Beneficial effects of voluntary wheel running on activity rhythms, metabolic state, and affect in a diurnal model of circadian disruption

The comorbidity between depression, cardiovascular diseases (CVD), and type 2 diabetes mellitus (T2DM) has been repeatedly described in the literature^1–4. Emerging evidence suggests that disruption of circadian rhythmicity may contribute to the development of these comorbidities, and we have proposed the term "circadian syndrome" for this combined pathological condition^5,6.

Circadian rhythms are the manifestation of an internal timekeeping system (the “circadian clock”) that is re-set (entrained) through daily environmental timing cues, especially light but also feeding and physical activity, allowing living things to anticipate periodic daily events, to orchestrate internal temporal programs of behavioral and physiological functions, and to flexibly set the order and scheduling of such functions to optimize fitness in the real world. In mammals, a “central” or “master” clock in the hypothalamus of the brain (suprachiasmatic nucleus [SCN]) synchronizes “peripheral” clocks found in cells, tissues, and organs throughout the brain and body to regulate metabolism and temporal physiology^7,8. This network of clocks is capable of adaptively re-aligning its oscillatory components under changing conditions, including to signals such as, e.g., temperature, blood glucose and oxygen levels, and glucocorticoids. Importantly, there is increasing recognition that pathological misalignment of network components—between the environment, behavior, SCN, and peripheral clocks (e.g., in shiftwork and jet lag) —has critical implications not only for an individual’s health and performance but also for the pathophysiology of aging and disease^9–12. Such circadian disruption in animal models, as well as in humans, may lead to the “circadian syndrome,” with depressive-like behavior¹³, CVD⁹, and T2DM¹³.

We have been investigating this syndrome in a unique animal model, the fat sand rat (Psammomys obesus). Sand rats are diurnally active in nature¹⁴, but when kept indoors under standard laboratory conditions they demonstrate an unstable, nocturnal phase preference, with low amplitude, and in some cases, no rhythm at all^5,15,16. Interestingly, similar findings have been reported for almost all diurnal species examined to date^17–22; conversely, such circadian disruption under normal laboratory conditions has not been documented in nocturnal rodents^16,22. The laboratory-associated disruption of circadian rhythmicity in diurnal rodents is even more pronounced when animals are kept under a short photoperiod regimen (5 h light:19 h dark)^15,23. Within approximately 8 weeks of short photoperiod acclimation with standard rodent diet, sand rats develop the "circadian syndrome," manifesting as glucose intolerance, elevated plasma insulin levels, cataracts, CVD, and depressive- and anxiety-like behaviors^5,6. Strong entrainment of the circadian system by keeping sand rats outdoors, where a wealth of biotic and abiotic variables cycle, or by bright light treatment, prevents the development of these disorders^15,23,24.

Notably, the depression- and anxiety-like phenotype induced by circadian disruption in mice and in sand rats is ameliorated upon provision of a running wheel^25,26, and wheel running is known to have synchronizing effects on the entrainment of circadian systems^25,27,28. Also, faster recovery of internal synchrony occurs following light/dark shift, and increased amplitude of SCN firing rates is seen compared to aged mice housed without a running wheel²⁹. The effects of wheel running are likely complex²⁹, but at least in part, effects on the circadian system may be mediated through the effects of exercise on skeletal muscles¹⁰. Skeletal muscle and bone have roles extending beyond regulation of locomotion and postural support, including the control of nutritional homeostasis, such as maintaining glucose and calcium levels. Feeding and exercise stimulate skeletal muscle tissues and change their functions, including the maintenance of tissue mass and metabolism³⁰. It is thought that through these interactions with skeletal muscles, exercise may regulate circadian factors that influence mental, metabolic, and cardiovascular health. For instance, deregulated circadian rhythms in skeletal muscles are associated with reduced glucose tolerance, as well as increased rates of diabetes and CVD^31,32.

To further explore these physiological effects on circadian rhythmicity and health, we have tested the influence of voluntary wheel running on the development of the full “circadian syndrome” in the fat sand rat model. Our results show that this rhythmic intervention has significant salutary effects on the metabolic and affective complications of the syndrome.

