As endurance training improves symptoms and quality of life and decreases mortality rate and hospitalization, it is increasingly recognized as a beneficial practice for heart failure (HF) patients. However, the mechanisms involved in the beneficial effects of exercise training are far from being understood and need further evaluation. Independent of hemodynamics effects, exercise training participates in tissue remodeling. While heart failure induces a generalized metabolic myopathy, adaptation to endurance training mainly improves energetic aspects of muscle function. In the present review, after presenting the main characteristics of cardiac and skeletal muscle energy metabolism and the effects of exercise training, we will discuss the evidence for the beneficial effects of endurance training on cardiac and skeletal muscle oxidative metabolism and intracellular energy transfer in HF.
These beneficial effects of exercise training seen in heart failure patients are also relevant to other chronic diseases (chronic obstructive pulmonary disease, diabetes, and obesity) and even for highly sedentary or elderly individuals [Booth F.W., Chakravathy M.V., Spangenburg E.E. Exercise and gene expression: physiological regulation of the human genome through physical activity. J Physiol (Lond) 2002;543:399–411]. Physical rehabilitation is thus a major health issue for populations in industrialized countries.
Alterations in cell energy metabolism participate in many pathophysiological processes including chronic heart failure (CHF). This syndrome is characterized by an augmented energy demand due to increased workload, a decreased energetic efficiency, and diminished energy metabolism leading to energetic imbalance of the myocardium [2–6]. HF is also characterized by a reduced ability to perform aerobic exercise and early fatigue. The reason for fatigue is attributable to persistent vasoconstrictor drive, endothelial dysfunction, and structural and functional abnormalities of skeletal muscle rather than to ventricular dysfunction per se[6–10]. Fatigue resistance in muscle is largely dependent on oxidative energy production and on control of energy fluxes [10–12]. Cardiac and skeletal muscle cells whose energy metabolism is high, fluctuating, and adapting to the special needs of the body exhibit sophisticated pathways for synthesizing, transferring, and utilizing energy in accordance with the needs of the body [11,13–15]. Thus, the view that HF is a paradigm of energetic failure of cardiac and skeletal muscles, leading to contractile failure, physical handicap, worsening and ultimately death, should be increasingly considered . On the other hand, exercise training improves endothelial function and coronary perfusion, decreases peripheral resistance, and induces cardiac and skeletal muscle cell remodeling, leading to increased oxygen uptake, substrate oxidation, and resistance to fatigue [16–19].
Although physical activity was avoided in CHF patients until the late 1980s, over the last decade, endurance exercise training in CHF patients has proven feasible , in terms of increased exercise capacity, quality of life [21,22], and to potentially reduce morbidity and mortality rates [23–25]. Adaptation to endurance training mainly involves energetic aspects of muscle function. To what extent endurance training in HF improves cardiac and skeletal muscle metabolism as well as how metabolic improvement contributes to the beneficial effects of exercise training is far from being understood.
This review will focus on the effects of exercise training in heart failure with a special emphasis on energy metabolism. We will first briefly summarize the current knowledge on cardiac and skeletal muscle energy metabolism in health and heart failure (see [2,4–6,26] for extensive reviews). In the second section, we will also summarize the effects of exercise training on muscle energy metabolism (see [11,12,27] for reviews). Finally, current knowledge on the effects of exercise training on energy metabolism in HF will be presented.
2. Cardiac and skeletal muscle energy metabolism in heart failure
2.1. Normal cardiac and skeletal muscle energy metabolism
Heart muscle must maintain blood circulation in states of both rest and high peripheral demand such as that triggered by exercise. A trained athlete is able to rapidly increase cardiac output as much as 6-fold from rest to high activity levels . The work done by the heart and hence energy yield must then be permanent, adaptable, rapidly regulated, and highly efficient.
