Effect of pulmonary rehabilitation on muscle remodelling in cachectic patients with COPD

Baseline patient characteristics

Cachectic and non-cachectic patients were well matched according to age and severity of airflow obstruction (table 1). However, cycle ergometer exercise tolerance, the 6-min walking distance and the total score of the St George’s Respiratory Questionnaire were all lower (p<0.05) in the cachectic patients (table 1). Vastus lateralis muscle fibre type distribution was not different between groups (table 2). In contrast, the mean muscle fibre cross-sectional area was lower (p = 0.035) in the cachectic patients (table 2). This difference was attributable to the lower (p<0.05) cross-sectional areas of type IIa and IIb fibres (table 2). Capillary/fibre ratio was not significantly different between groups (table 2).

Effects of training on exercise tolerance and quality of life

Mean training intensity sustained during the rehabilitation programme was not different between cachectic and non-cachectic patients (126±13% and 117±10% of baseline peak work-rate, respectively). Accordingly, peak work-rate on the cycle ergometer was increased by a similar percentage in both cachectic patients (by 16±2%; p = 0.011) and non-cachectic patients (by 18±3%; p = 0.001). Similarly, the 6-min walk distance increased (p<0.05) in both cachectic (by 53±16 m) and non-cachectic patients (by 42±13 m) with no difference between groups. However, none of the groups exhibited a clinically meaningful mean improvement (>54 m). The total St George’s Respiratory Questionnaire score was improved (reduced) by a clinically meaningful margin (-6±2 units) in non-cachectic patients only, whereas it remained unchanged (-1±1 units) in cachectic patients.

Vastus lateralis muscle fibre morphological adaptations

Following rehabilitation, vastus lateralis mean muscle fibre cross-sectional area was increased in both cachectic (by 264±66 μm²; p = 0.003) and non-cachectic patients (by 438±66 μm2; p<0.001); however, the increase was relatively greater (p = 0.039) in the non-cachectic (11±2%) compared to the cachectic patients (7±2%) (table 3). Although both groups increased cross-sectional areas of type IIa and IIb muscle fibres by a similar magnitude (p<0.01), only non-cachectic patients increased the cross-sectional area of type I fibres (p = 0.001) (table 3). In addition, whilst both groups equally decreased (p<0.005) the proportion of type IIb fibres, non-cachectic patients increased the proportion of type I fibres (p = 0.01), whereas cachectic patients increased the proportion of type IIa fibres (p = 0.012) (table 3). Furthermore, capillary/fibre ratio was increased (p<0.02) similarly in both groups (table 3).

Rehabilitation-induced molecular adaptations in peripheral muscle

Western blotting revealed that at baseline, cachectic and non-cachectic COPD patients exhibited different levels of expression among the remodelling markers tested (fig. 1a). Accordingly, cachectic patients showed significantly lower levels of expression of both IGF-I and MyoD. In contrast, the expression of TNF-α, total IκB-α and myostatin was similar between groups (fig. 1a). Total Akt protein expression was significantly greater in the cachectic patients (fig. 1a). The ratio of phosphorylated IκB-α to total IκB-α was higher in the cachectic group (fig. 1b). In addition, at baseline the ratio of phosphorylated Akt to total Akt was also significantly greater in cachectic patients (fig. 1c). Prior to the initiation of the pulmonary rehabilitation programme there were no significant differences in protein expression of total tyrosine nitration and MURF-1 between cachectic and non-cachectic patients. However, in the muscle of cachectic COPD patients, Atrogin-1 protein expression was significantly increased (by 145±54%) compared to non-cachectic patients.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

Factors associated with skeletal muscle hypertrophy and atrophy pathways at baseline. Protein expression of molecular markers in peripheral muscle homogenates from non-cachectic (□) and cachectic (▓) chronic obstructive pulmonary disease patients were studied by western blot at baseline for a) the level of tumour necrosis factor (TNF)-α, myostatin, insulin-like growth factor (IGF)-I, myogenic differentiation factor D (MyoD), total IκB-α and total Akt protein expression, b) the ratio of phosphorylated IκB-α/total IκB-α, and c) the ratio of phosphorylated Akt/total Akt. Results of densitometric analyses (mean±sem) for cachectic patients are presented relative to the mean baseline values of non-cachectic patients set at 100%. *: p<0.05 between groups.

