Perturbed diurnal rhythms are becoming increasingly evident as deleterious events in the pathology of metabolic diseases. Exercise is well characterized as a crucial intervention in the prevention and treatment of individuals with metabolic diseases. Little is known, however, regarding optimizing the timing of exercise bouts in order to maximize their health benefits. Furthermore, exercise is a potent modulator of skeletal muscle metabolism, and it is clear that skeletal muscle has a strong circadian profile. In humans, mitochondrial function peaks in the late afternoon, and the circadian clock might be inherently impaired in myotubes from patients with metabolic disease. Timing exercise bouts to coordinate with an individual’s circadian rhythms might be an efficacious strategy to optimize the health benefits of exercise. The role of exercise as a Zeitgeber can also be used as a tool in combating metabolic disease. Shift work is known to induce acute insulin resistance, and appropriately timed exercise might improve health markers in shift workers who are at risk of metabolic disease. In this Review, we discuss the literature regarding diurnal skeletal muscle metabolism and the interaction with exercise bouts at different times of the day to combat metabolic disease.
Skeletal muscle has an extensive network of clock-controlled genes, and dysregulation of its molecular clock can lead to deleterious metabolic consequences.
Physical strength and skeletal muscle mitochondrial function peak in the late afternoon, whereas low-energy sensitive signalling peaks in the morning.
Exercise is a robust Zeitgeber of skeletal muscle clocks, and exercise can reset the molecular circadian clock, thereby effectively ameliorating the negative effects of disrupted sleep patterns.
Optimizing the timing of exercise bouts could aid existing therapeutic interventions for the management of metabolic disease.
Divergent modalities of exercise can interact with the circadian rhythm, resulting in potent metabolic effects.
Mammalian cells possess an internal molecular clock that consists of transcriptional and/or translational autoregulatory feedback loops. At a physiological level, circadian clocks drive whole-body metabolism. At a molecular level, cell-autonomous circadian rhythms are produced by the activity of transcriptional activators CLOCK and BMAL1 and their target genes, which form a repressor complex that interacts with CLOCK and BMAL1 to inhibit transcriptional activity1,2. The feedback loop of the cell-autonomous core clock is highly regulated by several factors, including the activity of the master clock located in the hypothalamic superchiasmatic nucleus (SCN)1,2,3. Mounting evidence suggests that disruption of this machinery is highly detrimental to metabolism; for example, in several animal models disruption of the clock machinery results in obesity and insulin resistance4,5. Additionally, indirect disruption of the core clock machinery via poor sleeping patterns or variable shift work can have similar deleterious effects on metabolism in humans6,7.
Several studies demonstrate that exercise modifies the rhythm of the clock machinery in skeletal muscle8,9,10 (Fig. 1); however, the optimal timing of exercise for health and the potential of exercise training to ameliorate the effects of disrupted circadian rhythms have not been fully elucidated. Habitual exercise training has myriad health benefits and can be a therapeutic tool in both prevention and treatment of metabolic disease11. Metabolic diseases, such as obesity and type 2 diabetes mellitus (T2DM), are a growing, global health burden. The prevailing modern lifestyle interacts with underlying biology to create an environment in which metabolic diseases can flourish. Within this environment, exposure to artificial light, altered working and/or sleeping hours, diet, lack of physical activity and easily accessible, high-calorie foods are all contributory factors to a global rise in metabolic disease and T2DM1. One area of the modern lifestyle that has been scrutinized over the past decade has been the effect of disrupted circadian rhythms on health. The disruption of these diurnal rhythms is linked to an increased risk of developing metabolic disease6,7,12.
Diurnal rhythms can be disrupted by shift work, genetic mutations that give rise to divergent circadian rhythms or aberrant exposure to artificial light sources13,14,15. Additionally, patterns of eating and other behaviours can strongly modulate both sleep patterns and the internal cellular clock machinery that regulate circadian rhythms16. In short, the interaction between environmental factors and inherited biology can lead to perturbed daily patterns of behaviour and the molecular functioning of cells, which can disrupt the daily metabolic processes necessary to maintain health. Although exercise ameliorates many of the deleterious processes associated with these phenomena11, research on optimizing exercise timing is sparse. In other words, whether there is a right time of the day to trigger an optimal training response remains unclear.
Finding the optimum time for exercise is an area ripe for research as skeletal muscle has many clock-controlled genes17 and exercise capacity is known to fluctuate over the course of the day18. If the timing of exercise can be optimized to coincide with the greatest physiological and molecular response to exercise, perhaps its potency as a therapeutic tool might be increased even further. In addition to the metabolic health benefits of exercise, acute bouts of exercise can be a tool to improve sleep quality19. These beneficial effects of exercise on sleep might have relevance for treating disrupted sleep patterns, be they pathological, geographical (for example, seasonal variations in light) or a result of shift work. Conversely, there might also be negative effects of mistimed exercise. For example, it could be queried whether intense exercise is beneficial when performed at a time to which the participant is unaccustomed.
In this Review, we assess the current literature regarding the molecular mechanisms and therapeutic potential of exercise in regard to circadian rhythms. We summarize the current literature regarding the metabolic consequence of exercise at different times of the day and the associated beneficial health effects. The interaction of exercise with the molecular clock machinery and the role of exercise as a Zeitgeber and the putative health benefits are also discussed.
The skeletal muscle molecular clock
Intracellular circadian clocks can modulate physiological processes over the course of a day1. These molecular processes are regulated by an autoregulatory feedback loop composed of transcriptional activators CLOCK and BMAL1 and their target genes period (PER), cryptochrome (CRY) and NR1D1 (which encodes REV-ERBα), which accumulate temporally and form a repressor complex that interacts with CLOCK and BMAL1 to inhibit transcription1. Data from 2018 demonstrate that synchronized primary human skeletal muscle cells share several circadian characteristics with human skeletal muscle biopsy samples; however, rhythmic transcriptional activity has a greater magnitude in biopsy samples20. These data provide valuable insight into the isolated intrinsic skeletal muscle clock without interference from the SCN or behaviour. Additionally, the donors involved in this study demonstrated highly variable, interindividual rhythmic cellular gene expression, presumably owing to inherited factors. Putatively, these variable rhythms could derive from mutations within core clock genes, as demonstrated on a physiological level15.
Metabolic flux and the molecular clock
Disruption of the core clock can dysregulate skeletal muscle metabolism and alter how an individual responds to exercise. For example, a CRY1 polymorphism (rs2287161) interacts with increased carbohydrate intake to associate with insulin resistance in individuals who are homozygous for the minor C allele (CC)21. Although the mechanism of this interaction is unknown, this rs2287161 variant is thought to modulate the transcription factor binding site of CRY1 (refs21,22,23,24). In terms of skeletal muscle, muscle-specific BMAL1 depletion results in insulin resistance and obesity in mouse models, alongside decreased GLUT4 and TBC1D1 (ref.5). Additionally, the skeletal muscle core clock (BMAL1 and REV-ERBα) controls transcriptional programming of lipid and amino acid metabolism via direct binding to targets in these pathways25 (Fig. 1).
In primary human skeletal muscle cells, small interfering RNA (siRNA) targeting CLOCK dysregulated BMAL1 rhythmic transcriptional activity and impaired insulin-mediated glucose uptake20. In Bmal1−/− mice, anaerobic glycolysis, mitochondrial respiration and transcription of Hif1a target genes is reduced5. Furthermore, in human primary skeletal muscle cells, knockdown of CLOCK reduced the expression of HIF1α target gene VEGFA20. Thus, in skeletal muscle, the activity of the core clock machinery seems to be closely aligned to metabolic flux. Indeed, activity of AMPK, a key energy sensor in skeletal muscle, can reduce the stability of CRY1 directly and PER2 through casein kinase 1ε-mediated phosphorylation26,27,28. In addition, CRY1 and CRY2 interact with the lipid-sensing peroxisome proliferator-activated receptor-δ (PPARδ) to supress its activity28. CRY2 might be of particular importance in skeletal muscle as it also interacts with BCLAF1 to stabilize Tmem176b mRNA, a myocyte fusion-associated gene29.
Physical activity and the muscle clock
Physical activity modulates the molecular clock in skeletal muscle, affecting both the amplitude and phase of circadian rhythms9,10. The skeletal muscle circadian transcriptomic response clusters around the midpoint of the active phase in mice30. Additionally, a study of denervated skeletal muscle in rodent models demonstrated that the removal of motor neuron activation moderately dysregulates circadian transcriptional activity31. In humans, one-legged resistance exercise altered circadian gene expression and apparently induced a phase shift of core clock genes when compared with the contralateral control leg32. HIF1α target genes also have a temporally dependent response to strenuous exercise over the course of a circadian cycle in mice8.
These data support the assertion that the core clock machinery partly regulates the transcriptional response to exercise. Additionally, genetic ablation of Cry1 and Cry2 increased exercise capacity in mice and altered the exercise-induced gene signature28. In human studies, cultured primary myotubes from endurance-trained athletes had preserved rhythmic gene expression of SIRT1 and NAMPT, whereas myotubes derived from untrained lean or obese individuals, or patients with T2DM, did not33. Furthermore, the amplitude of NR1D1 gene expression correlated with insulin sensitivity and the exercise training status of the donor. Therefore, the status of the core clock machinery in skeletal muscle seems to be pivotal in regulating the molecular response to exercise, particularly in terms of gene expression. Whether these results are fully recapitulated in human physiology is unclear, but habitual exercise is associated with inherently altered amplitude of core clock genes in human myotubes33.
Mitochondria and the muscle clock
Another parameter that seems to relate metabolism and the skeletal muscle clock is mitochondrial function. Reduced oxidative capacity of skeletal muscle is associated with diminished exercise performance and the development of T2DM34,35. Isolated mitochondria from human skeletal muscle biopsy samples have increased oxidative capacity (state 3 oxygen consumption rate) at ~23:00 h compared with at 04:00 h, 08:00 h, 13:00 h and 18:00 h (ref.36). Nevertheless, the extent to which these data are the result of the intracellular molecular clock or external Zeitgebers is unknown.
The researchers who conducted this study36 also indicated that mitochondrial dynamics (fusion and fission of mitochondria) can oscillate in a circadian manner. However, mitochondrial content, as measured by mitochondrial DNA (mtDNA), levels of protein, mitochondrial mass or PGC1α expression, is not rhythmically expressed in human skeletal muscle36,37; mitochondrial biogenesis might be rhythmic in other cells or tissue types37,38. Therefore, diurnal changes in skeletal muscle mitochondrial function are probably a result of changes to mitochondrial morphology and/or mitochondrial dynamics and/or mitophagy, which are potential candidates for circadian regulation37.
Dysregulated mitochondrial fusion and/or fission (dynamics) might be important in the pathology of T2DM39, and mitochondrial dynamics seem to be regulated by circadian rhythms in cultured human macrophages40. Interestingly, cultured human fibroblasts display periodic mitochondrial dynamics driven by the molecular core clock41. This study identified DRP1 phosphorylation and activity as one possible mediator of rhythmic mitochondrial dynamics. When DRP1 was pharmacologically inhibited by 1 µM P110, the period length of BMAL1 transcriptional activity increased by >1 h, indicative of core clock modulation via mitochondrial metabolism. In a 2018 paper42, siRNA-mediated knockdown of GDAP1 (a protein involved in mitochondrial fission) in primary human skeletal muscle was shown to result in increased expression of NPAS2 (a paralogue of CLOCK) and decreased DBP expression. These data indicate that the mitochondrial dynamic machinery in skeletal muscle might also participate in retrograde signalling and modulation of the core clock.
Collectively, these findings41,42 are noteworthy as they suggest that the core clock in skeletal muscle can respond to alterations in mitochondrial dynamics in addition to metabolic stimuli while also driving metabolic outcomes. Exercise potently remodels mitochondrial morphology and dynamics, both acutely and chronically. Therefore, timing exercise bouts to coincide with the mitochondrial dynamic period might increase the acute effects of exercise in terms of substrate uptake and utilization. Furthermore, mitochondrial network remodelling could also potentially be amplified.
Exercise physiology and the muscle clock
Although it is clear that the molecular clock in skeletal muscle interacts with cellular physiology, characterizing the discrete ways in which the clock is involved in the regulation of human physiology is a challenge. Numerous factors (such as fatty acid and carbohydrate metabolic pathways, in addition to secreted factors) interlink in the regulation of metabolism and physiology by the molecular clock. Animal models are often vital in elucidating the molecular pathways that regulate physiology, and they are fundamental to advancing the field of circadian biology. However, the rodent models used to study how the molecular clock communicates with physiological responses are often unable to fully recapitulate human physiology. For example, rodents are primarily active at night, often fed ad libitum, have fragmented daytime sleep and generally do not produce melatonin17. In addition, although rodent models used in circadian biology are primarily active at night, their skeletal muscle molecular clock4 is not simply an inverse of the human counterpart20.
