Muscle Remodeling and the Exercise Physiology of Fish

McClelland, Grant B.

Exercise & Sport Sciences Reviews:
doi: 10.1097/JES.0b013e3182571e2c

Fish muscle responds to aerobic exercise training and cold acclimation with a more aerobic muscle phenotype than mammalian muscle but through both conserved and distinct molecular events. Differences from mammals in exercise metabolism and diversity in protein isoforms suggest that the regulation of muscle fuel use is more complex in fish. This review considers fish as powerful models for exercise and muscle physiology.

In Brief

Common and distinct responses of fish to exercise compared with mammals make them a powerful model for muscle and exercise physiology research.

Author Information

Department of Biology, McMaster University, Hamilton, Ontario, Canada

Address for correspondence: Grant B. McClelland, Ph.D., Department of Biology, McMaster University, 1280 Main St. West, Hamilton, Ontario, Canada L8S 4K1 (E-mail:

Accepted for publication: March 16, 2012.

Associate Editor: Benjamin F. Miller, Ph.D., FACSM

Article Outline
Back to Top | Article Outline


Exercise physiology has a long history of examining human performance and muscle function. Great strides have been made in understanding how to enhance athletic performance and what changes occur in muscle metabolic capacities with repeated contractions. These many years of research have enhanced our understanding of muscle physiology but also benefited our ability to help counteract many diseases. However, our knowledge is not complete, and new techniques and new approaches can provide unique insights into muscle metabolism. Here, I argue that using nonmammalian vertebrates, such as fish, can be a powerful approach to understanding basic principles that govern the physiology of vertebrate muscle. The goal of this article is not to provide a comprehensive review of exercise research in fish because this has been covered elsewhere (12). Indeed, two main features of fish will be highlighted: 1) the common and distinct mechanisms of mitochondrial biogenesis and muscle remodeling when compared with mammals and 2) the potential implications of an increased diversity in protein isoforms to the regulation of exercise metabolism. Also reviewed is our current understanding of how fish muscle powers both exercise and recovery.

Back to Top | Article Outline


The study of exercise metabolism in fish has been an active area of research for more than 50 yr (see (12) for a recent review). Many aspects of locomotion have been examined in fish — including the biomechanics of force generation, the biochemical properties of muscle, to the ecological implications of exercise capacity. From an applied perspective, exercise is of interest to some fish physiologists because it stimulates growth and, therefore, has applicability to the aquaculture industry. On the other hand, basic science researchers have long been fascinated with the diversity in locomotory ability between fish species: most notable is the extreme exercise performance involved in seasonal migrations (e.g., salmon (41)) and other high-performance phenotypes (e.g., tuna, billfish (3)). Fish species differ in their tolerance to varying environmental temperature, water salinity, and dissolved oxygen content, which impacts muscle characteristics and performance but this has been examined only on a limited number of species. Although not discussed in detail in this review, within a single species, there can be significant muscle specialization not only for locomotion but also for regional endothermy (e.g., heater organs of billfish), for the generation of electric currents (weak electric fish, electric eels), and for sound production (sonic muscle of gulf toadfish). There has been less research into the plasticity of muscle to either environmental or energetic stress. What little work that has been done reveals that fish muscle is very responsive to exercise and readily remodels both structurally and biochemically (24). As will be discussed in the next section, this plasticity of adult fish muscle is one aspect that makes fish a powerful model for exercise and muscle research.

Back to Top | Article Outline


For decades, it has been known that repeated muscle contractions lead to a remodeling of mammalian muscle. Principally studied in humans and rodents, many significant advances have been made in our understanding of the muscle response to exercise and limitations in muscle function imposed by disease. However, there remain many unanswered questions regarding the molecular signals responsible for these changes in muscle phenotype. Recently, there has been great interest in which molecular events result in an increase in muscle mitochondrial volume and muscle aerobic capacity. Powerful techniques have been developed to dissect the mechanisms (knockout, knock-in mice), and selection experiments with rodents have resulted in differences in complex traits such as aerobic capacity. However, these models also have their limitations. For example, the forward genetic techniques commonly used in invertebrates cannot be done economically in rodents. Can exercise science research questions be addressed using an alternate vertebrate experimental model that responds similarly to exercise as mammals but has added advantages? Model fish species such as zebrafish have been long favored for developmental studies because of their external fertilization, transparent embryos, and amenability to forward genetic techniques. However, adult fish of many species hold a number of advantages, especially for muscle physiology and exercise research. Two important developments have occurred to make fish even more attractive model organisms for muscle research: 1) the sequencing and annotation of the genomes of many fish species help simplify the elucidation of mechanisms, and 2) the accumulated evidence of many shared biological traits between fish and mammals support the use of these nonmammalian vertebrates as windows into human physiology and disease (39). Evolutionary conservation of biochemical pathways involved in carbohydrate, lipid, and protein metabolism suggests that fish may power aerobic exercise similarly and that fish muscle may respond in a similar fashion to stress, as do mammals. Any distinct responses of fish muscle to stimuli can not only inform us of basic physiological processes across taxa but also provide clues of putative alternate therapeutic targets for muscle disease. As I will review, fish do not necessarily adhere to the standard mammalian pattern of fuel use and appear to recruit fuel sources that do not play a major role powering locomotion in mammals. We also will see that fish muscle responds in common and distinct ways to stimuli to produce a more aerobic phenotype. Along with these novel biological responses, there are many other reasons to adopt fish as experimental models for exercise and muscle research (see Table).

1) Fish have a skeletal system consisting of cartilage and bone. Skeletal, smooth, and cardiac muscle have similar characteristics to homologous tissues in mammals.

2) Fish have a reduced effect of gravity on locomotion because they match their body density to that of water. This allows researchers to study nonload-bearing forces involved in swimming without the confounding effect of gravity.

