Ethical approval and consent to participate in this study were obtained from the National Institutes of Health. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Seoul National University (SNU-220506-2).

Mouse-derived C2C12 myoblasts (ATCC, VA, USA) were cultured in a growth medium containing 90% Dulbecco’s modified Eagle’s medium (DMEM; Gibco, NY, USA), 10% fetal bovine serum, and 100 unit/mL penicillin- streptomycin (PS; Gibco) at 37 °C and 5% CO₂. To resolve the overcrowding phenomenon caused by cell proliferation, subcultures were performed every 48 h to maintain an appropriate number of cells. To differentiate C2C12 myoblasts into myotubes, a differentiation medium containing 98% DMEM, 2% horse serum (HS), and 100 unit/mL PS was used for a week, with the medium refreshed every alternate day, as previously described [18].

On the 5^th day of differentiation, C2C12 myotubes were cultured in a medium containing 50% DMEM, 50% HS, and 100 units/mL PS for 2 h to induce serum shock and synchronize the periodic expression of genes. After synchronization, the cells were washed twice with 100% DMEM, and the medium was replaced with 98% DMEM, 2% HS, and 100 unit/mL PS. AICAR treatment group was treated with 1 μM AICAR (Sigma-Aldrich, MO, USA), and the Dex (Sigma-Aldrich, MO, USA) treatment group was treated with 100 nM Dex. AICAR + Dex treatment group was simultaneously treated with 1 μM AICAR and 100 nM Dex. Gene expression was measured in all four groups at 4 h intervals for 36 h.

Nine-week-old male C57BL/6N mice (n = 28) were obtained from Nara Biotech Co., Ltd. (Seoul, Korea). Two weeks of acclimatization was performed using the Seoul National University Institute of Laboratory Animal Resources. Room temperature was kept constant at 21 °C and humidity was maintained at 50 ± 5%. The mice were individually housed under standard conditions with food and water provided ad libitum. All mice were divided into four groups as follows: non-stress + sedentary (NS, n = 7), nonstress + exercise (NE, n = 7), light cycle stress + sedentary (LS, n = 7), and light cycle stress + exercise (LE, n = 7) groups. An activity wheel (Tecniplast) with a 35 cm diameter and 25.8 cm width was used to exercise the mice. Body weights of the experimental animals were measured every Monday at 3:00 PM, and the number of wheel rotations in the exercise group was automatically measured by digital count and checked every Monday/Friday at 3:00 PM. For the light stress regimen, we adopted the model of Opperhuizen et al. (2017) and modified it [19]. The non-stress group was housed under a 12 h:12 h light/dark cycle, and the light cycle stress group was housed under a 14 h:10 h light (300 lx)/dark cycle for the last three weeks of the six-week experimental period. Two hours of light at night was started at 6 pm and was turned off at 8 pm. For biochemical analysis, mice were anesthetized using isoflurane. Hypothalamus and skeletal muscles were immediately frozen in liquid nitrogen and stored at −80 °C after sacrifice.

CLAMS (Oxymax CLAMS, Columbus Instruments) is a non-invasive device used to measure the vertical and horizontal movements, food intake, drinking water, VO₂ consumption, VCO₂ production, and heat production of animals in real time. Room temperature was maintained at 23 °C. Of the 32 mice used in the experiments, 12 mice (three in each group) were kept for 96 h in a metabolic analysis cage. For accurate analysis, we excluded the adaptation period, which was the first 24 h. The following parameters were measured in the metabolic analysis cage: volume of oxygen consumed (VO₂, mL/kg·h), volume of carbon dioxide produced (VCO₂, mL/kg·h), respiratory exchange ratio (RER), heat (Cal/h), accumulated food (g), accumulated drink (g), XY total activity (all horizontal beam breaks in counts), XY ambulatory activity (minimum three different consecutive horizontal beam breaks in counts), and Z activity (all vertical beam breaks in counts). All parameters were recorded at 30 s intervals. During CLAMS analysis, all mice were housed in a single cage without voluntary wheels.

For GTT, overnight-fasted mice (n = 12, 3 in each group) were intraperitoneally injected with glucose (2 g/kg) after 8 h of fasting. Blood was collected from the tail vein at 15, 30, 60, and 120 min and glucose levels were measured.

Total RNA was extracted from the quadriceps muscles and hypothalamus of the mice using a total RNA extraction kit (RiboZol, Amresco). Then, cDNA was amplified for 15-35 cycles using gene-specific primers for mouse Bmal1, Per1, nuclear receptor subfamily 1 group D member 1 (Nr1d1), Cry, and Tbp (Table 1). Total RNA (100 ng) was amplified using PerfeCTa SYBR Green FastMix (Quanta Biosciences) and an ECO PCR system (Illumina).

All statistical analyses were performed using the Graph-Pad Prism 9 software (GraphPad Software Inc., San Diego, CA, USA). In cell experiments, multiple t-tests were performed to compare control vs. Dex and Dex vs. AICAR + Dex at all time points (36 h at 4 h intervals), and multiple testing corrections were not performed. A t-test was performed to compare two groups among the four groups (NS, NE, LS, and LE) in animal experiments. Comparisons between groups were performed using one-way analysis of variance forPCR analysis.

