Effect of early vs. late time-restricted high-fat feeding on circadian metabolism and weight loss in obese mice

Abstract

Time-restricted feeding (TRF) limits the time and duration of food availability without calorie reduction. Although a high-fat (HF) diet leads to disrupted circadian rhythms, TRF can prevent metabolic diseases, emphasizing the importance of the timing component. However, the question of when to implement the feeding window and its metabolic effect remains unclear, specifically in obese and metabolically impaired animals. Our aim was to study the effect of early vs. late TRF-HF on diet-induced obese mice in an 8:16 light–dark cycle. C57BL male mice were fed ad libitum a high-fat diet for 14 weeks after which they were given the same food during the early (E-TRF-HF) or late (L-TRF-HF) 8 h of the dark phase for 5 weeks. The control groups were fed ad libitum either a high-fat (AL-HF) or a low-fat diet (AL-LF). Respiratory exchange ratio (RER) was highest for the AL-LF group and the lowest for the AL-HF group. E-TRF-HF led to lower body weight and fat depots, lower glucose, C-peptide, insulin, cholesterol, leptin, TNFα, and ALT levels compared with L-TRF-HF- and AL-HF-fed mice. TRF-HF regardless whether it was early or late led to reduced inflammation and fat accumulation compared with AL-HF-fed mice. E-TRF-HF led to advanced liver circadian rhythms with higher amplitudes and daily expression levels of clock proteins. In addition, TRF-HF led to improved metabolic state in muscle and adipose tissue. In summary, E-TRF-HF leads to increased insulin sensitivity and fat oxidation and decreased body weight, fat profile and inflammation contrary to AL-HF-fed, but comparable to AL-LF-fed mice. These results emphasize the importance of timed feeding compared to ad libitum feeding, specifically to the early hours of the activity period.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-023-04834-4.

Introduction

The central circadian clock is located in the suprachiasmatic nuclei (SCN) of the anterior hypothalamus and generates endogenous rhythms of approximately 24 h [1]. Similar clocks are present in peripheral tissues, such as the liver, muscle, and adipose tissue [2]. The molecular clock consists of the transcription factors CLOCK and BMAL1 that heterodimerize and bind to E-box sequences to mediate transcription of a large number of genes, including Periods (Pers) and Cryptochromes (Crys). The protein products PERs and CRYs constitute part of the negative feedback loop and inhibit CLOCK:BMAL1-mediated transcription [3].

The circadian clock and metabolism are tightly linked. The clock regulates the expression and/or activity of certain metabolic enzymes, hormones, and transport systems [2]. In parallel, key metabolic regulators play a role in the core clock mechanism [4]. For example, reverse ERBα (REV-ERBα), retinoic acid receptor-related orphan receptor α (RORα), and peroxisome proliferator-activated receptor α (PPARα), regulators of lipogenesis and lipid metabolism, regulate Bmal1 transcription [5–7], and in turn, the CLOCK:BMAL1 heterodimer regulates the expression of Rev-erbα, Rorα, and Pparα [2]. In addition, adenosine monophosphate-activated protein kinase (AMPK), a sensor of low intracellular ATP levels, phosphorylates and, as a result, destabilizes the negative limb of the circadian clock, PERs and CRYs [8, 9]. Disrupted circadian rhythms lead to hyperphagia, diabetes, and obesity [10, 11].

Time-restricted feeding (TRF) in animals or time-restricted eating (TRE) in humans limits the time and duration of food availability without calorie reduction [12], i.e., food is available ad libitum for a few hours. Many physiological activities normally dictated by the SCN are altered by TRF [13, 14]. We and others have shown that TRF/TRE protected against diet-induced weight gain [1, 12, 15–19]. TRE also reduced body weight and fasting glucose, improved glucose tolerance, reduced blood pressure, and reduced atherogenic lipids in people with overweight and obesity [20, 21].

Although studies have shown that a high-fat (HF) diet leads to disrupted circadian rhythms [22, 23], timed high-fat diet without reducing caloric intake prevented metabolic problems [16, 24], emphasizing the importance of the timing component in the feeding regimen. In addition, it was recently shown that early TRE of 6 h for 5 weeks improved insulin sensitivity, β cell responsiveness, blood pressure, oxidative stress, and appetite in men with prediabetes compared to 12-h feeding [20]. Thus, it is well established that TRF/TRE is beneficial even if high-fat diet is consumed [19]. It was shown that TRF also stabilized and reversed the progression of metabolic diseases in mice with preexisting obesity and type II diabetes [17]. Although TRF/TRE has been found to be beneficial, comparison of early vs. late TRF/TRE has not been studied. Our aim was to study the effect of early vs. late time-restricted high-fat diet on circadian metabolism and body weight of diet-induced obese mice under conditions of 16 h of activity and 8 h of rest, mimicking the human daily routine.