The functional and structural diversity of mammalian skeletal muscles allows performances from strenuous fast strength production of locomotor muscles to long-lasting contractile activity of postural muscles. Energy metabolism plays a critical role in muscle function. Schematically, two extreme metabolic patterns have been defined [14,29–31]. Fast skeletal muscle fibers mainly rely on quickly mobilisable energy sources to develop strong and fast contractions but this is possible only for short periods of time because of limited cellular metabolic reserves (mainly phosphocreatine (PCr) and glycogen). These muscles are quickly fatigable and exhibit delayed recovery of their energy reserves through anaerobic glycolysis and less importantly mitochondrial oxidations. Slow muscles, on the contrary develop lower contractile force but are able to maintain long-term contractile activity because they rely on oxidative phosphorylations. They should have accurate adjustment of energy production to energy consumption. This metabolic specificity of muscle types generates a fiber-type specificity of substrate utilization, roughly oxidative fibers and cardiac muscle oxidize fatty acids and lactate, while glycolytic fibers mainly use glucose as substrate [29,32]. Creatine kinase (CK) and other phosphotransfer kinases (like myokinase) participate in energy shuttling within cardiac and skeletal muscle cells (reviewed in [14,33,34]). CK isoenzymes for example are located on sites of energy production (mitochondria or glycolytic complexes) and utilization (myosin and sarcoplasmic reticulum ATPases (SERCA)), ensuring fast and efficient energy transfer to ATPases, and signal transfer to energy producing sites (for review see [14,34,35]. This fine-tuning of energy fluxes in oxidative muscle and heart underlies fatigue resistance. Finally, direct energy channeling between mitochondria and sarcoplasmic reticulum (SR) or myofibrils in slow and cardiac muscle, also produces compartmentation of adenine nucleotides [36–38].
Another recent issue in energy metabolism concerns the transcriptional regulation and signaling pathways involved in the maintenance of energy homeostasis and mitochondrial biogenesis (for recent review see ). Mitochondrial biogenesis depends on the coordinated function of mitochondrial and nuclear genomes. The mitochondrial transcription factor (mtTFA) is encoded by the nuclear genome and activates transcription and replication of the mitochondrial DNA. MtTFA expression is controlled by nuclear respiratory factors (NRFs) that additionally stimulate expression of numerous nuclear-encoded mitochondrial proteins. NRFs expression and transcriptional activity are under the control of the transcriptional co-activator PGC-1α (transcriptional co-activator of peroxisome-proliferator-activated-receptor (PPAR) gamma) [39,40]. PGC-1α coordinates mitochondrial protein expression with substrate utilization by co-activating respectively PPARα that regulates fatty acid oxidation, and NRFs that activate transcription of nuclear and mitochondria-encoded proteins (Fig. 1). Creatine kinases are highly modulated in response to physiological stimuli, but surprisingly little is known concerning their transcriptional regulation in muscle cells .
2.2. Impacts of heart failure
2.2.1. Cardiac energy metabolism in heart failure
Early, the failing heart has been described as energy starved . Indeed, in the stage of heart failure, oxidative capacity of the myocardium decreases [42–44]. In addition, HF also impairs energy transfer and utilization. A generalized alteration of the creatine kinase system has long been observed with both cytosolic and mitochondrial isoenzymes being affected (for review see [4,5,26]). Furthermore, the failing heart has reduced mechanical efficiency, increasing the energy cost of contraction [3,45]. This energetic imbalance is the major cause for the lower PCr/ATP ratio that turned out to be a valuable predictor of mortality in CHF patients .
Decreased mitochondrial capacity in heart failure seems to be due to alteration in mitochondrial biogenesis, the whole PGC-1α/NRFs/Tfam transcription pathway being down-regulated [5,47]. Finally, pressure overload-induced hypertrophy results in deactivation of PPARα and subsequent dysregulation of fatty acid oxidation enzyme gene expression [48,49]. Which signals trigger the drop in creatine kinase in heart failure, remain to be established.
