Social housing in male mice - impact on experimental anxiety-related behaviour?

Several rodent models for anxiety disorders have been established to investigate the underlying mechanisms of normal and pathological types of anxiety. Inappropriate social housing conditions, which might be considered as chronically stressful for the animals, will affect the reliability of these models, since stress is known to influence emotionality fundamentally. However, social housing conditions still are a matter of controversial discussion. While the literature provides some indication that the animals’ response to social housing is especially pronounced in male mice due to their territorial behaviour, resulting in increased interindividual aggression, systematic research on the effects of social housing conditions on experimental behaviour is lacking. Here we present data from two studies in male inbred mice, frequently used to model anxiety disorders. We compared the behaviour and peripheral stress hormone levels of single housed individuals with that of animals housed in social groups. All animals were housed under distinct social housing conditions for at least two weeks. Subsequently, the animals were behaviourally tested in the modified hole board test for 5 minutes. This test has been shown to reliably monitor a variety of motivational systems in rodents, such as anxiety-related behaviour, locomotor activity, exploration and social affinity. In the first experiment, male C57BL/6 mice were housed either singly or socially in groups of 2, 3, 4 or 5 individuals per cage. Socially housed individuals were tested in the presence of their group-mates throughout. In singly housed mice we found no behavioural signs of isolation stress (e.g. latency until first board entry: one-way ANOVA, p = 0.674). Notably, basal corticosterone levels did not differ between groups (one-way ANOVA, p = 0.555). After exposition to the modified hole board test however, animals housed in groups of 3 – 5 individuals showed an increased corticosterone-response when compared to singly housed individuals (Student’s t-est; p < 0.0125). The results showed that housing in groups of 4 or 5 individuals increased the within-group variability for corticosterone levels when compared to the situation in singly- or group-housed (2 or 3 individuals) mice (Bartlett’s test, p< 0.05). Since the order of behavioural testing might have caused this variability we excluded this parameter in a, otherwise similar, second experiment by behaviourally testing only one animal per social group. Based on the results of the first experiment, groups of 3 individuals were compared with individually housed males of the strains C57BL/6, BALB/c and A. In singly housed mice we found neither behavioural (e.g. latency until first board entry: C57BL/6: Student’s t-test, p = 0.829; BALB/c: Student’s t-test: p = 0.393; A: Wilcoxon-Mann-Whitney-Test, p = 0.345; time on board: C57BL/6: Student’s t-test, p = 0.171; BALB/c: Wilcoxon-Mann-Whitney, p = 0.345; A: Student’s t-test, p = 0.414) nor hormonal signs of isolation stress (corticosterone-response: two-way ANOVA: housing condition effect, p = 0.157). In conclusion, individual housing appears to be non-stressful for male mice and might even reduce possible negative effects of group housing, such as increased within-group variability of experimental data