Mice that gorged during dietary restriction increased foraging related behaviors and differed in their macronutrient preference when released from restriction

Baseline

Sixty male MF1 mice (Harlan UK Limited, England), aged 14 weeks, were used in this study. After a period of acclimation they were separated into individual cages (M3 cage 48 × 15 × 13 cm; NKP Cages, Kent, UK), each with constant access to water. The mice were maintained at a temperature of 21 °C ± 0.5 °C under a photoperiod of 12 h light/dark. They had sawdust bedding and a cardboard tunnel for enrichment.

For the first 14 days of the study all of the 60 mice were provided with a weighed amount of approximately 10 g of standard chow, (CRM; Special Diets Services, BP Nutrition, UK), which is in excess of their daily dietary requirements. The food was placed directly onto the cage floor at the same time each day (middle of their light period), to avoid difficulties obtaining it from the hopper. Each mouse and its food were weighed every day at the time of food introduction and then 2 h later. Any remaining food from the previous day was removed from the bedding and weighed before the next day’s food was given. Previous studies with this strain of mouse on this diet indicated that loss of minor food items in the sawdust amounted to less than 2% of daily intake (Johnson, Thomson & Speakman, 2001).

Caloric restriction

From the 60 mice, 10 controls were selected which were matched for body mass with the remaining mice and these individuals continued to feed ad lib using the same regime as previously described. They were always fed at the same time as the restricted mice during the light phase. The remaining 50 mice were fed a restricted diet, calculated at 75% of each individuals average daily intake from the previous two weeks. As each individual was restricted according to its own baseline food intake, the level of treatment for all mice was therefore identical even although there will be variation between individuals in absolute terms. This diet restriction was continued for another 14 days, still with all 60 mice being weighed when the food was added and also 2 h later along with the remaining food. Near the end of the restriction period, we selected the top and bottom gorgers from the 50 restricted mice (n = 11 for gorgers and n = 9 for non-gorgers), and the other mice were removed from the study. Gorgers were determined as those that ate greater than 25% of their average daily intake in the first two hours after inclusion.

To assess the detailed behavior of the mice during the restriction period, we recorded footage of a subset of the mice. Three control mice, 3 mice showing gorging behavior and 3 mice showing non-gorging behavior were placed with their bedding in new cages which were identical in size to their previous cages (also M3 cage 48 × 15 × 13 cm; NKP Cages, Kent, UK) but were clear perspex with a dark screen behind them. Cardboard tubes were replaced with clear plastic igloos. After a 2 h period of acclimation to the new cage, a Hi-8 video camera (Sony CCD-VX1E/PAL) was used to record the mice for 3 h (one hour before food inclusion and two hours after food inclusion). While filming, the mice underwent the same regime as usual and were therefore weighed at the time of food inclusion, and two hours after as normal. Control mice also underwent their usual regime and had their food replaced at the same time as the restricted mice received their rations and were also weighed. The three-hour videos were converted to digital format and then analysed using HomeCageScan™ 2.0 (Clever Sys Inc., Virginia, USA). This software enabled a detailed analysis of the behaviours that mice conducted in the cage and has been validated to be over 90% accurate with respect to human scoring. The behaviour determined in each frame was recorded at a rate for 30 frames per second.

Diet choice recovery

For the final 14 days, the eleven gorgers, nine non-gorgers and ten controls were put into larger individual cages (1290D polypropylene cage, 42.5 × 26.6 × 15 cm; Techniplast) to allow for the provision of a diet choice. The food hoppers of the cages were divided into three separate compartments, where we placed 20 g of each of the three different foods. The three food choices were high in fat, protein and carbohydrate (high protein DO4080301, high fat D12492 and high carbohydrate D12450B; Research Diets, New Brunswick, USA). The feeding regime remained unaltered with the 30 mice weighed before 20 g of each food was added to the hopper and then the mouse and each diet was reweighed at 2 and 24 h after food inclusion.

To calculate the water and gross energy content of the four different food types used in this study, a sample of each diet was dried to constant mass and then analysed using bomb calorimetry (Adiabatic bomb calorimeter; Gallenkamp, Loughborough, UK) (Table 1).

Statistics

Normality tests (Anderson-darling) were completed on all data, which were normalised if required using a Box–Cox transformation before being subjected to ANOVA (Tukey’s), paired t-tests or General Linear Modelling (GLM). Minitab V16 (Minitab Inc, USA) was used throughout. Means are shown ± standard errors. The mean of the last 5 days of each time period, baseline, restriction, and diet choice are presented unless otherwise stated.

