects in skeletal muscle in people with RA remains unclear. In this pilot study we investigated the impact of a two-year health-enhancing physical activity (HEPA) program on protein proles in skeletal muscle biopsies from people with RA based on a unique set of interdisciplinary variables. Six people with RA participated. Muscle biopsies were taken from m. vastus lateralis, and perceived disease impact, anthropometrics and physical capacity tests were assessed at baseline and year 1 and 2. Mass spectrometry-based quantitative proteomics and immunohistochemistry analysis were employed to examine molecular changes in the skeletal muscle in response to long-term HEPA. The study participants had moderate to low adherence to the HEPA program, the perceived disease impact and anthropometric parameters were not altered. Among the physical capacity tests, the Timed-stands test showed a signicant improvement after one year. Mean pressure pain threshold was signicantly reduced after 1- and 2-year HEPA participation. Proteome analysis of the muscle biopsies showed an up-regulation of proteins in the mitochondrial respiratory chain and predicted increased ATP production and improved muscle contractility in response to HEPA. Molecular pathways identied by proteome analysis revealed correlations to pain variables. current subsample of a two-year intervention with HEPA aimed to promote HEPA in sedentary individuals with RA, we applied unbiased proteomics analysis of repeated muscle biopsies before and after one and two years. Although the physical activity was modest and decreased between the rst and second year, the effects on muscle proteome, e.g. on proteins involved in the mitochondrial respiratory chain and oxidative phosphorylation, were consistent suggesting a positive metabolic effect in muscle tissue of the long-term HEPA program. Similar effects with enrichment of proteins in the mitochondrial compartment, which include the electron transport chain have been reported from short-term training programs in men . The respiratory chain is fundamental for ATP production that is necessary to meet the energetic requirements of the muscle cells 30 . In healthy subjects who participated in those studies the muscle homeostasis improved and their capacity to supply ATP increased. Our ndings suggest that muscles in people with RA are capable of improving their mitochondrial function in a similar fashion as seen in healthy young men. In myositis, where the main pathological processes occur in the skeletal muscle, mRNA expression patterns after endurance training demonstrated up-regulation of genes involved in capillary growth, which has been associated with increased capillary density in muscle tissue 10 . We did not nd a signicant increase in the capillary density in the current study, which could be explained by the low number of participants in our study, although there was a tendency of higher number of CD31 positive endothelial cell per area after two years of HEPA. Our MS-based proteomics results acquired from muscle tissue proteome demonstrate a signicant up-regulation of several proteins in the mitochondrial respiratory chain: after one year of HEPA, four out of eight (COX7C, COX8A, NDUFS7, IDH3B) and after two years eight out of 21 (CX6B1, NDUFA4, CYC1, NDUA9, NDUA7, ATPA, NDUA2, ACAA2). This data corroborates predicted increased levels of ATP (p = 1,10E-07 and p = 1,20E-07 for year one and two HEPA, respectively), which synthesis requires augmented electron transport to create a longer lasting electrochemical gradient 31 . On a protein level we observed a 17% up-regulation of ATP synthase alpha (p < 0,05), and at the same time, in ve participants out of six a decreased level of IF1 was detected. IF1 inhibits the hydrolase activity of ATP synthase and increased expression of IF1 in tumours promotes the metabolic switch in cancer cells limiting the level of oxidative phosphorylation and promotes glycolysis 32 . In human skeletal muscle of obese patients with type 2 diabetes, respiratory chain predicting increased ATP production and muscle contractility. In addition, results indicated that HEPA applied to people with RA may have positive impacts on muscle tissues by decreasing muscle cell death.


Abstract
Background Rheumatoid arthritis (RA) induces loss of muscle mass and impairments, which discourage physical activity. Consequently, RA is associated with decreased levels of exercise. Positive effects of exercise on physical and mental health without increasing pain or disease activity have been demonstrated but the in uence of moderate-intensity physical activity on the molecular effects in skeletal muscle in people with RA remains unclear. In this pilot study we investigated the impact of a two-year health-enhancing physical activity (HEPA) program on protein pro les in skeletal muscle biopsies from people with RA based on a unique set of interdisciplinary variables.

