Housing mice in the individually ventilated or open cages—Does it matter for behavioral phenotype?

Individually ventilated caging (IVC) systems for rodents are increasingly common in laboratory animal facilities. However, the impact of such substantial change in housing conditions on animal physiology and behavior is still debated. Most importantly, there arise the questions regarding reproducibility and comparison of previous or new phenotypes between the IVC and open cages. The present study was set up for detailed and systematic comparison of behavioral phenotypes in male and female mice of three widely used inbred strains (C57BL/6JRccHsd, DBA/2JRccHsd, 129S2/SvHSd) after being kept in two housing environments (IVC and open cages) for 6 weeks (since 4 weeks of age) before behavioral testing. The tests addressed exploratory, anxiety‐like and stress‐related behavior (light‐dark box, open field, forced swim test, stress‐induced hyperthermia), social approach and species‐specific behavior (nest building, marble burying). In all tests, large and expected strain differences were found. Somewhat surprisingly, the most striking effect of environment was found for basal body temperature and weight loss after one night of single housing in respective cages. In addition, the performance in light‐dark box and open field was affected by environment. Several parameters in different tests showed significant interaction between housing and genetic background. In summary, the IVC housing did not invalidate the well‐known differences between the mouse strains which have been established by previous studies. However, within the strains the results can be influenced by sex and housing system depending on the behavioral tasks applied. The bottom‐line is that the environmental conditions should be described explicitly in all publications.

stress-induced hyperthermia), social approach and species-specific behavior (nest building, marble burying). In all tests, large and expected strain differences were found. Somewhat surprisingly, the most striking effect of environment was found for basal body temperature and weight loss after one night of single housing in respective cages. In addition, the performance in lightdark box and open field was affected by environment. Several parameters in different tests showed significant interaction between housing and genetic background. In summary, the IVC housing did not invalidate the well-known differences between the mouse strains which have been established by previous studies. However, within the strains the results can be influenced by sex and housing system depending on the behavioral tasks applied. The bottom-line is that the environmental conditions should be described explicitly in all publications. Among many factors contributing to these problems, environment in its broad sense is certainly one of the major issues. It is too often when the differences between the results by different laboratories are referred to be caused by environmental or procedural differences. As a matter of fact, this is not necessarily an unexpected or negative outcome. 4 There have been several attempts to standardize the research methods and protocols. 5 However, it appears that extreme standardization can yield the results idiosyncratic to particular laboratory, and therefore, standardization may not be the best way for improving external validity of animal models. 6,7 Nevertheless, complete reporting of the details of experimental design is an essential and mandatory part of scientific publications, although often neglected. 8,9 Despite all efforts made so far for improving the standards for reporting, the impact on the quality of reporting has been limited. 10 The full life history of the animals can be considered even more important determinant of the phenotype in comparison to the particular test protocol. In this respect, the environment of animal cage (and animal facilities in general) has been underestimated. Certainly, some aspects of housing conditions have been highlighted by earlier research 11 and in the current legislation regulating animal welfare-for example, social housing and appropriate enrichment. 12 Already these developments have caused some concerns in community. 13 In addition, during last 10-15 years most of the newly built or renovated animal facilities are equipped with the individually ventilated caging (IVC) systems. The IVC has been suggested to be useful for controlling the spread of infections between animals, to maintain immune and health conditions at required levels, but also beneficial for staff by reducing air pollution and allergens. 14,15 However, from the animal viewpoint this can be considered as a major difference and re-validation of the established phenotypes is warranted. Indeed, there are several reports published where behavioral phenotypes have been affected by moving the animals from the open cages (OCs) to IVC (see Table 1).
Therefore, the goal of the present study was a systematic evaluation of basic behavioral phenotypes in male and female mice of three commonly used inbred strains (C57BL/6JRccHsd, DBA/2JRccHsd, 129S2/SvHsd) housed in IVC or OCs. For comprehensive characterization, we designed a test battery with the focus on exploratory activity, social, species-specific and stress-related behavior. This approach was based on suggestions and recommendations by earlier studies addressing the concept of test batteries and the possible effect of repeated testing. 16 16,20 The mice were delivered in four batches with interval of 2 weeks (first and third batch-males, second and fourth batch-females; 12 animals per strain in each batch), and were randomly allocated (random number calculator, available at https://www.graphpad.com/quickcalcs/) to the groups of three animals per cage between two housing conditions: On the next day after arrival, the mice were ear punched for identification and the body weight was measured. In both housing conditions, enrichment was provided by bedding (aspen chips 5 × 5 × 1 mm, 4HP, Tapvei, Kiili, Estonia), nesting material (aspen strips, PM90L, Tapvei and a pressed cotton square, Nestlets; Ancare, Bellmore, New York) and aspen brick (100 × 20 × 20 mm, Tapvei).
Food (Global Diet 2916C, pellet 12 mm, Envigo, Horst, The Netherlands) and water (filtered and ultraviolet-irradiated) was available ad libitum. Room temperature was 22 ± 2 C and relative humidity 50 ± 15%. The lights were on between 6:00 and 18:00. The cages were cleaned once per week, suitable nesting material was transferred to the new cage in order to reduce aggressiveness and facilitate adaptation. 21,22 Animals were weighed before moving to the clean cage.
One C57BL/6J male mouse from the first batch (OC) was found dead next morning after arrival, and one C57BL/6J male mouse from the second batch (OC) was removed at the age of 9 weeks because of aggressive behavior towards the cage-mates. Accordingly, the number of mice in each sub-group (per strain, sex and housing condition) was twelve, except for C57BL/6J males in OCs (n = 10).

