Time-Restricted Feeding and Caloric Restriction: Two Feeding Regimens at the Crossroad of Metabolic and Circadian Regulation

Abstract

In addition to diet quality and quantity, the “timing” of food intake recently emerged as a third key parameter in nutritional and metabolic health. The link between nutrition timing and metabolic homeostasis is in part due to the regulation of daily feeding:fasting cycles and metabolic pathways by the circadian clock. Preclinical feeding regimen studies in rodents are invaluable to further define the modalities of this relationship and get a better understanding of its mechanistic underpinnings. Time-restricted feeding (TRF) and caloric restriction (CR) are examples of feeding regimen at the crossroads of metabolic and circadian regulation. Here we propose methods to implement TRF and CR highlighting the parameters that are relevant to the study of circadian and metabolic health. We also provide methods to determine their impact on the output of the circadian clock by analyzing diurnal expression profiles using 24 h time-series collection as well as their impact on metabolic homeostasis using a glucose tolerance test (GTT).

1. Introduction

A feeding regimen is a plan that specifies a diet type, amount, and schedule of nutritional intake. Beyond their role in cardiometabolic health, the quality, quantity, and timing of dietary intake can also affect the activity of the circadian clock and circadian rhythms. In this chapter, we highlight this relationship and propose methods to investigate it. The first section of the introduction will provide a short overview of the interaction between nutrition, metabolic health, and the circadian clock. In the second section of the introduction, we will describe experimental parameters that influence feeding regimen studies while highlighting the ones that can specifically affect the circadian clock. Then we will discuss different approaches to implementing feeding regimen studies. Subheading 3 describes possible protocols to implement manually two feeding regimens that lie at the interface between metabolic homeostasis and the circadian clock, namely time-restricted feeding (TRF) and calorie restriction (CR). We also describe methods to evaluate the activity of the circadian clock and metabolic health under these feeding interventions. In particular, we suggest a protocol for performing a glucose tolerance test that takes into account the challenges associated with the comparisons between groups under different feeding regimen.

1.1. Nutrition, Metabolic Health, and the Circadian Clock

Nutrition plays a crucial role in overall health and well-being. Good dietary habits are associated with lower risks of many chronic diseases that affect a large fraction of the World’s population [1, 2]. Reasonably, dietary guidelines emphasize the importance of the quality and composition of the food we consume as well as the quantity needed to maintain one’s caloric balance (https://www.dietaryguidelines.gov [3]). The “timing” of food intake recently emerged as a third key parameter [4–7]. The temporal aspect of nutrition encompasses both the duration of the daily eating window (number of hours between first and last caloric intake of the day) and the regularity of one’s eating habits. The link between nutrition timing and metabolic health is in part due to the regulation of daily feeding:fasting cycles and metabolic pathways by the circadian clock [8, 9].

There is a bidirectional relationship between the circadian clock and metabolic health. On the one hand, perturbations to the circadian clock function are associated with increased risk of developing metabolic disease. Indeed, epidemiological studies have shown that populations with disrupted circadian rhythms such as shift workers have increased risk of metabolic disease [10]. In addition, genetic disruption of the clock in a variety of animal models leads to metabolic dysfunction [11]. On the other hand, metabolic disease can be associated with alterations in the clock function. Landmarks studies have shown that one of the most widely used mouse preclinical model of obesity and metabolic disease, the diet-induced obesity (DIO) model—display dampened daily activity rhythms and altered daily feeding rhythms [12]. Further studies using timed feeding and time-restricted feeding (TRF) have then demonstrated that mistimed food consumption is associated with increased weight gain and metabolic disorders [13–16].