2.2.2. Skeletal muscle energy metabolism in heart failure
Changes in skeletal muscle morphology, metabolism and function in CHF patients include muscle atrophy, decreased vascularisation, fiber type shift towards faster phenotype, and decreased resistance to fatigue . The skeletal muscle metabolism of CHF patients exhibits greater changes and delayed recovery of PCr/ATP ratio following exercise . In animal models, defects in expression and activity of mitochondrial enzymes lead to decreased muscle oxidative capacity, and altered mitochondrial regulation . This supports that mechanisms underlying these energetic abnormalities could be related to a generalized disturbance of mitochondrial gene expression, and decreased PGC-1α expression as observed in experimental heart failure [47,52]. Creatine kinase expression is altered in skeletal muscles, in experimental heart failure in a muscle-type specific manner with the mi-CK subunit being the most affected , and in patients . Thus, metabolic alterations in heart failure affect both cardiac and skeletal muscles, suggesting a generalized metabolic myopathy in this disease .
Impairment in skeletal muscle mitochondrial function in patients deserves further consideration. Morphometric and biochemical studies of vastus lateralis from CHF patients revealed decreased volume density of mitochondria and oxidative enzymes in proportion with the decrease in peak oxygen consumption (VO2 peak) [55,56]. Recent studies have demonstrated however that in situ muscle oxidative capacity, mitochondrial ATP production and the transcriptional cascade PGC-1/NRFs/Tfam is identical in CHF patients and sedentary subjects [53,57,58]. One possible explanation is a beneficial effect of recent HF therapy . Indeed, angiotensin-converting enzymes inhibitors (ACEi) can be protective for mitochondrial function and biogenesis both in heart and skeletal muscle as suggested by results obtained in experimental heart failure [59,60]. Nevertheless, defects in creatine kinase and citrate synthase activity  are still observed suggesting the persisting metabolic defects in skeletal muscle in CHF patients.
The ageing processes in skeletal muscle shares many aspects in common with heart failure including muscle wasting, endothelial dysfunction, increased oxidative stress and muscle cytokines, apoptosis, and impaired mitochondrial function [61,62] participating in muscle energetic dysfunction. Thus the overlap of aging and CHF-associated changes may partially explain the more severe disabling consequences of the CHF syndrome in elderly patients .
3. Effects of physical training on cardiac and skeletal muscle metabolism
Regular exercise positively influences the cardiovascular system and is a determinant of health and quality of life. Endurance training improves heart and skeletal muscle energy metabolism and function.
3.1. Exercise training and cardiac metabolism
Adaptive responses of the heart to endurance training include resting and submaximal exercise bradycardia and increase in end-diastolic dimension. This leads to non-pathological cardiac hypertrophy, improved ventricular function and increase in the resistance of the heart to ischemic insult . However, whether changes in energy production and transfer occur to preserve or improve cardiac function is unclear.
In animal models, while some studies showed that regular endurance exercise increases glycolysis and oxidative metabolism [64,65], others demonstrated that adaptive responses result from increased muscle mass, rather than mitochondrial gene expression [66,67]. Exercise training in rats may also protect against aging-induced increase in oxidative stress . Data from previous studies suggest that the increase in cardiac mitochondrial capacity with training if any is subtle. Moreover, either increase or no change in fatty acid utilization capacity were reported [69,70]. These findings suggest that the exercise-induced cardiac hypertrophy mostly results in normal energy production from fatty acid metabolism. Little is known concerning creatine kinase expression and function in trained heart. Aerobic exercise training increases total myocardial CK activity and MB–CK content in canine left ventricular myocardium  but the physiological relevance of this observation is not obvious. Thus whether the heart adapts to repeated exercise by improving energy synthesis and energy fluxes or both await further investigation. Whether these results extend to human heart remains to be established.
3.2. Exercise training and skeletal muscle metabolism
Skeletal muscles adapt to repeated prolonged exercise by marked quantitative and qualitative changes in mitochondria, capillary supply, but only limited transitions in MHC isoforms [11,15,27].