3) Because fish have anatomically separated pure fast (white) and slow (red) muscle masses, true fiber type‐specific responses to acute and chronic exercise can be determined easily. Developmental differences have led to fuel use and biochemical differences between these tissue types. Red muscle exists as a wedge running along the lateral line in most species and is recruited for steady-state swimming (up to 70%–80% of critical swimming speed (Ucrit)) because of its high mitochondrial content (as indexed by citrate synthase (CS) (31)) and enhanced ability to take up and oxidize fatty acids probably because of higher carnitine palmitoyltransferase I (CPT-I) activities per unit mitochondria (25,35). The larger white muscle mass makes up the bulk of the animal’s trunk and is heavily recruited for burst exercise. This muscle mass has much lower mitochondrial density than red muscle and is specialized for carbohydrate metabolism (aerobic and anaerobic). This intertissue variation in aerobic capacity and metabolism can be used as a convenient model to examine the regulation of muscle phenotype. For instance, differences in expression of peroxisome proliferator‐activated receptor (PPAR) (25), PPAR-gamma coactivator 1α (PGC-1α) (16,17), CPT-I isoform (26) expression, and mitochondrial membrane fluidity (25) represent potential mechanisms for tissue-specific differences in muscle mitochondrial content and quality. Fish also are good models for muscle growth (31), which can be stimulated by exercise, and some species show either determinant growth, like mammals, or indeterminate growth with examples of either hypertrophic or hyperplastic growth.

4) Fish physiology and exercise physiology of fish have been active areas of research for over 50 yr. This provides an extensive database on many physiological aspects of key model species such as trout (Oncorhynchus mykiss), zebrafish (Danio rerio), stickleback (Gasterosteus aculeatus) and killifish (Fundulus heteroclitus). Many aspects of fish biology are well understood, such as the gill’s role in gas transfer, ionoregulation, and handling of nitrogenous wastes. The cardiovascular physiology of fish also is well studied, as is the biomechanics of burst and sustained swimming. Standardized performance tests for burst, sustained, and prolonged swimming have been firmly established (12).

5) There are tens of thousands of species of fish, many with novel life histories and physiologies, living in a variety of environments in the wild, that are ripe for exercise physiology hypothesis testing. Most work to date has been done on the salmonids, mainly trout, and this group tends to perform well in burst and steady-state exercise. Other groups and species are ideal models for alternate metabolic strategies. For example, the sharks and rays (elasmobranches) have extremely low to undetectable activities of CPT-I in muscle while expressing other enzymes in this pathway (23,34). Thus, elasmobranches such as dogfish (Squalus acanthias) are essentially natural knockout models that can be used to examine exercise metabolism with the loss of muscle lipid oxidation capacity. Muscle specialization for novel function also has led to variations in mitochondrial content in, for example, heat-producing organs and the fastest contracting skeletal muscle known in vertebrates, the toadfish sonic muscle. Skeletal muscle mitochondrial content varies 20-fold between relatively sluggish species (e.g., carp) to high-performance athletic species (e.g., tuna and billfish (27)). This diversity can be used to study the regulation of constitutive mitochondrial content. There also are many species with adaptive differences in thermal (4) and hypoxia tolerance (32), which affects muscle phenotype and capacity for phenotypic plasticity.

6) Current model species (e.g., zebrafish) are popular because of their low cost of housing, short generation times, comprehensive genomic databases, commercially available microarrays, and the ability to do large genomic and mutant screens. With increasing availability of genetic information for other fish species, such as pufferfish, trout, Atlantic salmon, three-spine stickleback and medaka, reverse genetic techniques such as RNAi, morpholinos, and transgenic generation now are being used routinely in many of these species. Therefore, there are few limitations to performing mechanistic physiology studies using a variety of fish species.

7) Swim training and even swim performance testing (24) can be carried out concurrently in large groups of animals, minimizing variation in experimental conditions. These many advantages (Table) make fish attractive models to compliment exercise and muscle research in mammals.

Back to Top | Article Outline


Response to Exercise

Fish muscles show a large range of phenotypic plasticity in terms of their metabolism and size both developmentally (10) and in response to energetic and environmental stress (e.g., 16,17,24). Changes in muscle also have been examined with increased use or disuse in developing zebrafish (39). Adult fish show significant muscle remodeling in response to chronic locomotion, with changes in muscle growth and metabolic phenotype. In fact, they respond similarly to low- to moderate-intensity swimming, as do mammals to endurance training, where repeated contractions result in a more aerobic muscle phenotype (17,24). Therefore, fish are a suitable model to examine basic principles of muscle remodeling in vertebrates and, more specifically, molecular events responsible for mitochondrial biogenesis. Many of the mechanisms of exercise-induced mitochondrial biogenesis in muscle are now known for mammals. Muscle contractions and increased adenosine triphosphate (ATP) turnover rates lead to bursts in cytosol Ca2+ that signal gene expression through activation of various kinases, including Ca2+/calmodulin-dependent protein kinase II (CAMK-II), CAMK-IV, CAMK-Kβ, and p38 mitogen-activated protein kinase (p38MAPK), which in turn activate p38 kinase. Secondary to the Ca-induced response are changes in adenosine monophosphate (AMP):ATP ratios, which activate the cellular energy sensor AMP-activated protein kinase (AMPK). Reactive oxygen species (ROS) and catecholamine stimulation also play roles in this process, and all these upstream activators result in an upregulation of PGC-1α, a cofactor to the nuclear receptors (PPAR) that regulate components of the lipid oxidation machinery, the nuclear respiratory factors (NRF) that regulate oxidative phosphorylation components, and the estrogen-related receptor-α (ERR-α) that regulates the Krebs cycle enzyme expression ((8) Fig. 1). Other isoforms in the PGC family of transcription factors, PGC-1β and the PGC-1‐related coactivator (PRC), which are expressed in mammals and fish muscle (18), also play roles in regulating mitochondrial biogenesis. Very few of these regulators of muscle phenotype had been studied in any detail in fish, especially in terms of exercise.