Our results show that voluntary wheel running for 11 weeks inhibits the development of the “circadian syndrome” in fat sand rats housed indoors in standard laboratory conditions. Daily rhythms of general locomotor activity and blood glucose are strengthened, blood glucose and plasma insulin levels are lowered, oral glucose tolerance is enhanced, and tests of depressive- and anxiety-like behaviors and memory are improved.

We found that sand rats kept without access to a running wheel were all behaviorally arrhythmic, whereas most sand rats kept with a running wheel showed a significant daily locomotor activity rhythm (7/10) and were either diurnal (5/10) or nocturnal (2/10), while only 3/10 were arrhythmic (Fig. 1). These results are consistent with those of an earlier study which showed that 3 weeks of voluntary wheel running had an enhancing effect on the activity rhythms of sand rats maintained under short photoperiod. Krivisky et al. (2015)²⁵ found that activity patterns of sand rats were more robust and less fragmented when they had free access to a running wheel, and the effect was more pronounced in animals acclimated to a short photoperiod. Of note, physical exercise elevates the arterial CO₂ tension (P_aCO2)³³, and Adamovich et al. (2019)³⁴ found that changes in carbon dioxide levels altered the expression of several core circadian clock genes and shifted circadian phase in cultured cells. We speculate that one mechanism by which voluntary exercise could influence the circadian system might be through the modulation of P_aCO2.

Voluntary wheel running had a significant effect on baseline blood glucose levels and on glucose tolerance, with the No wheels group showing significantly higher blood glucose levels than the Wheels group, both at baseline and 120 min after oral glucose administration in the oral glucose tolerance test (Fig. 2). Furthermore, the blood glucose rhythm demonstrated by the Wheels group was much more pronounced compared with the No wheels group (Fig. 3). Accordingly, plasma insulin levels were lower in the Wheels group compared to the No wheels group (Fig. 4). These results are in line with previous findings showing that treatments which enhance and strengthen the circadian system in sand rats prevent the development of T2DM that otherwise occurs under standard laboratory conditions. For example, Bilu et al. (2019)²⁰ showed that when kept outdoors in laboratory cages, sand rats were diurnal, demonstrating robust daily rhythms in blood glucose levels, and did not develop T2DM. This contrasted with sand rats kept indoors under similar conditions, which became nocturnal or arrhythmic, had constantly higher blood glucose levels with no significant daily rhythmicity, and developed glucose intolerance¹⁵. Along this line, Bilu et al. (2020)²⁴ found that sand rats subjected to morning bright light treatment, which enhances circadian rhythmicity^23,24, showed robust daily rhythms in clock gene expression and behavior, and were normoglycemic, with higher glucose tolerance than the control group.

Blood glucose homeostasis is manifested by daily rhythmicity in glucose tolerance and in blood glucose levels³⁵. At the beginning of the activity/feeding phase, when blood glucose is mainly of dietary origin³⁵, glucose tolerance and blood glucose levels are at their highest³⁶. During the resting/fasting period, glucose is recruited from endogenous gluconeogenesis in the liver, and glucose tolerance and blood glucose levels are lower³⁶. Daily rhythmicity in blood glucose levels is regulated directly by the SCN and by peripheral clocks in the liver, pancreas, muscle, and white adipose tissue³⁵. When the SCN is lesioned in rats, the daily fluctuation in glucose uptake and insulin sensitivity disappears³⁶. Furthermore, mutations of clock genes cause disruption of glucose homeostasis, with severely disturbed daily blood glucose rhythms, hyperglycemia, and hypoinsulinemia^36–38. However, rhythms in glucose tolerance and in blood glucose levels also emanate from behavioral rhythms in exercise and feeding^39,40. Skeletal muscle is the major tissue responsible for insulin-mediated glucose utilization. The plasma membrane GLUT4 content is correlated with glucose transport activity in both animal models and human skeletal muscle⁴¹. Regular exercise results in elevated insulin- and contraction-stimulated glucose transport capacity by augmenting skeletal muscle GLUT4 levels⁴², and GLUT4 activity demonstrates circadian rhythmicity with higher expression levels and translocation during the activity phase than the rest phase⁴³. In addition, feeding behavior modulates daily rhythms of blood glucose and glucagon concentrations⁴⁰. The classic function of glucagon is to increase hepatic glucose output when glucose concentrations decline during fasting⁴⁰. Ad libitum–fed rats show peaks in glucagon concentrations during the resting/fasting phase, shortly before the onset of the active/feeding phase⁴⁰. In contrast, rats that start fasting at the onset of their activity phase have decreased plasma glucagon concentrations at the end of the resting period, instead of the expected increase⁴⁰. Thus, further experiments will be required to determine the relative contributions of the SCN, peripheral clocks, and behavioral rhythms of exercise and feeding in accounting for the differences between the Wheels and the No wheels sand rats in glucose tolerance and blood glucose rhythmicity.

