Exercise Training Impacts Skeletal Muscle Clock in Adults with Prediabetes

Abstract

PURPOSE:

Disruption of the skeletal muscle molecular clock leads to metabolic disease, while exercise may be restorative, leading to improvements in metabolic health. The purpose of this study was to evaluate the effects of a 12-week exercise intervention on skeletal muscle molecular clock machinery in adults with obesity and prediabetes, and determine whether these changes were related to exercise-induced improvements in metabolic health.

METHODS:

Twenty-six adults (age: 66 ± 4.5 yrs; BMI: 34 ± 3.4 kg/m², FPG: 105 ± 15 mg/dL) participated in a 12-week exercise intervention and were fully provided isoenergetic diets. Body composition (DXA), abdominal adiposity (CT scans), peripheral insulin sensitivity (euglycemic-hyperinsulinemic clamp), exercise capacity (VO₂max), and skeletal muscle molecular clock machinery (vastus lateralis biopsy) were assessed at baseline and after intervention. Gene and protein expression of skeletal muscle BMAL1, CLOCK, CRY1/2, and PER 1/2 were measured by quantitative real-time PCR and Western blot, respectively.

RESULTS:

Body composition (BMI, DXA, CT), peripheral insulin sensitivity (glucose disposal rate; GDR), and exercise capacity (VO₂max) all improved (P<0.005) with exercise training. Skeletal muscle BMAL1 gene (fold change: 1.62 ± 1.01; P=0.027) and PER2 protein expression (fold change: 1.35 ± 0.05; P=0.02) increased, while CLOCK, CRY1/2 and PER1 were unchanged. The fold change in BMAL1 correlated with post GDR (r=0.43, P=0.044), BMI (r=−0.44, P=0.042), and body weight changes (r=−0.44, P=0.039) expressed as percent delta.

CONCLUSION:

Exercise training impacts skeletal muscle molecular clock machinery in a clinically-relevant cohort of adults with obesity and prediabetes. Skeletal muscle BMAL1 gene expression may improve insulin sensitivity. Future studies are needed to determine the physiological significance of exercise-induced alterations in skeletal muscle clock machinery.

INTRODUCTION

Disruption of circadian function is an emerging risk factor for metabolic disease (1). Shift workers are predisposed to obesity-related diseases, thought to be a detrimental consequence of being active at night (2–4). Troubling, this relationship persists even after controlling for traditional risk factors, including smoking, poor diet, and low physical activity (3). Highly controlled experimental studies in humans show that even short-term circadian disruption leads to metabolic derangements that mimic the pre-diabetic state (5). Animals with genetically disrupted molecular clocks develop obesity, hyperglycemia, and hepatic steatosis, characteristic of metabolic disease—even when fed a standard chow diet and when physical activity levels are maintained (6). These data and others (1, 7) open a new frontier, implicating that the origins of metabolic disease are directly linked to the circadian system beyond traditional lifestyle factors.

The circadian system encompasses a network of molecular clocks, including a central clock located in the brain, but also in peripheral tissues such as skeletal muscle (1, 8). The core molecular clock mechanism is a self-regulating transcriptional-translational feedback loop that includes BMAL1, CLOCK, CRY1, CRY2, PER1 and PER2 proteins. In skeletal muscle, this clock mechanism governs glucose metabolism (9). Skeletal muscle-specific BMAL1 knockout in animal models leads to glucose intolerance and hyperglycemia (10), findings which are paralleled in human primary myotubes isolated from patients with type 2 diabetes (11). These observations support the possibility that restoration of a disrupted skeletal muscle molecular clock may restore glucose homeostasis.

It is well accepted that exercise slows the progression of metabolic disease, in part by increasing insulin sensitivity (12). Given that the skeletal muscle molecular clock governs these metabolic processes (9), it may be that some of the beneficial effects of exercise are realized, in part, through actions on the skeletal muscle molecular clock. Exercise provides a time-setting cue, which allows the skeletal muscle clock to anticipate and prepare for upcoming energetic demands (13). Acute exercise increases skeletal muscle clock gene expression in humans (14, 15). Thus, exercise interventions may restore a disrupted skeletal muscle clock, contributing to improvements in whole-body metabolic health. Whether the skeletal muscle molecular clock is influenced by chronic exercise training, and how it may be linked to exercise-induced metabolic improvements is currently unknown.