Endurance training promotes an increase in mitochondrial volume density and mitochondrial proteins, in all three fiber types . A decreased mitochondrial permeability to ADP and enhanced functional role of mi-CK, occurs also in both glycolytic and oxidative fibers, favoring closer coupling between energy production and utilization [73,74]. The specific effects of increased contractile activity are caused by several factors, including nerve-dependent parameters, acute and sustained changes in intracellular free calcium concentration ([Ca2+]i), metabolic factors such as hypoxia, and/or mechanical load .
Whereas muscle responses to endurance training are now well described at the phenomenological level, much less is known about the intracellular pathways linking increased functional demand to changes in gene expression. PGC-1α plays a key role in regulating mitochondrial biogenesis in skeletal muscle during endurance training (Fig. 1) , and correlates with the training status of healthy individuals . Moreover, proteins involved in mitochondrial dynamics and protein assembly correlate with PGC-1α, extending the role of this co-activator to the proper assembly of protein subunits of mitochondrial and nuclear origin . A strong correlation has been observed between calcineurin activation, a calcium regulated phosphatase, and PGC-1α expression, suggesting that this metabolic transcriptional co-activator could be regulated by changes in intracellular calcium levels. However, calcineurin inhibition fails to block the exercise induced PGC-1α expression and activation of mitochondrial biogenesis either in rats  or in transplanted patients after cyclosporin treatment . Thus, other pathways like MAPK pathway and CaMKs, also activated during exercise, are likely involved in the control of muscle oxidative capacity [77,78]. The increased energy demand during sustained muscle contraction and the resulting activation of energy-sensing pathways also seem to be important for the induction of PGC-1α mRNA. The AMP sensitive protein kinase (AMPK), which is a sensitive indicator of reduced cellular energy status in skeletal muscle, is chronically activated during periods of repeated exercise , and is involved in PGC-1α transcription and mitochondrial biogenesis as demonstrated in animal models [80–82]. However the link between AMPK activation and transcription of PGC-1α remains to be determined. A decrease in intracellular PO2 occurs within the first seconds of muscle contraction. Local hypoxic areas may thus occur during prolonged exercises and in pathologies affecting peripheral circulation like heart failure. The hypoxia-inducible factor (HIF) family of transcription factors plays a critical role in mediating the hypoxic regulation of several genes  including the vascular endothelial growth factor (VEGF). Components of the HIF-1 pathway are activated by physical activity in healthy human skeletal muscle . The exercise-induced increase in VEGF gene transcription occurs especially within type IIb glycolytic myofibers, which are more susceptible to local hypoxia . This fine regulation may be of importance in HF where heart and skeletal muscle are prone to experience episodes of local hypoxia.
4. Relevance of exercise training to heart failure
Endurance training induces an improvement of muscle resistance to fatigue and justifies that endurance training programs could be considered as countermeasures of muscle weakness in heart failure patients (Fig. 2). Indeed, exercise training is able to oppose the deleterious effects of heart failure on skeletal muscle energy metabolism, although this is less clear for cardiac muscle.
4.1. Effect of exercise training on cardiac metabolism in heart failure
Whether consistently observed beneficial effects of exercise training of CHF patients result from cardiac improvements is still an open question . An array of arguments suggests that cardiac energy metabolism may be improved, or at least preserved. Some studies have been performed on animal models of myocardial infarction. A protective effect of training prior to myocardial infarction reduces infarct size , ventricular enlargement, improves remodeling, increases systolic function, and improves expression of cytochrome oxidase subunits, ventricular atrial natriuretic peptide, SR calcium ATPase and fatty-acid binding protein . Exercise training decreases apoptotic processes, and protects mitochondrial function from oxidative stress and other cardiac insults . In rats with myocardial infarction, aerobic endurance training attenuates ventricular and cellular hypertrophy and consistently restores contractile function, intracellular Ca2+ handling, and Ca2+-sensitivity in cardiomyocytes , thereby possibly increasing energetic efficiency.
