Answers are grounded in this paper and cite their sources.
Exercise and muscle physiologists have been interested in time-of-day or circadian properties of exercise performance and training since as early as the 1970s.1 There is a well-established diurnal variation in exercise performance components (strength and endurance) in both mice and humans.2,3 We and others have speculated that these time-of-day differences are partly driven by the inherent circadian clock-driven molecular changes that have been well characterized in skeletal muscle, reviewed previously here.4 Whether intrinsic muscle properties, motor unit firing rate or multi-tissue and/or organ interactions are the driving forces behind these diurnal differences in whole-body exercise performance is still understudied and to date, inconclusive.5,6 In their latest Function article,7 Fitzgerald et al. take a reductionist approach to assess whether intrinsic skeletal muscle contractile function, and contraction-stimulated glucose uptake show diurnal differences. An extensive set of ex-vivo experiments are presented from muscles with distinctly different contractile and metabolic phenotypes, as well as different loading and recruitment patterns. Extensor digitorum longus (EDL) and soleus muscles from young adult male and female C57BL/6NJ mice were put through 3 ex-vivo protocols, including force-frequency assessment to obtain specific tension values, a resistance to fatigue protocol and contraction stimulated glucose uptake. They performed these assays following muscle collection at 4 different times a day; zeitgeber times ZT1, ZT7, ZT13, and ZT19 (ZT12 = time of lights off). These times reflect the beginning and end of the light/inactive and dark/active periods for the mice. The outcomes of these experiments lead the authors to conclude that the assessed skeletal muscles do not display inherent time-of-day differences across all variables measured. We recognize the significant amount of work performed for all the assays described in this paper, however, we would like to respectfully challenge the authors’ conclusions. Circadian changes in physiology, unlike changes in gene expression, are commonly very small in magnitude but robust in daily repeatability. For example, circadian changes in human core body temperature and glucose tolerance are well established but they only display ∼2.5% intra-daily variation for body temperature and ∼12% for glucose tolerance.9 Circadian patterns of blood pressure are also clinically recognized (eg, dipper phenotype), and the normal average is about a 10%-20% difference over time of day. New clinical recommendations suggest that patients exhibiting <10% difference over time of day in their BP rhythms are now considered at higher risk for cardiovascular disease.10 Recognition of circadian skeletal muscle physiology is best described for human maximum isometric strength with diurnal differences ranging between 10%-15% from morning to afternoon.11 Biochemical measures of muscle mitochondrial function have identified ∼20% difference in ADP-stimulated (state 3) respiration between 1PM and 11PM and this day/night pattern is lost in muscle from Type II diabetics.12,13 These studies highlight that these small magnitude but robust variations in circadian physiologies are biologically relevant as well as clinically important. Thus, experiments to test for time-of-day physiologies need to carefully consider the parameters of the power analysis as well as the methods required for precise outcomes as seen in previous investigations.14 Our main concern with the Fitzgerald et al. paper is that we do not think the experiments presented are powered sufficiently to detect the significant differences that might be expected with muscle contractile function and contraction stimulated glucose uptake. We note that in their methods, the authors based their power calculations expecting a minimum of 20% difference which is higher than most known circadian physiologies as discussed above. In addition, visual assessment of the data presented in this publication suggest that the technical variability, particularly in the muscle mechanics measures, are higher than the variance used for their power calculations compounding the limits of this study. Thus, we propose that the results presented are insufficient to make any conclusion about time-of-day differences in muscle outcomes. We provide specific points in the sections below. Throughout the Function article, Fitzgerald et al. present muscle mechanics data from mice of similar age/weight in the same experimental group. As noted in the methods, they used standard deviations from a previous publication (see Figure 3B and 3I: Svensson et al. 2020) to calculate the power needed to detect a 20% difference. If one looks at Figure 1, panels E and H, and compares the variations around the means for specific tension, there is a much larger variance in this study compared to that reported in the Svensson et al. data.15 For example, in Figure 1E the variances around the mean for the EDL range from ∼15%-30% across the timepoints compared to ∼15% in the Svensson et al. data.15 The variability is larger for the soleus mechanics measures in Figure 1H ranging from ∼20% to >50% of the mean compared to ∼30% in the Svensson paper.15 With this range in variability with n = 6-8, it is not feasible to detect statistically significant differences that would be expected to be likely <20% across a day. For context, we note that age-associated declines in mouse muscle specific tension have been reported at ∼25% between 3 and 27 months of age in the EDL,16 and the decline in the dystrophic mdx mouse model is only ∼35% at 8-10 weeks of age in the tibialis anterior.17 We would expect any diurnal differences in specific tension of healthy muscle to be less than values considered to be more pathological. As an exercise, we performed power calculations using an estimated variation around the mean from data presented from the Fitzgerald et al. paper (20% or ± 56 kPa), and an expected difference to be a 15% change in specific tension over time of day and the expected specific tension being 280 kPa. With these parameters, we calculated that it would require a sample size of n = 14 muscles/timepoint to detect a significant change in specific tension. We recognize the significant amount of work performed in this study with challenging ex-vivo mechanics assays, but to discern 10%-15% differences over time of day requires a larger sample size and/or higher technical precision to support strong conclusions. As noted above, there are clear time-of-day differences in glucose tolerance test outcomes in humans and rodents.18 Work by Basse et al. (2018) performed time of day measures (4 timepoints, ZT1, ZT7, ZT13, and ZT19) on multiple different mouse skeletal muscles under basal and submaximal insulin stimulated conditions to test muscle intrinsic properties.19 In this paper, they report basal glucose uptake demonstrated significant time of day differences in the gastrocnemius and quadriceps muscles, but this was not detected in soleus or EDL muscles indicating muscle specificity of this property. We also speculate that as the rate of uptake is higher in these larger muscles, they may present better signal-to-noise outcomes for this assay. Insulin sensitivity showed time-of-day differences in the soleus and the EDL muscles with higher sensitivity at ZT19 vs. ZT7. The sample size for each timepoint was n = 6-8. In the Fitzgerald et al. paper, the authors focused on a different measure, time-of-day differences in contraction mediated glucose uptake. They conclude that there are no differences in contraction mediated glucose uptake in the soleus or EDL muscles. Again, looking at their data it is perhaps challenging to detect 20% differences in time-of-day contraction mediated glucose uptake when within single groups the delta ranged between ∼0.06 and ∼0.23 µmol/g/15 min, a difference of ∼50%-75% of the estimated mean value, (see Figure 2F, ZT7). Thus, we suggest that the results presented are not robust enough to conclude that there is no time-of-day difference in contraction mediated glucose uptake. To summarize, given our observations and based on existing circadian literature, we do not believe the authors’ conclusions are supported by the data presented when considering the design and technical variability/sensitivity of these assays. Whether differences in time-of-day exercise capacity are, as the authors discuss, related to exercise capacity test limitations, or whether these differences lie within biological factors such as intrinsic muscle properties, neural or hormonal factors, or organ system interactions are still to be determined but are essential biological and clinically relevant questions. The authors would like to acknowledge input from members of the Esser lab, as well as discussions with Dr Susan Brooks (University of Michigan). We also acknowledge funding support from Wu Tsai Human Performance Alliance AGR00023600 and NIH R01AR079220 to K.A.E. K.A.E. holds the position of Editorial Board Member for Function and is blinded from reviewing or making decisions for the manuscript.
No takes yet. Share an insight, caveat, or question.
Viggars et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: