Evaluation of effects of previous exposure to an acute stressor before testing for depression-like behaviours in mice

Abstract

Test batteries are an essential and broadly used tool for behavioural phenotyping, especially with regard to mouse models of particular diseases, such as depression. Facing the problem of an often limited number of mutant animals, it therefore seems crucial to develop and optimise such test batteries in terms of an ideal throughput of subjects. This study aimed to characterize several common stressors, which are used for the investigation of depressive-like features with regard to their capability of each of them to affect performance in a subsequent behavioural test. Here we investigated swim-, restraint- and footshock-stress in male C57/BL6 mice, focusing on post-stress corticosterone elevations as well as potential effects on the behavioural level.

The stressors increased circulating corticosterone levels when assessed 1 h after exposure. On the behavioural level, no test interactions could be detected, which suggests, that in general, combining these test conditions in experiments with a restricted availability of animals seems to be rather unproblematic.

• Stress
• test battery
• mice
• behaviour
• corticosterone

Introduction

In the age of genetic engineering, the analysis of behavioural phenotypes increasingly utilizes test batteries consisting of several paradigms. The reason for this is that the generation of conditional mutants requires complex breeding strategies. Limited colony sizes make mutant animals extremely valuable and increase the need to analyse each individual animal as extensively as possible. Thus, a single mouse is often exposed to a whole series of tests, which may be stressful, depending on the research focus. Therefore the results obtained in subsequent tests may be influenced by or even merely reflect stressful situations experienced during previous tests. Even though the suggestion to perform tests from least to most stressful is generally accepted and applied (van Gaalen and Steckler 2000; McIlwain et al. 2001), it is not trivial to determine the stressor intensity and rank the employed paradigms accordingly. Garcia et al. (2000) studied in rats the stress hormonal aspects of the effects of stressor intensity, stressor duration and previous stressor experience on the hypothalamic–pituitary–adrenal (HPA) axis activation, and showed that the more intense and longer the stress the greater the release of adrenocorticotropin (ACTH) and corticosterone. Surprisingly though, the recovery duration is not linked to the stressor duration, but, if the animal is repeatedly exposed to a stressor, the HPA-system recovery is accelerated, although the initial response is not blunted (Dal-Zotto et al. 2000; Garcia et al. 2000; Marti et al. 2001). Mercier et al. (2003) additionally analysed the effects of the most commonly applied stressors in rats (restraint stress, footshock stress and forced swim stress) in two activity-dependant paradigms after determining the stress reaction intensity by measuring serum corticosterone levels and body core temperature. The results were somewhat ambiguous, since the light extinction test, which was performed 24 h post-stress, revealed an altered activity level according to the previously experienced stressor intensity, while no differences were detectable in the locomotion patterns 48 h after stress exposure in the open-field paradigm (Mercier et al. 2003). In contrast, a decrease in locomotion has been reported to occur after footshock stress, which only became detectable 1 day after stressor exposure but stably lasted up to 14 days in male rats (van Dijken et al. 1992). In addition, the same study revealed long-lasting changes in the CNS, resulting in increased anxiety and an enhanced responsiveness of the HPA-axis to a novel stressor.

Although there are ample studies of stress effects at the hormonal, morphological and behavioural level, most have been published in rats (e.g. Akers et al. 2006; Thomas et al. 2007). However, since most transgenic experiments are performed in mice, it is essential to investigate stress-induced biological disturbances and the associated short-term behavioural reactions in this species as well, also because mice often react quite differently to stress (Jacobson and Cryan 2007), even with large inter-strain discrepancies. This is especially important when employing behavioural test batteries, which include acute stressors (often implicated in the test situation itself) that are: (i) advocated as animal models of affective disorders in man and (ii) could potentially modulate the outcome of the subsequent tests when performed in sequence. The purpose of the present study is therefore to establish an appropriate sequence within such behavioural test batteries (e.g. for phenotyping of often restricted mutant colony sizes), how to combine stressful test conditions (especially with regard to depression as a research focus) and to minimize the interaction between the most commonly utilized stress-incorporating depression-linked paradigms in mice. These are (1) swim stress (as used in the Porsolt forced swim test), (2) footshock stress (as in the learned helplessness paradigm) and (3) restraint stress (a widely used challenge of the HPA-system). In our study we performed all experiments in C57/BL6 mice, the most commonly used background strain with regard to mutations, and allowed inter-test intervals of approximately 24 h. In a first series of experiments, we examined the serum corticosterone levels 60 min after the animals were exposed to the respective stressor mainly as indicator of the occurrence and the intensity of the stress reaction. In a second set of experiments we investigated the effects on the behavioural level, thereby studying possible stressor interactions, with the final goal to determine the optimal sequence for a depression-linked murine test battery without evoking masking artifacts.