Baseline

Only data collected from the animals that were later selected as gorgers (n = 11), non-gorgers (n = 9) and controls (n = 10) was analysed. During the baseline period the mice that subsequently went on to become gorgers ate 0.5 g less than the other 2 groups (ANOVA F_2,27 = 9.07, P = 0.001). This meant that their energy intake was also significantly lower (Table 2). During baseline however there were no significant differences between the three groups for dry food intake two hours after introduction of food (ANOVA F₂, ₂₇ = 2.16, P = 0.14), and the mice ate an average of 2.9% of their total food intake within this time with a range of 0.4–7.8% indicating that they did not show any gorging tendencies prior to restriction. Body mass was not significantly different between the 3 groups and the rate of BM increase was on average 0.15 g/day which was not significantly different between the 3 groups even although there was a significant difference in daily food intake.

Restriction phase

Each restricted mouse received exactly 25% less food than it consumed when provided with food ad lib which was, as expected, a significant reduction in food and energy intake (Fig. 1). There was a significant increase in the daily food intake of control mice over the same time period as the restriction by an average of 0.72 ± 0.19 g/day (10.8%) (Paired t-test P = 0.02). Consequently the realised restriction relative to controls averaged 35%. Mice that showed gorging behavior were apparent after only 3 days of restriction with significant increases in 2 h food intake above the control and non-gorging mice occurring in some individuals, even on day 1. The extent of gorging behavior increased throughout the 14 day measurement period so that the increase in 2 h food intake was highly significant when averaged over the last 5 days of the restriction (ANOVA F₂, ₂₇ = 44.8, P < 0.001). Gorging mice ate an average of 51.6% ± 5.98% of their total food intake in 2 h, which was 3 times that eaten by non-gorgers in the same period and 21 times higher than the controls (Fig. 2). The ‘non-gorgers’ also increased their 2 h food intake during restriction but did not exceed the arbitrary limit to become defined as a gorger (25% of available food in 2 h) as they only consumed an average of 15.8% ± 1.98% of their total food intake in 2 h, which was significantly above the controls that ate 15 ± 0.33% of their 24 h food intake over the same 2 h period.

To compare body mass of the three groups for each phase, the last five days were averaged for all animals in each group. Gorgers and non-gorgers both significantly decreased their body mass during restriction by an average of 3.3 g and 1.2 g respectively (Paired t-test P < 0.001). The decrease in body mass observed in the gorgers compared to their baseline value was significantly greater than the non-gorgers (9.6% compared to 3.4%). The controls significantly increased their body mass over the same time period by 1.6 g (4.4%) (Paired t-test P < 0.01) (Fig. 3). Over the last 5 days of restriction the gorgers were still losing weight at a rate of 0.22 ± 0.04 g/day while the non-gorgers had stabilised their body mass as the average rate of weight loss was only 0.04 ± 0.05 g/day showing that they had almost reached energy balance. In contrast the controls were gaining weight at 0.14 ± 0.04 g/day. These values were significantly different between the 3 groups (ANOVA F_2,27 = 14.7, P < 0.001).

Behavioral analysis

During the restriction phase, there was a significant difference in the behavior of the three groups. For the hour before food was introduced the controls spent a significantly greater amount of time sleeping than both gorgers and non-gorgers, while non-gorgers slept more than gorgers (ANOVA F_2,8 = 50.8, P < 0.001). As expected, gorgers therefore showed greater amounts of general activity than the other 2 groups, (ANOVA F_2,8 = 80.1, P < 0.001). This was accounted for by the fact that gorgers spent more time foraging (looking through the bedding and around the cage) than non-gorgers who in turn spent more time than the controls (ANOVA F_2,8 = 3632.79, P < 0.001). Both restricted groups spent equal amounts of time climbing on the bars and rearing up on their hind legs which was more than the non-restricted controls (ANOVA climbing F_2,8 = 5968.27, P < 0.001; rearing F_2,8 = 14.23, P = 0.009). These behaviors both involve looking outside the cage, possibly to see when food would arrive. There was no difference in the amount of drinking or low level activity (general movement on the cage floor) between the groups (Table 3A).