Methods

Animals, treatments, and tissues

Ninety-two, 8-week-old C57BL/6 male mice were housed in a temperature- and humidity-controlled facility (23–24 °C, 60% humidity). Mice were entrained to a light–dark cycle of 16-h darkness and 8-h light (8:16 LD) for 10 days with food available ad libitum (AL). The 8:16 LD cycle was chosen to mimic the human LD cycle, 16 h of activity and 8 h of rest. Subsequently, mice were randomly assigned to either a control group (n = 20) or a high-fat group (HF) (n = 72), with food available ad libitum for 14 weeks. After 13 weeks, mice were placed on a 16-h fast and blood glucose was monitored to assure impaired glucose metabolism in the HF group. From week 14 onward, the HF group was randomly divided into three groups, an early time-restricted feeding group (E-TRF) (n = 24), a late time-restricted feeding group (L-TRF) (n = 24) or an ad libitum high-fat group (AL-HF) (n = 24) for 5 weeks. The control group received normal low-fat chow (Harlan Laboratories, Jerusalem, Israel) ad libitum (AL-LF) and the other three groups received a high-fat diet. Both TRF groups included an 8-h feeding window during the dark phase (the active hours): the E-TRF received food at the beginning of the dark period (ZT8–ZT16) (ZT0 is the time of lights on) and the L-TRF group at ZT16–ZT24 (Fig. 1A). Daily food intake was monitored three times a day: at ZT0, ZT8, and ZT16 throughout the experiment. Body weight was monitored once a week. At the end of the experiment, mice were anesthetized with isoflurane and blood, liver, epididymal fat, and the gastrocnemius muscle were removed every 4 h (n = 3–4 mice per time point) around the circadian cycle in total darkness (DD) under dim red light to avoid the masking effect of light. Tissues were immediately frozen in liquid nitrogen and stored at − 80 °C. For an accurate assessment of protein and mRNA oscillation or average daily levels, all measurements were either continuous or sampled at six time points throughout the circadian cycle. Mice were humanely killed at the end of the experiment. Animals received human care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication 86-23 revised 1985). The joint ethics committee (IACUC) of the Hebrew University and Hadassah Medical Center approved this study.

Fig. 1.

Effect of TRF-HF on body weight, food intake, and fat depots. A Experimental design. B Body weight. C Percent of food intake during the light or dark phase. D Final body weight. E Total daily caloric intake. F Distribution of the hourly caloric intake. G Respiratory exchange ratio (RER). H Average daily respiratory exchange ratio (RER). I Liver weight. J Fat tissue weight. K Locomotor activity. L Daily average locomotor activity. Mice were fed AL-HF and AL-LF for 19 weeks and E-TRF-HF and L-TRF-HF for 5 weeks after 14 weeks of AL-HF. Data are presented as mean ± SE. *p < 0.05, **p < 0.01, ***p < 0.0001

Assessment of metabolic data

Indirect calorimetry, energy expenditure, and locomotion and activity data of mice were obtained using a Promethion Metabolic Cage System (Sable Systems, Nevada, USA). Mice were habituated for 3 days in metabolic cages prior to recording. Access to water was allowed throughout the study. The control and HF groups received unlimited access to food, while the food hopper of E-TRF and L-TRF cages was programmed to provide access according to the study design (ZT8–ZT16, ZT16–ZT24, respectively). Light conditions were kept the same as in the home cages. Respiratory gases were measured with an integrated fuel cell oxygen analyzer and a spectrophotometric carbon dioxide analyzer. The respiratory exchange ratio (RER) was calculated as the ratio of carbon dioxide production to oxygen consumption. Metabolic data were assessed continuously during the last week of the experiment. The graphs are based on 5-min data records. Locomotor activity was calculated as the sum of all directed ambulatory locomotion of 1 cm/s or above within the x, y, z beam-break system.

Diet composition

Control diet was the usual low-fat (LF) chow diet and contained 53% cornstarch, 20% casein, 10% sucrose, 7% soybean oil, 5% cellulose, 4% mineral mix, 1% vitamin mix, and 0.3% methionine. The HF diet contained 38% cornstarch, 20% casein, 10% sucrose, 7% soybean oil, 15% coconut oil, 5% cellulose, 4% mineral mix, 1% vitamin mix, 0.3% methionine. Overall, the HF diet was based on soybean and coconut oil (42% kcal from fat) compared to 15.9% kcal from fat in the LF diet. The HF diet had 4.7 kcal/g vs. 3.95 kcal/g for the LF diet.

Homeostasis model assessment of insulin resistance (HOMA-IR)

Fasting blood glucose levels were determined using a glucometer (Optium Xceed; Abbott Laboratories, Maidenhead, UK). The insulin-resistance index from fasting serum insulin and plasma glucose levels was determined by the HOMA parameter: HOMA = fasting serum insulin (μU/ml) × fasting plasma glucose (mg/dl)/405.