Post-training, there was an increase (p<0.02) in the mRNA levels of IGF-I in both groups (table 3). Similarly, there was an increase (p<0.05) in the mRNA levels of MGF in both cachectic (from 1.18±0.31 to 2.24±0.71 mRNA copies·10⁴ GAPDH copies⁻¹) and non-cachectic patients (from 0.82±0.14 to 1.21±0.13 mRNA copies·10⁴ GAPDH copies⁻¹). The increase in mRNA levels of MGF was greater (p = 0.027) in cachectic patients. In these patients rehabilitation-induced changes in MGF mRNA expression were significantly correlated (r = 0.65, p = 0.034) with the respective changes in cycle ergometry peak work-rate. In non-cachectic patients rehabilitation significantly increased IGF-I protein expression two-fold (fig. 2b). In contrast, cachectic patients did not show increased IGF-I protein expression after the rehabilitation programme. In addition, the post-training IGF-I protein expression levels were significantly greater in non-cachectic than cachectic patients (fig. 2b).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2–

Insulin-like growth factor (IGF)-I and myogenic differentiation factor D (MyoD) expression following rehabilitation. a) Representative western blots for IGF-I and MyoD. b, c) Densitometric analyses of western blots from non-cachectic and cachectic chronic obstructive pulmonary disease patients for b) IGF-I and c) MyoD. Rehabilitation-induced changes in protein expression within each group are presented relative to the mean baseline value set at 100%. Results are presented as mean±sem. □: baseline; ▓: post-training. *: at p<0.05 between baseline and post-rehabilitation for each group; ^#: significant differences between groups.

Non-cachectic patients exhibited a significant increase in MyoD mRNA after training (table 3), consistent with the percentage increase in protein levels (fig. 3c). However, cachectic patients did not show increased post-training MyoD mRNA (table 3) or protein (fig. 2c) expression. After training, the level of MyoD mRNA and protein expression in non-cachectic patients was significantly higher than in cachectic patients (fig. 2c).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3–

Myostatin and tumour necrosis factor (TNF)-α expression. a) Representative western blots for Myostatin and TNF-α. b, c) Densitometric analyses of western blots from non-cachectic and cachectic chronic obstructive pulmonary disease patients for b) Myostatin and c) TNF-α. Rehabilitation-induced changes in protein expression within each group are presented relative to the mean baseline value set at 100%. Results are presented as mean±sem. □: baseline; ▓: post-training. *: p<0.05 between baseline and post-rehabilitation for each group.

Rehabilitation significantly reduced myostatin mRNA (table 3) and protein levels (fig. 3b) in non-cachectic patients. In contrast, this was not seen in cachectic patients (fig. 3b). TNF-α was not affected at either mRNA (table 3) or protein levels (fig. 3c) by rehabilitative exercise in either group. Post-training mRNA TNF-α expression was significantly greater in cachectic compared to non-cachectic patients (table 3).

Prior to rehabilitation, both groups of patients had the same amount of total cytoplasmic IκB-α (fig. 1a). After rehabilitation, the ratio of phosphorylated IκB-α to total IκB-α was significantly (p<0.05) decreased in both groups by the same amount (fig. 4b). Furthermore, prior to rehabilitation, cachectic patients had greater amounts of total cytoplasmic Akt (fig. 1a). After rehabilitation, the ratio of phosphorylated Akt to total Akt did not increase in either cachectic or non-cachectic patients (fig. 4d).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4–

Pulmonary rehabilitation decreases nuclear factor (NF)-κB activation and tends to increase Akt activation. Representative western blots for a) phosphorylated IκB-α, Total IκB-α and α-Actinin and c) phosphorylated Akt, Total Akt and α-Actinin. Densitometric analyses of blots studied in non-cachectic and cachectic chronic obstructive pulmonary disease patients for b) phosphorylated IκB-α and d) phosphorylated Akt. Rehabilitation-induced changes in protein expression within each group are presented relative to the mean baseline value set at 100%. Results are presented as mean±sem. □: baseline; ▓: post-training. *: p<0.05 between baseline and post-rehabilitation for each group.