In contrast to our previous study²⁰ reporting the development of mature cataracts in animals kept under short photoperiod conditions, we did not find a significant effect of wheel running on the development of mature cataracts (although 4/12 of the No wheels animals had mature cataracts, versus none in the Wheels group). This difference might be explained by the different length of the two experiments: Bilu et al. (2019)¹⁵ kept the sand rats under short photoperiod conditions for 20 weeks, whereas the current experiment lasted only 11 weeks, a time which might not be sufficient for the development of mature cataracts in these animals.

Another contrast with previous studies was the difference in heart weight, heart/body weight, and left ventricular wall thickness between the groups. In earlier studies we found that short photoperiod acclimated sand rats had a larger heart weight and heart/body weight ratio than neutral photoperiod acclimated sand rats¹⁵, and that animals with diabetes had a larger heart weight than non-diabetic sand rats⁵. We ascribed these results to the pathological cardiovascular effects of hyperglycemia, such as elevated aortic stiffness and reduced myocardial metabolism, which alters diastolic function in patients with diabetes mellitus. Furthermore, higher fasting blood glucose levels and glycated hemoglobin are associated with abnormal left ventricular relaxation in people with diabetes⁴⁴. However, in the current study, the heart weight and heart/body weight ratio were larger, and the left ventricular wall was thicker, in the Wheels group and not in the diabetic No wheels group (Fig. 5). These results, which suggest left ventricular hypertrophy (LVH) in the Wheels group, could be explained by a physiological morphological adaptation to exercise ("athlete's heart")⁴⁵, rather than the pathological LVH seen in the previous studies of diabetic sand rats^5,15,16. In this case, since the Wheels group ran for about 5 h a day on their wheels, it is possible that the LVH we found is similar to that found in athletes participating in sporting disciplines involving intensive isometric exercise, like long-distance running⁴⁵. These athletes may exhibit substantial physiological increases in average ventricular parietal mass, and in left ventricular (LV) parietal thickness, without impairment of contractile strength or LV performance⁴⁵. However, in order to differentiate between a purely physiological, adaptive process vs. a pathological hypertrophic cardiomyopathy, further experiments will be needed to examine aspects of the sand rat heart's anatomical and functional characteristics, such as LV cavity size, LV diastolic function, and the absence of T-wave inversion on electrocardiography.

The higher liver weight in the No wheels group compared to the Wheels group (Fig. 6) is consistent with earlier findings of an increased liver size and abnormalities of liver enzymes in patients with diabetes mellitus⁴⁶. During periods of hyperglycemia, glucose freely enters the hepatocytes, driving glycogen synthesis, and eventually causes the accumulation of excessive amounts of glycogen in the hepatocytes (glycogenosis), causing hepatomegaly and elevated aminotransferases. These abnormalities are readily reversible with sustained euglycemic control⁴⁷. The other major cause of hepatomegaly in people with diabetes is steatosis. This is a function of the body habitus and state of insulin resistance, rather than glycemic control, and it may progress to fibrosis and cirrhosis⁴⁷. In order to distinguish between these two causes of hepatomegaly, histological studies will be needed to assess for swollen hepatocytes containing excess glycogen in the cytoplasm, and often also in the nucleus. These are indicative of glycogenosis⁴⁸. There was no significant difference in liver/body weight between the groups.

There has long been discussion regarding the ameliorating effects of exercise on cognitive functions and affect. However, exercise has only recently received the attention of the scientific community with the key interest in its effects on cognitive functions, affective disorders, spatial learning and memory, and as a non-drug method to maintain brain health⁴⁹. Cross-sectional studies associate high self-reported levels of regular exercise with improved mental health and low depression⁵⁰. Moreover, self-reported high levels of habitual physical activity correlate with fewer symptoms of both anxiety and depression and better mental health⁵¹. The neurobiological effects of exercise, acting as a mood elevating agent, appear to influence several neural mechanisms related to depression and anxiety⁵¹. There is evidence that physical activity causes physiological changes in monoamine levels⁵², upregulates neurotrophic factors⁵³, alters the levels of the stress hormone cortisol⁵⁴, and leads to adaptations in limbic structures implicated in depressive and anxiety disorders.