Methods
Six people with RA participated. Muscle biopsies were taken from m. vastus lateralis, and perceived disease impact, anthropometrics and physical capacity tests were assessed at baseline and year 1 and 2. Mass spectrometry-based quantitative proteomics and immunohistochemistry analysis were employed to examine molecular changes in the skeletal muscle in response to long-term HEPA.

Results
The study participants had moderate to low adherence to the HEPA program, the perceived disease impact and anthropometric parameters were not altered. Among the physical capacity tests, the Timed-stands test showed a signi cant improvement after one year. Mean pressure pain threshold was signi cantly reduced after 1-and 2-year HEPA participation. Proteome analysis of the muscle biopsies showed an up-regulation of proteins in the mitochondrial respiratory chain and predicted increased ATP production and improved muscle contractility in response to HEPA. Molecular pathways identi ed by proteome analysis revealed correlations to pain variables.

Conclusions
Our results indicate that long-term physical activity in people with RA may up-regulate proteins in the mitochondrial respiratory chain predicting increased ATP production and muscle contractility similar to what has been reported in healthy young men and might have a positive impact on muscle tissue by decreased muscle cell death.

Background
Planned and structured exercise has long been recommended to people with rheumatoid arthritis (RA) in order to prevent or reduce disability that may partly be attributed to a vicious circle of systemic and local in ammation, muscle wasting and reduced levels of physical activity 1 . More recently, the value of exercise to reduce risk factors for cardiovascular disease that are associated with RA has been recognized [2][3] . Public health recommendations for health-enhancing physical activity (HEPA) to reduce cardiovascular disease risk include at least 30 minutes of daily, at least moderate-intensity, physical activity and, in addition, twice weekly strength training [4][5] . HEPA in an everyday environment has also proven to be bene cial in people with RA 6-7 . However, whether HEPA affects molecular pathways in muscle tissue behind systemic effects to the same extent as supervised high-intensity exercise is largely unknown.
A recent study revealed broad-ranging molecular alterations regarding pro-in ammatory markers, transcriptional pro les and metabolic signatures in skeletal muscle of people with RA compared to matched healthy individuals 8 . Furthermore, disease activity, disability and physical inactivity in people with RA were associated with altered levels of the muscle pro-in ammatory mediators and metabolic intermediates. Also, oxidative stress induced modi cations in actin and promoted muscle weakness in people with RA by breaking actin lament stability and disrupting myosin interactions 9 .
Our group has reported that endurance exercise affects molecular pathways that regulate in ammation, aerobic capacity, capillary growth and remodelling of skeletal muscle in patients with polymyositis and dermatomyositis 10 . Whether in ammatory signalling and dysregulated remodelling in skeletal muscle in RA may be mitigated by combining pharmacological treatment with physical activity over a long period of time is not known.
A two-year program of Physical Activity in RA (PARA) 2010 was designed to investigate its feasibility and to study changes in self-reported variables, physical capacity measurements, in ammation, pain modulation and cardiovascular risk 11 . Positive changes during the two-year study period were observed for HEPA participants concerning symptoms and physical capacity [12][13] and cardiovascular risk factors [14][15] , while pain modulation did not change 16 . In this context we performed a pilot study with focus on protein adaptations in muscle tissue in a subgroup of people with RA who consented to repeated muscle biopsies. The aim of the present study was to investigate potential changes in protein expression patterns in response to a two-year HEPA program applying proteomic analysis in repeated skeletal muscle biopsies in relation to clinical variables in a subgroup of patients from the PARA 2010 study.

Participants and samples
Participants in the PARA 2010 study were recruited from the Swedish Rheumatology Quality registers (SRQ). They were asked to participate in a two-year program aimed to increase and maintain HEPA levels among people with RA in line with public health recommendations [11][12][13] For this sub-study six women who consented to donate muscle biopsies at baseline, at one-and two-year follow-ups were included. The participants were independent in daily living, interested in physical activity support, uent in Swedish, not already obtaining HEPA as per recommendations and without major diseases preventing HEPA 11 . The median age was 64.5 (range 54-73) years and median disease duration 6.5 (range [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] years. Information about disease activity, physical activity and treatment is displayed in Table 1. Skeletal muscle biopsies and blood samples were collected at Karolinska University Hospital, Stockholm, Sweden and at Sunderbyn Hospital, Luleå, Sweden. Ethical approval was obtained from Stockholm Regional Research Ethics Committee (2010/1232-31/1). Each participant gave written informed consent to participate in the study.

HEPA program
The support program aimed at adoption and maintenance of HEPA according to recommendations 4 .
The program consisted of three parts: moderate-intensity physical activity, strength training at public gyms, and bi-weekly support group meetings based on behaviour change theory 17 . Participants were expected to perform both the moderate-intensity physical activity and the gym training as part of their daily life at times that were convenient to them 11 .
At the start a physiotherapist, specialised in rheumatology, gave instructions and set up HEPA programs tailored to each participant's capacity. Thereafter physiotherapists were available for consultations in the gym for one hour weekly. Biweekly, one-hour support group meetings were organized by physiotherapists in order to educate, give feedback and improve the participants' self-e cacy for HEPA. Groups consisted of 5-10 persons and focused on how to incorporate and maintain HEPA in their daily schedule based on the Transtheoretical Model of Change (TTM) 18 . These gatherings were also important occasions to share experiences and enable social support, in accordance with the social cognitive theory (SCT) 17 . All participants received a pedometer and free access to a website (www.keepwalking.se) to monitor and stimulate physical activity.
During the second year the participants maintained the individualized HEPA programs, without physiotherapist support and support group meetings were optional. More information about the HEPA program is described by Nordgren et al. 11 .

Clinical assessments
Disease Activity Score 28 (DAS 28) 19 , Rheumatoid Factor (RF)/Anti-Citrullinated Protein Antibody (ACPA) positivity, C-Reactive Protein (CRP) and Erythrocyte Sedimentation Rate (ESR) and information on drug treatment were retrieved from the SRQ and patient records.
Data from self-reported questionnaires and physiotherapy assessments were collected before starting the HEPA program, after one and two years and included: -Assessments of body mass index (BMI), waist circumference, and blood pressure.
-Tests of physical capacity including maximal aerobic capacity (VO 2max) estimated from a submaximal bicycle ergometer test 21 , lower limb function with the Timed-Stands Test (TST) 22 , maximum and average grip strength with the Grippit device 23 and maximum voluntary contraction force (MVC) of the right knee extensors using a Biodex Multi-Joint System 4 Pro dynamometer were collected by physiotherapists.
Sensitivity to pressure pain was evaluated using pressure algometry 24 . Global pain intensity was rated on a visual analogue scale (VAS, range 0-100) before pressure pain thresholds (PPTs) were assessed bilaterally at three sites (mm.supraspinatus, mm.gluteus maximus and the lateral epicondyles). The mean of the stimulus intensity (kPa) of the six sites was calculated for each participant and used as the individual PPT mean 24 . Suprathreshold pressure pain sensitivity (SPPS) was assessed at the same six sites. The participants were asked to indicate when the pressure reached an intensity rated as 4/10 and 7/10, respectively on the Borg CR-10 scale 24 .
Adherence to the HEPA program assessed by two weekly text messages; one on the number of circuit training sessions performed in the past week and the other on the number of additional days with physical activity performed on at least a moderate intensity level for ≥ 30 minutes 25 . The total number of days with circuit training and/or moderate-intensity physical activity is reported.