The purpose of this study was to examine the effects of a 12-week exercise intervention on the molecular clock machinery in human skeletal muscle and investigate potential links between changes in skeletal muscle clock machinery and exercise-induced metabolic improvements. Relative changes (pre to post) in gene and protein expression were assessed for core components of the molecular clock, including BMAL1, CLOCK, CRY1, CRY2, PER1 and PER2. Secondarily, we compared changes in both gene and protein data to whole-body changes induced by exercise training, including body composition, in vivo peripheral insulin sensitivity, and exercise capacity. It was hypothesized that skeletal muscle molecular clock machinery would increase after 12-weeks of exercise training, and that these changes would be related to improvements in specific metabolic health outcomes including insulin resistance and body composition.

METHODS

Study Design

Skeletal muscle tissue samples were available from a previous exercise intervention. The initial study was designed as a randomized controlled trial to test the effects of a lifestyle intervention on cardiometabolic risk factors in adults with obesity (16). From this initial cohort, 26 sets of skeletal muscle samples were available for molecular analysis and were included in this study. Pre- and post-metabolic assessments were conducted over a 3-day metabolic control period in a Clinical Research Unit, which allowed activity and diet to be closely monitored. Participants were provided an isoenergetic dinner meal the evening prior to metabolic testing equivalent to one third of their basal metabolic rate. The macronutrient breakdown of the meal was 55% carbohydrate, 30% fat, and 15% protein. All metabolic testing was conducted after a 12 hour fast. All post-intervention metabolic tests were performed approximately 16 hours after the final exercise bout.

Participants

As previously described, participants were recruited from the local community using study advertisements. Participants were required to be between 60-80 years of age, BMI between 25-40 m/kg², weight stable (<2 kg weight change within the past 6 months), and sedentary (<20 minutes of exercise twice per week) prior to enrollment. Participants underwent screening procedures consisting of a medical history, physical examination, and clinical labs. Participants were excluded for smoking (within the past 5 years), excessive alcohol consumption (> 5 drinks per week), heart, kidney, liver, thyroid, intestinal and pulmonary disease or taking medications known to affect study outcomes. A resting 12-lead echocardiogram and submaximal exercise stress test was used to exclude participants with contraindications exercise. All women were postmenopausal and not using hormone replacement therapy. These criteria were described in the parent study and no other criteria were used for participant inclusion in this study. The study was approved by the Cleveland Clinic Institutional Review Board. All participants gave informed consent prior to participation (16).

Exercise Intervention

Participants completed a 12-week aerobic exercise intervention. During the intervention, participants were provided isoenergetic diets, under the supervision of a registered dietitian. Energetic (kJ) assignments for each participant were determined by multiplying basal metabolic rate (BMR) by a sedentary factor of 1.25. BMR was assessed using indirect calorimetry during the fasted state, immediately upon waking (Vmax Encore, Viasys, Yorba Linda CA). Diets were fully-provided during the 12-week intervention. Participants were randomly assigned to receive either a low glycemic index diet (40 arbitrary units) or a high glycemic index diet (80 arbitrary units). Macronutrient composition and fiber were matched between groups. Dietary adherence was ensured via daily food-container weigh backs plus weekly counseling with a dietitian. In the current study, N=11 were assigned to low glycemic-index diet, and N=15 were assigned to a high glycemic-index diet.

Exercise was supervised by an exercise physiologist and was performed 5 days per week at ~85% of heart rate max on a treadmill for 60 minutes. Compliance to the exercise prescription was confirmed during each session via Polar Heart Rate monitoring. Maximal oxygen consumption (VO₂max) was measured during an incremental treadmill test at four time points over the course of the study: pre-intervention, 4 weeks, 8 weeks, and 12 weeks. Accordingly, exercise training intensity was adjusted in 4-week increments based on the obtained heart rate during each maximal test. An exercise test was considered maximal if participants achieved at least 3 of the following criteria: 1) oxygen consumption plateau (<150 ml/min), 2) heart rate within 15 beats of age-predicted max, 3) respiratory exchange ratio >1.15, and/or 4) volitional exhaustion. On average, 95 ± 5% of exercise sessions were attended by participants throughout the 12-week intervention.