Serum separation analyses and enzyme-linked immunosorbent assay (ELISA)

Blood was collected from the inferior vena cava at the end of the experiment. Blood was kept at room temperature for 30 min for clotting and consequently centrifuged at 2000×g for 15 min. Serum was collected and stored at − 20 °C. Serum alanine aminotransferase (ALT/SGPT), aspartate aminotransferase (AST/SGOT), HDL, total cholesterol, and triglycerides were measured by American Medical Laboratories (AML, Herzliya, Israel). Serum LDL-cholesterol was calculated using the equation: LDL = total cholesterol − HDL − (triglycerides ÷ 5). Serum adiponectin was measured using ELISA (Merck-Millipore, Billerica, MA, USA) according to the manufacturer’s instructions. Serum hormone levels were determined for insulin, C-peptide, leptin, and TNFα using the Mouse Luminex Discovery Assay, MILLIPLEX^® based on the Luminex^® technology according to the manufacturers’ instructions, and analyzed using the Milliplex Analyst software (Merck-Millipore).

Histological and pathological examination

Liver tissue was embedded in paraffin and serial sections (3–5 μm thick) were cut from each block and stained with hematoxylin and eosin (H&E) for general morphology and by Sirius Red (SR) for fibrosis. Histological slides were prepared by Gavish Research Services (Ness Ziona, Israel), as was described [25]. Histopathological changes were scored by a board-certified toxicologic pathologist, using semi-quantitative grading using 5 grades (0–4), according to the severity of the changes. Evaluation of inflammation and fatty degeneration was performed in the H&E-stained sections. For inflammation, a semi-quantitative analysis was carried out, using a severity-scoring grade of 0–4 (normal to severe) as follows: grade 0, no inflammation; grade 1, very mild inflammation (very mild perivascular infiltrations); grade 2, mild inflammation (a few mild to moderate perivascular infiltrations); grade 3, moderate inflammation (perivascular and multifocal infiltrations); grade 4, severe inflammation (multifocal to diffuse infiltrations). The inflammation score was determined by counting the scope of perivascular infiltration of immune cells. Micro-/macro-vesicles were analyzed using a semi-quantitative severity-scoring grade of 0–4 (normal to severe) as follows: grade 0, no vesicles; grade 1, few vesicles in hepatocytes (< 10% of the specimen); grade 2, a few vesicles in hepatocytes (10–25% of the specimen); grade 3, moderate number of vesicles in hepatocytes (26–75% of the specimen); grade 4, a high number of vesicles in hepatocytes (> 76% of the specimen). The SR-stained sections were examined (× 4 magnification) and scored by a semi-quantitative scoring system for the presence of fibrosis, as follows: grade 0, no fibrosis; grade 1, very mild fibrosis; grade 2, mild fibrosis; grade 3, moderate fibrosis; grade 4, severe fibrosis. The fibrosis score was determined by evaluating the extent (or perimeter) of fibrotic tissue. Pictures were taken using Olympus microscope (BX60, serial no. 7D04032) equipped with a camera (Olympus DP73, serial no. OH05504) at objective magnifications of × 10.

RNA extraction and quantitative real-time PCR

RNA was extracted from liver, fat, and muscle tissues using TRI Reagent (Sigma, Rehovot, Israel). Total RNA was DNase I-treated using RQ1 DNase (Promega, Madison, WI, USA) for 2 h at 37 °C, as was previously described [12]. 2 μg of DNase I-treated RNA were reverse-transcribed using MMuLV reverse-transcriptase and random hexamers (Promega). One-twentieth of the reaction was then subjected to quantitative real-time PCR using SYBR Green Supermix (Quanta Biosciences, Beverly, MA, USA) and primers spanning exon–exon boundaries and the ABI Prism 7300 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Gene expressing was normalized to actin. Reaction conditions were as follows: 3 min at 95 °C, 10 s at 95 °C, 45 s at 60 °C. The fold change in target gene expression was calculated by the 2^−ΔΔCt relative quantification method (Applied Biosystems).

Western blot analysis

Liver, fat, and muscle tissue samples (~ 200 mg) were homogenized in lysis buffer, as was described [26]. Samples were run onto a 10% SDS–polyacrylamide gel and transferred onto nitrocellulose membranes, as was described [26]. Blots were incubated overnight at 4 °C with an antibody against AMP-activated protein kinase (AMPK) and its phosphorylated form (pAMPK) (Cell Signaling Technology, Beverly, MA, USA), fatty acid synthase (FAS) (Cell Signaling Technology), acetyl CoA carboxylase (ACC) and its phosphorylated form (pACC) (Cell Signaling Technology), BMAL1 (Cell Signaling Technology), REV-ERBα (Santa Cruz Biotechnologies, Santa Cruz, CA, USA), NFκB (Santa Cruz Biotechnologies), uncoupling protein 2 (UCP2) (Santa Cruz Biotechnologies), CRY1 (Santa Cruz Biotechnologies), or CLOCK (Santa Cruz Biotechnologies). After three washes with TBST buffer (50 mM Tris–HCl, pH 7.5; 150 mM NaCl; 0.1% Tween-20) at room temperature, blots were incubated with horseradish peroxidase-conjugated secondary antibody (Pierce, Rockford, IL, USA). Anti-mouse antibody (Santa Cruz Biotechnologies) was used to detect actin, the loading control. The immune reaction was detected by enhanced chemiluminescence (Santa Cruz Biotechnologies, Santa Cruz, CA, USA). Finally, bands were quantified by scanning and densitometry and expressed as arbitrary units.

