Timed Running and Eating During Daytime Increase Running Speed on a High-Fat Diet Without Changing Total Daily Running

Physical activity and food intake are closely intertwined, and the correct timing of these behaviours plays a significant role in the regulation of energy metabolism and homeostasis. When the timing of food intake and physical activity is not in line with the day–night cycle, which occurs during shift work, insomnia, or jet lag, it can lead to a situation in which peripheral and/or central clocks are no longer synchronized, also known as circadian misalignment. Circadian misalignment is known to impair glucose tolerance in shift workers, which can contribute to the development of type 2 diabetes mellitus [1]. In mouse models, which are nocturnal animals, food intake during their innate inactive phase (i.e., the light phase) also causes insulin resistance and exacerbates weight gain [2,3].

Current-day society is characterized by a dietary environment that is ‘unnaturally’ energy-dense and hyperpalatable, comes in virtually unlimited quantities, exploits the limbic system, and supersedes peripherally derived satiation and adiposity signals [4]. The well-described deleterious effects of a high-fat diet (HFD) consumption on the development of obesity can be further amplified by the erroneous timing of caloric consumption. Overconsumption of an HFD not only increases body weight, adiposity, glucose, and insulin levels [3] but also modifies the daily profile of wheel running behaviour and circadian synchronization to light [5]. Previously, we demonstrated in male Wistar rats that consumption of a free-choice high-fat high-sugar diet restricted to their innate inactive phase (i.e., the light phase) advanced hepatic clock rhythms [6] but did not reduce the amplitude of some soleus clock rhythms as observed with a chow diet restricted to the light phase [7]. Thus, both timing and nutritional factors seem to play a significant role in regulating clock gene expression rhythms of core metabolic organs (i.e., the liver and muscle).

Voluntary wheel running (VWR) has beneficial effects on several physical parameters related to metabolic health, such as body weight, fat content, and lean mass, and these beneficial effects can be further enhanced by a correct timing of physical activity [8,9,10,11]. In mice, VWR advanced the phase of the hepatic clock by 4 h in both chow- and HFD-fed conditions [12]. In rats, VWR in the inactive light phase did not shift the hepatic clock but did reduce amplitude of the soleus clock rhythms, whereas simultaneous time-restricted access to both a running wheel and a standard chow diet during the light phase successfully prevented the dampening of the soleus clocks and shifted both the hepatic and soleus clock rhythms by ~12 h, thus keeping them aligned [13].

Here, we investigated whether this combination of time-restricted voluntary wheel running and time-restricted feeding could ameliorate the negative metabolic health effects of a high-fat diet, such as increased body weight and fat gain, as well as prevent HFD-induced dysregulation of clock gene expression rhythms in liver and muscle.

Here, we report that time-restricted feeding in combination with time-restricted running is not only beneficial during consumption of a chow diet (as published before [13]), but also when animals are eating a high-fat diet (HFD). The dark-phase running and feeding group on a high-fat diet (DRDF-H) had a lower body weight and fat mass percentage not only compared to the sedentary group on a high-fat diet (NR-H) and the light-phase running and feeding group on a high-fat diet (LRLF-H), but even compared to the sedentary group on a chow diet (NR). Remarkably, the LRLF-H group displayed a larger initial running peak than the LRLF group on chow, although the total daily running distance during the time restriction period did not differ (Figure S1I). No such differences in running activity were observed between the DRDF-H and DRDF groups. The hepatic clock gene expression patterns of the time-restricted HFD groups were similar to those of our previous time-restricted cohorts on a chow diet [13], suggesting that the combination of LR and LF prevented the reduction in the clock gene oscillation amplitude in hepatic clock genes caused by light-phase feeding (LF) in combination with HFD, as observed in our previous study [6] (Figure S2). In contrast to our previous chow time-restricted feeding and running study that was performed with the exact same experimental setup and design as the current study [13], the acrophase and amplitude of some soleus clock genes were affected by the time-restricted protocol under HFD conditions (Figure S3). The mean relative expression of the lipid droplet-associated Plin5 gene was increased by HFD and decreased by the combination of time-restricted feeding and running in both chow and HFD rats. The expression rhythm of Plin5 was shifted in the light phase time-restricted feeding and running groups under both chow and HFD conditions (See Figure 5 for an overview of our findings).

While the weight gain of LRLF-H was higher than that of chow NR, the fat gain of LRLF-H did not differ from that of NR, which indicates an increased fat-free mass (e.g., muscle) gain in LRLF-H compared to NR. The time-restricted feeding and exercise groups on a high-fat diet (DRDF-H and LRLF-H) showed no difference in caloric intake compared to the sedentary high-fat diet group (NR-H). This contrasts to our previous chow results [13], in which the dark phase restricted (DRDF) group consumed significantly more calories than the NR group, while the chow light phase-restricted (LRLF) group consumed fewer calories than the DRDF group. In our previous time-restricted feeding study using a free-choice high-fat high-sucrose (fcHFHS) diet, we found that time-restricted feeding reduced the caloric intake of chow, but did not affect total fat and sucrose intake compared to ad libitum feeding [6]. This preference for high caloric food potentially explains the absence of an effect of time restriction on total caloric intake in the current HFD study in contrast to our previous chow study [13].

