Evidence-based severity assessment of the forced swim test in the rat

Abstract

The forced swim test (FST) is a traditional assay, which has been used for more than 40 years to assess antidepressant effects of novel drug candidates. In recent years, a debate about the test has focused on the assumption that the FST is highly aversive and burdening for the animals because of the earlier anthropomorphic interpretation and designation as a "behavioral despair test". The Directive 2010/63/EU and the German Animal Welfare law require a prospective severity classification of the planned experimental procedures. Still, an objective examination of the animals' burden in this test has not been performed yet. To fill this gap, we conducted an evidence-based severity assessment of the forced swim test in rats according to a 'standard protocol' with a water temperature of 25°C. We examined parameters representing the physiological and the affective state, and natural as well as locomotion-associated behaviors in three separate experiments to reflect as many dimensions as possible of the animal's condition in the test. Hypothermia was the only effect observed in all animals exposed to the FST when using this standard protocol. Additional adverse effects on body weight, food consumption, and fecal corticosterone metabolite concentrations occurred in response to administration of the antidepressant imipramine, which is frequently used as positive control when testing for antidepressant effects of new substances. We conclude that this version of the FST itself is less severe for the animals than assumed, and we suggest a severity classification of 'moderate' because of the acute and short-lasting effects of hypothermia. To refine the FST according to the 3Rs, we encourage confirming the predictive validity in warmer water temperatures to allow the rats to maintain physiological body temperature.

Fig 1

(A-D): Experimental design. All three experiments were conducted in male and female animals. (A) Timeline of the FCM experiment. We used n = 8/group in the first male cohort and n = 10/group in the second male cohort. The first female cohort consisted of n = 10/group and the second n = 8/group. CC = Cage change training. (B) Sampling scheme in the FCM experiment. The detailed chronological sequence of the collection intervals is shown exemplarily for the days 19 to 22. The ‘acute response’ interval went from 10 h to 13.5 h after the FST, leading to collection times from 7 pm to 10:30 pm. The ‘delayed response’ interval was from 10:30 pm to 9 am the next day, i.e., 13.5–24 h after the FST. ‘Recovery response’ samples were timed 28–34 h after an FST session, therefore between 1 pm and 7 pm (C) Timeline of the Burrowing experiment. The group size was n = 12 for both sexes. Because of poor burrowing behavior in male animals, only data for female rats are available. BURROW = burrowing, IBT = inner body temperature. (D) Experimental timeline of the homecage parameters and the open field. We used n = 12/group in both sexes. LINT = latency until interaction with nest material test, SPT = saccharin preference test, NEST = nest score.

Fig 2

(A-D) Physiological parameters. (A) Body weight change. The percentage change in body weight from the previous day during the week of the FST is shown as a boxplot, separated by sex. Animals of the FST-Imipramine group show a mild but significant loss of body weight after the FST and Imipramine injections. Data from all cohorts were considered. Sample size of male animals n = 30 for all treatment groups. Sample size of female animals was n = 42 for each treatment group. (B) Food intake. The mean and standard error of the daily food consumption per 100 g body weight is displayed separately for male and female rats. Rats treated with Imipramine showed a significantly reduced food consumption on the days of the injections. The intake of the FST and FST-Sal groups did not change. Sample size of male animals: n = 5 cages per treatment group. Sample size of female animals: n = 8–20 cages per treatment group. (C) Inner body temperature. Boxplots showing the internal body temperature measured directly after the forced swim test/the handling procedure on both FST days. The forced swim test led to a significant drop in temperature. Only data for female animals are available. Sample size: n = 12 per treatment group. (D) Recovery to physiological body temperature. Already after 60 min the female rats could regain their normal body temperature. Sample size: n = 5.

Fig 3

(A-B) Fecal corticosterone metabolite concentrations are shown as boxplots. (A) FCM response in male animals. The panels show the FCM concentrations of all tested days separated by the collection intervals ‘acute response’, ‘delayed response’, and ‘recovery response’. Animals of the FST-Imi group displayed higher concentrations at the time points FST1, FST2, and P1 in the ‘acute response’ interval and at FST1 and FST2 during the ‘recovery response’. No other treatment group showed significantly elevated stress hormone metabolite levels. The sample size was n = 17-18/group. (B) FCM response in female animals. The FCM concentration in female animals showed substantial cohort differences. Therefore, the panels are separated by cohort and collection interval. In the first cohort (FCM I), we could detect elevated FCM levels after the second FST in the FST-Imi group during the ‘recovery response’. In cohort two (FCM II), animals displayed much higher FCM concentrations, especially during the baseline measurement. In the ‘acute response’ interval, the Control-group showed higher FCM values than the FST-Imi group at P2, two days after the FST. No significant time:treatment interactions could be detected in this cohort. The sample size in the first cohort was n = 10/group, and in the second, n = 8/group.

Fig 4

(A-D) Species-specific behaviors. (A) Saccharine preference test results over 24 h are shown as a boxplot. No significant time:treatment interaction was observed in the 24 h saccharin preference test. N = 6 cages per treatment group. (B) Burrowed gravel in the burrowing assay. The mean amount of burrowed gravel (± SEM) is depicted from the last training day (D15). We did not find significant treatment effects or time:treatment interactions. Sample sizes were: Control n = 8, FST n = 6, FST-Imi n = 9, FST-Sal n = 8. (C) Latency until interaction with nest material test displayed as a boxplot. No effects of the FST were observed. The sample size for all groups was n = 12. (D) Nest Score. Baseline nest scores on days 13 and 14 are shown together with the scores on the mornings after the two FST days (day 20 and day 21) as boxplots. The sample size for all groups was n = 6 (cage-based parameter) for each sex.

Fig 5

(A-D) Open field. All parameters are shown as boxplots and separately for female and male animals. Results from the baseline week (day 15) and after the FST (day 22) are displayed next to each other. We did not see effects in any of the open field parameters. The group size was n = 12 per sex. (A) Total distance moved over the time course of 30 min. (B) Center time. (C) Relative movement. (D) Velocity.