Statistical analyses

All results are expressed as means ± SE. One-way ANOVA (time of day) test was performed to analyze the circadian pattern of clock and metabolic genes and proteins with several time points. Tukey HSD (honestly significant difference) was performed as a single-step multiple comparison procedure and statistical test in conjunction with ANOVA for the evaluation of significant differences among the groups in average daily expression levels of metabolic genes and proteins. A Student’s t test was performed for the evaluation of significant differences between two groups. For all analyses, the significance level was set at p < 0.05. Statistical analysis was performed with JMP software (version 16) (SAS Institute Inc., Cary, NC, USA). Further analysis of circadian rhythmicity was performed using Acro software (version 3.5, Circadian Rhythm Laboratory, University of South Carolina, Walterboro, SC, USA) and Circwave software (version 1.4) (Circadian Rhythm Laboratory, University of Groningen, Groningen, The Netherlands).

Results

To examine the impact of early vs. late time-restricted feeding on circadian metabolism and obesity, mice were fed ad libitum a high-fat diet for 14 weeks, after which they were given the same food during the early (E-TRF-HF) or late (L-TRF-HF) 8 h of the dark phase for 5 weeks (Fig. 1A). The control groups were fed ad libitum either a high-fat (AL-HF) or a low-fat diet (AL-LF) for the whole duration of the experiment (19 weeks) (Fig. 1A).

Timed high-fat diet leads to reduced body weight, food intake, and fat depots

AL-HF and AL-LF gained weight during the first 14 weeks of feeding with a greater body weight in the AL-HF group (Fig. 1B). Food consumption was higher during the dark period and lower in the light period in the AL-HF group compared to the AL-LF group (Fig. 1C). TRF-HF for 5 weeks led to weight loss in both the early and late groups (Fig. 1B), E-TRF-HF lost 7.8% and L-TRF-HF lost 4.9% of body weight. Of all HF groups (AL-HF, E-TRF-HF, L-TRF-HF), the final body weight of only the E-TRF-HF was not significantly different from the AL-LF group (Fig. 1D). Caloric intake was significantly lower in the TRF groups compared to the AL-HF group, but higher compared to the AL-LF group (Fig. 1E, F). Thus, the E-TRF-HF group weighed the same as the AL-LF group although their caloric intake was higher. Respiratory exchange ratio (RER) was highest for the AL-LF group and the lowest for the AL-HF group (Fig. 1G, H). The overall daily RER of the TRF groups was significantly lower than the AL-LF group, but the E-TRF-HF group had significantly higher RER than of the AL-HF (Fig. 1G, H). When comparing the RER of the fasting or eating periods during the active phase, higher RER was observed during fasting in the E-TFF-HF group compared to the L-TRF-HF group (Supplementary Fig. S1A–C). Similar RER was observed during the feeding hours of both TRF groups (Supplementary Fig. S1A–C). Interestingly, during the non-active phase only the E-TRF-HF group displayed lower RER compared to the AL-LF group (Supplementary Fig. S1D). As expected from their lower body weight, fat tissue, but not liver, weight was lower in the TRF groups compared with AL-HF group and the E-TRF-HF group was not significantly different compared with the AL-LF group (Fig. 1I, J). Changes in body weight did not result from a change in activity, as locomotor activity did not differ among all four groups (Fig. 1K, L). Taken together, these results show that timed HF diet leads to lower body weight and fat depots compared with AL-HF-fed mice with even better results if feeding is timed to the early dark phase.