To address this gap in the literature, we examined the effect of exercise on muscle remodeling and mitochondrial biogenesis in nonmammalian vertebrates using zebrafish. Adult zebrafish were made to swim against a current and in large groups (∼60 fish) 6 d wk-1 for 4 wk at an intensity equivalent to 25% to 30% of their maximal sustainable swimming speed (Ucrit, see later section for a description of performance testing in fish). A mixed tissue sample composed of red and white muscle was sampled and analyzed for changes in maximal in vitro enzyme activities (apparent Vmax) as indices of the capacity for lipid and carbohydrate oxidation, glycolytic flux, and mitochondrial density. Because the regulation of this muscle plasticity with exercise had not been examined previously at the molecular level, we also measured changes in messenger RNA (mRNA) levels for the enzymes and transcriptional regulators shown to be important for muscle remodeling in mammals (8). We found that adult zebrafish muscle was highly responsive to repeated contractions, most notably the induction of CS and cytochrome oxidase enzyme activities, suggesting quantitative and qualitative changes in mitochondria. This confirmed that there is a conserved response between mammals and fish to low-intensity endurance training, the stimulation of mitochondrial biogenesis. What emerged from these data was that some of the putative mechanisms regulating these responses appear to be common in mammals and fish but that fish also show distinct changes in mRNA expression that contribute to the exercise-induced phenotype. Notably, a regulator of respiratory gene expression, NRF-1, increased in expression by four times, which correlated with a two times increase in CS mRNA expression, a response in common with mammals. Other important transcription factors that regulate mitochondrial biogenesis and phenotype in mammals, PGC-1α and the PPAR family of nuclear receptors, behave quite differently in fish. In mixed muscle of zebrafish, mRNA levels of PPAR-α, PPAR-β1, and PGC-1α either significantly decreased or did not change significantly with training (Fig. 2). This is intriguing because all three PPAR isoforms (α, β/δ, γ) found in mammals and PGC-1α are elevated with exercise in humans. We also found that zebrafish red and white muscles show distinct responses during an 8-wk course of training. Peak mRNA expression of PGC-1α occurred early in training in both muscle types, whereas PGC-1β only increased in red muscle and early in training (Fig. 2). We also tested the hypothesis that multiple individual isoforms of the PPARs that are found only in fish might respond uniquely to exercise. However, we found that neither PPAR-β1 nor PPAR-β2 expression was induced in either muscle fiber type by training (17). Interestingly, zebrafish red and white muscles have distinct responses to a single bout of moderate exercise. By 24 h of recovery after a single 3-h swimming session of moderate intensity, CS mRNA increased significantly. In red muscle, PGC-1α was increased immediately after exercise, whereas it took 2 h to peak in white muscle. However, PGC-1β mRNA expression was not induced by this single exercise session (17). Taken together, the acute and chronic exercise responses in fish muscle suggest that PGC-1α is not as directly involved in the regulation of CS expression as it is in mammals. This is an important aspect of fish muscle remodeling given the more direct role of PGC-1α in regulating mitochondrial biogenesis in mammals. Many questions still remain regarding the regulation of CS and other respiratory gene expression in fish. For instance, are there additional regulatory steps that explain the temporal separation between peak PGC-1α and CS expression or do fish use a novel regulatory pathway to control the induction of CS gene expression? PGC-1α does seem to coordinate regulation by the PPAR but not NRF-1 (16,18), and NRF-1 expression correlates well with that of CS in zebrafish (24) as it does in mammals.

Other known regulators of muscle metabolic remodeling (e.g., PRC, silent mating-type information regulator 2 homolog 1, NRF-2, p53) have yet to be examined in fish, at least in terms of exercise. Moreover, there has been increased interest in the role of hypoxia inducible factor-1α in response to endurance training and also acute and chronic hypoxia signaling during exercise in mammals (19). Given the evolutionarily conserved nature of this transcription factor, fish may respond in a similar manner to mammals to stressors that induce cellular hypoxia such as intense exercise. However, it remains to be seen if some or all of these regulatory pathways play common or distinct roles in muscle remodeling across taxa.

Exercise is a growth stimulant in fish, and it has been argued that it is this feature that makes them a valuable exercise model (31). Endurance exercise training can result in the hypertrophy of the different muscle types, with an increase in red muscle mass in some fish species. However, after 4 wk of training in zebrafish, there was no change in the absolute or relative red muscle mass assessed by histochemistry or any change in mRNA expression of either fast or slow myosin heavy chain, at least in a mixed muscle sample (17,24). Exercise also causes cardiac growth in fish (5), and zebrafish are known to regenerate damaged cardiac tissue (11), making them an attractive model for studies of physiological cardiac hypertrophy or to examine the potential role of exercise in muscle repair pathways.