After food inclusion, controls spent a significantly greater amount of time sleeping than non-gorgers who in turn, slept more than gorgers (ANOVA F_2,8 = 692.3, P < 0.001). Gorgers were more generally active spending a significantly greater amount of time foraging (ANOVA F_2,8 = 71.85, P < 0.001), rearing (ANOVA F_2,8 = 71.27, P = 0.006), eating (ANOVA F_2,8 = 614.99, P < 0.001) and conducting low level activity (ANOVA F_2,8 = 34.85, P < 0.001) than non-gorgers and controls who did not differ from each other in these activities (Table 3B). The differences between controls and non-gorgers was apparent when comparing climbing behavior as non-gorgers spent significantly more time climbing than the controls (ANOVA F_2,8 = 47.08, P = 0.001). There was no difference in drinking activity between the three groups.

Diet choice

After providing all three groups with a choice of diets high in fat, carbohydrate and protein, the controls daily food intake significantly decreased so that over the last 5 days of measurements it was 3.8 ± 0.13 g which was 2.1 g less than just prior to diet choice provision (Table 2). They ate some of all three diets and although the energy value for each diet was greater than the previously fed chow (Table 1), the energy intake calculated from the total dietary intake was reduced by 31.2 kJ/day (or 26.9% lower) than prior to providing the diet choice. This difference was significant (Paired T-test T = 5.50, P < 0.001). They managed, however, to maintain a similar rate of body mass increase even with the reduced energy intake. This is because there were increases in digestive efficiency on the 3 diets compared to the standard chow (Table 1 previously measured by J Kagya-Agyeman, 2009, unpublished data) which meant that the energy assimilated was not significantly different from the baseline period (Paired t-test T = 1.19, P = 0.26) (Fig. 4).

For the mice that had been subjected to dietary restriction, there was an initial marked hyperphagia over the first 3 days of free diet choice before food intake stabilised at a level that did not differ significantly from controls (ANOVA F_2,27 = 0.96, P = 0.40). Mice that had been gorgers continued to have higher 2 h intake than either non-gorgers or controls when provided with the ad lib diet choice. On average they consumed over double that of the non-gorgers and controls in 2-h after feeding (ANOVA day 1–3 F_2,27 = 23.28, P < 0.001). This pattern persisted throughout the 14 days of measurement after being given ad lib access to food (day 10–14: F_2,27 = 5.73, P = 0.008). During day 1–3 of the diet choice, 24 h energy intake in the gorgers was significantly higher (25%) than their baseline energy intake (P < 0.001). Non-gorgers, however, matched their energy intake to that of the baseline period (Paired T-test T = 0.25, P = 0.81). Both the gorgers and non-gorgers fully recovered their lost body mass in the first few days (Fig. 3) so that they were not different to the controls (ANOVA F_2,27 = 0.10, P = 0.91). BM was above the initial baseline measurements for the gorgers (Paired T-test T = 4.04, P = 0.002) although did not reach significance for the non-gorgers (Paired t-test T = 2.21, P = 0.058). The energy intake over the final 5 days of diet choice, like the controls, was significantly lower than the baseline measurement for these 2 groups (Paired T-test P < 0.001) although they still maintained a positive weight gain at the same rate as during the baseline period. As with the controls this was due to the increased digestive efficiency resulting in an energy assimilation which was not significantly different than baseline for gorgers (Paired T-test T = 0.01, P = 0.99). The non-gorgers had a slight but significant reduction in assimilated energy by 10 kJ/day (13%) relative to baseline (Paired T-test T = 3.37, P = 0.01).

The different amounts of each diet consumed (high protein, high fat or high carbohydrate) over either the first 1–3 days or the last 5 days of diet choice were compared. In all groups, the high fat diet was preferred above the high carbohydrate diet and least preferred was the high protein diet (Fig. 5). Both groups of animals that had been on food restriction had a significantly greater intake of high carbohydrate diet than controls over the first 3 days (ANOVA F_2,27 = 5.19, P = 0.012; Fig. 5A). In addition, the gorgers also had a significantly greater intake of high protein diet than both non-gorgers and controls (ANOVA F_2,27 = 3.70, P = 0.038). High fat diet consumption didn’t differ between the groups. The amount of energy consumed for each of the three macronutrients across all 3 diets was calculated (Table 2). The restricted mice did consume more carbohydrate when combining all 3 diets than the controls while only the gorging mice increased their intake of protein. By the end of the diet choice period, the intakes of each diet had normalised to the levels of the controls (Fig. 5B).