Timed high-fat diet improves serum parameters

After 14 weeks of HF feeding, the AL-HF group had significantly higher glucose levels compared to the AL-LF group (Fig. 2A). After 5 weeks of TRF, both E-TRF-HF and L-TRF-HF groups showed significantly lower fasting glucose levels compared to the AL-HF group. The L-TRF-HF group led to a reduction in fasting glucose to a level that was not significantly different from the AL-LF group (Fig. 2B). Interestingly, the E-TRF-HF showed a greater reduction in fasting glucose and reached a level that was significantly lower compared to the AL-LF group and the L-TRF-HF group (Fig. 2B). Insulin levels in both early and late TRF groups were lower than the AL-HF group and did not differ from the AL-LF group (Fig. 2C). AL-HF mice exhibited high homeostatic model assessment for insulin resistance (HOMA-IR) levels (Fig. 2D), which are indicative of reduced insulin sensitivity. HOMA-IR was lower in both early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group (Fig. 2D). When compared to each other, the E-TRF-HF group had lower HOMA-IR levels compared to the L-TRF-HF group (Fig. 2D). Similar results were achieved when C-peptide, which mirrors insulin production, was measured. Both TRF groups had lower levels compared to the AL-HF group and the E-TRF-HF group had lower levels than the L-TRF-HF (Fig. 2E). Serum cholesterol, HDL-cholesterol and LDL-cholesterol all showed lower levels in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group except for the LDL cholesterol levels which showed lower levels (Fig. 2F–H). Compared to the AL-HF, serum triglycerides were only significantly reduced in the E-TRF-HF group, but did not differ compared to AL-LF (Fig. 2I). Serum adiponectin levels were significantly lower in the AL-HF group compared to the AL-LF group (Fig. 2J). The TRF regimens led to higher levels of adiponectin compared to the AL-HF group, but still lower levels compared to the AL-LF group (Fig. 2J). Serum leptin showed lower levels in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group (Fig. 2K). Serum TNFα levels were lower in both TRF groups compared to AL-LF and AL-HF groups (Fig. 2L). Liver enzyme AST (aspartate aminotransferase) levels did not significantly differ among the groups (Fig. 2M). However, liver enzyme ALT (alanine aminotransferase) levels were lower in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group (Fig. 2N). Taken together, these results show that timed HF diet leads to lower glucose, insulin, cholesterol, leptin, TNFα and ALT levels compared with AL-HF-fed mice with even better results if feeding is timed to the early activity phase.

Fig. 2.

Effect of TRF-HF on serum parameters. A Glucose levels after 14 weeks of HF or LF diet. B Final glucose levels. C Insulin levels. D HOMA-IR. E C peptide levels. F Cholesterol levels. G HDL cholesterol levels. H LDL cholesterol levels. I Triglyceride levels. J Adiponectin to fat mass ratio. K Leptin levels. L TNFα levels. M AST levels. N ALT levels. Mice were fed AL-HF and AL-LF for 19 weeks and E-TRF-HF and L-TRF-HF for 5 weeks after 14 weeks of AL-HF. Serum was collected every 4 h for 24 h. Data are presented as mean ± SE. *p < 0.05, **p < 0.01, ***p < 0.0001

Timed high-fat diet improves liver histology and metabolism

Histological analyses of the liver revealed that liver inflammation and fatty degeneration were higher in the AL-HF group compared to the AL-LF group (Fig. 3A–E), as was expected. Fatty degeneration or ballooning degeneration could be indicative of a progression from simple steatosis to steatohepatitis. Fibrosis severity in the liver was highest in the AL-HF group, as it showed fibrotic tissue, in contrast to significantly lower fibrosis in both TRF groups and an even lower grade of fibrosis in the AL-LF group (Fig. 3A, D). The reduction in the extent of fibrotic tissue was significant in both TRF group compared to the AL-HF group (Fig. 3A, D). Histological analyses of the liver also revealed that liver micro- and macro-vesicles were higher in the AL-HF group compared to the AL-LF group (Fig. 3D, E). When the amount of total fat was analyzed, levels were lower in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group (Fig. 3F). Similarly, when NFκB was measured, a cellular marker of inflammation, levels were lower in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group (Fig. 3G). We next measured the key metabolic factor AMPK, whose activation leads to induction of catabolic pathways and that has been shown to play an important role in the core clock mechanism [10, 27]. Both TRF regimens led to increased ratio of liver pAMPK/AMPK (Fig. 3H), indicating intracellular low energy levels and inhibition of fatty acid synthesis. When fatty acid synthase (FAS) was measured, both AL-HF and L-TRF-HF showed increased levels compared to the AL-LF group, while E-TRF-HF levels did not differ from the AL-LF group (Fig. 3I). Taken together, these results demonstrate that timed HF diet leads to reduced inflammation and fat accumulation compared with AL-HF-fed mice.

Fig. 3.

Effect of TRF-HF on liver histology and metabolism. A Histopathological photographs of H&E stain (a) or SR (b) staining. 1a-b AL-LF, 2a-b AL-HF, 3a-b E-TRF-HF, 4a-b L-TRF-HF. B Inflammation score. C Micro-vesicles score. D Fibrosis score. E Macro-vesicles score. F Fat weight. G NFκB daily average protein levels. H pAMPK/AMPK daily average protein level ratio. I FAS daily average protein levels. Mice were fed AL-HF and AL-LF for 19 weeks and E-TRF-HF and L-TRF-HF for 5 weeks after 14 weeks of AL-HF. Liver samples were collected every 4 h for 24 h. Proteins were analyzed by Western blotting. Yellow and black arrows in the H&E staining designate micro- and macro-vesicles, respectively. Black arrows in the SR staining designate fibrosis. Data are presented as mean ± SE. *p < 0.05, **p < 0.01, ***p < 0.0001. Pictures were taken at objective magnifications of × 10