We designed an experiment (Figure 1) to investigate whether exercise affects light stress-induced metabolic disruption. Light is also a potent circadian stressor. Mice in the LE and LS groups were raised in a 14 h light-on state for the last three weeks of the total 6-week experimental period (Figure 1A). Mice in the NE and LE groups were given free access to a running wheel for six weeks. At the beginning of week 4, the light cycle stress group was subjected to a 14 h:10 h light/dark cycle. During the 3 weeks of light stress, the running distance of the light-cycle-stressed mice was significantly decreased compared to that of the non-stressed mice (Figure 1B). There was no significant weight variance among the groups (p > 0.05) at week 0. Body weight gain was similar (p > 0.05) in the NS and NE groups during the 6-week intervention period (Figure 1C). However, there was significant weight gain (p < 0.05) in both the LE and LS groups compared with the NS group (Figure 1C). There was no significant difference (p > 0.05) in the total food intake (data not shown) among the groups. GTT indicates the ability of the body to regulate blood sugar levels. At 60 min after glucose injection, there was a significant difference (p < 0.05) in the GTT in the NE group compared to that in the other groups (Figure 1D). In the absence of stress, exercise reduces blood glucose levels at specific time points. However, at 90 min, there was a significant difference (p < 0.05) only in the LS group compared with the NE group (Figure 1D). There were no significant differences (p > 0.05) among the groups at the other time points (0, 30, and 120 min). Similarly, the glucose area under the curve during the GTT was not significantly different between the groups. VO₂ consumption and heat production increased during the night cycle but not under light cycle stress conditions

We assessed the differences in metabolic parameters among the groups using metabolic chambers. Distinct pattern changes were observed in the dark:light cycles of all groups (Figure 2A). Separate analyses of each phase showed that the RER level was significantly lower (p < 0.05) in the LS group than in the NS group during the dark phase (Figure 2A). However, this stress effect was blunted (p > 0.05) in the LE group compared to that in the NS group (Figure 2A). In the dark cycle, there was a significant increase in the x-axis activity compared to that in the light cycle in both the NS (p < 0.001) and NE (p < 0.05) groups, but there was no significant difference between the dark and light cycles (p > 0.05) in both the LS and LE groups (Figure 2B). However, there was no significant difference (p > 0.05) in the total RER during the dark cycle or day in the metabolic chambers among the groups (Figure 2C). In the dark cycle, there was a significant increase in VO₂ consumption (p < 0.05) compared with that in the light cycle in both the NS and NE groups, but there was no significant difference between the dark and light cycles (p > 0.05) in both the LS and LE groups (Figure 2D). Similarly, heat production also significantly increased in the dark cycle in the NS and NE groups compared to that in the light cycle, but not in the LS and LE groups (Figure 2E). These data imply that three weeks of light cycle stress was sufficient to inactivate metabolic parameters, even in voluntarily exercised mice.

Expression levels of circadian genes are shown in Figure 3. The hypothalamus is responsible for synchronizing internal rhythms. To examine the effect of exercise on circadian-related gene fluctuations induced by light-cycle stress, we measured the mRNA expression levels of Baml1, Per1, Cry1, and Nr1d1 in the hypothalamus. There were no significant differences in hypothalamic gene expression between the groups. Circadian genes within the skeletal muscles also play critical roles in the regulation of energy metabolism [20]. To reveal how exercise and light cycle stress influence circadian gene expression in skeletal muscles, we compared the relative mRNA expression levels of Baml1, Per1, Cry1, and Nr1d1 in quadriceps muscles. Per1 expression was significantly reduced in the quadriceps muscles of the LS group compared to that in the muscles of the NS group. However, there was no significant difference between the NS and NE or LE groups. These results underscore the possibility that exercise partially affects the light cycle stress-interrupted homeostasis of circadian-related transcription in skeletal muscles.

High serum concentrations induce the circadian expression of various genes whose transcription also oscillates in living animals [21]. In our experimental settings, differentiated C2C12 cells were subjected to the same regimen of serum treatment, and the levels of several mRNAs were monitored after treatment with an exercise mimetic (AICAR), a stress hormone mimetic (Dex), or both AICAR and Dex (AICAR + Dex). Comparative analysis showed that Bmal1 expression was significantly higher in the AICAR + Dex treatment group than in the control group at 16 h (Figure 4A). The mRNA expression level of Per1 was altered in the AICAR + Dex treatment compared to that of the control at 12 and 24 h, whereas Dex treatment alone significantly downregulated compared to that of the control at 16 h (Figure 4B). At 20 and 24 h, the mRNA level of Cry1 was significantly reduced in both the Dex-only and AICAR + Dex treatments compared to that in the control. However, at the 28 h time point, the expression level of Cry1 only changed in the Dex treatment compared to that in the control (Figure 4C). Compared to the control, the mRNA expression levels of Nr1d1 were upregulated in the Dex treatment group at 24 and 28 h (Figure 4D). Overall, these results imply that exercise-induced AMP-activated protein kinase (AMPK) activation could not fully reverse the effects of stress, but is a potential regulator that plays an important role in circadian gene oscillation disrupted by light cycle stress mimetic conditions.