Back to Top | Article Outline
Response to Temperature

There are other stimuli known to induce remodeling in mammalian muscle, including shivering thermogenesis, chronic electrical stimulation, and hyperthyroidism. Although there are differences in how these stimuli induce remodeling, they all cause hypermetabolism in mammals. In contrast, nonmammalian ectothermic vertebrates undergo muscle remodeling in response to chronic cold exposure, which may coincide with a decrease in metabolic rate. Fish respond to a chronic decrease in temperature by increasing mitochondrial content. This helps compensate for the deceleration effect of low temperature on metabolism because it enhances diffusion rates by decreasing diffusion distance (29). In fact, in many fish species, cold-induced skeletal muscle remodeling occurs to the same extent as exercise training (e.g., 24). It is likely that this temperature-induced muscle remodeling occurs through a separate set of molecular events than those involved in contraction-induced remodeling seen in exercised fish and endurance-trained mammals (16,24,29). When zebrafish are exposed to a 10°C drop in ambient temperature, a more aerobic muscle phenotype emerges (24). Although mitochondrial quantity appeared to increase by an equivalent amount to that seen in exercise, mitochondrial quality differed as cold acclimation increased the capacity for fatty acid oxidation, whereas chronic exercise did not (24). However, this is not true for all fish species because exercise training can increase β-oxidation enzyme levels in trout (9). Cold-induced and exercise-induced changes in gene expression were distinct in zebrafish. After 4 wk of cold exposure, zebrafish showed no changes in mRNA for several metabolic genes, including CS, and only a modest 2.3 times increase in NRF-1 mRNA compared with a 4 times increase with exercise. Cold also caused a large decline in PPAR-α but no change in the PPAR-β1 isoform (which declined with exercise). A similar gene expression response to cold has been seen with a 6-wk acclimation to 4°C in goldfish (Carassius auratus (16)). Longer cold exposures in threespine stickleback did result in an increase in CS mRNA but only after more than 60 d of cold acclimation (29), well after zebrafish showed increased CS enzymatic activity (24).

The upstream events of temperature-induced remodeling are unclear but most likely are distinct from those involved in exercise-induced remodeling (Fig. 1). It is unlikely that temperature acts via Ca2+-associated signaling, as seen with prolonged contractions. Although the importance of these pathways for exercise-induced muscle remodeling in fish remains underexplored, a protein phosphatase important for muscle remodeling in mammals, calcineurin, does not seem to play a major role in muscle hypertrophy associated with swim training in trout. In fact, calcineurin’s target protein NFAT2 has been found to decrease relative to levels in untrained controls (22). Neither calcineurin nor other key components of this signaling pathway have not been examined in cold-acclimated fish, but other regulators of muscle phenotype have been implicated in response to temperature. It has been proposed that ROS increases as temperature decreases, and this signal can activate the observed muscle mitochondrial biogenesis. However, cold exposure in ectotherms may actually reduce mitochondrial ROS production, and cold is known to lead to the increased expression of genes for ROS-scavenging enzymes in zebrafish (24). An alternative hypothesis suggests that, at low temperatures, turnover of mRNA and proteins decreases or that translational efficiency increases, leading to greater mitochondrial volumes (38); but the data in support of this hypothesis remain equivocal. Perhaps because cold acclimation can decrease energy production in fish, ADP:ATP ratios may rise and trigger changes in AMPK, affecting transcription factor activity. AMPK has been shown to respond to changes in cellular energy status in fish (14), as it does in mammals, but its role in either temperature- or exercise-induced muscle remodeling is currently unclear. Thus, a great number of unanswered questions remain regarding the signaling pathways that lead to muscle remodeling in fish. However, the ability of fish muscle to respond to divergent stimuli that likely act through different and novel pathways to those pathways used in mammals makes fish an attractive model for muscle research. In addition, cold acclimation represents a simple experimental perturbation that readily will induce muscle remodeling and does not require any specialized equipment such as swim flumes or treadmills and exercise ergometers as in the case of humans.

Back to Top | Article Outline


Ancient whole genome duplication events in vertebrates have resulted in multiple copies of certain genes. In fact, additional duplication events in the ray-finned fish give them the potential of expressing eight genes for every one seen in invertebrates and for every four genes seen in mammals (26). These duplicated genes represent a major source of biological diversity and the potential for more complex metabolic regulation in the fish compared with mammals. However, not all duplicated genes are retained after duplication, and these extra genes may undergo 1) pseudogenization, where the extra copies become functionless or unexpressed; 2) subfunctionalization, where the duplicated copies each adopt only part of the functions of the ancestral gene; or 3) neofunctionalization where, as the name suggests, the duplicated gene evolves a novel function (26). For example, whereas mammals express three isoforms of CPT-I (α, β, C), three isoforms of PPAR (α, δ/β, γ), and two PGC-1 (α and β), after fish and mammals diverged, there were several lineage-specific genome duplication events in fish, resulting in multiple isoforms of these and other proteins for which mammals may have fewer or even a single copy. Many genes, which may have undergone neofunctionalization or subfunctionalization, have not been identified, fully sequenced, or kinetically characterized in fish, but three important ones that have are CPT-I (26), PGC-1 (18), and some of the PPARs (17). What these studies reveal is that some regulators probably do not retain extra copies of the gene (e.g., PGC-1α), whereas others have undergone subfunctionalization or neofunctionalization (e.g., PPAR). CPT-I is another gene that also has retained extra copies of duplicated genes in fish. Past work on trout showed that the kinetics of this enzyme are different in fish muscle and liver compared with the same tissues in mammals. Trout red muscle CPT-I is more sensitive to inhibition by malonyl-CoA than liver CPT-I (25), opposite to the pattern seen in mammals. Trout express five isoforms (named CPT-Iα1a, α1b, α2, β1, β2 (26)) and, in contrast to mammals where isoform expression is tissue specific, all five isoforms are expressed across all tissues in trout. Sensitivity of CPT-I to its allosteric regulator can be attributed to binding site structural changes and its interaction with the protein’s transmembrane domains. Trout isoform differed in the position of transmembrane domains and the predicted amino acid substitutions of these different proteins when compared with how mutations in mammalian CPT-I affect enzyme regulation suggest that the isoform composition of a given tissue would impart it with unique fatty acid oxidation kinetics. Moreover, muscle CPT-I isoform composition is malleable across development (26) and fasting in trout and with migration in the case of Pacific salmon (Morash, AJ, Yu, W, Le Moine, C. et al., unpublished data, 2012). However, the consequences of multiple isoforms of other key regulators, transporters, and enzymes of muscle metabolism have not yet been studied in detail in fish. Although fish may recapitulate mammalian muscle phenotypes in response to energetic cues, the increased cellular complexity caused by neofunctionalization and subfunctionalization of protein isoforms suggests that they may use a novel regulation of metabolism. For example, fish may serve as natural models to examine the integrated effects of amino acid substitutions on the activity and regulation of naturally occurring isoforms. Moreover, changing isoform composition with exercise could reveal important aspects regarding altered kinetics of key components of muscle metabolic pathways.

