Main

Environmental enrichment strategies aim to improve the welfare of laboratory animals by providing opportunities for them to engage in naturalistic behaviors. The natural behavioral repertoire of mice includes digging, burrowing and burying behaviors, which are used in the wild for defense1 or for the construction of burrows2. In the wild, burrows provide rodents with shelter from changes in ambient temperature, concealment from predators and a safe place to breed3,4. Laboratory mice engage in digging and burrowing behaviors similar to wild mice regardless of age and gender5, suggesting that these behaviors are essential to the rodents' wellbeing. Furthermore, digging and burrowing behaviors have been used in various studies as a reliable tool to assess pain6,7,8 and cognitive function9,10,11,12,13, supporting the need for laboratory mice to be able to engage in these behaviors not only for welfare reasons but also for scientific purposes.

In order to satisfy their behavioral needs, laboratory mice should be provided with a substrate in which to dig or burrow, such as cage bedding3,14. We reviewed all publications of the journal Lab Animal (NY) between 2010 and 2013 and found that the depth of cage bedding provided to laboratory mice varied from ∼0.5 cm to ∼1 cm. This is consistent with Deacon's findings15 that providing cage bedding with a depth of ∼1 cm is common practice in the husbandry of laboratory mice; however, the author suggested that this depth of bedding is not sufficient to promote distinct digging and burrowing behavior15. The effect of cage bedding amount on the body temperature of mice16 and on the concentration of ammonia within the cage17 have been evaluated, but the preferences of mice for particular amounts of bedding have not yet been compared.

Because digging and burrowing are vital components of the behavioral repertoire of all mice18,19, it is reasonable to hypothesize that laboratory mice would generally prefer greater amounts of cage bedding in which to dig and burrow. Studies of consumer demand behavior have indicated that laboratory mice are highly motivated to gain access to burrowing substrates18,19. When suitable substrate such as soil is provided, mice build complex burrows20. Mice engage in digging and burrowing behaviors even when they have access to already constructed burrows18, indicating that engaging in these behaviors is essential for mice and constitutes a behavioral need. Preference tests have been used in mice to detect their preferences for and aversions to different environmental factors21, such as nesting materials22,23, bedding structures24 or ambient temperatures25, to gain more information about their needs. The aim of the present study was to carry out a preference test to determine whether laboratory mice prefer cages with greater amounts of bedding.

Methods

Animals

All housing of animals and experimental procedures were carried out in accordance with the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes and the Animal Welfare Act of Germany.

We obtained 72 female BALB/cByJRj mice and 72 female C57BL/6NRj mice, all 6 weeks of age, from a commercial breeder (Janvier Labs, Le Genest Saint Isle, France). The specific pathogen–free (SPF) status of the mice was documented in a health certificate provided by the breeder. We monitored the health status of the mice after our experiments according to the Federation of European Laboratory Animal Associations (FELASA) recommendations for the health monitoring of rodent and rabbit colonies in breeding and experimental units26. We weighed the mice upon arrival.

For each strain, we divided mice into 18 experimental groups consisting of four mice each. Each experimental group was housed together in a cage system consisting of two Makrolon Type III cages (375 mm × 215 mm × 150 mm; Bioscape, Castrop-Rauxel, Germany) connected by a Perspex acrylic tube (with an outer diameter of 30 mm and an inner diameter of 24 mm; Rudolf Brinke KG, Hannover, Germany), which allowed the mice to move freely between the two cages. The cage system was constructed in cooperation with FBI Science (Essen, Germany) according to the requirements of our institute. Cage systems were kept on an open cage rack. We provided the same amount of pelleted diet (Altromin No. 1324; Altromin, Lage, Germany) and tap water ad libitum to each cage. We provided 1 l of coarse-grained aspen chip bedding (ABEDD; LAB & VET Service GmbH, Vienna, Austria) to each cage, which provided a bedding depth of ∼1 cm. Cage bedding and water were changed weekly. The mice were housed in conventional holding rooms at a temperature of 22 ± 2 °C with relative humidity of 55 ± 10% and a circulation of 10–16 air changes per hour. The room was maintained on a 12-h:12-h light:dark cycle with lights (140 ± 10 lux; located on the cage rack 1 m above the floor) turned on at 6:00 a.m.

Two days after their arrival, we anesthetized each mouse with isoflurane and implanted two subcutaneous transponders (ISO 11784/85 FDX-B; Planet ID, Essen, Germany), one in the shoulder area and one near the base of the tail. We allowed the mice to acclimate to their environments for 2 weeks before the experiments began. All mice were 8 weeks of age. The mean weight of BALB/cByJRj mice was 20.6 ± 1.7 g, and the mean weight of C57BL/6NRj mice was 18.6 ± 2.3 g.