These issues contribute to a somewhat convoluted model, which is made more complex by external Zeitgebers such as feeding and light exposure. Although the majority of circadian studies admirably control for these issues, practical realities prevent the creation of a perfect model. Potential solutions to counteract the practical realities include the creation of more humanized mouse models43,44, although these models often do not recapitulate the full spectra of human physiological phenomena45.
Another solution is to use a species with physiology that is more closely aligned to that of humans; an elegant attempt has been made to resolve some of the aforementioned issues by using a primate (baboon) model17, which should prove an invaluable resource for this field. Initial findings from this study demonstrate that skeletal muscle has the highest number of circadian cycling genes of all tissues in baboons (3,182 cycling genes versus 2,615 in mice; 649 of these genes overlap), whereas the liver has the most in mice (3,700 versus 529 in baboons; 150 of these genes overlap). These findings highlight the importance of skeletal muscle in primate circadian biology and help inform our interpretation of circadian rodent studies. Indeed, there was no correlation (R2 = 0.086, P = 0.38) between the number of tissue-specific cycling genes in mice and baboons. Furthermore, following an analysis of all tissues, several core clock genes had opposite rhythms in mice and baboons, including Bmal1, Per1 and Cry1. By combining resources such as primate models with well-controlled and well-designed rodent models and new technology in well-controlled human studies, the field of circadian biology can shine further light on key questions, such as when the best time to exercise is.
As stated previously, one important, underexplored area is how the skeletal muscle molecular clock interacts with exercise in terms of metabolic health outcomes. As it stands currently, the literature cannot support the assertion that the human skeletal muscle molecular clock directly modulates the diurnal exercise response, although accumulating data from rodent models partially corroborate this hypothesis8,28.
Exercise physiology and circadian timing
The health outcomes of exercise have not been extensively studied with regard to the optimal diurnal timing; however, data do exist with regard to the interaction between diurnal timing and exercise performance (Fig. 2). For example, more world records are broken by athletes competing in the early evening, even when environmental conditions and scheduling bias are partially controlled46. Increased strength, power and endurance are often observed in the afternoon and evening compared with early morning18,47. Disruptions of diurnal rhythm can also negatively affect athletic performance. Eastward trans-meridian travel has a greater negative effect on intermittent sprint performance and psychological indicators of fatigue than westward travel48. The chronotype of an athlete can also have a role in exercise performance at different times of the day47,49. As competing athletes aim to optimize training and preparation, interest in the interaction between the circadian rhythm, exercise performance and exercise response is growing; however, this interaction should also be appraised when considering exercise as a treatment or preventive clinical tool.
Different modalities of exercise
Differing modalities of exercise result in varying metabolic perturbations and signalling outcomes, including the type of skeletal muscle fibre recruited during the exercise bout. In high-intensity or resistance exercise, a greater amount of type II fibres are recruited than in low-intensity endurance exercise, which predominantly recruits type I fibres50. Type II fibres are fast-twitch, fatigable and more glycolytic than highly oxidative type I fibres50. The circadian gene expression pattern of the core clock is similar in these fibre types; however, the distinct fibre types display unique expression of most other diurnally cycling genes51. Differential recruitment of fibre types during exercise might influence circadian gene expression in an exercise-specific manner.
Resistance exercise is the most susceptible form of exercise to diurnal rhythms. Daytime peak force is nearly always demonstrated as being highest in the afternoon and evening (16:00–20:00 h), and lowest in the morning (06:00–10:00 h), with the discrepancy ranging from ~3% to 18% depending on experimental design18,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73 (although the majority of these studies found that the discrepancy clusters around ~8%). The response to resistance exercise is thought to be particularly susceptible to time of day, partly owing to diurnal neuromuscular performance, which can be somewhat ameliorated by acclimation59,65.
In short-duration, high-intensity exercise, a similar daily pattern of performance to that of resistance exercise is observed. Peak and mean power output vary by ~8% and ~11%, respectively18,66,70, with peaks and troughs in performance and capacity noted at similar times to resistance exercise (peaks in the afternoon and evening and troughs in the morning).
Moderate and endurance exercise
For long-duration exercise, the time-of-day effect on performance is equivocal compared with high-intensity and resistance training18,71. Although some studies do demonstrate higher aerobic and endurance performance in the afternoon and evening than in the morning70, these effects seem to be moderate and are not consistent between studies. A finding in some studies of diurnal fluctuations in long-duration exercise is that although overall performance is not markedly changed, physiological parameters, such as heart rate, are altered70,72,73, although directionality of these physiological parameters is variable. The equivocal findings in these studies could be related to a statistical power issue, and a meta-analysis might elucidate this further.