Materials and methods

Animals

For all experiments 12-week old male C57/BL6(N) mice were used. The mice were individually housed in conventional macrolon cages (Type II) with tissue nesting material and acclimatized for 2 weeks to a reversed 12-hour dark–light cycle (lights on: 18:00–06:00 h). Each mouse was initially exposed to one specific stressor, followed by one of the applied behavioural tests. Stressed groups (n = 8/group) were matched and subsequently compared with respective unstressed controls (n = 8/group). Mice received a standard pellet diet and water ad libitum. All animal procedures were approved by the German animal welfare office of the Regierungspräsidium Karlsruhe, Germany.

Experimental design

To test the effects of the respective tests on physiological and behavioural parameters, we divided our study in single coordinated experiments and arranged the conditions as shown in the design plan below:

Table

	Stressor	Read-out paradigm
Experiment 1	(a) Swim stress	Corticosterone (1 h post-stress)
	(b) Footshock stress
	(c) Restraint stress (HPA-axis challenge)
Experiment 2	(a) Swim stress	Open-field (basal and 24 h post-stress)
	(b) Footshock stress
	(c) Restraint stress
Experiment 3	(a) Swim stress	Two-way avoidance (24 h post-stress)
	(b) Restraint stress
Experiment 4	(a) Footshock stress	Forced swim test (24 h post-stress)
	(b) Restraint stress

Applied stressors

All stressors were applied during the dark phase, i.e. the activity period of the animals.

Swim stress

The mice were placed into a glass cylinder (23 cm high, 13 cm diameter) filled with water (22°C) to a depth of 10 cm, for 6 min.

Footshock stress

For footshock stress the mice were exposed to a transparent plexiglas shock chamber (18 × 18 × 30 cm), equipped with a stainless steel grid floor (Coulborn precision regulated animal shocker, Coulborn Instruments, Düsseldorf, Germany), through which they received 360 footshocks (0.150 mA) on 2 consecutive days. The footshocks applied were unpredictable with varying shock duration (1–3 s) and interval episodes (1–15 s), amounting to a total session duration of approximately 52 min. This shock protocol was used since it induces a state of learned helplessness in the two-way avoidance paradigm (Chourbaji et al. 2005).

Restraint stress

Mice were restrained for a period of 30 min in a plastic tube (length: 12 cm; diameter: 2.85 cm) with 8 holes (0.6 cm diameter) to ensure free respiration, following a standard procedure for the challenge of the HPA system (Ridder et al. 2005).

Behavioural tests

All behavioural testing was performed with illumination during the dark phase of the activity cycle.

Open-field

Activity was monitored in a square, white open-field apparatus, measuring 50 × 50 cm² and illuminated from above by 25 Lux. The mice were placed individually into the arena and monitored for 10 min by a video camera (Sony CCD IRIS). The data were analysed using an image processing system (EthoVision 2.3, Noldus Information Technology, Wageningen, Netherlands). For each sample, the system recorded position, object area and the status of defined events. Parameters assessed for the present study were total distance moved, velocity and time in centre, which was defined as the area 10 cm from the walls. Open-field activity was measured 24 h after stress exposure.

Forced swim test

The mice were placed into a glass cylinder (23 cm high, 13 cm diameter) filled with water (22°C) to a depth of 10 cm, for 6 min. The onset and the percentage of time spent immobile were recorded. Immobility was defined as motionless floating in the water, with only movements necessary for the mouse to keep its head above the water. In contrast, swimming was defined as time spent with active escape or struggling movements (Porsolt 2000).

Two-way avoidance

Twenty-four hours after the respective stress procedure, changes in response-behaviour were assessed by testing shuttle box performance in the two-way avoidance paradigm (Graphic State Notation, Coulborn Instruments, Düsseldorf, Germany). The shuttle box consisted of two equal-sized compartments (18 × 18 × 30 cm) separated by a gate (6 cm wide, 7 cm high). The shuttle box also contained a grid floor, through which current was applied, and a signalling light at the top of both compartments. Spontaneous initial shuttles from one compartment to the other were counted during the first 2 min by infrared-light beams at the bottom of each of the two divisions. Performance was analysed according to the behaviour during 30 shuttle escape trials. Each trial started with a light stimulus of 5 s, announcing a subsequent footshock (intensity: 0.150 mA) of maximum 10 s duration. The inter-trial interval was 30 s. Behavioural responses were defined as follows: avoidance as adequate reaction to the light stimulus by moving to the other compartment immediately, escape as shuttling to the other section in reaction to the electric shock and failure, when no attempt to escape was made. Escape latency was measured as the time taken to escape. Shuttles in between the trials were recorded to determine inter-trial activity. Total time of testing for helplessness was about 20 min, the exact time period depending on the animal's ability to learn the paradigm.