Experimental procedure

Mice are highly social animals27,28,29; therefore, we tested them in groups rather than individually. We tested six experimental groups (n = 24 mice) from each mouse strain in one of three test conditions. In each test condition, we provided different amounts of bedding to the two cages in the cage system in one of the following combinations: 0.5 l (bedding depth of ∼0.5 cm) of bedding and 1.5 l (bedding depth of ∼1.5 cm) of bedding, 0.5 l of bedding and 6 l (bedding depth of ∼6 cm) of bedding, and 1.5 l of bedding and 6 l of bedding. Bedding volumes of 0.5 l and 6 l represent the most and the least bedding, respectively, that can be provided in a Type III cage: a volume of <0.5 l would not be adequate to cover the cage floor, whereas a volume of >6 l would make it difficult to close the cage lid and would cause the water bottle to leak. The bedding volume of 1.5 l is the volume of cage bedding commonly provided in our facility. In the test conditions in which 6 l of bedding was provided in one of the cages, the tube was attached to the cage wall at a higher level above the level of the bedding; therefore, we placed the other cage in this cage system on a platform so that the tube was in a horizontal position (Fig. 1).

Figure 1: Test cage system consisting of two cages connected by a tube.
figure 1

Two sensors at each end of the tube connecting the two cages detected the activity of the mice via implanted transponders. In the test cond`itions in which 6 l of bedding were provided to one cage (right), the cage containing the smaller bedding amount (left) was placed on a platform.

Full size image

An experimental group of four mice was placed in the cage system on day 0 of the experiment. Using the transponders implanted in the mice, we continuously tracked the movements of the mice between the two cages of the cage system with an automated computer tracking system (FBI Science GmbH, Essen, Germany) for 1 week. Two sensors at each end of the tube connecting the two cages detected the crossing of a mouse from a cage into the tube and from the tube into a cage. We used these detected crossings to measure the time spent by each mouse in each of the two cages and in the tube, as well as to measure the number of times each mouse crossed between the two cages. Prior to this study, we carried out experiments to validate the tracking system by comparison with video recordings of mouse activity in the cage system. The data obtained with the automated tracking system were in 98.5% accordance with video analysis, and the times spent in the right-hand cage and in the left-hand cage as determined by the automated tracking system and by video analysis were significantly correlated (r2 = 1 and r2 = 0.989, respectively). These findings suggest that the automated tracking system provided reliable information about the activity and locations of the mice.

We weighed the food and water in each of the two cages within a cage system before and after the experiment to measure an experimental group's overall consumption in each of the two cages.

In each of twelve 1-week experiments, three experimental groups were tested at one time, each in one of the three test conditions. In order to reduce the potential influence of the external environmental conditions on the results, after each experiment we rotated the cage systems of each test condition between three shelves and switched the two cages of a system between the two positions on each shelf. Consequently, for each mouse strain, each bedding volume in each test condition was tested once in each of six possible positions.

Statistical analysis

We analyzed the tracking data collected between day 1 and day 6 of the 7-d experiment using the StatView computer program (Version 5.0; SAS Institute Inc., Cary, NC). We used a paired t-test to analyze the differences in time spent in the cage with more bedding and time spent in the cage with less bedding. The effects of light phase were analyzed by comparing the times spent in each cage during the light phase (6:00 a.m. to 6:00 p.m.) with the times spent in each cage during the dark phase (6:00 p.m. to 6:00 a.m.).We analyzed the effect of mouse strain and the effect of the test condition using an analysis of variance (ANOVA) followed by a Scheffé test. The level of statistical significance was set at < 0.05.

Results

Experimental group comparisons

We first analyzed the data from each experimental group of four mice. C57BL/6 experimental groups spent a significantly greater percentage of time in the cage with more bedding than in the cage with less bedding in all three test conditions (P < 0.0001 when 0.5 l bedding and either 1.5 l of bedding or 6 l of bedding were provided; P = 0.0001 when 1.5 l of bedding and 6 l of bedding were provided). BALB/c experimental groups spent a significantly greater percentage of time in the cage with 1.5 l of bedding (P < 0.0001) or 6 l of bedding (P = 0.0598) when the other cage in the cage system contained 0.5 l of bedding, but in the test condition in which the two cages contained 1.5 l of bedding and 6 l of bedding, the mice showed no significant difference in the percentages of time spent in the two cages (P = 0.1589).

Owing to the small number of samples when analyzing the experimental groups rather than individual mice, we next determined whether there was a variance in the activity of individual mice within an experimental group. Using the data of one BALB/c and one C57BL/6 experimental group, we calculated the coefficients of variance (CVs) for the times spent in the two cages and for the crossing frequency between the cages for each hour of the experiment. We compared the overall mean of these CVs for both lighting phases with the mean CV during dark phase hours and the mean CV during light phase hours. The results showed that the times spent in each cage and the number of crossings differed between the mice of a given experimental group within a 24-h period, suggesting that each mouse within an experimental group moved independently between the two cages. This was true for both C57BL/6 mice and BALB/c mice. Therefore, we decided to continue our analysis using the data from individual mice rather than from the experimental groups as a whole.