Exercise capacity and body temperature
The SCN of the hypothalamus contains the main clock components1. The main clock components function as the master circadian clock, whereby they synchronize and maintain the rhythms of the peripheral clocks3. Several signals derived from the SCN can influence peripheral tissue clocks, including the skeletal muscle clock, in addition to influencing daily variations in body temperature, the levels of secreted factors (such as insulin) and activity of the autonomic nervous system3,5,74. Of these, a key factor that can influence exercise performance at specific times of day is body temperature, but more specifically core body and skeletal muscle temperature64, which peak in the late afternoon or early evening. Explicitly, the core body temperature increases by ~0.8 °C in the afternoon or early evening and the temperature of skeletal muscle increases by ≥0.35 °C at these times72,75.
The temperature of skeletal muscles has a multifactorial effect on local metabolic processes and contractile efficiency. ATP turnover, phosphocreatine consumption and fibre conduction velocity are elevated with increased skeletal muscle temperature (passive heating) during maximal power output (6-second maximal sprint, cycle ergometer)76,77. A close relationship also exists between blood temperature, blood perfusion and aerobic metabolism in exercising limbs78. Temperature is a conserved entraining agent, acting as a Zeitgeber in the majority of mammals79, and the core clock machinery in skeletal muscle is strongly responsive to synchronization by temperature, controlled by the SCN80.
The thermoregulatory response to exercise also oscillates over the circadian cycle, with an apparent reduced ability to dissipate core body heat in the morning as compared with the afternoon72. This reduced ability to dissipate heat could be a driving factor for the generally increased exercise performance in the afternoon compared with the morning. In support of this theory, central and peripheral fatigue, often manifesting in reduced voluntary contractile activation, are closely associated with core (and to some extent, peripheral) temperature81. An active or passive heating of core and peripheral tissues is known to acutely improve strength and power (although heating above optimal levels of ~38.5 °C is not very effective and can be detrimental to performance)82. However, a 2018 study demonstrated that passive warming to increase core body and skeletal muscle temperature did not completely ablate diurnal variation in repeated sprint performance82. These data suggest that other factors, independent of heat per se, have a role in regulating the circadian exercise performance, particularly in regards to power and/or strength.
Exercise response and hormonal fluctuation
Some other crucial determinants of exercise performance are fluctuations in hormonal secretion and metabolism. For example, 1,000-metre time-trial performance for nine male recreational cyclists was improved in the evening (18:00 h) as compared with the morning (08:00 h) by ~7%70. Furthermore, the researchers showed that oxygen uptake and aerobic mechanical power output were higher in the evening trial than in the morning trial70. Interestingly, in this trial noradrenaline response to exercise was higher in the morning trial than in the evening trial, indicating that hormonal response to exercise is altered at different times of the day.
Plasma levels of testosterone and cortisol also display diurnal variations pre-exercise and post-exercise, although the magnitude of response to exercise seems similar in evening and morning exercise83. In healthy young men, the plasma concentrations of testosterone and cortisol are higher at 08:00 h than at 22:00 h, whereas the testosterone:cortisol ratio is higher at 20:00 h than at 08:00 h. It could be speculated that fluctuations in circulating testosterone and/or cortisol are partly responsible for the acute diurnal fluctuations in response to resistance exercise; however, the physiological relevance of these fluctuations to exercise outcomes is debatable84.
Diurnal substrate metabolism and exercise
Another key factor affecting diurnal fluctuations in exercise capacity, performance and response is the metabolism and availability of energy substrates. Food and feeding are Zeitgebers per se, and intrinsic molecular clocks seem to respond primarily to habitualized, rather than intrinsically conserved, feeding behaviours85. Given this factor, distinguishing between teleological and empirical summations of diurnal feeding patterns and determining the effect these have on circadian rhythms is complex. The modern western tradition of three meals per day is not necessarily pervasive across modern and historical cultures86,87,88. These lines of evidence indicate that human physiology is highly adaptive to different feeding timings, which should be taken into consideration when interpreting the interaction between feeding and the circadian rhythm.