Metabolic Assessments

Body Composition.

Height and weight were each measured three times using a stadiometer and calibrated scale. Participants wore hospital gowns, and measurements were conducted after a 12-hr overnight fast. Body mass index (BMI) was calculated from these values. Body composition was assessed using dual X-ray absorptiometry (iDXA, Lunar, Madison WI) in order to assess relative and absolute fat amounts, as well as fat-free tissue.

Abdominal adiposity was measured using computed tomography (CT) scans at the L4 region in 4 mm slices. Images were analyzed for: abdominal circumference, subcutaneous and visceral adipose depots, and total abdominal depot (sum of subcutaneous and visceral adipose tissue depots).

Glucose Metabolism.

In vivo peripheral insulin sensitivity was measured using the gold standard euglycemic-hyperinsulinemic clamp technique, as described previously (17). While in the supine position, plasma glucose was maintained at approximately 90 mg/dL, using variable dextrose (20%) infusion via intravenous catheter (YSI 2300 STAT Plus analyzer). Venous blood samples were arterialized by warming the hand to 60°C. Concurrently, a primed-continuous infusion of insulin was administered into the contralateral arm at a dose of 40 mU/m²/min⁻¹ for 120 minutes. The glucose disposal rates (GDR) presented herein were determined from steady state during the 90-120-minute period of the clamp procedure.

Skeletal Muscle Biopsies

Tissue Collection:

These methods have been described previously (18, 19). In brief, skeletal muscle biopsies were obtained from the vastus lateralis muscle between 8-9 am after an overnight fast. The muscle tissue was immediately dissected free from fat and connective tissue and immediately frozen and stored in liquid nitrogen until further analysis.

RNA Isolation:

As previously described (19), muscle RNA was extracted from 10-20 mg of skeletal muscle tissue. These samples were homogenized in TRI Reagent^® (Sigma, St Louis, Missouri) and further isolated using an RNA isolation kit, according to the manufacturer’s instructions (Bio-Rad, Hercules, CA). RNA concentration and purity were assessed by measuring absorbance at 230, 260, 280 nm, using the NanoDrop ND-1000 Spectrophotometer (Thermo Scientific, Wilmington, Delaware). Random assessments of RNA integrity were performed using an Agilent bioanalyzer (Agilent, Santa Clara, CA).

cDNA Synthesis:

RNA samples were treated with DNaseI (Invitrogen, Carlsbad, CA), and total RNA was reverse transcribed into cDNA according to manufacturer’s instructions (iScript cDNA synthesis kit, Biorad, Hercules, CA). Primer pairs for target genes were obtained from previous literature, shown in Table 1.

Table 1.

Primers sequence for Real-time PCR

Gene Target	Forward	Reverse
BMAL1	5’-GTA ACC TCA GCT GCC TCG TC-3’	5’-TAG CTG TTG CCC TCT GGT CT-3’
CLOCK	5’-TCT TGA CCT TAT GCC ATT CCA-3’	5’-CTT ATG CTT TGT TGC TGT CAA cc-3’
PER1	5’-ACA AGC AAA TTT GGC AGC ATC-3’	5’-CCT GCT TCA GCA CAG AGG TCA-3’
PER2	5’-CGT TGG AAC CAC CCA GAC ATC −3’	5’-ATG CAG TCG CAA GCT GTC AGA-3’
CRY1	5’-TCT GGC ATC AGT ACC TTC TAA TCC-3’	5’-CTG TGT GTC CTC TTC CTG ACT AG-3’
CRY2	5’-GAC CAG GTT GCA GTG GCG TA-3’	5’-GCC CTG GAA GCC AAC AGA ATA A-3’
GAPDH	5’- CAC CAA CTG CTT AGC ACC CC-3’	5’-TGG TCA TGA GTC CTT CCA CG-3’

BMAL1: Brain and Muscle ARNT-Like 1; CLOCK: Circadian Locomotor Output Cycles Kaput; PER: Period Circadian Regulator 1; CRY: Cytochrome; GAPDH: Glyeraldehyde-3-phosphate dehydrogenase.

