The effect of food timing on torpor propensity and characteristics in laboratory mice during a common food restriction paradigm

Introduction

Food restriction and fasting are common techniques used in behavioural, metabolic, and circadian studies, amongst others. However, many of these studies do not account for the induction of torpor in response to limited food availability in mice. Torpor is a hypometabolic and hypothermic state employed by certain species to conserve energy, but it also profoundly alters central (Huang et al., 2021; von der Ohe et al., 2006) and peripheral (Melvin & Andrews, 2009) physiology. Laboratory mice readily enter torpor in response to fasting (Hudson & Scott, 1979; Swoap et al., 2006) but the long-term effects of chronic torpor induction on subsequent behaviour and physiology are largely unknown.

Food timing is rarely reported in studies using food restriction, and feeding schedule may vary depending on when a task is performed, or sample taken. However, torpor propensity has been shown to be under circadian control and is strongly influenced by food timing (Oelkrug et al., 2011; van der Vinne et al., 2018). In an accompanying publication (Wilcox et al., 2024d), we reported that torpor was readily induced during a chronic food restriction paradigm and that the likelihood of torpor occurrence was closely related to decreased bodyweight. Subsequently, we hypothesised that differences in food timing within and between studies may result in variation in torpor characteristics and contribute to variations in physiology and behaviour due to differences in prior torpor history. Here, we investigated the effect of food timing on torpor characteristics. All mice studied (n=8) readily entered torpor in response to food restriction, regardless of feeding time, and feeding time significantly affected the latency to torpor, timing, and duration of torpor bouts.

Methods

Animals and recording conditions

C57BL/6J mice were used in this study (University of Oxford Biomedical Services in-house colony; N=8, aged ~11 weeks). Male mice were chosen for this study to ensure data were representative of the current research landscape, which primarily uses males. However, we expect that these data will be applicable to females, as torpor has been shown to be readily induced in female mice (Oelkrug et al., 2011; Zhang et al., 2020). The C57BL/6J strain was chosen as they are commonly used for behavioural studies, or as a background for genetic mutants.

Mice were individually housed in wire-top cages (48x15x13cm) on a 12:12h light-dark (LD) cycle for the duration of the experiment. Cages were housed in custom light-tight chambers (LTCs; Figure 1B), with four cages per chamber (Fisher et al., 2012; Tam et al., 2021). LTCs were illuminated by cool white LED strips (Maplin, UK; 200 lux at the cage floor). Ambient room temperature and relative humidity were maintained at 21±1°C and 60±1%, respectively. Water was provided ad libitum throughout. Bedding material (sizzle nest) was kept to a minimum, and enrichment items, such as tunnels, were removed to prevent mice from being obscured from view of the thermal cameras, although ~10g of nesting material was provided to align with the thermal preference of mice (Gaskill et al., 2011).

Figure 1. Experimental design and raw temperature traces.

(A) Mice were assigned to either the morning-fed group, or the night-fed group (n=4). Night-fed animals were kept on a reverse light-dark cycle (lights on at 7 pm), to allow for feeding to be completed in parallel. During food restriction, all mice were fed once daily; morning-fed mice were fed one hour after lights on, whilst night-fed mice were fed one hour after lights off. Mice were weighed ~2 hours after food was provided; additional food was provided as necessary at this time to ensure bodyweight was maintained at ~85% of free feeding bodyweight. (B) Mice were individually housed in standard wire-top cages, placed under continuous thermal imaging cameras (two cages per camera), in light-tight chambers. Cameras were placed ~20 cm above the cage floor and positioned to ensure mice were in view. (C) Thermal images of a mouse during euthermia (left), and a mouse during torpor (right). Skin temperature was measured by recording the warmest pixel in view every 1 s (1 Hz). Ta represents measurement of the ambient temperature (coldest pixel every 1 s). (D) Representative skin temperature (Tskin) trace of a torpor bout during food restriction. Torpor bouts were detected by determining when Tskin had decreased by >3 standard deviation below the median euthermic temperature, for >1 h. Red circles indicate where torpor was determined to begin, and black circles indicate where torpor was determined to end using these criteria. (E) Raw Tskin temperature traces over consecutive days of food restriction (days 8–17) from a representative morning-fed mouse (left) and a representative night-fed mouse (right). Feeding time is indicated by the pink line, whilst the time at which mice were weighed is shown by the black line.