Observations indirectly suggest that exercise training can improve myocardial energy metabolism. Exercise improves coronary blood flow and endothelium-dependent vasodilatation, and increases resistance vessel sensitivity to the metabolic mediator adenosine, ameliorating oxygen and substrate supply to the failing heart in coronary artery diseases in humans  and pacing induced heart failure . In humans, better left ventricular diastolic filling, increased ejection fraction and stroke volume, decreased end diastolic LV dimension and increased cardiac output at peak exercise, suggest improved cardiac load and efficiency for a given work rate [16,23,91,92]. Moreover the deleterious neuro-endocrine and pro-inflammatory cytokine burden of CHF are decreased by exercise training [93–95]. This “anti-inflammatory” effect of exercise training in CHF may further decrease the oxidative stress within cardiac tissue and therefore improve intracellular energetics.
However, direct evidences for beneficial effects of exercise training on cardiac energy metabolism are sparse. In experimental heart failure, physically active animals did not modify oxidative capacity, oxidative enzymes and energy transfer enzyme activity compared to their sedentary counterparts, but irreversible aortic stenosis may have precluded the potential beneficial effects of exercise training on cardiac metabolism . One recent study in CHF patients with non-ischemic cardiomyopathy measured oxidative metabolism by [11C] acetate and positron emission tomography . Training induced a coordinated decrease in myocardial O2 uptake indicating improved forward work efficiency. Therefore, exercise training in CHF seems to improve cardiac energetic efficiency.
Because clear beneficial effects of training in terms of ventricular remodeling, efficiency and even patients survival has been shown after myocardial infarction , exercise rehabilitation has become standard of care in this setting , provided that exercise testing and training are not initiated within the first few days of the acute event when arrhythmias and adverse remodeling can occur . It remains to be established whether this results from the effects of training on ventricular geometry and heart rate only, from better energetic efficiency of the cardiomyocyte, or from improved intrinsic oxidative capacity and intracellular energy transfers in human failing cardiomyocytes or a combination of these factors.
4.2. Effect of exercise training on skeletal muscle metabolism in heart failure
Effects of exercise training on skeletal muscle of CHF patients are more documented. Increases in VO2 peak have been consistently reported with exercise training, together with improvements in global indices of muscle metabolism like lesser PCr decrease, lower Pi/PCr ratio and higher PCr resynthesis rates for a given workload after training [101–104]. This improved overall metabolic balance for a given workload can result from improved O2 and substrate supply to skeletal muscle tissue, increased tissue oxidative capacity, and/or more efficient energy utilization.
4.2.1. O2 and substrate supply to skeletal muscle
One of the first studies that examined the effects of training in CHF patients showed increased peak leg blood flow to active skeletal muscles and more efficient peripheral oxygen extraction after training, without improved submaximal exercise leg blood flow . Nevertheless, the CHF-induced decrease in capillarity is improved after training [106,107]. Moreover, the decreased endothelium-mediated vasodilatory properties of the skeletal muscle during CHF are at least partly corrected, at the local and systemic levels [8,17,108]. This improvement in flow mediated vasodilatation by exercise training in CHF is accompanied by enhanced expression of eNOS  and/or increased activity of vascular antioxidative enzymes . Thus, exercise training has been proposed as an effective antioxidant and antiatherogenic therapy at the vascular level which is able to correct and/or improve endothelial dysfunction . Therefore, despite unchanged leg blood flow, a favorable intramuscular redistribution of nutritive flow and oxygen may occur after training. It remains that these hypotheses still await more conclusive evidence.
4.2.2. Skeletal muscle oxidative capacity and energy transfer
In chronic heart failure patients, endurance training reduces phosphocreatine depletion and ADP increase during exercise, and enhances the rate of phosphocreatine resynthesis after exercise indicating a substantial improvement of skeletal muscle oxidative capacity . Increase in mitochondrial volume density positively correlates with changes in VO2 peak and anaerobic threshold exercise  or changes in type I fiber percentage, and negatively with muscle venous lactate for a given workload . As exercise training in heart failure patients increases mitochondrial size , it is likely that it will also increase muscle oxidative capacity in skeletal muscle, resulting in increased exercise tolerance. In older human skeletal muscle, exercise enhances mitochondrial activity, which is likely related to the concomitant increase in mitochondrial biogenesis . In CHF rats, muscle oxidative capacity increases with physical activity in muscles recruited during exercise  and in vastus lateralis muscle of patients after cardiac transplantation . Whether this increase in mitochondrial density and oxidative capacity with training in CHF occurs because of increased coordinated transcription of nuclear and mitochondrial genes is presently undemonstrated. However, this hypothesis is likely because of the coordinated changes in PGC-1α expression, muscle oxidative capacity, VO2 peak and the training status in healthy individuals and in CHF patients .