Serum corticosterone determination

For corticosterone analysis, trunk blood was obtained by conscious decapitation within 30 s of removing the animal from the cage. Anaesthesia was not used to avoid effects on corticosterone secretion. Serum corticosterone concentrations were determined by radioimmunoassay (ICN Biomedicals, Eschwege, Germany); assay sensitivity was 7.7 ng/ml, and intra-assay coefficient of variation was 5.0%. Mice were killed 1 h after the termination of stress (n = 8 for each stress condition, n = 8 controls without stress exposure).

Statistical analysis

Statistical analysis was carried out using the XLstat statistics program. One-way ANOVA was performed for the comparison between the stressor and control groups. Results are reported as means ± SEM (standard error of the mean). Significance was evaluated at a probability of 5% or less ( < 0.05).

Results

Experiment 1: Serum corticosterone levels are affected by restraint and swim stress

At 60 min post-stress, all types of stress—footshock, restraint and swim stress—increased the serum corticosterone concentration in the stressed mice in comparison to their unstressed controls; the effect in footshocked mice almost reached statistical significance (Figure 1).

Figure 1 Serum corticosterone concentrations 60 min after stressor exposure. Controls were not exposed to a stressor. Footshock stress showed the weakest effect (p = 0.056), while restraint and swim stress significantly increased corticosterone concentration (restraint stress: **p = 0.0001; swim stress: ***p = 0.002). n = 8 mice per group.

Experiment 2: Open-field activity is not influenced by any stressor used

General locomotion, i.e. total distance moved and velocity as well as centre time, as measured in the open-field paradigm on the day after stress exposure, did not differ significantly between the different stressor groups, neither before nor after stress (Figure 2).

Figure 2 Open-field activity. Locomotion, indicated as total distance moved in the open-field, was not changed significantly after exposure on the previous day to any of the stressors, i.e. footshock, restraint or swim stress. Columns represent means ± SEM. n = 8 mice per group.

Experiment 3: Two-way avoidance behaviour (helplessness) is not changed by restraint or swim stress

In the 2-way avoidance test, no significant differences were detectable between the animals 24 h after being exposed to restraint stress or swim stress and their respective controls (Figure 3). Interestingly, none of the stressed groups reached the criterion for “helpless”, i.e. more than 6 failures/30 shuttle escape trials and an escape latency lasting longer than 4.75 s, as previously defined (Chourbaji et al. 2005).

Figure 3 Two-way avoidance behaviour. Number of failures (A) and escape latency (B), the main indicators for the definition of helpless behaviour, were not significantly altered, neither by swim nor restraint stress. Columns represent means ± SEM. n = 8 mice per group.

Experiment 4: Neither restraint nor footshock affects the forced swim test

When the mice were observed in the forced swim test the day after being exposed to the footshock stress procedure or restraint stress, no significant differences were detectable between the groups in the two most relevant parameters of the test, latency to start floating and total immobility time (Figure 4). Merely the animals that previously experienced restraint stress displayed a tendency to a decreased latency to start floating.

Figure 4 Porsolt forced swim test. Latency to start floating (A) and total immobility (B), were not significantly influenced by either footshock or restraint stress on the previous day (p = 0.063). Columns represent means ± SEM. n = 8 mice per group.

Discussion

Working with animals that are exposed to different experimental conditions requires knowledge about possible interactions of the employed tests. Referring to the findings of earlier studies of the effects of stress (van Dijken et al. 1992; Dal-Zotto et al. 2000; Mercier et al. 2003), mainly in rats, it seemed essential to further characterize in mice frequently employed stressors (restraint, footshock and swim stress), which are postulated to influence for example fear conditioning in rats (Shors et al. 1992; Conrad et al. 1999; Rau et al. 2005), since they are also commonly used for the analysis of depressive-like behavioural features in mice (Willner 1991; Cryan and Mombereau 2004; Ridder et al. 2005). The C57/BL6 strain is one of the most frequently used background strains for genetic engineering (while genetic manipulation of rats is still rare) and it is used in many studies of depression (Wellman et al. 2007). Therefore our experimental design comprised male 12 week old adult mice of this strain, but it should be noted that other strains of mice may react differently to the battery of stressors and behavioural paradigms we used.