Semi-Quantitative Relative PCR Analysis:

mRNA expression analysis was performed in quadruplicate, as previously described (20). The threshold cycle (Ct) of each gene target was subtracted from the Ct of human GAPDH, serving as the internal assay standard (∆Ct). Relative changes in mRNA abundance (pre to post-intervention) were calculated using the comparative ∆∆Ct method, in which fold induction was calculated as an exponential of the negative value of the difference between ∆Ct pre and post.

Protein Analysis:

Western blot technique was applied to measure differences in skeletal muscle protein expression before and after exercise training in 11 of the participants for which sample was available. Approximately 2 mgs of skeletal muscle tissue were immersed in 500 ul of RIPA buffer (Thermo Scientific) with protease and phosphatase inhibitor. Samples were further homogenized with a tissue homogenizer (FastPrep-24 5G, MP Biomedical), followed by centrifugation at 15,000 g for 10 minutes at 4°C in order to remove debris. Protein concentrations were assessed by BCA (Thermo Scientific Pierce BCA Protein Assay Kit). Protein samples were further denatured in Lameli Buffer with 2% ß-mercaptoethanol in 100°C for 10 min. The denatured samples were separated using a 4-20% Novex Tris-Glycine SDS-PAGE (Invitrogen) at 125 volts for 90 minutes at room temperature. Samples were then transferred to a PVDF membrane, which was activated with methanol, at 100 volts for 90 minutes at 4°C. Ponceau staining was done to visually confirm sample transfer to the PVDF membrane. Membranes were blocked at room temperature using 5% BSA in TBS-T and incubated overnight in primary antibodies in 5% BSA in TBS-T - BMAL1 (Cell Signaling Technology, 1:1000, catalog no. 1402S), CLOCK (Cell Signaling Technology, 1:1000, catalog no. 5157S), PER 1 (Santa Cruz Technology, 1:500, catalog no. sc-398890), PER2 (Abcam, 1:500, catalog no. ab179813), CRY1 (Santa Cruz Technology, 1:1000, catalog no. sc-101006), CRY2 (Santa Cruz Technology, 1:200, catalog no. sc-293263) and Hsc70 (Santa Cruz Technology, 1:2000, catalog no. sc-7298 hrp). Membranes were subsequently incubated in appropriate secondary antibodies in 5% BSA in TBS-T for approximately 1 hour: anti-Rabbit (GE Healthcare Life Sciences, 1:2000, catalog no. NA930V) and anti-mouse (GE Healthcare Life Sciences, 1:2000, catalog no. NA9931V). Immunoreactive proteins were developed and visualized using chemiluminescence reagent (ECL Prime: MF). Protein expression was quantified by densitometric analysis using ImageJ. Hsc70 expression was used as a loading control for normalization on all blots.

Statistical Analysis

Data are presented as mean (“☐”) standard deviation. GraphPad Prism v7 was used to perform statistical analyses. Differences between pre- versus post-intervention were determined using a Student’s paired t-test. Based on the previous study design, changes in skeletal muscle clock machinery were analyzed for effects of dietary glycemic index during the 12-week intervention (low glycemic index vs. high glycemic index) using a two-way ANOVA (group x time). Statistical outliers were identified using a Grubbs test ($\alpha$=0.001). Pearson’s correlation coefficient was calculated to assess relationships between changes in whole-body metabolic outcomes versus changes in skeletal muscle molecular clock machinery. Correlational data were tested for normality using a Shapiro-Wilk test. Significance was accepted at P<0.05.

RESULTS

Exercise Intervention

Participant characteristics before and after the 12-week exercise intervention are presented in Table 2. Improvements in body composition were observed, evidenced by reductions in body weight (−9.2 ± 4.7%; P<0.0001), BMI (−10.0 ± 4.6%, P<0.0001), and DXA-assessed %body fat (−10.3 ± 9.8%; P<0.0001). Total abdominal adiposity, assessed by abdominal CT scans, also decreased by 20.8 ± 13.0% (P=0.012), which occurred as a result of reductions in both subcutaneous (17.2 ± 12.3%, P=0.008) and visceral (34 ± 26%, P=0.014) adipose tissue depots. Exercise capacity, assessed by VO₂max testing, increased by 28.3 ± 14.3% (P<0.0001). Peripheral insulin sensitivity improved, as glucose disposal rate significantly increased by 91.0 ± 72.1% (P<0.0001), assessed during the 90-120 min steady state period of the euglycemic-hyperinsulinemic clamp procedure.