The physiological regulation of glucose and triglycerides is an important factor in diurnal rhythms in general and in the diurnal exercise response11. The levels of glucose and triglycerides are partly controlled by intrinsic clocks in several tissues. For example, in mice, liver-specific ablation of Bmal1 induces hypoglycaemia during the fasting phase89, and skeletal-muscle-specific ablation of Bmal1 robustly modulates glucose uptake and metabolism90. Therefore, peripheral clocks have critical roles in maintaining glucose homeostasis. The SCN also drives diurnal variations in postprandial triglyceride uptake into skeletal muscle and brown adipose tissue in rats, probably via a REV-ERBα-mediated pathway91.
Improving our knowledge regarding the interaction between diurnal feeding patterns and exercise intervention strategies is important to better understand metabolic regulation. For example, different exercise intensities seem to profoundly alter postprandial triglyceride metabolism over the course of the subsequent day92,93. Using high-throughput metabolomics, one study demonstrated that skeletal muscle and plasma have unique metabolic signatures that are robustly reprogrammed by divergent nutritional challenges94. Putatively, manipulating timing in exercise intervention protocols to protect against subsequent excursions in postprandial metabolites could be a preventive strategy in combating metabolic disease. Furthermore, a rapidly growing area of research involves the interaction between periodized nutrition and exercise responses95,96,97,98. Many periodized nutrition protocols involve manipulating carbohydrate availability before, during or after exercise bouts, for example, by performing an intense workout in the evening with subsequent low carbohydrate intake resulting in lowered carbohydrate availability (muscle and liver glycogen) followed by sleep. This method has demonstrated some benefits in exercise performance and skeletal muscle signalling response of the lipid oxidation pathway95,96,97,98. This approach has been primarily researched in an athletic context, but these studies should also focus on prevention and intervention strategies for health.
Chronic circadian timing of exercise
Many studies have assessed the acute effect of diurnal rhythms on exercise physiology18,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73; one noteworthy cross-sectional study conducted by Maria Küüsmaa and colleagues was published in 2016 (ref.99). The researchers investigated the effect of 24 weeks of combined strength and endurance training conducted in the morning or evening on physical performance, muscle hypertrophy and serum hormone concentrations. Specifically, 42 (mean of groups range from 30.8 ± 5.0 to 36.1 ± 6.5 years) males were assessed for chronotype (no participants had an extreme morning or evening chronotype), matched and assigned to four groups (morning or evening training and endurance before strength training or strength training before endurance training). The researchers reported that the evening training groups gained more muscle mass than the morning training groups99. Interestingly, diurnal rhythms in testosterone and cortisol remained unaltered by training. These data suggest that inherent differences exist in the response to training in the evening versus morning that acclimation cannot completely ameliorate. We would like to point out, however, that although this study was rigorous and participants were well matched, conventional scientific process dictates that confirmation of this finding can come only from a crossover study design — although, in reality, conducting such a study on this scale would be impractical.
Another study by Sedliak and colleagues65 assessed 25 young (age of participants in the morning group was 23 ± 2 years and the age of participants in the afternoon group was 24 ± 4 years), previously untrained males who were randomly divided into a morning exercise group (07:30–08:30 h), an afternoon exercise group (16:00–17:00 h) and a sedentary control group. Voluntary muscle force increased in both the morning and afternoon exercise group after 11 weeks training65. Compared with the morning group, mean force was consistently higher in the afternoon exercise training group before, after and during training; however, these differences did not reach statistical significance.
The reason for the difference in morning versus evening training-induced gains in muscle mass between these studies65,99 is unknown; however, study cohort size and length of training might account for the differences. For example, Küüsmaa and colleagues99 included a larger total cohort and the duration of training was longer than in the study by Sedliak and colleagues65, which might increase the ability to detect diurnal effects. Furthermore, the exercise protocols were different between studies. Nevertheless, Sedliak and colleagues65 reported that phosphorylation of p70S6KThr421/Ser424, an exercise-responsive signalling molecule involved in protein synthesis and cell growth, was only acutely responsive to morning exercise after 11 weeks of training and was not responsive to afternoon exercise training.
These data indicate that differences in functional training outcomes between morning or afternoon exercise can be small once an individual has become acclimated to morning exercise training; however, signalling differences related to protein synthesis can persist. Putatively, the effect on protein synthesis might be mediated by exercising in varying nutritional states. In addition, the skeletal muscle core clock interacts with mTOR and downstream signalling100,101,102, which might link functional training outcomes to the core clock machinery. In rats, the fractional synthesis rate of skeletal muscle protein synthesis103, and mTOR and p70S6K activity104, peak during the light phase in these nocturnal animals. Assuming the inverse holds true for humans (as discussed previously, this is not always the case) might mean that protein synthesis rates peak in the evening.
