Whole-body and muscle responses to aerobic exercise training and withdrawal in ageing and COPD

Background Chronic obstructive pulmonary disease (COPD) patients exhibit lower peak oxygen uptake (V′O2peak), altered muscle metabolism and impaired exercise tolerance compared with age-matched controls. Whether these traits reflect muscle-level deconditioning (impacted by ventilatory constraints) and/or dysfunction in mitochondrial ATP production capacity is debated. By studying aerobic exercise training (AET) at a matched relative intensity and subsequent exercise withdrawal period we aimed to elucidate the whole-body and muscle mitochondrial responsiveness of healthy young (HY), healthy older (HO) and COPD volunteers to whole-body exercise. Methods HY (n=10), HO (n=10) and COPD (n=20) volunteers were studied before and after 8 weeks of AET (65% V′O2peak) and after 4 weeks of exercise withdrawal. V′O2peak, muscle maximal mitochondrial ATP production rate (MAPR), mitochondrial content, mitochondrial DNA (mtDNA) copy number and abundance of 59 targeted fuel metabolism mRNAs were determined at all time-points. Results Muscle MAPR (normalised for mitochondrial content) was not different for any substrate combination in HO, HY and COPD at baseline, but mtDNA copy number relative to a nuclear-encoded housekeeping gene (mean±sd) was greater in HY (804±67) than in HO (631±69; p=0.041). AET increased V′O2peak in HO (17%; p=0.002) and HY (21%; p<0.001), but not COPD (p=0.603). Muscle MAPR for palmitate increased with training in HO (57%; p=0.041) and HY (56%; p=0.003), and decreased with exercise withdrawal in HO (−45%; p=0.036) and HY (−30%; p=0.016), but was unchanged in COPD (p=0.594). mtDNA copy number increased with AET in HY (66%; p=0.001), but not HO (p=0.081) or COPD (p=0.132). The observed changes in muscle mRNA abundance were similar in all groups after AET and exercise withdrawal. Conclusions Intrinsic mitochondrial function was not impaired by ageing or COPD in the untrained state. Whole-body and muscle mitochondrial responses to AET were robust in HY, evident in HO, but deficient in COPD. All groups showed robust muscle mRNA responses. Higher relative exercise intensities during whole-body training may be needed to maximise whole-body and muscle mitochondrial adaptation in COPD.


Study design
Figure S1. Study protocol followed by all volunteers. Biopsy, microbiopsy of vastus lateralis; CPET Incremental , symptom-limited incremental cardiopulmonary exercise test; CPET Submax , submaximal cardiopulmonary exercise test at workload corresponding to 65% of the workload at VȮ2 PEAK in baseline CPET Incremental ; PFT, pulmonary function test; strength, quadriceps maximal voluntary contraction; PA, physical activity monitoring for 7 days.
Training intensity was reset at week four if workload at VȮ2 PEAK had increased.

Baseline assessments
Baseline assessments were performed over three visits. At visit 1 measures of anthropometry (height, body mass, body composition by DEXA), pulmonary function and quadriceps strength were performed as was a familiarisation symptom-limited incremental cycling cardiopulmonary exercise test (incremental CPET). At the second visit (minimum 48 hours after visit one) a further incremental CPET (to verify the preceding test) was performed Training followed by a submaximal cycling exercise tests (separated by >30 minutes resting time). A minimum of 7 days later, at the third visit, a quadriceps muscle biopsy was performed (week 0). Habitual physical activity was assessed over seven days prior to intervention and repeated during the first and last weeks of the exercise withdrawal period.

Anthropometry, pulmonary function, quadriceps strength and physical activity monitoring
Measurements of pulmonary function including lung volumes by plethysmography [S1], body mass index (BMI), body composition (dual energy x-ray absorptiometry, DEXA; Lunar Prodigy, GE Healthcare, Buckinghamshire, United Kingdom), and quadriceps isometric strength at 90° knee and hip flexion (Cybex II Norm: CSMi, Stoughton, USA) were performed at baseline.
Habitual physical activity was monitored using a triaxial accelerometer (SenseWear; BodyMedia, Pittsburgh, USA) worn during waking hours on 7 consecutive days before baseline, and during the first (week 9) and final (week 12) week of exercise withdrawal. A minimum of 8 hours of data per day was required to be included in the analysis. The mean step count of the initial 8 hours after waking was calculated for each individual.