Time spent in each cage

In each of the three test conditions, BALB/c mice and C57BL/6 mice spent a significantly greater percentage of time in the cage with more bedding than in the cage with less bedding (P < 0.0001; Fig. 2). C57BL/6 mice spent a significantly greater percentage of time in the cage with more bedding than did BALB/c mice overall (P = 0.0010) and in the test conditions in which 6 l was the larger bedding volume (versus 0.5 l of bedding (P = 0.0135) or 1.5 l of bedding (P = 0.0120)). Mice of both strains spent significantly greater percentages of time in the cage with more bedding in test conditions in which the smaller bedding volume was 0.5 l (versus 1.5 l of bedding (P = 0.0005) or 6 l of bedding (P = 0.0065)) than in the test condition in which 1.5 l was the smaller bedding volume (versus 6 l of bedding).

Figure 2: Percentage (mean ± s.d.) of time spent in the cage with more bedding and in the cage with less bedding in each test condition.
figure 2

(a) BALB/c mice. (b) C57BL/6 mice. *P < 0.0001.

Full size image

Effect of light phase on time spent in each cage

We analyzed the effect of light phase on the percentage of time spent in the cage with more bedding versus the cage with less bedding. During both light and dark phases, BALB/c mice (Fig. 3) and C57BL/6 mice (Fig. 4) spent a significantly greater percentage of time in the cage with more bedding than in the cage with less bedding in all three test conditions. Mice of both strains spent a greater percentage of time in the cage with more bedding during the light phase than they did during the dark phase. On average, BALB/c mice and C57BL/6 mice spent ∼17% more time and ∼33% more time, respectively, in the cage with more bedding during the light phase than during the dark phase. During the dark phase, C57BL/6 mice spent a greater percentage of time in the cage with more bedding than did BALB/c mice, although this difference was not significant.

Figure 3: Percentage (mean ± s.d.) of time BALB/c mice spent in the cage with more bedding and in the cage with less bedding in each test condition.
figure 3

(a) During the light phase. (b) During the dark phase. *P < 0.05. **P < 0.0001.

Full size image
Figure 4: Percentage (mean ± s.d.) of time C57BL/6 mice spent in the cage with more bedding and in the cage with less bedding in each test condition.
figure 4

(a) During the light phase. (b) During the dark phase. *P < 0.0001.

Full size image

Number of crossings between the two cages

Overall, mice of both strains crossed between the two cages a greater number of times during the dark phase than during the light phase (P < 0.0001). BALB/c mice crossed between the two cages a greater number of times than did C57BL/6 mice during the light phase (P < 0.0001), whereas C57BL/6 mice crossed between the two cages a greater number of times than did BALB/c mice during the dark phase (P = 0.3798).

The test condition had a significant effect on the number of times the mice crossed between the two cages, both during the light phase (P = 0.0366) and during the dark phase (P = 0.0218). In the test condition in which the cages contained 0.5 l of bedding and 1.5 l of bedding, the mice on average crossed between the two cages a greater number of times (163.23 ± 124.42 times during the light phase and 548.67 ± 609.10 times during the dark phase) than they did in test conditions in which the cages contained 6 l of bedding and either 0.5 l of bedding (116.29 ± 75.82 times during the light phase and 326.96 ± 99.34 times during the dark phase) or 1.5 l of bedding (132.11 ± 88.09 times during the light phase and 448.86 ± 260.15 times during the dark phase).

Differences in food and water consumption

In the test conditions in which the cages contained 0.5 l of bedding and either 1.5 l of bedding or 6 l of bedding, experimental groups of BALB/c mice and C57BL/6 mice consumed more food (P = 0.0283 and P ≥ 0.05, respectively) and more water (P < 0.0001 and P = 0.0278, respectively) in the cage with more bedding. In the test condition in which the cages contained 1.5 l of bedding and 6 l of bedding, experimental groups of BALB/c mice and C57BL/6 mice consumed more food and water in the cage with 1.5 l of bedding, although these differences were not significant.

Discussion

We provided different amounts of bedding to two cages in a cage system in three combinations to test the preferences of female BALB/c mice and female C57BL/6 mice for different bedding amounts. In each combination we tested, mice of both strains spent significantly more time in the cage with more bedding, suggesting that these mice strongly prefer cages with more bedding to cages with less bedding. The preference for the cage with more bedding was more pronounced in test conditions in which the smaller bedding amount was 0.5 l than in the test condition in which the smaller bedding amount was 1.5 l, suggesting that 0.5 l may be an inadequate amount of bedding for cages housing BALB/c and C57BL/6 mice. The results are consistent with previous studies that have shown the importance of adequate cage bedding for thermoregulation16,25 and for species-appropriate husbandry3,18. A positive effect of deeper cage bedding on the welfare of rodents other than mice has been reported; for example, golden hamsters housed in cages with 80 cm of bedding engaged in significantly less wire-gnawing behavior than those housed in cages with 10 cm of bedding30.