Rodent preclinical models of obesity, diabetes, and cardiometabolic disease have been vastly used in nutrition and metabolism research to understand the mechanistic underpinnings of the diseases and explore therapeutics approaches. Now they are also used to study the interaction between food timing, circadian clock, and metabolic health. In this chapter, we focus on time-restricted feeding (TRF) and caloric restriction (CR), two feeding regimens that are at the crossroad of metabolic and circadian regulation. TRF is a feeding regimen wherein food intake is limited to a consistent 8–10 h daily window without changes in nutritional quality or quantity. TRF of a high fat diet has been shown to protect and reverse obesity and associated metabolic disorders [13–15, 17–21]. From the clock standpoint, TRF prevents dampening of the circadian clock associated with the consumption of an energy-dense diet and restores high amplitude diurnal rhythms in clock gene expression and physiological output of the circadian clock [13, 15, 22]. CR is the most established and well-characterized intervention to slow down aging and prevent chronic disease. For example, a 15–40% daily caloric reduction can delay age-related diseases and increase life span in worms, flies, rodents, and non-human primates [23]. Notably, CR in mice can be seen as an extreme form of TRF since mice usually eat all their food within 3 h. This temporal restriction on eating could be in part responsible for the health benefits of CR. CR can also affect the activity of both the central and peripheral clocks [24–26]. Whether the clock is necessary for the benefits of CR is still under investigation but at least in C57Bl/6 mice, BMAL1, a master component of the clock, has been shown to be required for the benefits of CR [27]. Thus, CR and TRF are two feeding regimens that have strong metabolic and health benefits but also impinge on the activity of the circadian clock.

1.2. Feeding Regimen Studies

As for any research using mouse models, studies that involve modifications in the feeding regimen should mention the strain, genotype, age, and sex of the animals being used. In addition, when conducting feeding regimen intervention studies, other parameters that need to be specified are: (1) composition and energy content of the diet, (2) duration of the feeding intervention, (3) age of the mice upon initiation, (4) housing information such as animal density and bedding materials, and (5) timing of food access relative to the light:dark (L:D) cycle when applicable. The latest point is key when studying feeding-clock interaction since the timing of food intake can affect the activity of the circadian clock. This will be discussed in further details with the example of TRF and CR in Subheading 3.

There are a variety of methods to implement feeding regimen studies and monitor feeding behaviors in rodents that can broadly be divided into automatic versus manual approaches. Automatic systems to control food availability, monitor food consumption or both are available from commercials vendors (Columbus Instruments, TSE Systems, Sable Systems, Research Diets BioDAQ, etc.) and have also been developed in research laboratories lab [26, 28]. Commercial systems are usually integrated with other technologies such as activity or calorimetry analysis into full “metabolic phenotyping units/cages” which allow for high frequency and simultaneous measurements of food consumption, activity, and energetics patterns (CLAMS, PhenoMaster, Promethion).

These systems are relatively expensive and often require single housing which make them not easily amendable to large-scale studies. Some laboratories have developed custom-made food dispensers that are much cheaper and allow the delivery of a specific amount of food at a given time and frequency (without recording the actual amount of food eaten). On the other hand, manual approaches can also be used to perform feeding regimen studies. They are labor intensive when aiming to monitor body weight and food consumption and might be less precise that automated approaches but they are much easier to adopt with large cohort studies.

In this chapter we provide possible methods to implement TRF and CR manually in mice as well as to evaluate the activity of the circadian clock and metabolic health under these feeding interventions. There are many ways of implementing feeding regimen studies and the optimal design and approach depends on the research question and the feasibility at one’s research institution. Our goal is to highlight some of the parameters that are especially relevant to TRF and CR in the context of investigating feeding-clock interactions. Since these parameters might influence the study outcomes, they should ideally be controlled as much as possible in the approach and precisely described in Subheading 3. We feel that providing a detailed description of the protocol implemented goes a long way in allowing the reproduction of research findings between research institutions. Ultimately, we hope that this description will help researchers in the rigorous design and performance of these types of studies as well as their reproducibility between different research teams, for the purpose of scientific advancement.

2. Materials

2.1. Manual Implementation of TRF and CR

When implementing TRF or CR during the dark phase, having the mice on a reverse L:D cycle can dramatically increase the feasibility of a long-term intervention while minimizing the disruption to the experimenters own circadian rhythms among others. This is achievable by housing the animal cohort either in a reverse L:D cycle housing room or in a light-tight cabinet in which the lighting schedule can be controlled automatically (“circadian cabinets”). Other materials needed to implement TRF and CR manually are:

1. Diet.

2. Scale to record body weight and food weight.

3. Diet containers.

2.2. Twenty-four Hour Time-Series Collection (See Note 1)

The materials needed for a 24 h time-series collection are:

1. Dissection tools and board.

2. Liquid nitrogen for fast freezing of the collected organs.

3. Pre-labeled collection tubes.

4. Light-proof cage/animal container to prevent light exposure during the dark phase if the mice are being relocated for collection.