Our standard test battery to examine basal behaviour includes the assessment of locomotion, exploration and anxiety, in which the order of the tests is arranged from the least to the most stressful as suggested by others (van Gaalen and Steckler 2000; McIlwain et al. 2001), followed by the analysis of depressive-like characteristics implying particular types of defined stress (often produced by the testing itself). Whether the same animal may be exposed to several of these stressful tests without producing artefacts in subsequent experiments has so far been unknown, but this information is essential for the interpretation of results, especially in mutant animals, in which behavioural as well as morphological effects may be induced by the genetic manipulation or possibly the stress history of the animal.

All the applied stressors in this study led to corticosterone release as examined in Experiment 1. Swim stress showed the greatest increase, while footshock stress just failed to stimulate statistically significant release of corticosterone. The difference in the post-stress corticosterone increases following the different stressors (1 h after termination of the respective stress) may be due to the different amount of time that elapsed since stress onset. The shortest interval from stressor onset to blood collection was for swim stress (duration 6 min) and this showed the strongest effect, whereas the footshock stress protocol had the longest duration (52 min), and showed the weakest effect. Furthermore, mice were exposed to the footshock protocol twice, with a 24 h interval, before blood collection; these mice might have shown habituation. Nonetheless, with regard to the focus of phenotyping studies, the differences in the measured corticosterone responses among the stressors should be considered, especially since these stressors did not modulate performance in the behavioural paradigms, i.e. Two-way avoidance, forced swim test and open-field. Since forced swimming evoked the largest increase in corticosterone secretion, this seems to be a suitable acute stressor to embed in a chronic (mild) stress paradigm as described by Willner (1997). At the behavioural level, however, none of the three stressors investigated significantly affected behaviour in any of the subsequent tests, suggesting a practicable sequential combination of these paradigms within a testing battery.

In Experiment 2, we did not detect any effects on general locomotor activity 24 h after stress exposure (i.e. total distance moved), nor regarding the time the animals spent in the centre of the arena. This is important, since changes of locomotor activity could also influence the behaviour in other tests, e.g. in Two-way avoidance, in which changes of activity may produce artefacts in the shuttle box paradigm, or in the forced swimming test. These results also suggest that the open-field test as described here is robust and not affected by acute stressor exposure 24 h before the test. It remains to be shown whether mutant strains show the same lack of effect of previous stressor exposure in this behavioural test.

As assessed by Experiment 3, Two-way avoidance was not influenced by either swim- or restraint stress, which underlines the specificity of stressor exposure necessary to provoke learned helplessness in mice in our previously published experimental design (Chourbaji et al. 2005). We found previously that helpless behaviour is solely induced by exposure to 2 days of unavoidable, unpredictable and uncontrollable footshock stress while the same amount of footshock stress applied in a predictable form did not produce helplessness. The present results provide information about other stressors and re-emphasize that learned helplessness is a specific cognitive depression paradigm and not merely a reflection of exposure to any type of stressor.

In the forced swim test (Experiment 4), we did not detect any modulation by restraint or footshock stress. Other studies, however, describe a sensitisation of male NMRI mice to this test situation induced by social defeat, to which the animals were exposed 9 days earlier (Keeney et al. 2006). Nevertheless, due to the strong elevation in corticosterone evoked by swim stress, we rate this type of acute stressor as the strongest among those applied in this study, and therefore propose use of the forced swim test as the last one in a range of paradigms included in a test battery comparable to ours, especially if corticosterone action is of importance.

In summary, considering the several different findings that are described in rats, e.g. an increase in corticosterone secretion after restraint stress (Vollmayr et al. 2004) or changes in activity produced by footshocks (van Dijken et al. 1992), it seemed crucial to evaluate these types of stressors also in mice, especially in the C57/BL strain representing the major background strain for targeted mutagenesis. The present study adds novel information about the effects of specific acute stressors in mice on subsequent behavioural performance in established tests to assess depression.

Acknowledgements

This work was supported by a grant to P.G. from the Deutsche Forschungsgemeinschaft (SFB636-B3). M.A.V. had a scholarship from the GK791, University of Heidelberg).