When comparing the current HFD study with our previous chow study, the DRDF-H and DRDF groups showed similar body weight increase, but fat gain was still significantly higher in DRDF-H than in DRDF. Therefore, DRDF could not completely prevent the effects of HFD. In other words, the correct timing of food intake and physical activity may partly, but not fully, reduce the negative metabolic effects of HFD. On the other hand, fat mass and gain of fat mass were similar between the LRLF and DRDF-H groups, indicating that eating healthy but being active at the wrong time of the day is as detrimental as eating unhealthy and being active at the right time of the day. This is in line with human studies reporting that exercise during the late active phase (at 18:30, maximum of 1 h) only partly reversed HFD-induced changes in the metabolic profile [15]. Studies in male Swiss mice fed an HFD also reported that only time-restricted feeding during the dark phase (TRF, ZT 16–0) combined with exercise (ZT0–1), but not time-restricted feeding alone, improved respiratory exchange rate, energy expenditure, mitochondrial respiration, decreased lipogenic and glucogenic gene expression, and attenuated HFD-induced changes in the insulin signaling pathway in the liver [16,17]. In addition, voluntary wheel running did not fully counteract the negative metabolic effects of HFD, including adiposity and free cholesterol levels in mice [18]. Thus, both time-restricted feeding and exercise are effective methods to combat the negative effects of an HFD but may not be able to completely compensate for these effects depending on the metabolic measurement of interest.

In our previous study, the soleus Per1 and Reverb-a rhythms were dampened by light running (LR), and the combination of LR and LF partially prevented this dampening [13]. The current study shows that under HFD conditions, the LRLF combination successfully prevented this dampening effect of eating at the wrong time of day. However, this “rescuing” effect of LRLF does not work for all genes, as the soleus Per2 rhythm was still dampened in LRLF-H. Considering studies reporting that both LF and HFD reduces soleus Per2 expression [19], and both HFD-dark-phase restricted feeding and HFD-ad libitum feeding prevent dampening of soleus Per2 [7], soleus Per2 is likely strongly regulated by both dietary content and timing. As Per2 directly regulates PPARγ activity, which is a nuclear receptor essential for lipid metabolism, based on the adiposity and other metabolic outcomes, it is not surprising that the “rescuing” effect of wheel running was limited when performed at the wrong time of the day (i.e., LRLF and LRLF-H) [20].

In the liver too, the results of the LRLF-H group revealed a protective effect of the LRLF combination on the dampening of some core clock gene rhythms triggered by HFD. A previous study showed that HFD in combination with LF either dampened (Per2 and Reverba) or caused a small phase advance (Bmal1 and Cry1) in core clock gene expression rhythms in the liver [6]. However, we found no dampening or small phase shifts in LRLF-H compared to the LRLF chow group, suggesting that the combination of LR and LF has a stronger influence on liver clock gene rhythms than HFD. This is in line with the observation that TRF prevented a loss of diurnal rhythmicity that occurred during HFD-induced obesity in mice [21], also suggesting that the influence of TRF on diurnal rhythmicity is more powerful than that of HFD consumption.

In our previous study with a chow diet, the acrophase of DRDF was similar to that of NR and was also in antiphase with LRLF in both the liver and the soleus muscle. In this study, a phase advance of DRDF-H compared to NR-H was observed in three out of the seven soleus clock genes tested. Moreover, the remaining four genes showed a trend toward a phase advance. As a result, the majority of clock genes no longer showed a clear antiphase between LRLF-H and DRDF-H. In the liver, a phase advance of DRDF-H compared to NR-H was only observed for Bmal1, while most other genes retained the antiphase between DRDF-H and LRLF-H despite the HFD.

The current study demonstrated the gene specific response of clock genes towards HFD in combination with time-restricted running and feeding. An increasing number of studies have provided evidence that clock gene responses are dependent on organs, tissue type, feeding timing, diet, exercise timing, the combination of both, and many other factors [22,23,24]. One study reported that TRF alone can affect the expression or rhythmicity of 80% of all genes in at least one tissue [25]. Further studies on clock gene expression rhythms and their specificity are needed to provide the necessary insights on how to prevent or repair circadian misalignment to treat or prevent its negative health effects.

A previous study has reported that LF enhances running endurance via myocyte Plin5 and induces a daily rhythm of muscle Plin5 expression in a clock-dependent manner in mice [14]. In the current study, we only found a higher activity peak in the LRLF-H versus LRLF group at the beginning of the light period, which may be linked to increased running endurance. No differences in total amount of running were found between LRLF-H and DRDF-H, although running in the LRLF-H animals was concentrated in 4 h versus 10 h in the DRDF-H group. In addition to the mouse versus rat difference, the major difference is that while Xin et al. (2023) utilised a treadmill and a forced running model to assess maximal running capacity, we used a voluntary running model [14,26]. Although several studies have shown that voluntary running increases running endurance [27,28,29,30], our observations may not strictly increase endurance, but rather increase running intensity, as it does not challenge the maximal exercise capacity of rats.