Therapeutic potential for timing exercise
Epidemiological evidence suggests that T2DM and obesity are associated with loss of sleep quality105,106, although whether this is a causative phenomenon is unknown. Furthermore, obesity or T2DM might intrinsically disrupt the core clock machinery33. Ageing is associated with a gradual increase in period length of core clock genes, in addition to the dysregulation of other rhythmic biological processes107. As exercise is known to re-set clock genes in skeletal muscle and other tissues, it could be hypothesized that appropriately, and recurrently, timed exercise can help to re-set the daily clock and improve pathologically deteriorating circadian rhythms. Improving these dysregulated daily rhythms might help ameliorate negative metabolic consequences.
Exercise and sleep quality
Sleep surveys have documented regular physical activity as a variable associated with improved overall sleep quality19,108. Several mechanisms are thought to mediate this effect, including negative-feedback regulation of body temperature after exercise77. In addition, metabolic perturbations induced by exercise might regulate the neurotransmitter systems. For example, high-intensity exercise increased the plasma concentrations of the sleep-promoting molecule adenosine in rats76. In humans, acute exercise performed before the late evening (before 22:00 h) has been consistently demonstrated to boost sleep quality109; however, exercise performed shortly before going to bed might induce a stress response that attenuates this improvement and might even be detrimental to sleep quality110.
Habitual exercise is thought to improve sleep quality, even in the absence of acute exercise110. In addition to the physiological effects of exercise that promote sleep quality, performing exercise outside and increasing daylight exposure might be an additive enhancer of sleep111. Indeed, exposure to sunlight in the morning improved sleep quality and circadian entrainment in office workers112. It might be speculated that in terms of optimizing sleep hygiene specifically, outdoor exercise could be performed in the morning to maximize the beneficial results of daylight exposure.
Exercise and shift work
In addition to the exercise response being regulated by the circadian clock, exercise is also an effective Zeitgeber. Approximately 13–20% of workers in Europe and the USA are engaged in shift work that includes some time working at night113. Epidemiological studies and acute studies have shown that shift work increases risk factors for developing metabolic disease6,7. Just 4 days of simulated shift work can reduce insulin sensitivity7, and this deleterious effect might interact with genetic disposition to increase the risk of developing T2DM6.
The biological processes that mediate the increased risk of insulin resistance and T2DM resulting from disrupted circadian rhythms are multifactorial. For example, environmental factors interact with circadian biology; one study reported that reduced meal frequency and increased snacking are observed in night-shift workers14. In addition, the normal metabolic response to food is altered in shift-workers14. Lack of sleep due to changing diurnal patterns might also have an important role in the increased risk of insulin resistance resulting from disrupted circadian rhythms, and one study reported that just 1 week of sleep deprivation reduced insulin sensitivity114.
As with many aspects of metabolic disease pathology, individual differences seem to determine susceptibility to the negative consequences associated with shift work14. Increased amplitudes of the circadian rhythm might result in increased tolerance to shift work. A training-induced increase in amplitude of the skeletal muscle clock33 could be one method by which habitual physical activity could improve resistance to shift-work-induced metabolic phenomena. Alternatively, using acute or chronic exercise to improve sleep quality109 might also aid adaptation to shift work and reduce sleep deprivation.
Designing adequate intervention strategies to improve the acute and chronic health of shift workers is a key issue in combating the rise in metabolic diseases. Putatively, correct timing of exercise bouts might ameliorate some deleterious results of acute and chronic shift work. As mentioned previously, body temperature and thermoregulatory response have a robust circadian rhythm. In one study115, participants cycled for 15 minutes every hour during the first three of eight consecutive night shifts. Exercise facilitated temperature rhythm phase delays, which better aligned with daytime sleep. Although this exercise protocol is impractical for the majority of shift workers, it is proof of principle that correctly timed exercise can aid sleep patterns in regard to shift work.
Furthermore, timed exposure to a 5,000-lux bright light (22:10–23:40 h) might have an additive beneficial effect on hormonal diurnal rhythms, more specifically on levels of sulfatoxymelatonin, and the adaptation to shift work when combined with 90 minutes of moderate-intensity exercise (04:10–05:40 h)116. Given that short-duration bouts of exercise improved the adaptation to shift work115, another promising area of research that could be applied to shift workers is the concept of breaking up sedentary time.