Timed high-fat diet advances clock gene expression

As has been previously shown [22–24], the AL-HF group had altered circadian rhythms compared to the AL-LF group (Supplementary Fig. S2). We next measured the effect of early and late TRF-HF diet on clock gene expression in peripheral tissues. Our analyses revealed that liver clock genes (Bmal1, Clock, Per1, Cry1, Rorα and Rev-erbα) showed more advanced and higher amplitude rhythms in the E-TRF-HF compared to the L-TRF-HF group (Fig. 4A). Similar results were achieved in fat tissue (Fig. 4B). In the muscle, clock genes were more advanced, but not all clock genes showed higher amplitude rhythms in the E-TRF-HF compared to the L-TRF-HF group (Fig. 4C). Daily average protein levels of liver BMAL1, CRY1 and REV-ERBα in the E-TRF-HF group were higher compared to the AL-HF group (Fig. 5A, Supplementary Fig. S3A). In fat tissue, daily average protein levels of BMAL1 and CRY1 were higher, whereas those of and REV-ERBα were lower in the E-TRF-HF group compared to the AL-HF group (Fig. 5B, Supplementary Fig. S3B). In muscle tissue, daily average protein levels of BMAL1, CLOCK and CRY1 were higher in the E-TRF-HF group compared to the AL-HF group (Fig. 5C). Daily average protein levels of liver and fat tissue CLOCK, CRY1 and REV-ERBα were higher in the E-TRF-HF group compared to the L-TRF-HF group (Fig. 5A, B, Supplementary Fig. S3A, B). Daily average protein levels of muscle CLOCK and CRY1 were higher, whereas those of REV-ERBα were lower in the E-TRF-HF group compared to the L-TRF-HF group (Fig. 5C, Supplementary Fig. S3C). Taken together, these results show that early TRF leads to advanced rhythms with higher amplitudes and daily protein expression levels.

Fig. 4.

Effect of TRF-HF on clock gene mRNA expression in peripheral tissues. A Liver tissue. B Adipose tissue. C Muscle tissue. Mice were fed AL-HF and AL-LF for 19 weeks and E-TRF-HF and L-TRF-HF for 5 weeks after 14 weeks of AL-HF. Tissues were collected every 4 h for 24 h. mRNA was determined by real-time PCR

Fig. 5.

Effect of TRF-HF on clock gene daily average protein levels. A Liver tissue. B Adipose tissue. C Muscle tissue. Mice were fed AL-HF and AL-LF for 19 weeks and E-TRF-HF and L-TRF-HF for 5 weeks after 14 weeks of AL-HF. Liver samples were collected every 4 h for 24 h. Proteins were analyzed by Western blotting. Data are presented as mean ± SE. *p < 0.05, **p < 0.01, ***p < 0.0001

Timed high-fat diet improves fat and muscle metabolism

We next measured the effect of early and late TRF-HF diet on metabolic enzymes in fat and muscle tissue. The E-TRF-HF group had a higher pAMPK/AMPK ratio compared to the AL-HF group (Fig. 6A), suggesting increased activation of the catabolic AMPK pathway. In parallel, the ratio of the pACC/ACC in the E-TRF-HF group was also higher compared to the AL-HF group and compared to the L-TRF-HF group (Fig. 6B) suggesting decreased fatty acid synthesis, although the levels of FAS were not significantly lower (Fig. 6C). Another indication for increased fatty acid oxidation was also seen with the higher levels of UCP2 [28] in the E-TRF-HF group (Fig. 6D). Pparγ transcript levels were higher in both TRF groups (Fig. 6E) suggesting storage of fatty acids, insulin sensitivity, and reduced inflammation. Indeed, NFκB levels were lower in the E-TRF-HF group compared to the AL-HF group and to the AL-LF group (Fig. 6F). Similar results were achieved in muscle tissue, i.e., a higher ratio of pAMPK/AMPK and pACC/ACC with increased UCP2 protein expression in the E-TRF-HF group compared with the AL-HF group (Fig. 6G–I, Supplementary Fig. S3C). The L-TRF-HF showed higher ratio of pACC/ACC compared with the AL-HF group (Fig. 6G–I). Taken together, these results indicate that although both early and late TRF improve the metabolic state, a better improvement is achieved with the early feeding.

Fig. 6.

Effect of TRF-HF on metabolism in fat and muscle tissue. A Adipose tissue pAMPK/AMPK daily average protein levels. B Adipose tissue pACC/ACC daily average protein levels. C Adipose tissue FAS daily average protein levels. D Adipose tissue UCP2 daily average protein levels. E Adipose tissue Pparα average mRNA levels. F Adipose tissue NFκB daily average protein levels. G Muscle pAMPK/AMPK average protein levels. H Muscle pACC/ACC average protein levels. I Muscle UCP2 average protein levels. Mice were fed AL-HF and AL-LF for 19 weeks and E-TRF-HF and L-TRF-HF for 5 weeks after 14 weeks of AL-HF. Fat and muscle tissues were collected every 4 h for 24 h. Proteins were analyzed by Western blotting. mRNA was determined by real-time PCR. Data are presented as mean ± SE. *p < 0.05, **p < 0.01, ***p < 0.0001

Discussion

Most TRF studies initiate TRF early at the onset of the activity phase [16–18, 29]. This is the optimal time for TRF as studies have shown the importance of early food consumption for glucose tolerance and insulin sensitivity in humans [30–32]. This study was conducted under 8:16 lighting conditions that mimic human daily routine. These conditions allowed the division of food consumption to the early or late 8 h of the activity cycle without delaying food consumption to the light phase. In addition, this study was the first to measure the effect of different timed HF regimens (early and late) in a metabolic system that measures gas exchange and calculates RER.