Incremental CPET and a submaximal cycling exercise tests
Incremental CPET was performed in accordance with international guidelines [S2]  The intensity of the sub-maximal test corresponded to the initial training intensity and measures of heart rate, VĖ, and respiratory exchange ratio (RER) indicate the physiological stress experienced during training.

Muscle sampling and processing
Muscle biopsy samples were obtained from the vastus lateralis muscle of the dominant leg at mid-thigh level using a needle micro-biopsy technique [S4]. Briefly, after the skin was sterilised with Betadine solution, local anaesthetic (lignocaine) was injected subcutaneously and to the depth of the fascia. A small incision (5 mm) was made in the skin and any subcutaneous adipose, through which a 12 g micro-biopsy needle was inserted (Bard Magnum, Arizona, USA). Four passes were performed, each harvesting ~20 mg of tissue.
Approximately 40 mg of freshly isolated vastus lateralis muscle tissue was finely diced on a cooled glass plate, and weighed for mitochondrial function and content measurements. The remaining muscle tissue was immediately dissected free of visible adipose and connective tissue, snap frozen and stored in liquid nitrogen for subsequent DNA and mRNA analyses. The biopsy site was dressed with a butterfly closure, waterproof sterile dressing and a compression bandage to apply light pressure in order to minimise the risk of bleeding or bruising. Subsequent biopsies were performed 2.5 cm from the preceding incision site.

Muscle Mitochondrial measurements
Each sample was homogenised on ice for 3 min in a buffer solution (pH 7.0, KCl 100mM, KH2PO4 50 mM, Tris 50 mM, MgCl2 5mM, EDTA 1 mM, ATP 1.8 mM) using a Teflon pestle homogeniser. The crude homogenate was then centrifuged at 650 g for 3 min at 4 o C, and the resultant supernatant was transferred to a test tube and centrifuged at 15,000 g for 3 min at 4 o C. The resulting pellet formed contained the mitochondria. Following this, the supernatant was removed and discarded before resuspension of the pellet in 300 μl of the original homogenisation buffer. This was then centrifuged at 15,000 g for 3 min at 4 o C. After removal of the supernatant, the pellet was re-suspended in a re-suspension solution (pH 7.0, human serum albumin 0.5 mg/ml, sucrose 240 mM, monopotassium phosphate 15 mM, magnesium acetate tetrahydrate 2mM, EDTA 0.5 mM). The differential centrifugation isolated mitochondrial suspension was then kept on ice immediately prior to measurement of mitochondrial ATP production rates (MAPR).

Mitochondrial ATP production rates (intrinsic mitochondrial function)
Following the method of Wibom et al [S5], 2.5 µl of diluted mitochondrial suspension was added to each well of a luminometer plate.

Citrate synthase activity (mitochondrial content)
Muscle citrate synthase (CS) maximal activity was determined at 37°C on the isolated mitochondrial solution using a kinetic spectrophotometric method to follow the change in absorbance of a 5,5'-dithiobis (2-nitrobenzoic acid)(DNTB) buffered solution as previously described (E6). Briefly, 15 μl mitochondrial suspension was added to 185 μl homogenisation buffer (95% extraction buffer containing 1% triton) and was homogenised using a glass pestle at 200 rpm for 2 min (5). The homogenate was centrifuged at 24,000 g (Eppendorf, Hamburg, Germany) before CS was determined spectrophotometrically in the supernatant [S5]. Acetyl-CoA is formed at a rate determined by the quantity of CS protein present.
Relative mitochondrial DNA copy number Genomic DNA (nDNA) and mitochondrial DNA (mtDNA) were extracted from skeletal muscle Qiagen using a QIAamp® DNA Mini kit according to the manufacturer's instructions. Briefly, the procedure involved initial tissue lysis in a buffer containing proteinase K, incubation for 3 hrs at 56°C to digest the myofibril proteins