Table 2.

Participant Characteristics Before and After a 12-Week Exercise Intervention

	Pre (n=26)	Post (n=26)	P value
Age (yrs)	66 ± 4	-	-
Sex (M/F)	13/13	-	-
Body Weight (kg)	98 ± 15	89 ± 12	< 0.0001
BMI (kg/m2)	34 ± 4	31 ± 4	< 0.0001
Body Fat (%)	43± 7	40 ± 8	< 0.0001
FPG (mg/dL)	105 ± 15	95 ± 9	0.001
HOMA-IR	3.5 ± 1.4	2.8 ± 1.0	0.001
GDR (mg/kg/min)	2.6 ± 1.1	4.6 ± 1.4	< 0.0001
VO2max (mg/kg/min)	21.2 ± 3.7	27.3 ± 6.3	< 0.0001
Total Abdominal Perimeter (cm2)	120 ± 10	115 ± 12	0.012
Subcutaneous Adipose Tissue (cm2)	490 ± 143	441± 138	0.008
Visceral Adipose Tissue (cm2)	111 ± 84	79 ± 52	0.014

Data are displayed as mean ± standard deviation. BMI: body mass index; FPG: fasting plasma glucose; FPI: fasting plasma insulin, GDR: glucose disposal rate from euglycemic-hyperinsulinemic clamp during 90-120 min time increment. VO₂max: maximal oxygen uptake.

Skeletal Muscle Molecular Clock Machinery

Gene expression of the core components of the skeletal muscle molecular clock were assessed, including BMAL1, CLOCK, CRY1, CRY2, PER1, and PER2. Relative changes pre- to post-intervention were quantified. Protein expression of the core components of the skeletal muscle molecular clock were also assessed in a subset of participants (n=11).

Gene Expression:

BMAL1 was significantly increased after the 12-week exercise intervention (Figure 1; Panel A; P=0.027). The effect of dietary glycemic index (low GI vs. high GI) was not significant (F=2.22, P=0.15). All other genes were unchanged (Figure 1, Panels B-F; CLOCK: P=0.108, CRY1: P=0.195, CRY2: P=0.837, PER1: P=0.774, and PER2: P=0.322). A single outlier was identified within the CRY1 and PER2 data sets, and removed from the analysis. It should be noted that inclusion or exclusion of these outliers did not influence summary statistics.

Figure 1.

Effect of a 12-week exercise intervention on skeletal muscle molecular clock gene expression, from the vastus lateralis muscle in adults with obesity and prediabetes. Relative changes (pre-post) in gene expression were measured using quantitative-real time PCR analysis, and calculated using the delta Ct method, in which the post-condition was compared to a normalized pre-condition (Panels A-F). Data are expressed as mean ± SD. Final sample sizes for pre/post pairs differed for among targets (BMAL1: n=22, CLOCK: n=22, CRY1: n=19, CRY2: n=24, PER1: n=22, and PER2: n= 19).

Protein Expression:

PER2 expression was significantly increased (Figure 2; Panel G; P=0.02). The effect of dietary glycemic index (low GI vs. high GI) was not significant (F=3.10, P=0.12). Representative images for 3 pre-post pairs for all protein targets, including the loading control (Hsc70) are presented in Figure 2, Panel A. All other protein targets were unchanged (Figure 2; Panels B-F; BMAL1: P=0.16, CLOCK: P=0.52, CRY1: P=0.95, CRY2: P=0.39, and PER1: P=0.84).

Figure 2.

Effect of 12-week exercise intervention on skeletal muscle molecular clock protein expression, from the vastus lateralis muscle in adults with obesity and prediabetes. Panel A: Representative Western Blots images for 3 pre/post pairs, for each molecular target. Panels B-G. Densitometric analyses of relative changes in protein expression, in which the post-condition is expressed a fold change relative to pre-condition. Data are expressed as mean ± SD. Final sample sizes for pre/post pairs differed for among targets (BMAL1: n=11, CLOCK: n=11, CRY1: n=11, CRY2: n=11, PER1: n=11, and PER2: n=10).