Results

The change in daily mean Tskin compared to baseline (day 0, the last day of ad libitum feeding) was calculated and used for comparisons due to absolute Tskin values varying between individuals (van der Vinne et al., 2020). Tskin decreased over days of food restriction as mice spent more time in torpor (Figure 2A-B). Day of food restriction significantly affected Tskin in both groups (F_(3.04,18.3)=32.3, p<0.0001), but no significant group effect was found (F_(1,6)=0.413, p=0.5441), nor was there a significant interaction between time and group (F_(17,102)=0.4898, p=0.9525).

Figure 2. Effect of food timing on torpor characteristics.

(A) Change in mean daily Tskin, relative to baseline, over days of food restriction (mean ± SEM). Tskin significantly decreased as food restriction progressed for both groups (F_(3.04,18.3)=32.3, p<0.0001; Two-way ANOVA with repeated measures). (B) Comparison of mean change in Tskin on the last 6 days of recording, when temperature and bodyweight decline had stabilised (mean ± SEM). There was no difference in change in Tskin between groups (p=0.5860, t(6)=0.5753; unpaired t-test). (C) Mean number of torpor bouts per day over days of food restriction. The number of daily bouts significantly increased during food restriction (F_(3.15,18.9)=8.85, p=0.0006; Two-way ANOVA with repeated measures). (D) Comparison of the mean number of bouts entered by each group across the final 6 days of food restriction when bodyweight had stabilised (mean ± SEM). The night group entered significantly more torpor bouts per day compared to the morning group (p=0.0125, t(6)=3.52; unpaired t-test). (E) Change in the latency to the first torpor bout after feeding over days of food restriction (mean ± SEM). Latency to torpor decreased over time. (F) Comparison of latency to torpor in stabilised conditions (mean ± SEM). No difference was found between groups (p=0.0928, t(6)=1.997; unpaired t-test). (G) Comparison of relative torpor timing over the 24h recording period for the final 6 days of food restriction (mean ± SEM). On average, torpor bouts occurred later in the day during morning-feeding compared to the night group but this difference was not significant (p=0.0933, t(6)=1.99; unpaired t-test). (H) Change in the mean duration of torpor bouts over days of food restriction (mean ± SEM). Torpor bouts increased in length over time for both groups (F_(3.06,18.4)=7.87, p=0.0013; Two-way ANOVA with repeated measures). (I) Comparison of mean torpor bout length on the final 6 days of recording (mean ± SEM). Morning-fed bouts tended to be longer on average (p=0.0558, t(6)=2.37; unpaired t-test). (J) Comparison of the longest bout per day over the final 6 days of food restriction (mean ± SEM). No difference was found between groups (p=0.4473, t(6)=0.813; unpaired t-test).

Raw temperature traces suggested potential differences in torpor timing and length between groups (Figure 1E). Unsurprisingly, the number of torpor bouts increased as food restriction progressed before stabilising in line with bodyweight (F_(3.15,18.9)=8.85, p=0.0006; Figure 2C). Feeding time was a significant source of variation, but there was no interaction between group and day of food restriction. Comparison when bodyweight stabilised found that the night group entered significantly more torpor bouts per day compared to the morning fed group (Morning: 2.8±0.2, 95% CI [2.17–3.42]; Night: 4.8±0.5, 95% CI [3.10–6.49]; t(6)=3.52, p=0.0125; Figure 2D).

Latency to torpor was highly variable between animals up to day 8 of food restriction, although mean latency decreased, before stabilising from day 9 for both groups (Figure 2E). No significant group difference was found for stabilised latencies, although average latency was longer for the morning-fed group (Morning: 9.6±0.8h, 95% CI [6.91–12.2]; Night: 6.7±1.2h, 95% CI [2.87–10.5]; p=0.0928, t(6)=1.997; Figure 2F). Additionally, no statistical difference in the relative timing of torpor over each 24h recording period, although night-fed mice entered torpor earlier on average (Morning: 14.2±0.4 ZTh, 95% CI [13.1–15.3]; Night: 12.2±0.9 ZTh, 95% CI [9.23–15.2]; p=0.0933, t(6)=1.99; Figure 2G). However, torpor timing varied greatly in each group, especially in the night-fed group. As such, a significant difference may be found with a larger sample size.