Muscle mRNA expression linked to fuel metabolism
RNA was extracted from ~30 mg snap-frozen muscle as previously described [S7]. First strand cDNA was synthesised from 1 µg of total RNA, using Superscript III reverse transcriptase (Invitrogen Ltd, Paisley, UK) and random primers (Promega, Southampton, UK) and stored at -80 o C until analysis. TaqMan low density arrays were performed using an ABI PRISM 7900HT sequence detection system, and data analysed using SDS 2.1 software (Applied Biosystems, USA). Data were further analysed using RQ Manager software (Applied Biosystems, USA), where the threshold level was normalised across all plates before Ct values were calculated for each gene target and sample. Relative quantification of mRNAs of interest was measured using the 2 -ΔΔCt method with hydroxymethylbilane synthase (HMBS) as the endogenous control as it was unaffected by exercise intervention (data not shown). A total of 59 transcripts known to be involved in muscle carbohydrate and lipid metabolism were targeted for analysis in the present study (Table S1). Target selection was led by published data involving high-throughput and targeted RT-PCR approaches from our research group which identified muscle transcripts responsive to exercise intervention [S8], insulin resistance [S9] and changes in fuel metabolism with nutritional [S10] and pharmacological intervention [S11].
Additionally unpublished muscle transcript data from our group from research involving limb immobilisation in healthy volunteers was accessed. To associate altered biological functions to the targeted probe sets, Ct values were uploaded to Ingenuity Pathway Analysis (IPA) software (QIAGEN, Hilden, Germany) for pathway analysis of gene expression data. The overall outcome of IPA (e.g. cellular function) is predicted by calculating a regulation Z-score and an overlap p-value, which are based on: 1, the number of regulated target genes' function; 2, the magnitude of expression change; 3, the direction of expression change; and 4, their concordance with the IPA database, which is constructed from an extensive curated literature database. The overlap p-value was calculated by IPA to identify significantly enriched function pathways from the submitted list of significantly changed genes. These pvalues were generated from the right-tailed Fisher's Exact Test, and a significance threshold of p<0.05 was used to assess the statistical significance of the function pathways. In order to control for any enrichment of false positive results when undertaking multiple comparisons (type II errors) IPA utilises Bonferroni's corrected p-value set at p< 0.05.

Exercise training intervention
Participant underwent supervised training on an electrically braked cycle ergometer (Lode Corival, Groningen, The Netherlands). Three supervised sessions of 30 min duration were performed per week. Individuals who were unable to compete 30 min continuously were permitted to rest (~5 min) before resuming the session over a total permissible duration of 60 min. Exercise at the prescribed intensity (workload corresponding to 65% VȮ2 PEAK ) is known to increase muscle lactate accumulation, pyruvate dehydrogenase complex (PDC) activation and flux [S12], and rates of mitochondrial carbohydrate and lipid oxidation from both plasma and muscle sources [S13] well above the resting state, thereby providing a robust stimulus to muscle metabolic adaptation.

Statistical analysis
A power calculation performed on MAPR data (glutamate and succinate) using G-Power software (version 3.1.9.2, Dusseldorf University, Germany) for ANOVA one-way fixed effects given α = 0.05, number of groups = 3, power = 0.9 and effect size = 0.6 using the data from Barany et al. [S14] recommended n=9 for the healthy control group, which we rounded up to n=10 for healthy volunteers. Given the inherently variable nature of physiological responses in patients with COPD, this number was increased to n=20 in the COPD group.

Baseline
Fourteen HO and 15 HY volunteers consented to participate with 10 from each group completing all study measures. Twenty-seven patients with COPD consented, 20 of whom completed the training intervention and week eight assessments with one drop-out during the exercise withdrawal period leaving 19 COPD patients in the week 12 analysis.  There were no significant within-group changes over time in physical activity assessed by daily step count (Table S5). Heart rate during steady state exercise was reduced from baseline after eight weeks training and after four weeks exercise withdrawal in all groups (Fig. S2 A; all p < 0.01) and VĖ was reduced at the same time points in HO (p < 0.01) and HY (p < 0.05) but was unchanged in COPD ( Fig. S2 B).  Figure S2. Physiological responses to steady-state sub-maximal exercise at a 65% of the work load achieved at VȮ2 PEAK in the baseline test. A, Change in heart rate (HR) after 8 weeks training, and after 4 weeks exercise withdrawal where subjects returned to habitual physical activity levels. B, Respiratory exchange ratio (RER) at the same time points as above (A).
Within group change p < 0.05 for: HO, *; HY, †; COPD, ‡. Values are mean (SEM). Figure S3. Differentially regulated muscle mRNAs associated with lipid metabolism following 4 weeks exercise withdrawal compared to baseline in healthy older, healthy young and COPD groups. Abbreviated gene names are defined in Table S2.
Healthy Young Healthy Older COPD Lipid Metabolism (After 4 Weeks Exercise Withdrawal) Figure S4. Differentially regulated muscle mRNAs associated with carbohydrate metabolism following 4 weeks exercise withdrawal compared to baseline in healthy older, healthy young and COPD groups. Abbreviated gene names are defined in Table S2.
Healthy Young Healthy Older COPD Carbohydrate Metabolism (After 4 Weeks Exercise Withdrawal)