| Behavioral tests
Testing started when the animals were 10 weeks old (ie, after 6 weeks of adaptation) and a battery of behavioral tests was carried out during 2 weeks in the order described below. According to arrival, testing was also carried out in four blocks by two experimenters and sequence of animals for separate tests was randomized. For all tests the animals were moved to the testing room at least 30 minutes before the start and testing order was randomized each time. Experiments were carried out during the light period (between 8:00 and 16:00) in the following order ( Figure 1A)-light-dark (LD) box (day 1); open field (OF, day 3); social exploration (SOC, day 5); marble burying test (MBT, day 8); Nest (day 8-9); stress-induced hyperthermia (SIH, day 9); forced swimming test (FST, day 12). The number of fecal boli was counted by experimenter after the end of the trial.

| Open field
The same arena as for the LD test was used without the dark insert, illumination of the arena was approximately 150 lx. Animals were released in the corner of the arena and monitored for 20 minutes. For analysis, the arena was divided into the center and periphery, the peripheral zone defined as a 6 cm wide corridor along the wall.

| Marble burying test
Test was performed in the IVC cage with a thick layer (4-5 cm) of dampened bedding. The cage was covered by a perforated transparent polyvinyl chloride-lid. Twenty glass marbles were placed on top of the bedding (arranged in 4 × 5) and mice were allowed to explore the cage and interact with the marbles for 30 minutes. Thereafter, the mice were returned to the home cage, the test cages were labeled and the marbles were counted into three categories (covered fully [hidden], less than 50% or more than 50% of the marble visible) by experimenter blinded regarding the identity of the subjects.

| Nest construction
In the afternoon (2 hours before dark onset) the mice were individually moved to the new cages (the same type as for group housing), with food and water available ad libitum and one intact Nestlet (5 cm square, approximately 2.5 g) as nesting material. Thus, the nesting material was familiar for the animals. Quality of the nest was assessed and scored next morning (2 hours after beginning of the light period, that is, 16 hours after beginning the trial) according to 5-point scale. 23

| Stress-induced hyperthermia
After visual inspection and assessment of the nest, the mice were removed from the cage and rectal temperature was measured. Then the body weight was measured and animal was returned to the cage, 10 minutes later measurement of rectal temperature was repeated.
The difference between two measurements is defined as a stressinduced hyperthermia. Thereafter, the mice were returned to their original home cages in groups of three.

| Forced swimming test
The mouse was placed for 6 minutes in the glass cylinder ( 18 cm, height 25 cm) filled with water at 23 ± 1 C to the height of 15 cm.
The time of immobility (passive floating, when the animal was motionless or doing only slight movements with tail or one hind limb, whereas the animal was judged to be active when struggling, climbing or swimming using all four paws) was measured in 2-minute intervals.
Behavior was video recorded and immobility episodes were detected by Ethovision XT 10.0 (Noldus) software.