Our results show that timed HF diet leads to a unique metabolic phenotype of higher caloric intake, but similar total activity and body weight to that of the AL-LF group. TRF-HF for 5 weeks led to weight loss in both the early and late groups, but the E-TRF-HF group lost more than the L-TRF-HF and the final body weight was not significantly different from the AL-LF group. Thus, the timed HF diet leads to lower body weight and fat depots compared with AL-HF-fed mice with even better results if feeding is timed to the early activity phase. The available evidence suggests that > 60% energy restriction on 2–3 days per week or on alternate days and TRF, limiting the daily period of food intake to 8–10 h or less, produce equivalent weight loss when compared to continuous energy restriction [33, 34]. Our results support these findings, but with a greater benefit for the early hours of the activity period.

Metabolic switching is the ability to shift substrate oxidation from glucose to fatty acids throughout feeding and fasting periods within the course of a day. RER can be used as an indicator to assess metabolic switching during the 24-h day cycle. Studies showed that during the inactive phase of TRF-HF regimens, the RER decreases during the fasting period and increases above 0.8 during feeding [14]. In our study, only the E-TRF-HF group showed significantly lower RER during the non-active phase compared to the active phase, which could indicate metabolic switching, i.e., burning internal fat stores, as is also corroborated by the increased UCP2 levels in fat tissue and the decrease in fat tissue in the E-TRF-HF group. Interestingly, the RER during the active vs. non-active phase in the two TRF groups was different, suggesting that the prolonged fasting, specifically during the late part of the active phase, as in the E-TRF-HF, could lead to a more efficient metabolic switching.

Five weeks of time-restricted-HF diet were sufficient to reduce fasting glucose levels compared to the AL-HF group. These levels were not significantly different in the L-TRF-HF group compared to the AL-LF group, but they were significantly lower in the E-TRF-HF. These results were mirrored by lower levels of C-peptide, insulin and insulin resistance index in both TRF groups and specifically in the E-TRF-HF group. In a similar study, in which TRF was given early or late during the activity phase, TRF reduced weight and fat mass vs. AL, with a greater reduction in E-TRF vs. L-TRF [35]. However, in that study [35], TRF improved glucose tolerance and protected mice from high-fat diet-induced hepatosteatosis vs. AL, but with no difference between E-TRF and L-TRF. The differences between our results and those of Regmi et al. could stem from the diet given during TRF, HF in our study vs. LF in Regmi et al. In addition, the late TRF in the Regmi et al. study, slid into the light phase, which could attenuate the metabolic benefits. Our results are supported by previous findings in which improved insulin sensitivity and β cell responsiveness was found in men with prediabetes that had E-TRF (6-h feeding, with dinner before 3 p.m.) compared to a control group (12-h feeding) for 5 weeks [20]. Increase in CRY1 protein levels in liver, muscle and fat tissue could contribute to the improved glucose homeostasis, as it was shown that CRY1 inhibits gluconeogenesis and improves glucose sensitivity [36–38].

Improved metabolism was also seen in serum cholesterol, HDL-cholesterol and LDL-cholesterol, which showed lower levels in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group. Serum adiponectin levels, which are lower in obesity [39, 40] were indeed significantly lower in the AL-HF group compared to the AL-LF group, but TRF led to higher levels of adiponectin as has been reported in weight loss studies [41]. In addition, the levels of leptin, which functions as a satiety signal whose levels are correlative with fat tissue mass [42], were lower in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group.

Serum TNFα levels did not differ between the AL-LF and AL-HF groups, but TRF led to reduced levels as was previously reported [12, 24]. Similarly, ALT levels were lower in both TRF groups compared to AL-LF and AL-HF groups. As these serve as markers of inflammation, our findings indicate that timing can attenuate the detrimental inflammatory impact of HF diet. Indeed, histological analyses of the liver revealed that liver inflammation, micro-vesicle, macro-vesicle, and fibrosis scores were higher in the AL-HF group compared to the AL-LF group. These scores showed reduced levels in the TRF groups compared to the AL-HF group. Similarly, the levels of NFκB, a cellular marker of inflammation [43], were lower in both the early and late TRF groups compared to the AL-HF group and did not differ from the AL-LF group. Thus, timed HF diet leads to improved metabolism and reduced inflammation compared with the AL-HF-fed mice with even better results if feeding is timed to the early dark phase. These levels are similar to and sometimes even better than those of AL-LF-fed mice.