Although the preferences shown in each of the three test conditions were consistent, they differed somewhat between the two strains of mice, especially in the test condition in which the cages contained 1.5 l of bedding and 6 l of bedding. Although mice of both strains preferred the cage with 6 l of bedding in this test condition, C57BL/6 mice showed a stronger preference than BALB/c mice. Previous studies have reported differences in the behavior of these two strains, possibly as a result of adaptation to the different habitats of their wild ancestors31,32: whereas C57BL/6 mice engage in more digging and burrowing behavior, BALB/c mice engage in more nest-building behavior. Nevertheless, burrowing constitutes a behavioral need for BALB/c mice as well18.

The preferences of the mice for cages with more bedding appeared to be more pronounced during the light phase than during the dark phase. Like other nocturnal animals, mice tend to rest during the daytime and thus may prefer to sleep in cages with more bedding. Comparable to the findings of other studies27,33, the preferred cage of the mice during the light phase was also the cage in which they preferred to sleep (data not shown).

The cage system comprised a relatively large area compared with a typical mouse cage, which may have encouraged the mice to move between both cages regularly34. Previous studies showed a higher level of running wheel activity for C57BL/6 mice in comparison to BALB/c mice35,36; therefore, we expected that C57BL/6 mice would show an overall greater number of crossings between the two cages than BALB/c mice. Instead, C57BL/6 mice were more active than BALB/c mice during the dark phase only in the two test conditions in which 6 l was the larger bedding volume provided. It is possible, however, that C57BL/6 mice were more active within a single cage of the cage system. In the same two test conditions, the mice crossed between the two cages a fewer number of times. The large volume of bedding provided in these test conditions may have promoted intensive digging and burrowing behavior, limiting the activity of the mice to a single cage. The experiments were not recorded by video, however, so the behaviors of the mice within each cage are not known.

Although it has been reported that food and water consumption are higher near the location of the nesting site of mice33, other preference tests have shown that food and water consumption are not necessarily higher in the preferred location22,24,37. In this study, the relationship between food and water consumption and the preferred cage was not clear. In test conditions in which the cages contained 0.5 l of bedding and either 1.5 l of bedding or 6 l of bedding, mice of both strains consumed more food and more water in the cage with more bedding. In the test condition in which the cages contained 1.5 l of bedding and 6 l of bedding, however, mice of both strains tended to consume more food and more water in the cage with 1.5 l of bedding, even though they spent more time overall in the cage with 6 l of bedding. It is possible that in this test condition, the mice used their time in the cage with less bedding especially for feeding and drinking.

Adaptation can influence the preferences of animals to varying degrees33,38. In a study in which the preference of mice for two different types of cage bedding was compared, previous experience with a bedding type was shown to influence not the type of bedding preferred but rather the strength of the mice's preference for that bedding type24. Therefore, it may be predicted that mice previously housed in cages with a larger amount of bedding might show a stronger preference for cages with that amount of bedding. Although we provided 1 l of bedding to all cages during the acclimatization period, all mice preferred 6 l of bedding to 0.5 l of bedding or 1.5 l of bedding. Because the volume of cage bedding provided initially did not differ between the groups, no reliable statement can be made whether the use of larger cage bedding amounts during acclimatization would have affected the outcome of this study.

Although the mice created small hollows in the cage bedding, they were not able to build complex burrows, regardless of the bedding amount. Therefore, carrying out health checks of the mice was not problematic in any of the test conditions. Like any other form of environmental enrichment39,40,41, the bedding amount might influence physiological parameters, such as body weight, relative organ weights, hematological data or behavioral parameters. It is currently unknown how different bedding amounts can affect experimental results, but this should be evaluated in future studies. Furthermore, published studies should specify the amount of bedding provided to animals' cages, as well as the resulting depth of the bedding in the cages, in order to improve the reproducibility and comparability of experiments.

The results of the present study highlight the preference of female BALB/c and C57BL/6 mice for greater amounts of cage bedding. Although the importance of suitable husbandry of laboratory mice is undisputed, the practical implementation of such husbandry must be considered. Providing 6 l of bedding to every mouse cage of this size would increase enormously the workload of personnel as well as financial costs. Therefore, we do not necessarily recommend the use of this bedding amount, but rather suggest that the amount of bedding provided to mouse cages should be increased as much as practically possible to ensure that mice are provided with a more naturalistic environment. This may be especially necessary for mice included in experiments for which any other type of environmental enrichment is barred.