Back to Top | Article Outline


Linking muscle phenotype with whole body performance is an important goal of integrative physiology. Exercise capacity can be quantified in a number of ways in swimming fish. Separate standardized tests have been developed to assess burst, sustained, and prolonged swimming. Burst performance to fatigue does not require specialized equipment and often consists of a forced high-intensity exercise bout (chasing) lasting on average about 20 s and is powered almost entirely by anaerobic metabolism. Prolonged swimming tests last from 2 to 200 min, depending on the speed, and end with the fish reaching exhaustion. Sustained swimming lasts more than 200 min without muscular fatigue and is powered primarily by aerobic metabolism. The critical swimming test, termed Ucrit, akin to V˙O2max in mammals, is often used to assess sustained swimming performance. It involves a progressive increase in water velocity to the highest speed that can be sustained by the fish. Specially designed swim flumes (two common types are Brett style and Beamish style) also can be adapted for different sized fish, be used for large-scale training and exercise testing, or for high-throughput testing (31). These flumes also allow for the measurement of oxygen consumption (V˙O2) and collection of other respiratory gases (i.e., CO2, NH3). Assessment of Ucrit allows for the standardization of training intensity or relative exercise intensity (as % Ucrit) for submaximal exercise tests in diverse groups or species. Fish allow one to link nonload-bearing exercise with the metabolic characteristics of topographically separate red and white muscle masses that are recruited under different locomotory behaviors (i.e., burst vs sustained swimming).

Back to Top | Article Outline


The metabolism of exercise has been examined in some detail in fish, but few studies have looked at the effect of training on fuel use. The energetic cost of exercise and V˙O2 patterns with changing swimming intensity and temperature are known for many species. The patterns of use and the sources of metabolic fuels catabolized during exercise are less studied mainly because of some technical challenges involved in using an aquatic animal. However, it is known that many aspects of fuel mobilization, cytosolic and membrane transport, circulatory convection, and mitochondrial oxidation are similar between fish and mammals. Indeed, fish express glucose transporters, possibly express fatty acid translocase (FAT/CD36, based on inhibition by sulfo-N-succinimidyloleate (35)) and monocarboxylate transporters (28). Fish use plasma albumin for fatty acid transport and express fatty acid binding protein in myocytes for cytosolic fatty acid transport (23). However, there are some important differences in the types and proportions of metabolic fuels used during exercise. Fish rely on protein to a greater extent and may use significantly higher amounts of lipoproteins than mammals during exercise. Indeed, ignoring the contribution of proteins to exercise metabolism, as is often done for mammals, could lead to large errors because they can be important for powering locomotion in some fish species during specific tasks (e.g., salmon rely almost exclusively on proteins to power swimming at the end of their migration (41)).

Back to Top | Article Outline
Indirect Calorimetry

The use of metabolic fuels in exercising fish has been studied in vivo, but only a handful of studies have made measurements of fuel kinetics and oxidation during an exercise bout. It generally is agreed that fish power burst activity recruiting mainly the large white muscle mass and using the breakdown of intramuscular glycogen (reviewed in (12)). Surprisingly, training does not seem to increase intramuscular glycogen stores in fish as it does in mammals (9). Early studies on aerobic exercise suggested that it was supported principally by protein and lipid oxidation, with carbohydrates playing only a minor role (reviewed in (13)). These studies mostly used the compositional approach, basing fuel use on depletions of fuel stores during exercise. Later, an instantaneous approach was developed measuring the three respiratory gases (O2, CO2, and ammonia-N + urea-N) in water. This method allows for the measurement of fuel use during submaximal exercise when changes in body composition do not necessarily reflect their proportional use (13). It was concluded that lipid is the main fuel for submaximal aerobic exercise in fish with a significant role for carbohydrates and a minor but significant role for proteins. Training has been shown to increase the use of lipids to power exercise, a response commonly seen with training in mammals when individuals are compared at the same absolute speed (23). It is well established that humans and many other mammals follow a predictable pattern of fuel use during submaximal aerobic exercise where, at low-intensity exercise, lipids provide the majority of energy. As intensity increases, carbohydrate oxidation covers more and more of the total energy expenditure and eventually contributes 100% to muscle ATP production at V˙O2max (23). Some studies show fish roughly following the mammalian pattern (13,15). For example, juvenile rainbow trout that swam at 45% and 75% of their maximum sustained swimming speed (Ucrit) showed lipid oxidation declining whereas carbohydrate oxidation increased as a percentage of the total V˙O2 (Fig. 3 (13)). However, even at the highest intensity tested, carbohydrate only contributed 25% of total energy consumption. The importance of proteins to overall exercise metabolism is high in fish, accounting for up to 30% of total V˙O2 in some cases. So it appears that, qualitatively, fuel use patterns are similar in fish and mammals but that lipids and proteins support a greater proportion of ATP turnover in swimming fish.