To test the effect on overall metabolism at the cellular level, we measured expression and activity of key metabolic regulators, i.e., AMPK, ACC, and UCP2. Under TRF, mice were devoid of food for 16 h leading to increased AMP levels and subsequent AMPK activation, which leads to induction of catabolic pathways and that has been shown to play an important role in the core clock mechanism [10, 27, 44]. Both TRF regimens led to increased ratio of pAMPK/AMPK in liver, indicating intracellular low energy levels and inhibition of fatty acid synthesis. However, in muscle and fat tissue, only the E-TRF-HF group had a higher pAMPK/AMPK as well as a higher pACC/ACC ratio compared to the AL-HF group, suggesting increased activation of the catabolic AMPK pathway and decreased fatty acid synthesis, although the levels of FAS were not significantly lower. Another indication for increased fatty acid oxidation was also seen with the higher levels of muscle and fat tissue UCP2. Fat tissue Pparγ transcript levels were higher in both TRF groups suggesting storage of fatty acids, insulin sensitivity and reduced inflammation [45]. Thus, at the cellular levels, timed HF diet led to the activation of catabolic pathways. These findings support the reduced body weight of TRF-fed mice.

The circadian clock has been shown to be important for regulated metabolism. Mutations or knockouts of clock genes has led to metabolic disturbances [8, 11, 46–48]. Moreover, we and others have shown that high-fat diet leads to changes in behavioral rhythmicity and correlate with disrupted circadian gene expression within hypothalamus, liver, and adipose tissue [22, 23, 49–52]. In addition, it has recently been shown that TRF profoundly impacts gene expression, i.e., nearly 80% of all genes show differential expression or rhythmicity under TRF [53]. Liver and fat tissue clock genes showed more advanced and higher amplitude rhythms in the E-TRF-HF compared to the L-TRF-HF group. In the muscle, clock genes were more advanced, but not all clock genes showed higher amplitude rhythms in the E-TRF-HF compared to the L-TRF-HF group. Indeed, it has been established in the literature that timed feeding is dominant and can lead to changes in the phase of gene expression [13, 54, 55]. We surmise that the early hours of the activity period, which come immediately after the rest period, are crucial for the metabolic switch. During the early hours of the activity phase, feeding serves as the signal. As the key metabolic factors, such as AMPK, SIRT1, PGC1α, are all regulators of the clock mechanism [2], the initiation of feeding at the change from the rest phase to the activity phase is crucial to synchronize the clock and overall metabolism. This happens late if feeding is delayed to the last 8 h of the activity period, leading to less optimal metabolism compared to E-TRF. However, having the timing component in the L-TRF is sufficient to improve metabolism compared to ad libitum feeding, as has been shown in many TRF studies [1, 16, 18, 20, 24].

A limitation of this study is the rather short (5 weeks) of TRF. At the end of the experiment, body weight of the E-TRF-HF was not significantly different from the AL-LF group. It is possible that a longer duration of E-TRF, even though HF, would lead to a lower body weight compared to the AL-LF group. In addition, morphology of muscle and fat tissue should be examined under early and late TRF, to determine changes in these important metabolic tissues.

Conclusion

Our findings show that timed HF diet, preferably limited to the early hours of the activity phase, leads to increased insulin sensitivity and fat oxidation and decreased body weight, fat profile and inflammation contrary to AL-HF-fed, but comparable to AL-LF-fed mice. It was reported that mice fed a HF diet during the light phase gain significantly more weight than mice fed only during the dark period [56]. Our results emphasize the importance of timed feeding compared to ad libitum feeding [12] and specifically to the early hours of the activity period. The 8:16 lighting conditions, which is more relevant to the human modern activity pattern, may allow to better planning of timed meals that would lead to improved metabolism.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ACC

Acetyl CoA carboxylase

AL

Ad libitum

AMPK

AMP-activated protein kinase

DD

Constant dark

FAS

Fatty acid synthase

HDL

High-density lipoprotein

HF

High fat

HOMA-IR

Homeostasis model assessment of insulin resistance

LD

Light–dark

LDL

Low-density lipoprotein

LF

Low-fat

PPARα

Peroxisome proliferator-activated receptor α

TG

Triglycerides

TNFα

Tumor necrosis factor α

TRF

Time-restricted feeding

ZT

Zeitgeber Time

Author contributions

Conceptualization, OF and ST; methodology, ST and NC; validation, OF, ST, and NC; formal analysis, ST; investigation, ST and NC; resources, OF; data curation, ST and NC; writing—original draft preparation, ST; writing—review and editing, OF; supervision, OF; funding acquisition, OF. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Israel Science Foundation (grant no. 1675/19). Shani Tsameret is a fellow of the Ariane de Rothschild Women Doctoral Program.

Data availability

The data presented in this study are available on request from the corresponding author.

Declarations

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Ethics approval

The animal study was reviewed and approved by the joint ethics committee (IACUC) of the Hebrew University and Hadassah Medical Center.

Consent to participate

Not applicable.

Consent to publish

Not applicable.