Muscle biopsies
Eighteen muscle biopsies were collected from six participants at the annual physical exams (before start of the exercise program, after one and

Data analysis
Acquired MS raw les were analysed with Proteome Discoverer 1.3 (ThermoFisher Scienti c) and peak lists were searched using Sequest algorithm against the UniProtKB/SwissProt human database (containing 42163 canonical and isoform sequences, 2019-09-24) and ltered to a 1% FDR using Percolator. Precursor mass tolerance was set to 10 ppm and fragments mass tolerance was set to 0.02 Da for HCD-FTMS. The search algorithm assessed tryptic peptides with maximum 1 missed cleavage; carbamidomethylation (C) and TMT 10-plex (K, N-term) were set as xed modi cations and oxidation (M) as variable modi cation. Only unique peptides in the data set were used for quanti cation, while reporter ions were quanti ed by Proteome Discoverer on HCD-FTMS tandem mass spectra (integration tolerance 20 ppm).
Biological context and molecular networks were analysed using PANTHER Classi cation System and Ingenuity Pathway Analysis (IPA; Ingenuity Systems, www.ingenuity.com).
Comparisons were made between baseline and year one, and between baseline and year two.

Immunohistochemistry staining
Cryostat sections (7 µm) of the muscle biopsies were xed in 2% formaldehyde-PBS solution washed in PBS and stored in the freezer at -80°C.
Staining of cryosections to detect endothelial cells with monoclonal mouse antibodies anti-CD31 (clone EN4, Monosan, Uden, Netherlands) and anti-CD34 (clone QBend/10, BioGenex, The Hague, Netherlands) was performed according to a standard protocol 28 . Isotype-matched irrelevant antibody was used as negative control (clone DAK-GO1, Dako, Glostrup, Denmark). Stained sections were analysed with a Polyvar II microscope (Reichert-Jung, Vienna, Austria) and photographed with a digital Leica camera 300F (Leica, Cambridge, UK) ( Figure 4). CD31 and CD34 positively stained capillaries were counted with the Quantimet 600 image analyser (Leica) and the results were expressed as numbers of capillaries per area (mm²) re ecting the capillary density.

Statistical Analysis
Statistical analysis of self-reports, physical capacity, PPTs and HEPA was performed with STAT View for Windows (version 5.0.1.0; SAS Institute Inc). Data were analysed with non-parametric Wilcoxon's signed rank test. P-values <0.05 were considered statistically signi cant. Quantitative proteomics data was analysed by paired T-test in order to investigate the signi cance of the difference in expression between the base level and year one or two. Clinical outcome measures were correlated to the results of the muscle proteome analysis applying Pearson's test. In IPA, the activation z-score, an indication for an activated/inhibited status of certain transcription regulators or processes was calculated on the basis of experimentally observed protein levels and ndings in current literature. Associations between proteins and pathways or functions were calculated with a right-tailed Fisher exact test (p<0.05).

Clinical data
The achieved frequency of HEPA recorded for the participants was an average of 287 days during the rst year, and 179 days for the second year (Table 1). TST improved signi cantly compared to baseline after one year but was not sustained at two-year assessments (Table 2). Generally, perceived disease impact, pain, fatigue, health-related quality of life and activity limitation did not change from baseline to one-or two-year follow-ups, neither did BMI, waist circumference or blood pressure. No changes in VO 2max , MVC of the knee extensor or grip strength when compared to baseline were recorded. Among the pain-related variables, the mean PPT was signi cantly reduced after one and two years while SPPS was signi cantly decreased after one year (Table 2).

Muscle Biopsy Data-Capillary Density
Capillary density in skeletal muscle was investigated using immunohistochemistry. The mean density of CD31 and CD34-positive capillaries was not signi cantly altered after one or two-year following the HEPA program ( Supplementary Fig. 1A and 1B).