Back to Top | Article Outline
Biochemical Approach

A second approach to understanding exercise metabolism in fish has been a biochemical approach to assess changes in fuel stores, metabolic intermediates, enzyme activation, and cellular energy charge using serial sampling at various time points during exercise. Although it is not possible to serially sample from individual fish, true fiber type-specific responses can be assessed by sampling both pure red and white muscle masses during the exercise period. This method only is effective for submaximal swimming measurements if great care is taken so the fish do not perform any burst activity before euthanasia. These studies have found that changes in metabolites and enzyme activation reflect changing muscle recruitment patterns as swimming intensity increases (33). In red muscle, pyruvate dehydrogenase (PDH) is activated initially and declines as swimming progresses. PDH activation correlates with exercise intensity, with higher activities as intensity increases to 90% Ucrit. Levels of acetyl-CoA and acetyl-carnitine increase, whereas free creatine decreases as swimming intensifies, demonstrating that the absolute rates of fatty acid oxidation rise. The concentrations of the metabolic intermediate malonyl-CoA decline with exercise duration when lipid oxidation rates are increasing, suggesting that regulation of CPT-I activity is similar in fish and rodents, but not humans (33). The biochemical method supports the conclusions of indirect calorimetry measurements for a significant role for lipids in supporting low and high swimming rates. The fueling of high-intensity exercise also has been studied using the biochemical approach and shows that early stages of burst swimming are fueled by phosphocreatine hydrolysis and glycogen depletion, with large transformations in active glycogen phosphorylase, especially in white muscle, which is recruited heavily at these intensities (33). Thus, in terms of burst exercise, fish behave similarly to mammals but also allow for muscle fiber-specific monitoring immediately after exercise and during muscle recovery.

Back to Top | Article Outline
Continuous Infusion

One of the biggest advances in the study of fuel use in exercising fish has been the development of the continuous isotopic infusion technique for use in these animals (6,7). Previously, bolus injections of isotopes or blood concentration changes were used to infer metabolite turnover, muscle uptake, and oxidation of fuels, which can lead to misleading results (7). The continuous infusion technique can be used easily if fish are of a sufficient size (>500 g) to allow for dual cannulation of the dorsal aorta, with the two cannula inserted to different lengths into the vessel. One cannula is used for the infusate delivery and the other cannula is used for blood sampling. Fish also must be sufficiently large for serial blood sampling during exercise. Careful validation of this technique in swimming fish shows that it can be used for measurements in nonsteady-state conditions. The resulting data show some interesting kinetics of fatty acids and glycerol (1), lipoproteins (20), and lactate relative to mammals. Unlike some mammals, rates of lipolysis and fatty acid turnover rates are very high at rest in trout but do not increase with the onset of exercise (1). Also unique to trout and probably fish in general are high resting turnover rates of plasma triacylglycerol (TAG), higher than postabsorptive mammals, and turnover remains elevated with exercise (20). It, therefore, has been suggested that fish may rely on plasma TAG to a greater extent than even albumin-bound free fatty acids because they could cover all of the fuel requirements for exercise (20). This hypothesis is supported by three observations: 1) migrating salmon deplete plasma lipoproteins during the course of migration (41); 2) trout maintain lipoprotein concentrations approximately four times higher than mammalian values (∼90% of their circulating lipids); and 3) prolonged aerobic locomotion activates lipoprotein lipase in the red muscles of trout (21). Exercise lactate kinetics have been measured using bolus injection techniques in fish and show only modest increases in turnover with prolonged swimming (40). Recently, continuous isotopic infusion techniques have been used to examine lactate kinetics at rest and show that 1) resting rates of lactate production are higher than rates measured with bolus injection and 2) with hypoxia exposure, tissues (mainly white muscle) take up and sequester lactate to reduce plasma lactate load (30). Some technical challenges remain when using these techniques. Primarily, it is difficult to collect expired isotopically labeled CO2 (e.g., 14CO2) as it enters the large volumes of water necessary to house and swim fish. This can limit direct measurement of circulatory fuel oxidation rates and distinguish the relative roles of fuels coming from outside and inside the muscles to power locomotion. This does not detract from the fact that these studies represent considerable advances in accurately measuring metabolite turnover in fish. These results suggest that fish have the potential to be valuable models for studying fuel kinetics, especially fuels used to a greater extent in fish during exercise (e.g., lipoproteins) than by mammals.

Back to Top | Article Outline
Recovery From Exercise

The ability to recover quickly from an exercise bout has ecological relevance to many fish species. The faster animals can recover from intense exercise, the quicker they can escape additional predators or resume foraging activities. For this reason, many fish species recover very rapidly from exhaustive and submaximal aerobic exercise (36), and the return of muscle to the resting state follows a similar pattern to that seen in mammals. Briefly, metabolic pathways are activated to resynthesize ATP, phosphocreatine, and glycogen. However, unlike mammals, lactate appears to be retained mostly in white muscle for in situ glycogen repletion (36) and little, if any, is exported to the liver or oxidized to power other recovery processes (33). Similar to mammals, fish power the replenishment of glycogen stores after exercise by increasing lipid catabolism (33).

Although many aspects of exercise metabolism are similar between fish and mammals, the regulation of fuel pathways may be more complicated because of potentially multiple novel isoforms of key proteins. As discussed for CPT-I previously, one can speculate that multiple but yet unidentified isoforms of other enzymes and transport proteins could result in muscle fiber type-specific regulation of fuel delivery and oxidation with different sensitivities to allosteric and covalent regulators. Moreover, there is relatively little known regarding the cellular localization of proteins known in mammals to translocate to the plasma membrane (glucose transporter-4) or mitochondria (CD36/fatty acid translocase) or both (CD36/fatty acid translocase) with muscle contraction and/or training. This increased complexity represents a valuable resource to uncover alternate modes of metabolic regulation not used by mammals.