Rehabilitation induced a significant increase in muscle protein nitration in cachectic patients only (fig. 5a). Interestingly, following exercise training, Atrogin-1 and MURF-1 protein expression was significantly increased in cachectic patients, whereas it was significantly decreased in non-cachectic patients (fig. 5c and d).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5–

Pulmonary rehabilitation increases nitrotyrosine, Atrogin-1 and muscle ring finger protein (MURF)-1 protein expression in cachectic patients. a, c, d) Densitometric analyses of blots studied in non-cachectic and cachectic chronic obstructive pulmonary disease patients for a) protein tyrosine nitration, c) Atrogin-1 and d) MURF-1. b) Representative western blots for Atrogin-1, MURF-1 and α-Actinin. Rehabilitation-induced changes in protein expression within each group are presented relative to the mean baseline value set at 100%. Results are presented as mean±sem. □: baseline; ▓: post-training. *: p<0.05 between baseline and post-rehabilitation for each group; ^#: significant differences between groups.

Baseline muscle remodelling markers

The biochemical pathways engaged in the development of muscle atrophy are thought to result from an imbalance between protein synthesis and breakdown; however, the exact mechanisms involved are not fully understood 17. Accordingly, molecular factors regulating muscle remodelling were analysed at baseline and compared between patient groups. Our study shows for the first time that baseline protein expression of Atrogin-1, a ligase regulating ubiquitin-mediated protein degradation 9, was significantly increased in the muscle of cachectic patients, whereas IGF-I and MyoD protein expression were lower in cachectic patients (fig. 1). Other studies have compared IGF-I and MyoD expression between COPD patients and healthy age-matched subjects 20 and found that the expression of these two anabolic markers were reduced in hospitalised and clinically stable COPD patients compared to healthy controls. A study by Crul et al. 20 suggested that diminished physical activity may be the main mechanism accounting for decreased muscle IGF-I and MyoD levels in COPD. In support of this notion, physical inactivity has been shown to be associated with decreased IGF-I mRNA levels and alterations in the MyoD/myogenin expression ratio in a mouse muscle-disuse model 21, 22. In contrast to IGF-I and MyoD protein expression, we found that the phosphorylated Akt to total Akt ratio was significantly increased in cachectic COPD (fig. 1c), an indication previously interpreted as a failed attempt to maintain or restore muscle mass in patients with COPD 23.

However, greater muscle fibre atrophy in cachectic compared to non-cachectic patients could also be explained by greater activation of the NF-κB transcription factor complex 6, 7. Although at baseline cachectic and non-cachectic patients did not differ in terms of the level of expression of total IκB-α (fig. 1a), we found that the phosphorylated IκB-α to total IκB-α ratio was two-fold greater in cachectic compared to non-cachectic patients (fig. 1b). However, interpretation of these findings has to be carried out with caution because the IκB-α ratio utilised is an indicator of the NF-κB activation and, as such, further experiments measuring activation of NF-κB in the nucleus were not possible due to muscle tissue limitation.

NF-κB can be activated by a number of cachexia-associated factors including TNF-α 7. Although we did not find significant differences in baseline local muscle mRNA or protein TNF-α expression between study groups (fig. 1a), we can not exclude the possibility that the greater degree of muscle wasting seen in cachectic COPD was instigated by increased local muscle expression of cytokines (other than TNF-α) which are involved in inflammation, such as interleukin (IL)-1β and IL-6, and which are thought to trigger muscle wasting via NF-κB activation 24. Nevertheless, a recent study 24 demonstrated muscle atrophy due to systemic inflammation-dependent chronic activation of NF-κB in the muscle of mice without increased TNF-α, IL-6 or IL-8 mRNA levels. Consistent with a recent study 16, at baseline we found no differences in muscle protein tyrosine nitration between cachectic and non-cachectic patients.