Proteome pro les in repeated muscle biopsies
Changes in protein expression in skeletal muscle after endurance exercise were monitored using mass spectrometry based proteomics. Proteome pro les of skeletal muscle biopsies resulted in quanti cation of 962 proteins in the samples obtained at years zero, one and two. Multivariate statistics applied for the analysis of obtained proteomics data demonstrated separation between muscle biopsies proteomes of patients at zero, one and two years (Supplementary Figure 2). The protein levels for the follow-up years were compared to the patient's own values in the biopsy taken before the intervention (year zero). Eight proteins were signi cantly (p<0.05) altered after one year of the HEPA program ( ve up-and three down-regulated) ( Table 3). After two years of the HEPA program 21 proteins were signi cantly altered (p<0.05) (18 proteins were up-and three proteins down-regulated) compared to the baseline level (Table 4). Data analysis also revealed up-regulation of the ATP synthase alpha protein (p<0,05), and at the same time, in ve participants out of six the protein mitochondrial ATPase inhibitor (IF1) was reduced but did not reach statistical signi cance at group level. Moreover, we found that the protein, Carnitine Palmitoyltransferase 1B (CPT1B) involved in fatty acid betaoxidation pathway in muscle mitochondria, was signi cantly up-regulated while the protein Otubain-1 (OTUB1) involved in mitochondrial apoptosis was signi cantly lower both after one year and two years of the HEPA program.

Gene Ontology Analysis And Pathway Analysis
Gene ontology (GO) enrichment analysis and Overrepresentation Test using PANTHER Classi cation System performed on the basis of the whole human genome revealed overrepresentation of the biological processes related to muscle contractility, energy metabolism and mitochondrial function (Fig. 1).
The analysis with IPA indicated that our dataset of proteins was enriched in proteins involved in the following molecular and cellular functions:  Table 3).

Correlation analysis between clinical outcome measures and results of proteomic analysis
Further IPA analysis revealed that several disease and function annotations predicted to be altered after the exercise program (Table 5). Most importantly, the HEPA program seemed to be associated with changes that predicted increased ATP production, another indication of elevated mitochondrial activity. ATP is essential for contractility of the skeletal muscles and also contractility was predicted to be upregulated, and this was signi cant in the biopsies taken after one year of HEPA but not after two years. Furthermore, the HEPA program was associated with protein changes that predicted reduced brosis after one year of HEPA and decreased muscle cell death after both one and two years of HEPA (Table 5). The p-value indicates the signi cance of the results while the absolute z-score needs be 2 or higher to detect a relevant effect on this function. Data were retrieved from Ingenuity Pathway Analysis.
Signi cant negative correlations were revealed between PPTs and mitochondrial proteins at one-year follow-up (Table 6). PLS-DA of data from baseline and year one resulted in a model where principal components PC1 and PC2 combined described 57% of the variation in data. The model was not signi cant (CV-ANOVA p > 0.05) and could not be used to predict before/after training status based on the protein pro le. The same approach was used when comparing baseline and the second biopsy after two years (PLS-DA: R2X(PC1 + PC2) = 67%, CV-ANOVA > 0.05). Prediction by the model might be improved after variable selection. However, the aim with this study was not to obtain a robust prediction model. Instead this model was used to select possible variables for further investigation. Variables with VIP (variable importance in the projection) above 1.0 were selected and subjected to pathway analysis. IPA demonstrated that proteins in uencing separation between the sample groups in the PLS-DA model were related to organ in ammation, which was predicted to decrease (z-score=-2.082, p = 4.62E-04 at year two).