In heart failure patients, abnormal local generation of NO by over-expression of iNOS, increases production of reactive oxygen species and enhances cytokine levels creating local deleterious conditions for optimal mitochondrial functioning, thus decreasing the functional oxidative capacity of muscle tissue [110,116,117]. Exercise training partially reverses these cellular events by decreasing tissue iNOS, cytokine production, local ROS generation and AngioII-mediaed vasoconstriction [118,119], and by enhancing expression of antioxidative enzymes , thus having anticatabolic effects on one hand and restoring better environmental conditions for mitochondrial energy production on the other. Indeed, training decreases local cytokines and partially reconstitutes oxidative defenses, resulting in decreased muscular damage and apoptosis . Nevertheless the direct demonstration of the effects of oxidative stress on functional mitochondrial oxidative capacity is still lacking.
As stated above, intracellular energy transfer through creatine kinase is also limited in heart failure [51,57]. These abnormalities persist in ACEi treated patients [53,57]. In heart transplant recipient, a significant increase in mi-CK activity is observed with training without changes in M–CK . In rats, voluntary wheel running increases both M–CK and mi-CK in soleus muscle, but mi-CK is still lower than in control rats . Interestingly, running performance strongly correlates with mi-CK activity both in rats  and CHF patients  further underscoring the importance of tight integration between oxidative energy synthesis and energy utilization in determining muscle performance [73,74]. Nevertheless, direct assessment of exercise training on mitochondrial function and energy transfer in heart failure still deserve further consideration.
Considering the similarity between muscle disability in aging and heart failure, it is not surprising that healthy or CHF elderly highly benefit from exercise training in terms of exercise tolerance, contractile, mitochondrial and endothelial functions, cardiovascular endurance and oxidative defenses [61,62,115,121].
Heart failure induces a metabolic myopathy affecting both heart and skeletal muscles. This manifests itself mainly by decreased oxidative capacity, shift in substrate utilization and altered energy transfer by phosphotransfer kinases. In skeletal muscle, endurance exercise capacity is mainly conditioned by increased oxidative capacity, increased lipid utilization, and improvement of energy fluxes and better coupling between energy production and utilization. Prolonged exercise is thus able to counteract these deleterious effects by improving oxygen and substrate delivery, as well as metabolic remodeling of cardiac and skeletal muscles. Although beneficial effects of endurance training in heart failure are indubitable, further work is needed to delineate the pleiotropic effects of physical activity on cardiac and skeletal muscle functions. This issue is of interest for clinical output, especially for rehabilitation of patients with heart failure.
If the beneficial effects of endurance training in CHF patients are well established in terms of exercise capacity, quality of life and even morbi-mortality, there is increasing evidence that it is at least not harmful and may even be beneficial for the failing heart itself. If by further studies this proves to be true in human patients, training should be implemented in the care standards, together with β-blockers and medication antagonizing the renin–angiotensine–aldosterone system. Further research examining the mechanisms of these beneficial effects at the whole organ and cellular levels will be of vital importance to identify potential advantageous effects of pharmacological and physical therapy. The specific effects of other kinds of exercise such as resistance exercise, alone or in combination with endurance exercise on muscle mass and energy metabolism also should be investigated in the future.
R.V.-C. is supported by the Centre National de la Recherche Scientifique. This work was supported by an INSERM PROGRES, an “Association Française contre les Myopathies” and a “Fondation de France” grants.