In muscle, NF-κB can also inhibit mRNA and protein expression of MyoD 14, 25, an essential transcription factor in myogenesis 9. Downregulation of MyoD expression can also be induced by over-expression of myostatin, a strong negative regulator of muscle growth 15. However, at baseline we found that myostatin protein expression was not different in cachectic patients compared to non-cachectic patients, thereby suggesting that the lower MyoD expression in cachectic patients is probably due to the increased activation of NF-κB 6, 7.

Exercise-induced effects on exercise capacity and muscle remodelling markers

Cachectic and non-cachectic COPD patients were selected to control for differences in patient phenotype. At the beginning of the study, cachectic patients were characterised by poorer exercise capacity and greater peripheral muscle fibre atrophy (tables 1 and 2). Following rehabilitation, exercise capacity was increased similarly in both cachectic and non-cachectic patients, most probably as a result of exercising both groups at a similar overall training load.

In addition, confirming previous findings in non-cachectic COPD 5, we have shown for the first time that rehabilitation induces a significant increase in vastus lateralis muscle mean fibre cross-sectional area and capillarisation in cachectic patients. In addition, rehabilitative exercise reduced the proportion of IIb fibres in favour of the more oxidative fibres (i.e. type IIa). These findings suggest that peripheral muscle remodelling in response to exercise training is, at least partially, preserved in cachectic patients with COPD. Indeed, we observed that muscle fibre hypertrophy was accompanied by significant upregulation of local muscle IGF-I and MGF mRNA expression that are known to play an important role in the hypertophic adaptation of muscle to overload 23, 26. In our study the magnitude of increase in IGF-I mRNA expression after rehabilitation in cachectic COPD (by 88%) was similar to that reported for the non-cachectic patients in the present (by 79%) and previous (by 75%) studies 5. However, the increase in IGF-I mRNA expression after rehabilitation was translated at protein level in non-cachectic COPD only (fig. 2b); this may explain the greater degree of peripheral muscle fibre hypertrophy found in this group and highlights a potential post-transcriptional modification in cachectic patients. In addition, mRNA degradation may also act as an important mechanism in the regulation for the expression of muscle IGF-I 27.

Another potential factor that may account for the lower degree of peripheral muscle fibre hypertrophy observed in cachectic patients is their inability to upregulate local muscle MyoD in response to rehabilitative exercise (table 3, fig. 2c). MyoD is expressed in muscle satellite cells and mature myofibres and has been implicated in mediating the process of cell proliferation and differentiation for subsequent muscle regeneration and hypertrophy 14.

Within our study we investigated the effects of exercise training on two factors that are known to modulate the expression of local muscle MyoD; namely TNF-α and myostatin 14, 15. The downregulation of myostatin in response to exercise training in healthy humans has been shown to be essential for muscle growth 28. In non-cachectic COPD patients, resistance training has been shown to induce a modest reduction in myostatin mRNA abundance 29. Therefore, the results of the present study expand those in non-cachectic COPD 29 by demonstrating that endurance training induced significant reductions in both mRNA and protein myostatin expression (table 3, fig. 3b). Rehabilitation-induced reduction in local muscle myostatin expression in non-cachectic patients was accompanied by a significant upregulation of both mRNA and protein expression for MyoD, thereby confirming the results of a previous study 5. In contrast, in cachectic patients neither mRNA nor protein expression for myostatin was significantly reduced (table 3, fig. 3b). This finding may explain the lack of significant training-induced upregulation of local muscle MyoD expression in cachectic COPD and account for the lower degree of fibre hypertrophy.

Relationship between markers of inflammation and training responses

Muscle TNF-α mRNA and protein expression were not significantly affected by exercise training in cachectic patients (table 3, fig. 3c), thus expanding the results of previous studies carried out in patients without peripheral muscle wasting 5, 30. Interestingly, we found that the phosphorylated IκB-α to total IκB-α ratio was significantly downregulated in both groups following rehabilitation, most probably indicating reduced activation of the NF-κB transcription factor (fig. 4b).