Discussion
In the current subsample of a two-year intervention with HEPA aimed to promote HEPA in sedentary individuals with RA, we applied unbiased proteomics analysis of repeated muscle biopsies before and after one and two years. Although the physical activity was modest and decreased between the rst and second year, the effects on muscle proteome, e.g. on proteins involved in the mitochondrial respiratory chain and oxidative phosphorylation, were consistent suggesting a positive metabolic effect in muscle tissue of the long-term HEPA program.
Similar effects with enrichment of proteins in the mitochondrial compartment, which include the electron transport chain have been reported from short-term training programs in healthy men [29][30] . The respiratory chain is fundamental for ATP production that is necessary to meet the energetic requirements of the muscle cells 30 . In healthy subjects who participated in those studies the muscle homeostasis improved and their capacity to supply ATP increased. Our ndings suggest that muscles in people with RA are capable of improving their mitochondrial function in a similar fashion as seen in healthy young men. In myositis, where the main pathological processes occur in the skeletal muscle, mRNA expression patterns after endurance training demonstrated up-regulation of genes involved in capillary growth, which has been associated with increased capillary density in muscle tissue 10 . We did not nd a signi cant increase in the capillary density in the current study, which could be explained by the low number of participants in our study, although there was a tendency of higher number of CD31 positive endothelial cell per area after two years of HEPA.
Our MS-based proteomics results acquired from muscle tissue proteome demonstrate a signi cant up-regulation of several proteins in the mitochondrial respiratory chain: after one year of HEPA, four out of eight (COX7C, COX8A, NDUFS7, IDH3B) and after two years eight out of 21 (CX6B1, NDUFA4, CYC1, NDUA9, NDUA7, ATPA, NDUA2, ACAA2). This data corroborates predicted increased levels of ATP (p = 1,10E-07 and p = 1,20E-07 for year one and two HEPA, respectively), which synthesis requires augmented electron transport to create a longer lasting electrochemical gradient 31 . On a protein level we observed a 17% up-regulation of ATP synthase alpha (p < 0,05), and at the same time, in ve participants out of six a decreased level of IF1 was detected. IF1 inhibits the hydrolase activity of ATP synthase and increased expression of IF1 in tumours promotes the metabolic switch in cancer cells limiting the level of oxidative phosphorylation and promotes glycolysis 32 . In human skeletal muscle of obese patients with type 2 diabetes, the impairments in the balance between mitochondrial ATP-synthase and increased expression of IF1 lead to alterations in lipid metabolism and induction of proin ammatory cytokine TNF-alpha 33 . As seen in Table 5 physical activity in our participants lead to elevated levels of mitochondrial proteins that may increase ATP production improving lipid metabolism and decreasing cell death in muscle tissue. In combination with improved muscle strength these processes could prevent loss of function and muscle mass, moreover they are able to reverse the effects of rheumatoid cachexia 34 .
Surprisingly, we found a negative correlation between four mitochondrial proteins and pressure pain sensitivity at one-year follow-up (Table 6). Therefore, our results do not support the hypothesis that improved mitochondrial respiratory capacity and decreased lipid accumulation in the muscle tissues due to physical activity lead to decreased pain sensitivity. On the other hand, our results tally the previous study that demonstrated that, in people with RA, the HEPA program improved clinical outcome and decreased global pain intensity, but had no positive effect on pain sensitivity 16 .
Multivariate analysis followed by IPA demonstrated that proteins in uencing separation between the sample groups in the PLS-DA model were related to organ in ammation, which was predicted to decrease (p = 4.62E-04 at year two). Serum levels of C-reactive protein (CRP) in our investigated patient group were generally low (Table 1) and probably re ecting ongoing immunosuppressive treatment and low disease activity and did not change during the intervention. Bene cial effects of physical activity on in ammation and possible mechanisms of antiin ammatory effects of exercise have been reported in other studies and one suggested explanation being by reduction in body fat 35-36 . The improved fatty acid metabolism and enhanced ATP production due to changes in the mitochondrial machinery demonstrated in our work might also contribute to the anti-in ammatory effect of physical activity.
Our study has clear limitations such as the small sample size and decreased physical activity during the second year. We therefore consider our study as exploratory and further validations of the results are needed. However, the proteomic analysis was performed in repeated muscle biopsies allowing us to follow the molecular changes in muscle for each participant thus eliminating the problem with inter-individual variations.

Conclusions
Our results indicate that long-term HEPA in people with RA up-regulates proteins in the mitochondrial respiratory chain predicting increased ATP production and muscle contractility. In addition, results indicated that HEPA applied to people with RA may have positive impacts on muscle tissues by decreasing muscle cell death.

List Of Abbreviations
List of abbreviations For current study the ethical approval was obtained from Stockholm Regional Research Ethics Committee (2010/1232-31/1). Each participant gave written informed consent to participate in the study.

Consent for publication
Not applicable.
Availability of data and materials The datasets used in the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare no con icts of interests.  Figure 1 Biological processes overrepresented in the muscle proteome after training period in RA patients. The size of the circles corresponds to number of genes associated with the biological process. Cut-off levels are limited by dash lines.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryFiguresPARA.docx