Torpor bouts increased in duration as food restriction progressed for both groups (F_(3.06,18.4)=7.87, p=0.0013), before stabilising in line with bodyweight stabilisation at 85% (Figure 2H). On average, torpor bouts were longer in the morning-fed group compared to the night-fed group, and was trending towards statistical significance (Morning: 4.4±0.7h, 95% CI [2.02 6.72]; Night: 2.6±0.1h, 95% CI [2.16–3.03]; p=0.0558, t(6)=2.37; Figure 2I). There was no statistical difference in the length of the longest torpor bout per day between groups (Morning: 7.3±1.3h, 95% CI [3.08–11.5]; Night: 6.0±0.8h, 95% CI [3.51–8.55]; p=0.4473, t(6)=0.813; Figure 2J).

Discussion

Previously, we reported that torpor is readily induced in response to food restriction with reliable torpor bouts occurring daily when bodyweight reaches 85-90% of free feeding bodyweight (Wilcox et al., 2024d). Here, we investigated the effect of feeding time on torpor characteristics. We report that feeding time influenced some torpor characteristics, namely the number of torpor bouts being entered each day, with night feeding resulting in more torpor bouts. Although bout duration was not significantly altered by feeding time in this study, night-fed bouts were shorter on average, therefore the increased number of bouts may be to compensate and to maximise energy savings. Other studies have found that torpor is primarily under circadian control, which can be influenced by food timing (Northeast et al., 2020; van der Vinne et al., 2018). As such, there may be competing forces contributing to torpor timing and characteristics (i.e., LD cycle, food timing) which may explain why some characteristics were altered but others were not.

Torpor bout timing and duration between individuals and between days within individuals were found to be highly variable (Figure 1E). Moreover, our sample size was reduced due to it being deemed inappropriate to swap treatment groups as originally planned. Unfortunately, these factors limited our ability to compare between groups. We believe that with larger sample sizes, additional group differences would be found, and consider the current report preliminary.

The variation observed highlights the potential for torpor to confound experimental data due to differing torpor history. This may be an important consideration for behavioural tasks, as behavioural performance can be significantly affected by an animal’s experience (Milinski et al., 2021). One approach for minimising the effect of torpor on subsequent performance would be to control for feeding time by ensuring that food is provided at consistent times between days and for each animal. Often, time of feeding is not reported in the published literature; therefore, we recommend that time of feeding is clearly reported to facilitate replication of results and comparisons between studies.

It may be possible depending on the study to provide food at a time that will minimise the impact of torpor on subsequent experimental interventions. For example, food could be provided in the morning to shift torpor bouts to the dark phase, thus providing a window of opportunity to perform experimental manipulations. Where this is not possible, we recommend that mice are closely monitored for torpor bouts, for example using continuous thermal imaging, so that tasks are not being performed immediately following a torpor bout. Moreover, we recommend conducting tasks when torpor history, including number and depth of bouts, is comparable between individuals and days to minimise within- and between-individual variation.

These recommendations are relevant to all fields where food restriction is routinely used in mice and have the potential to refine food restriction procedures used in these fields. From a scientific perspective, we have identified that torpor is readily induced in laboratory mice in response to common food restriction protocols, and that time of feeding effects torpor characteristics. In relation to the 3Rs, remote measurement of Tskin via non-invasive thermal cameras will reduce stress in the animals. Moreover, continuous monitoring of Tskin may allow for early detection of unexpected or pathological hypothermia, enabling appropriate action to be initiated sooner. These recommendations will also contribute to improved data output with less variable data, resulting in reduced sample sizes required in future studies.

Author contributions

S.L.W, S.N.P, D.M.B and V.V.V designed experiments. S.L.W performed measurements. S.L.W and V.V.V performed analysis. S.L.W wrote the manuscript with input from all other authors.