Back to Top | Article Outline


Adopting fish as a model for muscle and exercise physiology will allow researchers to exploit the many strengths of these organisms to compliment and contrast to mammalian models. Although the embryonic and larval stages have been used frequently to study muscle development, plasticity of adult fish muscle is a relatively unexplored area of research.

There are few limitations to using reverse genetic techniques in popular fish models, and some species have the added advantage of the visualization of molecular processes in vivo. The use of transgenic zebrafish has been very successful, especially when linked to the expression of green fluorescent protein for easy visualization of cellular responses. For example, zebrafish transgenic for myostatin II have been used to study its role in hyperplasia and show a significant increase in muscle fiber number but not fiber size (10). Applied environmental researchers have developed transgenic zebrafish as sentinels for aquatic pollution by linking gene response elements for metals, aromatic hydrocarbons, and electrophiles to drive a luciferase reporter gene (2). This is an attractive approach to examine pathways turned on or off by exercise in swimming fish or mitochondrial biogenic pathways stimulated by cold acclimation. Forward genetics and behavioral genetic screens that are impractical to perform in mammals can be performed easily in fish. For example, this type of screen has identified a zebrafish mutant model for PDH defects to study its role in neurological dysfunction (39). Similar mutant genetic screens could be performed for differences in exercise performance to uncover important genes involved in muscle metabolism.

There are a number of limitations to reverse genetic approaches because knockout models often do not reflect complex phenotypes. In part, this issue has begun to be addressed through artificial selection for running ability in rodents. In contrast, because of their huge natural diversity, individual fish species can be compared to study different locomotory phenotypes that are the result of complex changes in physiology over evolutionary time. As mentioned earlier, the difference in lipid metabolism between the elasmobranchs (sharks and rays) and teleost fish has been used as a model of alternative muscle metabolic strategies during exercise. The lack of mitochondrial fatty acid oxidation in muscle has resulted in a reliance on ketone bodies, produced through liver fatty acid oxidation, to fuel muscles during exercise and recovery (34). Many other examples of natural “knockout” fish models exist, for example, the myoglobinless or hemoglobinless icefish (37). These fish compensate for lack of O2 transport proteins with large blood volumes, heart masses, and cardiac outputs. They also have high mitochondrial densities and fat content to aid in O2 diffusion (29). The diversity among fish species is ripe for mining to discover new and unique models for exercise research.

Back to Top | Article Outline


Overall, fish respond to endurance exercise training in a similar manner as mammals with muscle remodeling to produce a more aerobic phenotype. The putative molecular mechanisms that underlie this remodeling have common and distinct features compared with mammals. Notably, the so-called “master regulator” PGC-1α in mammals does not seem to play a direct role in muscle phenotypic plasticity in fish. Genome duplication events in the evolution of the fish lineages have diversified the number of isoforms for enzymes and transport proteins involved in muscle metabolism and their regulators. This suggests that the regulation of muscle metabolism and fuel use during exercise in fish may be more complex than in mammals. Overall, these features make fish a very interesting and powerful model for exercise physiology.

This research was supported by the Natural Sciences and Engineering Council of Canada.

The author declares no conflict of interest.