| Statistics
All statistical analyses were conducted using IBM SPSS Statistics 25.
Data from behavioral analysis were subjected to Univariate Analysis  Figure 1B,C). There was no

| Light-dark box
The latency to enter the light compartment (Figure 2A

| Open field
Similar to LD-test, there was no overall difference in total activity (distance travelled, Figure 3A,

| Social approach
Total activity during 10 minutes of testing ( Figure 4A

| Marble burying and nest building
Burying activity (number of marbles completely covered by bedding material after 30 minutes, Figure 5A) was not affected by housing conditions (F 1,130 = 3.33, P = 0.07), and although significant strain difference was found (F 2,130 = 70.81, P < 0.0001, B6 > D2 > 129), there was no interaction between environment and strain (F 2,130 = 0.12, Nest construction ( Figure 5B

| DISCUSSION
With the present study we wanted to address in a systematic manner the possible modifying effect of IVC housing on mouse behavioral phenotype. In order to do so, we tested the male and female mice of three common inbred strains after housing them for 6 weeks (starting from the age of 4 weeks) in the IVC or OCs.
There is no doubt that the housing conditions have a robust influence on the behavior and physiology of laboratory mice. 11 However, the observed effects can depend on the sex and strain of the animals on the one hand, and on the cage and laboratory environment on the other hand. [24][25][26] For the latter part, the major difference can arise from the housing system used in the animal facilities-either IVC or, nowadays explicitly on a regressive basis, conventional OCs. [27][28][29] A major variable between the different housing conditions is a microenvironment and ventilation affecting on it. In the IVC, forced ventilation has an effect on ammonia and carbon dioxide concentrations, relative humidity and temperature because of the differences in ventilation rate. 14,30 The IVC seem to keep the ammonia level in a more tolerable concentration compared to OCs. 14 Being exposed to a reduced oxygen concentration may induce hypoxia causing impairments in animal physiology. 31 Mice in the IVC can be disturbed by high ventilation rates, 32 suffer from noise and vibration originating from the ventilation system 33 and limited climbing possibilities. 34 Acoustic environment, which differs between the housing conditions, may have a different effect on mouse strains with different hearing sensitivity. 35 It has been suggested that ventilation, which varies among caging systems, has an effect on mice growth and behavior. 36 In the present study, no difference in weight gain was detected because of housing conditions, whereas controversial results have been reported by others. Significant differences in body weight development have been reported between different IVC systems 14 and after changing the housing system from OC to the IVC. 37 On the other hand, the housing system did not have any impact on growth in male rats. 38 The diet may have a direct influence on the growth, and there is a growing and compelling evidence on the role of gut microbiota on brain and behavior. 39 In our study, the mice were housed in the specific pathogen free housing conditions, except for the time spent in the center which was slightly increased in the IVC group, pointing to possible reduction in anxiety-like behavior (opposite to the light-dark test). However, it is important to note that the latter finding was affected by genetic background -the B6 mice from OCs displayed reduced activity in the center, whereas the D2 mice from OCs showed increased activity. The discrepancy between the results of the LD box and OF (and in general, between different "unconditioned," "exploratory" tests for anxiety-like behavior) may be because of differences in the strength and type of aversive stimulation (bright light, openness, height etc.). 55 We did not see any significant difference in expression of directed exploration towards conspecific (a same-sex, unfamiliar mouse) during the social approach test, whereas others have shown that history of housing in the IVC increased social activity. 48 The observed strain differences in activity and social preference were largely corresponding to the earlier comparisons. 56 Neither marble burying behavior nor nest construction was influenced by housing condition. The differences between the strains for these parameters were consistent with the previous reports. 57,58 For testing the nest building, the mice were single-housed overnight in a clean cage with food and water freely available. Single-housing can be viewed as a stressor, although the effects of this kind of manipulation have been usually monitored after several days or even weeks. One result of the acute stress can be a loss of body weight. 59 Interestingly, we found that overnight weight loss was more severe in the OCs. Another unexpected outcome was a higher basal temperature measured in mice housed in the OCs. Based on these two findings it could be speculated that acute single-housing is more stressful in OCs than in the IVC's, although further experiments are needed for validation of these effects. Nevertheless, the SIH was not different between two housing conditions. As a last procedure in the battery, we conducted the forced swim test for measuring the coping in a situation of acute stress. It appeared, that housing condition did not have any effect on the immobility time ("behavioral despair"). However, the strain ranking (D2 mice floating less than B6 or 129) was in line with the earlier findings. 19,53,60 In summary, we have characterized some basic behavioral phenotypes in three commonly used inbred mouse strains after long-term housing in individually ventilated cages and compared it to housing in OCs, otherwise maintained in the same environment. It can be concluded that under these conditions the IVC housing did not change the well-known differences between the strains. Therefore, the external validity of research findings is most likely not compromised by this change in housing conditions. However, within the strains the differences may occur, dependent on the specific tasks applied. Therefore a critical re-evaluation of the phenotypes in genetically modified mouse strains is warranted. In future studies, the caging system used needs to be explicitly mentioned in the methods.

ACKNOWLEDGMENTS
This study was supported by Jane and Aatos Erkko Foundation and Biocenter Finland. We want to thank the personnel of Laboratory Animal Center for assistance.