Several studies agree that in male Wistar rats, running at 22–25 m/min without a slope is considered high-intensity, and 22 m/min for 50–60 min is commonly used for an endurance treadmill training model that corresponds to a maximal oxygen consumption (VO2max) range of 65–75% [31,32,33,34,35]. In the present study, the LRLF-H group max. ran at 450 m/h on the running wheel, which is equivalent to 7.5 m/min. Running distances differ depending on the type of running wheel, for instance, the average running distance for an angled wheel is 10–20 km/day, while for an upright running wheel, it is 8–10 km/day in mice [29]. Taken together, the intensity of voluntary running in the LRLF-H group was equivalent to more than 15 m/min on a treadmill, which is considered moderate exercise. However, differences in size, material, voluntary vs. forced running, and several other factors make it challenging to compare the exercise load between a treadmill and a running wheel.

Previous studies have also reported that HFD upregulates Plin5 expression in the heart [36], liver [37,38], and muscle tissues [39,40]. In this study, we confirmed a significant increase in muscle Plin5 expression in HFD NR compared to chow NR. In addition, we found that under both chow and HFD conditions, relative Plin5 expression was downregulated by the combination of time-restricted running and feeding compared to the non-runners. These results are in line with previous studies in both mice and humans. A study in mice demonstrated that strength training decreases Plin5 level in heart and skeletal muscle tissues [41]. Also, human muscle biopsy results from Xin et al. (2023) confirmed that exercise reduces Plin5 expression in a time-of-day-dependent manner in both young and adult cohorts [14].

Despite the increased wheel running intensity in the early light phase in the LRLF group on an HFD, we did not find the time-of-day-dependent increase in muscle Plin5 expression reported in rats. To our knowledge, there have been no studies on the combined effect of LF and exercise on Plin5 expression, where the stimulatory effect of LF on Plin5 expression may be antagonized by the inhibitory effect of exercise. Based on the result of the current study, it is likely that the effect of LR is stronger than that of LF. However, another possibility is the experimental duration. Plin5 in myokines has been reported to drastically increase FGF21 expression in muscle and elevate serum FGF21 concentration [42]. However, this effect seems to be dependent on the experimental duration. While acute exercise increases circulatory FGF21 levels, chronic exercise lasting four or more weeks has the opposite effect [43]. Under HFD conditions, it has been demonstrated that wheel running causes no change in Plin5 mRNA nor protein expression [37], which is different from our current results. This discrepancy could be due to a difference in the composition of the high-fat diet or species differences. Lack of intensity or duration of the wheel-running activity is unlikely to be the contributing factor. The duration of our experiment was about seven weeks, while the experiment of Tuikka et al. lasted ten weeks. Regarding running distance, in the study of Tuikka et al. the maximum running distance between chow and HFD did not differ in the first part of the experiment, while the general running distance was double in chow than those of HFD in the 2nd half of the experiment [37]. This difference is similar to that observed between running in the active-phase and inactive-phase [13]. The fact that the decrease in Plin5 in our study was observed in both LRLF and DRDF groups under both chow and HFD conditions suggests that the running distance in the experiment of Rinnankoski-Tuikka was likely sufficient.

We found that the combination of time-restricted feeding and time-restricted running ameliorates high-fat diet-induced weight gain and body fat accumulation, ameliorates LF-induced disturbances in the soleus clock more effectively than under chow conditions, and has a protective effect against high-fat diet-induced changes in liver clock gene expression. Taken together, the combination of time-restricted feeding and time-restricted running helps to partially prevent the negative effects of an HFD on both metabolic health and clock gene expression rhythms.

We acknowledge Maurice Heemskerk and Dirk van der Werf from the Mechatronics department of the NIN for their assistance and support for the running wheel setup and the software for recording and regulating. We also thank Kim Falize, and the IWO unit for their help during the experiment.

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms26157658/s1↗.

Conceptualization, A.K.; Methodology, A.K.; Validation, A.K.; Formal analysis, A.S.; Investigation, A.S. and R.T.; Resources, A.K.; Data Curation, A.S. and R.T.; Writing—Original Draft Preparation, A.S.; Writing—Review and Editing, A.K., J.D.M., D.J.S., C.-X.Y. and A.S.; Supervision, A.K., J.D.M., D.J.S., C.-X.Y. and E.F.; Funding Acquisition, A.K., D.J.S. and J.D.M.; Guarantor, A.K., J.D.M., D.J.S. and C.-X.Y. All authors have read and agreed to the published version of the manuscript.

All experimental procedures were performed in accordance with the European guidelines for laboratory animals (EU Directive 2010\63\EU) and were approved by the Dutch Central Committee for Animal Experiments (CCD permits: AVD80100202216157) and the Agency for Animal Welfare (IvD) committee of the Netherlands Institute of Neuroscience (Royal Dutch Academy of Sciences).

Data will be made available on reasonable request.

The authors have no conflict of interest.

A.S. was supported by BioClock Consortium (by the NWA-ORC program of the Dutch Research Council (NWO; project number 1292.19.077)). C.X.Y. was supported by the Netherlands-Canada Type 2 Diabetes Research Consortium (ZonMW 459001021).