A relationship between circadian rhythmicity and affective disorders has been observed at the biochemical, molecular, and clinical levels. Manipulation of light exposure affects various neurotransmitter systems related to mood⁵⁵, including adrenaline and serotonin⁵⁶. Antidepressants and mood stabilizers affect systems related to circadian rhythmicity⁵⁷, and modifications of genes related to circadian rhythms influence affective-like behaviours. However, the mechanisms underlying the relationship between circadian rhythmicity and affective disorders are still not yet fully understood. In the present experiment, we found that animals maintained without running wheels showed higher anxiety- and depressive-like behaviors compared to animals maintained with running wheels (Fig. 7a, b). These results replicate previous findings, showing that in sand rats maintained under short photoperiod conditions, under which they display increased anxiety- and depressive-like behaviors⁵⁸, these disorders could be ameliorated by voluntary wheel running²⁵. Similar effects of exercise on anxiety- and depressive-like behaviors were previously described in other animal models of depression, such as mice⁵⁹, and rats⁶⁰, as well as humans⁶¹. Whereas it is possible, at least in part, that the ameliorating effects of running wheels on the FST reported here could be the result of better physical fitness, this explanation seems unlikely to account for the entire profile of effects. We found no correlations within the Wheels group between the level of general activity and the measures obtained in the behavioral tests, arguing against the level of activity per se as an explanation for the behavioral results.

In addition to lower anxiety- and depressive-like behaviors, sand rats with access to running wheels showed higher preferential exploration of the novel object in the NORT, indicating better recognition memory (Fig. 7c). Notably, microarray studies have demonstrated that wheel running elevates hippocampal BDNF mRNA and protein levels, as well as the levels of its high-affinity receptor TrkB⁶². Upregulation of these molecules augments neurogenesis in the dentate gyrus of the hippocampus, a brain structure that is crucial for memory function, thus improving spatial learning and memory⁶³. BDNF, its receptors, and epigenetic modulators are involved in the pathophysiology of affective disorders⁶⁴, T2DM⁶⁵, and circadian system function⁶⁶. In an earlier study we found that acclimation to short photoperiod resulted in a diminished circadian rhythm of BDNF mRNA expression levels in the frontal cortex and SCN compared to neutral photoperiod. The sand rats that showed diminished BDNF circadian rhythms also demonstrated higher blood glucose and insulin levels, as well as significantly higher anxiety- and depressive-like behaviors compared to animals acclimated to neutral photoperiod. We speculated that BDNF may, at least in part, mediate the effects of circadian disruption on the development of the “circadian syndrome” in sand rats via the SIRT1-BDNF-Trkb pathway⁶⁷.

In conclusion, our results demonstrate that voluntary wheel running ameliorates the metabolic and affective complications of the “circadian syndrome” in a diurnal animal model of circadian disruption, supporting an important role of rhythmic physical activity in modulating pathological features of the syndrome. The mechanisms mediating the beneficial effects of wheel running on components of the syndrome are likely complex and multifactorial, but we hypothesize that stabilization of disrupted circadian clock and rhythm function plays a key role. We suggest that the utilization of a diurnal rodent animal model offers an effective way to further analyze the metabolic, cardiovascular, and affective behavioral changes related to circadian disruption and their underlying mechanisms.

This research was supported by a grant from the Israel Science Foundation (ISF) (Grant No. 866/17) to Noga Kronfeld-Schor, and by a grant from the United States -Israel Binational Science Foundation (BSF) (Grant No. 2015118) to William J. Schwartz and Noga Kronfeld-Schor.

C.B., N.K.S, P.Z., W.J.S., and H.E. conceived and designed the study. C.B. and V.V.D. performed the study, C.B., N.K.S., and H.E. analyzed the data. C.B., N.K.S, P.Z., W.J.S., and H.E. wrote the manuscript.

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

The authors declare no competing interests.