Emerging evidence suggests that NF-κB is involved in one of the most important signalling pathways linked to the loss of skeletal muscle mass in various physiological and pathophysiological conditions. A recent study using an animal model has provided unequivocal evidence that specific modulation of NF-κB activity can prevent skeletal muscle loss 31. It has recently been recognised that NF-κB can be activated not only by cachectic factors (such as TNF-α) but also by muscle disuse 9, 32. Muscle disuse triggers an NF-κB activation pathway that is distinctly different from that triggered by a cytokine (e.g. TNF-α) 32. If this is correct, one would expect that exercise training would decrease the phosphorylation ratio of the inhibitory IκB-α, which is in fact what our study demonstrates (fig. 4b), thereby indicating NF-κB as a possible target for exercise training. Thus, it is possible that when disused, atrophic peripheral muscles of COPD patients undergo exercise training and NF-κB is progressively de-activated, as indicated by the decreased phosphorylation of IκB-α, allowing a number of downstream signalling pathways to be activated and promote muscle hypertrophy. Conversely, as baseline NF-κB activity was twice as high in cachectic patients compared to non-cachectic patients (fig. 1b), and since the phosphorylated IκB-α to total IκB-α ratio was downregulated in both groups following rehabilitation by the same magnitude (fig. 4b), it is conceivable that NF-κB activity was still greater in cachectic patients following rehabilitation, thereby limiting the degree of muscle remodelling.

The limited degree of muscle remodelling in cachectic compared to non-cachectic patients can also be justified by our findings that exercise training caused a greater induction of muscle protein nitrotyrosination compared to non-cachectic patients 16, most likely facilitating degradation through the ubiquitin/proteasome pathway 6. Our study is the first to provide proof for this notion since we found that two specific regulators of ubiquitin-mediated muscle proteolytic pathway (Atrogin-1 and MURF-1) were upregulated following rehabilitation in cachectic patients only. In contrast, in non-cachectic patients, Atrogin-1 and MURF-1 protein expression was downregulated, whereas growth and regeneration factors were unregulated; thus, justifying the greater degree of muscle fibre hypertrophy in these patients. The diverse response to exercise training between cachectic and non-cachectic patients resembles that recently documented in COPD patients and age-matched healthy controls where the protein degradation pathway was upregulated in COPD only 33. Therefore, it is likely that exercise training predominantly produces an anabolic influence among non-cachectic COPD, while in cachectic COPD regular muscular activity simultaneously activates anabolic and catabolic pathways, thereby limiting the magnitude of the training effect.

Study limitations

The invasive nature of the muscle biopsies precluded a parallel, age-matched, healthy control training group which would have allowed comparisons in training-induced changes in local muscle anabolic and catabolic factors between healthy subjects and COPD patients with or without muscle wasting. In addition, for the same reason, a control COPD group was not studied, i.e. not undertaking rehabilitation. Such a group would have distinguished which of the effects of rehabilitation were due to a true effect of training and which are due to between-occasion muscle sampling differences. Furthermore, absence of quadriceps measurements of strength or mid-thigh cross-sectional area is acknowledged as a limitation in our study. However, recent studies 23, 34 have shown reduced mid-thigh cross-sectional area, as well as strength in cachectic patients compared to non-cachectic COPD patients and healthy controls. Furthermore, both in vivo and in vitro muscle function measurements would have provided more adequate evidence to link the observed molecular/cellular changes with the clinical outcomes of pulmonary rehabilitation in cachectic COPD. Another potential limitation in our study was that cachectic patients were also those showing worse lung function. Indeed, it is well known that weight loss and increasing severity of COPD are often associated with each other and both lead to poor prognosis of the disease 2.

In conclusion, although skeletal muscle training produces greater peripheral muscle fibre phenotypical adaptations in non-cachectic patients compared to cachectic patients with COPD, cachectic patients retain the capacity for peripheral muscle remodelling in response to rehabilitation and are able to increase exercise capacity as much as those without cachexia.