Correlational Analyses

Skeletal Muscle Molecular Clock Machinery:

Significant increases in BMAL1 gene expression correlated with post glucose disposal rate during the 90-120 min steady state period of the euglycemic-hyperinsulinemic clamp (Figure 3; Panel A; r=0.43, P=0.044), as well as declines in body weight (Figure 3; Panel B; r=−0.44, P=0.039) and BMI (r=−0.44, P=0.042), expressed as percent delta.

Figure 3.

Panel A: Correlational analyses among BMAL1 gene fold induction (pre-post) and post glucose disposal rate during the 90-120 min steady state period of the euglycemic-hyperinsulinemic clamp, after a 12-week exercise intervention. Panel B: Correlational analyses among BMAL1 gene fold induction (pre-post) and reductions in body weight after a 12-week exercise intervention, expressed as % delta.

DISCUSSION

The primary findings of this study are that skeletal muscle BMAL1 gene expression and PER2 protein expression significantly increased after a 12-week exercise intervention in adults with obesity and prediabetes. Exercise-induced increases in BMAL1 gene expression positively correlated with clinically meaningful outcomes, including enhanced peripheral insulin sensitivity, assessed by the gold-standard euglycemic-hyperinsulinemic clamp technique, as well as reductions in BMI and body weight. While this level of evidence is correlative, it is tempting to speculate that chronic exercise training may mitigate metabolic disease, in part, by impacting skeletal muscle molecular clock machinery.

Lifestyle interventions consisting of diet and exercise are first-line therapies for metabolic diseases, such as type 2 diabetes (21). The mechanistic underpinnings of the extensive health benefits realized from regular exercise are not completely understood, and it remains possible that the circadian system may be involved. Exercise inherently impacts skeletal muscle tissue, and recent literature suggests that exercise is capable of influencing the molecular clock mechanism (13). Few studies have investigated the effects of exercise on human skeletal muscle BMAL1 gene expression. A single bout of resistance exercise (10 sets of 8 repetitions at 80% of a predetermined repetition maximum) increased BMAL1 gene expression by approximately 1.2-fold assessed 6 hours after exercise in 4 untrained healthy men (14). Similarly, a single bout of aerobic exercise (70 minutes at 70% VO₂max) increased BMAL1 gene expression by 1.6-fold 4 hours after exercise, which further increased to 3.5-fold 8 hours after exercise, in 10 endurance-trained men (15). In agreement with recent literature, we observed a 1.6-fold increase in skeletal muscle BMAL1 gene expression, assessed approximately 16 hours after the final bout of the 12-week exercise intervention. While trending (Figure 2, Panel: B, P=0.19), we did not observe a mirrored increased in BMAL1 protein expression, and it may be that we were underpowered to detect an effect. Nonetheless, we add to the emerging evidence base by demonstrating that chronic exercise training also increases skeletal muscle BMAL1 gene expression in adults with obesity and prediabetes.

Previous work in preclinical models demonstrates that skeletal muscle BMAL1 governs glucose metabolism by regulating GLUT-4 protein translocation, insulin sensitivity (9), and substrate utilization (22). Skeletal muscle BMAL1 disruption leads to metabolic impairments that are observable at the whole-body level, such as glucose intolerance (10). The data herein are the first to extend this relationship to humans, as seen in a positive correlation between BMAL1 gene fold change and post-intervention peripheral glucose disposal rate (r=0.434, P=0.044) in a clinically-relevant cohort of adults with obesity and prediabetes. In addition, increases in BMAL1 also correlated with improvements in body composition, including reductions in BMI and body weight, providing further associative support for relationships between exercise intervention-induced increases in skeletal muscle BMAL1 and clinically meaningful outcomes. These data suggest that some of the variation in exercise-induced improvements in peripheral insulin sensitivity and body composition may be related to BMAL1 (approximately 19%). Future studies will be needed in order to determine the physiological significance of skeletal muscle BMAL1 in relation to exercise-induced improvements in metabolic health.

One of the novel findings in this study is the significant increase in PER2 protein expression in human skeletal muscle after the exercise intervention. PER2 is a core component of the molecular clock that works in concert with BMAL1 (23). Consistent with our work here, a previous study in rodent skeletal muscle found PER2 mRNA was increased approximately 1 hour after an acute bout of aerobic exercise (24). Together, these studies suggest that both acute and chronic exercise training impact PER2. The relationship between PER2 and exercise may be reciprocal in nature. While exercise increases PER2 expression, PER2 influences exercise capacity in animal models (25). Notably, PER2 knockout mice exhibit reduced exercise capacity at low and moderate intensities, when performed late during their active period (25). These current data add to mounting evidence of an interconnected relationship between molecular clocks and exercise capacity, a topic that warrants further study.