Preliminary research suggests that for office workers, brief periods of activity over the course of a day have a beneficial effect on health outcomes, which might be over and above those observed from scheduled exercise per se117. One might speculate that regular bouts of interventional breaks to sedentary time also have similar beneficial results in shift workers. Another intervention that has shown promising health benefits is time-restricted feeding16; this intervention is based on a defined daily feeding–fasting rhythm with a determined window of food consumption, often around 8 h (ref.118). This approach could be another putative method of ameliorating the negative health consequences of shift work, potentially implemented in parallel to physical activity interventions.
Seasonal disruption of circadian rhythms
Given that daylight is one of the most powerful Zeitgebers for the majority of tissues, it is interesting to assess the effect of the change in seasonal daylight variation in northern latitudes. Sleep quality and daytime fatigue showed stronger seasonal variations in individuals from Norway (69°) than in individuals from Ghana (5°)119. Seasonal variations in daylight might also have a role in metabolic regulation. Single-nucleotide polymorphisms (SNPs) at CRY1, CRY2 and MTNR1B seem to interact with seasonal variations to modulate glucose homeostasis120. Although lack of daylight in the winter might negatively affect sleep and metabolic health, levels of physical activity generally increase in countries with northern latitudes in the summer121. Introducing winter exercise strategies, matched with optimal daylight exposure, could be an important therapeutic intervention for metabolic health.
As a side note, the human thermoregulatory system also exhibits seasonal variation as a result of ambient temperature acclimatization122. As diurnal changes in the thermoregulatory system seem to be a key factor in the exercise response in terms of circadian rhythm, it could be important to take note of thermoregulation seasonality when designing exercise interventions.
The majority of studies assessing high-intensity or strength training report that exercise performance is increased in the afternoon and evening compared with early morning18,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73. Several factors might influence this finding, including neuromuscular regulation59,65, circadian thermoregulation72,75, hormonal metabolism70,83,84, nutritional status11,90,91 and the skeletal muscle molecular clock8,28, among others. How diurnal exercise performance interacts with the acute exercise response and health outcomes remains unclear. Finding an answer to this question is important, particularly given the challenges of scheduling regular exercise into a modern lifestyle. Furthermore, exercise might be a useful therapeutic tool to treat poor sleeping patterns. Using correctly timed exercise as a therapeutic tool might apply to individuals who work shifts or have other disturbances to sleep hygiene. Putatively, synchronizing exercise and nutrient interventions to the molecular circadian clock might maximize the health-promoting benefits of exercise to prevent and treat metabolic disease.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.`
The authors are supported by grants from the Novo Nordisk Foundation (NNF14OC0011493, NNF14OC0009941 and NNF18CC0034900), the Wenner-Gren Foundation, the Swedish Research Council (2015–00165), the European Research Council (233285) and the Strategic Research Programme in Diabetes at Karolinska Institutet (2009–1068). The authors are grateful to B. Atkins for his contribution to figure design.
- Core clock
A set of protein-coding genes (active in nearly all mammalian cells) that oscillate in expression and activity in a circadian manner.
Repeated bouts of exercise resulting in physiological adaptations.
A diurnal cycle is any pattern that recurs every 24 hours, not necessarily biological or intrinsic.
A rhythmically occurring natural phenomenon that acts as a cue in the regulation of the body’s circadian rhythms.
- Oxygen consumption rate
The amount of oxygen consumed by metabolic processes in tissues, cells or organelles. When applied to measuring mitochondria, different metabolic states (states 1–5) of the mitochondria are used.
The interindividual differences in the circadian phase of activity patterns and sleep–wake cycles.
- Maximal power output
Maximal intensity of exercise or skeletal muscle contraction measured by power output (Watts).
- Maximal sprint
A short burst of intense exercise after which the individual is momentarily unable to continue owing to fatigue.
- Cycle ergometer
A fixed cycling machine often used in fitness testing to estimate exercise intensity.
- Periodized nutrition
The strategic combined use of exercise training and nutrition, or nutrition alone, with the overall aim to improve the physiological response to exercise training.
- Voluntary muscle force
Skeletal muscle contraction force produced as a result of endogenous activation of motor neurons.
- Acute exercise
A single exercise bout, rather than exercise training.
- Sleep hygiene
Habits and practices that are conducive to sleeping well on a regular basis.