2.3. ipGTT

The materials needed to perform an ipGTT are:

1. Clean cages for fasting—paper bedding recommended.

2. Scale to record body weight.

3. Sharp razor blade to make a small cut on the tail.

4. Digital glucose reader and strips to measure blood glucose.

5. d-Glucose to prepare glucose solution to inject.

6. Insulin syringes to inject glucose bolus.

7. Timer to monitor blood sampling time.

3. Methods

3.1. Manual Implementation of TRF Studies in Rodents (See Note 2)

1. Set the duration of the daily feeding window.

   Classically, mice on TRF have ad libitum access to food for a period of 8–12 h daily 7 days a week. The amount of food provided every day should be ample enough that the animals don’t run out of food during the feeding interval.

2. Set the timing of the feeding window relative to the light:dark cycle in the animal housing room.

   Per convention, Zeitgeber Time 0 (ZT0) denotes the time at which lights turn on in the housing room (see Note 3). For TRF during the light phase, food will be present in the cage for a chosen number of hours (step 1) between ZT0 and ZT12. For TRF during the dark phase, food will be present in the cage between ZT12 and ZT24. Classically, for 9 h TRF during the dark phase, food is provided between ZT13 and ZT21 [13, 17].

3. Prepare the food.

   If monitoring food consumption is part of the experimental design, an accurate measure of weekly intake can be obtained by weighing the amount of food placed in the cage and the amount of food left at the end of a week. The amount of food required for 1 week can be evaluated from ad libitum cages on the same diet and should be adapted to the housing density.

4. Implement TRF daily.

   There are different methods to implement TRF. Two of them are presented below:

   Method 1:

   1. Add food in the hopper at the set time-of-day (see step 2).
   2. Take out food at the end of the chosen feeding duration (see step 1).
   3. When monitoring food consumption, use food from a container dedicated to each cage to take food in and out from the cage. Prepare the food container for each cage that is being monitored with at least a week worth of food accurately weighed and recorded. Weigh the food left in every container after 1 week. Weekly food consumption can be measured in grams by subtracting the amount of food left to the amount of food put in. Daily food consumption in kcal per mouse can be calculated based on the diet caloric density, the number of animal in the cage, and the actual number of days of food monitoring.
   Method 2 (see Notes 4 and 5):

   1. Place mice in a cage that contains food at the set time-of-day (see step 2), the “feeding cage.”
   2. At the end of the chosen feeding duration (see step 1), transfer mice in a cage without food, the “fasting cage.”
   3. When monitoring food consumption, place at least a week worth of food accurately weighed and recorded in the “feeding cage” and weigh the food left in the cage after 1 week.
   4. This method involves twice-a-day animal handling that should be reflected in the controls.

3.2. Manual Implementation of Caloric Restriction (CR) in Rodents (See Note 2)

1. Set the percentage of caloric restriction.

   Classically, mice on CR are fed every day with 60–70% of the total daily caloric intake of mice fed ad libitum. The amount of food to be provided every day can be evaluated prior to the start of the intervention from monitoring food consumption in similar mice fed ad libitum with the same diet. The exact quantity of food to be provided should be adapted to the housing density. When switching an ad libitum fed cohort to CR, adopting a weekly 10% reduction approach over 2 or 3 weeks can be helpful in preventing an abrupt drop in body weight. If an ad libitum control cohort is run in parallel, weekly monitoring of food consumption in this group can allow to fine tune the amount of food to give to the CR group—with a 1 week delay. A decrease in the caloric intake is expected as the mice age.

2. Set the timing and frequency of feeding.

   Mice on CR can be fed the entire amount of food at once daily—this is the most current approach—or by portions across the day. The later usually requires the use of automated food dispensers. These two approaches are associated with very distinct feeding patterns but also activity profiles reflecting that they act differently on the activity of the central clock. When feeding the mice at once, the time-of-day of food addition is also critical, since food anticipatory activity (FAA) is different when the food bolus is provided during the light or the dark phase (see Note 6 for more details). In any case, most mice on CR eat the entire amount of food provided within 2–3 h, leading to an extended period of fasting [26].

3. Prepare the food.

   If monitoring food consumption is part of the experiment, an accurate measure of daily intake can be obtained by weighing the amount of food placed in the cage and the amount of food left at the end of a day. In most cases, it should be 0.