Other components of the molecular clock were unaltered by our assessments, and this is consistent with a previous study that reports changes in some clock, and clock-associated, genes in skeletal muscle after a 6-month lifestyle intervention in a cohort of adults with type 2 diabetes and coronary artery disease (26). The expression of human skeletal muscle clock genes rhythmically oscillates across a 24-hr period (27, 28). The skeletal muscle biopsy approach is advantageous for interrogation of tissue-specific clock machinery in skeletal muscle, as it allows for direct molecular assessment in humans; however, single biopsies are temporally limited as they do not provide information on rhythmic oscillations. Utilizing serial biopsies during a standardized time period would provide greater detail regarding the impact of exercise on the rhythmic patterns of skeletal muscle molecular clock machinery. One reason for the lack of these data is likely due to the inherent methodological challenges that come with evaluating skeletal muscle circadian function in humans, as serial sampling is burdensome on participants and recruitment. Newly emerging non-invasive assessments of in vivo skeletal muscle metabolism measurements, including ³¹P-magnetic resonance spectroscopy or near infrared spectroscopy techniques as well as in vitro assessments of temporal skeletal muscle clock gene expression may overcome this barrier.

Our study is not without limitations. Variables such as chronotype (one’s inherent preference for morning versus evening) (29), and seasonal variation may have influenced the effects of exercise on skeletal muscle clock machinery. Given the relative infancy of the circadian field with respect to skeletal muscle biology, it is difficult to discern the extent to which these factors may have influenced the outcomes of the current study. Time of day in which exercise is performed is gaining attention as a modifier of muscular responses to exercise(30). One the other hand, morning versus evening exercise did not influence changes in glycemic control and circadian rhythm wrist skin temperature after a 12-week multimodal exercise intervention in adults with overweight non-T2D and T2D(31). The current study did not standardize between-subject time of day in which exercise was performed, but within subject time of day was very consistent. Whether time of exercise contributes to response variability in skeletal muscle clock machinery is unknown. From a nutritional perspective, the timing, distribution, and macronutrient composition of meals may impact circadian-related metabolic responses (32–34). In this study we controlled macronutrient composition and fiber by providing all foods for the diets, which were distributed across 3 main meals (breakfast, lunch, dinner) plus snacks, but did not control for timing of meal consumption or snacking pattern. Tightly controlling dietary intake timing, distribution, and macronutrient composition may influence circadian outcome variability and may be of interest for future research in this field. Another key to interpreting findings is the consideration of time of day in which skeletal muscle biopsies were performed. In the current study, samples were collected between 8-9 am, the extent to which small differences in time of collection may contribute to response variability remains to be determined. Skeletal muscle clock gene expression fluctuates over the course of 24 hours (35), and larger effect sizes may be capturable by coordinating muscle biopsy times with peak diurnal gene expression. Whether changes in skeletal muscle molecular clock machinery were driven by exercise training per se or weight loss induced from the intervention cannot be determined from the current study design. Nonetheless, this study provides insight into the effects of chronic exercise training on skeletal muscle molecular clock machinery assessed in the morning in patient populations, and our findings can be used to guide the development and design of future studies in this area.

In conclusion, a 12-week exercise intervention influences skeletal muscle molecular clock machinery in humans, including significant increases in skeletal muscle BMAL1 gene expression and PER2 protein expression. BMAL1, a core component of the molecular clock, is necessary for whole-body glucose tolerance in preclinical models (10). We provide correlative data linking exercise-induced increases in skeletal muscle BMAL1 gene expression to clinically meaningful outcomes, including enhanced peripheral insulin sensitivity, as well as reductions in BMI and body weight, in a group of adults with obesity and prediabetes. Our findings provide evidence that the skeletal muscle molecular clock mechanism is targetable by chronic exercise training. Future studies will be needed in order to determine if the established metabolic benefits of chronic exercise are mediated, in part, through actions on the skeletal muscle molecular clock.