Back to Top | Article Outline


1. Bernard SF, Reidy SP, Zwingelstein G, Weber JM. Glycerol and fatty acid kinetics in rainbow trout: Effects of endurance swimming. J. Exp. Biol. 1999; 202: 279–88.
2. Carvan MJ 3rd, Dalton TP, Stuart GW, Nebert DW. Transgenic zebrafish as sentinels for aquatic pollution. Ann. N. Y. Acad. Sci. 2000; 919: 133–47.
3. Dalziel AC, Moore SE, Moyes CD. Mitochondrial enzyme content in the muscles of high-performance fish: Evolution and variation among fiber types. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2005; 288: R163–72.
4. Dhillon RS, Schulte PM. Intraspecific variation in the thermal plasticity of mitochondria in killifish. J. Exp. Biol. 2011; 214: 3639–48.
5. Gamperl AK, Farrell AP. Cardiac plasticity in fishes: Environmental influences and intraspecific differences. J. Exp. Biol. 2004; 207: 2539–50.
6. Haman F, Weber JM. Continuous tracer infusion to measure in vivo metabolite turnover rates in trout. J. Exp. Biol. 1996; 199: 1157–62.
7. Haman F, Powell M, Weber JM. Reliability of continuous tracer infusion for measuring glucose turnover rate in rainbow trout. J. Exp. Biol. 1997; 200: 2557–63.
8. Hock MB, Kralli A. Transcriptional control of mitochondrial biogenesis and function. Annu. Rev. Physiol. 2009; 71: 177–203.
9. Johnston IA, Moon TW. Exercise training in skeletal muscle of brook trout (Salvelinus fontinalis). J. Exp. Biol. 1980; 87: 177–94.
10. Johnston IA, Bower NI, Macqueen DJ. Growth and the regulation of myotomal muscle mass in teleost fish. J. Exp. Biol. 2011; 214: 1617–28.
11. Jopling C, Sleep E, Raya M, Marti M, Raya A, Belmonte JCI. Zebrafish heart regeneration occurs by cardiomyocyte dedifferentiation and proliferation. Nature. 2010; 464: 606–11.
12. Kieffer JD. Perspective — Exercise in fish: 50+ years and going strong. Comp. Biochem. Physiol. A. 2010; 156: 163–8.
13. Kieffer JD, Alsop D, Wood CM. A respirometric analysis of fuel use during aerobic swimming at different temperatures in rainbow trout (Oncorhynchus mykiss). J. Exp. Biol. 1998; 201: 2123–33.
14. Lau GY, Richards JG. AMP-activated protein kinase plays a role in initiating metabolic rate suppression in goldfish hepatocytes. J. Comp. Physiol. B. 2011; 181: 927–39.
15. Lauff RF, Wood CM. Respiratory gas exchange, nitrogenous waste excretion, and fuel usage during aerobic swimming in juvenile rainbow trout. J. Comp. Physiol. B. 1996; 166: 501–9.
16. Le Moine CM, Genge CE, Moyes CD. Role of the PGC-1 family in the metabolic adaptation of goldfish to diet and temperature. J. Exp. Biol. 2008; 211: 1448–55.
17. Le Moine CM, Craig PM, Dhekney K, Kim JJ, McClelland GB. Temporal and spatial patterns of gene expression in skeletal muscles in response to swim training in adult zebrafish (Danio rerio). J. Comp. Physiol. B. 2010; 180: 151–60.
18. Le Moine CM, Lougheed SC, Moyes CD. Modular evolution of PGC-1α in vertebrates. J. Mol. Evol. 2010; 70: 492–505.
19. Le Moine CM, Morash AJ, McClelland GB. Changes in HIF-1α protein, pyruvate dehydrogenase phosphorylation, and activity with exercise in acute and chronic hypoxia. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2011; 301: R1098–104.
20. Magnoni L, Vaillancourt E, Weber JM. High resting triacylglycerol turnover of rainbow trout exceeds the energy requirements of endurance swimming. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2008; 295: R309–15.
21. Magnoni L, Weber JM. Endurance swimming activates trout lipoprotein lipase: Plasma lipids as a fuel for muscle. J. Exp. Biol. 2007; 210: 4016–23.
22. Martin CI, Johnston IA. The molecular regulation of exercised-induced muscle fiber hypertrophy in the common carp: Expression of MyoD, PCNA and components of the calcineurin-signaling pathway. Comp. Biochem. Physiol. B. Biochem. Mol. Biol. 2005; 142: 324–34.
23. McClelland GB. Fat to the fire: The regulation of lipid oxidation with exercise and environmental stress. Comp. Biochem. Physiol. 2004; 139: 443–60.
24. McClelland GB, Craig PM, Dhekney K, Dipardo S. Temperature- and exercise-induced gene expression and metabolic enzyme changes in skeletal muscle of adult zebrafish (Danio rerio). J. Physiol. (London). 2006; 577: 739–51.
25. Morash AJ, Kajimura M, McClelland GB. Intertissue regulation of carnitine palmitoyltransferase I (CPT-I): Mitochondrial membrane properties and gene expression in rainbow trout (Oncorhychus mykiss). BBA-Biomembranes. 2008; 1778: 1382–9.
26. Morash AJ, Le Moine CMR, McClelland GB. Genome duplication events have led to a diversification in the CPT I gene family in fish. Am. J. Physiol. 2010; 299: R579–89.
27. Moyes CD. Controlling muscle mitochondrial content. J. Exp. Biol. 2003; 206: 4385–91.
28. Ngan AK, Wang YS. Tissue-specific transcriptional regulation of monocarboxylate transporters (MCTs) during short-term hypoxia in zebrafish (Danio rerio). Comp. Biochem. Physiol. B. 2009; 154: 396–405.
29. O’Brien KM. Mitochondrial biogenesis in cold-bodied fishes. J. Exp. Biol. 2011; 214: 275–85.
30. Omlin T, Weber JM. Hypoxia stimulates lactate disposal in rainbow trout. J. Exp. Biol. 2010; 213: 3802–9.
31. Palstra AP, Tudorache C, Rovira M, et al.. Establishing zebrafish as a novel exercise model: Swimming economy, swimming-enhanced growth and muscle growth marker gene expression. PLoS One. 2010; 5: e14483.
32. Richards JG. Physiological, behavioral and biochemical adaptations of intertidal fishes to hypoxia. J. Exp. Biol. 2011; 214: 191–9.
33. Richards JG, Heigenhauser GJ, Wood CM. Lipid oxidation fuels recovery from exhaustive exercise in white muscle of rainbow trout. Am. J. Physiol. 2002; 282: R89–99.
34. Richards JG, Heigenhauser GJ, Wood CM. Exercise and recovery metabolism in the pacific spiny dogfish (Squalus acanthias). J. Comp. Physiol. B. 2003; 173: 463–74.
35. Richards JG, Bonen A, Heigenhauser GJ, Wood CM. Palmitate movement across red and white muscle membranes of rainbow trout. Am. J. Physiol. 2004; 286: R46–53.
36. Schulte PM, Moyes CD, Hochachka PW. Integrating metabolic pathways in post-exercise recovery of white muscle. J. Exp. Biol. 1992; 166: 181–95.
37. Sidell BD, O’Brian KM. When bad things happen to good fish: The loss of hemoglobin and myoglobin expression in Antarctic icefishes. J. Exp. Biol. 2006; 209: 1791–802.
38. Storch D, Lannig G, Pörtner HO. Temperature-dependent protein synthesis capacities in Antarctic and temperate (North Sea) fish (Zoarcidae). J. Exp. Biol. 2005; 208: 2409–20.
39. Taylor MR, Hurley JB, van Epps HA, Brockerhoff SE. A zebrafish model for pyruvate dehydrogenase deficiency: Rescue of neurological dysfunction and embryonic lethality using a ketogenic diet. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 4584–9.
40. Weber JM. Effect of endurance swimming on the lactate kinetics of rainbow trout. J. Exp. Biol. 1991; 158: 463–76.
41. Weber JM. The physiology of long-distance migration: Extending the limits of endurance metabolism. J. Exp. Biol. 2009; 212: 593–7.

mitochondria; carnitine palmitoyltransferase; cold acclimation; genome duplication; lipoproteins; PGC-1

©2012 The American College of Sports Medicine