4. Implement CR daily with a one-time daily addition of food

   1. Add set amount of food (see step 1) in the hopper at the set time-of-day (see step 2).
   2. Take out food left—if any—after 24 h (see Note 7).
   3. When monitoring food consumption, use food from a container dedicated to each cage to take food in and out from the cage. Record the amount of food left daily—if any—to compute weekly food consumption for each cage. Daily food consumption in kcal per mouse can be calculated further based on the diet caloric density, the number of animals in the cage, and the actual number of days of food monitoring.

3.3. Evaluation of the Activity of the Molecular Circadian Clock in Rodents from 24 h Time-Series Collection (See Note 8)

To assess the activity of the clock at the molecular levels, samples are collected at specific times across a day—a 24 h time-series collection. These samples can then be analyzed by targeted or untargeted omics approaches (see Chapter 21). Guidelines for the experimental design to obtain and analyze biological rhythms in genome-scale datasets are available in the landmark consortium paper from Michael E. Hugues [29]. In particular, this manuscript highlights the complexity of choosing “the right combination of replicates and temporal resolution for their intended application.” This manuscript is a must read when deciding on the parameters described below.

1. Set the sampling frequency.

   Most circadian time-series studies in rodents are performed with a sampling interval of 3–4 h (i.e., 6 or 8 time points per day respectively). Biological replicates at each time point are also recommended. For example, for a discovery study comparing the effect of ad libitum to TRF feeding on the diurnal transcriptome in various clock mutants, we chose to use a sampling interval of 3 or 4 h with biological duplicates at each time point [17]. That is 12–16 animals per feeding group per genotype. Thus the number of animals required for a circadian collection is significant, usually exceeding the number required for a steady state analysis.

2. Choose the approach to cover 24 h.

   There are two approaches for covering a 24-h cycle: (1) all in 1 day, (2) in multiple shifts. The numbers of mice and organs to be collected as well as available personnel are key parameters to inform this choice. Regardless of the approach, 24 h time-series collections are strenuous and require careful planning.

3. Collect.

   Perform traditional collection and flash freezing of the samples collected at each time point. Limit mice exposure to light prior to collection during the dark phase time points.

3.4. Performing an Intra-peritoneal Glucose Tolerance Test (ipGTT) in Mice Under Different Feeding Regimen

Most feeding regimen studies include the measurements of readouts of metabolic health such as serum glycemia, insulinemia, and lipidemia as well as kinetics of glucose processing and insulin action using glucose or insulin tolerance tests, respectively. The results of these assays are highly dependent on the fasting state of the mice. Recommendations exist to guide the investigators when performing these assays in order to standardize them and increase reproducibility [30]. However, especially when comparing different groups of mice undergoing chronic fasting intervention, the standard procedure might not be applicable (see Note 9).

For GTT assays, 14–18 h long “overnight fast” is often used. In most cases the overnight fast occurs during the dark phase, which is the phase at which mice consume the most of their food, thus resulting in the depletion of energy stores, especially hepatic glycogen. In the case of TRF, mice undergo 15–16 h long fast daily, of which 12 h are happening during the light phase. For mice under CR, most animals undergo >20 h of fasting daily. Here we propose an ipGTT protocol to compare mice under ALF and TRF groups.

1. Fast the mice.

   We recommend fasting the ALF group at the same time that the TRF group is fasted daily so both groups experience the same fasting duration prior to the assay.

   1. Transfer the mice to a clean cage with water and without food.
      A clean cage is preferred to the home cage without food to prevent coprophagy. Non-corncob bedding (that could be eaten by the mice during fasting) is also preferred.
   2. Properly identify the fasting cages according to local institutional animal care and use committee (IACUC) requirements.

2. Set the time-of-day of the experiment.

   Since insulin sensitivity varies according to the time-of-day, it’s highly recommended to perform all metabolic studies at the same ZT time especially if longitudinal measurements are presented (see Note 3). The time chosen should be indicated in the methods section. For GTT we recommend setting the time of the glucose bolus at a time where the TRF group is usually fed.

3. Perform GTT at the set time-of day.

   1. Weigh the mice and make them identifiable (e.g., tail tag with a marker).
   2. Measure basal glycemia with a strip glucometer from tail blood obtained after a tail tip cut with a sharp sterile blade.
   3. Inject glucose at a dose of 1–2 g/kg body weight.
   4. At 15, 30, 60, and 120 min following the IP injection, blood glucose levels are measured. Between sampling times mice will be returned to their cage and monitored for any changes in behavior.