Significant outcomes
-
In test batteries for affective-like behaviour, healthy outbred intact mice demonstrate within test but not across tests consistency of behaviour.
-
For intact outbred mice, sex differences in tests for affective-like behaviours appear to be limited.
-
Additional work is needed to explore consistency of behaviour in test batteries after manipulations or drug treatment.
Limitations
-
The study tested only two strains of mice and only in a limited number of behavioural tests and models.
-
The study is limited to animals that were not exposed to any manipulation or drug, and therefore, mice represent intact outbred mice population.
-
Not all tests were used in all experiments, and they were not performed in the same order. This limits the possibility for comparisons across experiments.
Introduction
Exploring human psychopathology, especially affective disorders, is difficult due to many reasons, among others, the complexity of representing the mental symptoms and the underlying physiology in animal models (Flaisher-Grinberg & Einat, Reference Flaisher-Grinberg, Einat and Gould2009; Einat, Reference Einat2014; Nestler & Hyman, Reference Nestler and Hyman2010). Animal models are repeatedly stated to be critical in exploring the biology of mental disorders and the development of new treatments but, at the same time, are repeatedly criticized for major problems of validity and reproducibility (Young & Einat, Reference Young and Einat2019; Kafkafi et al., Reference Kafkafi, Agassi, Chesler, Crabbe, Crusio, Eilam, Gerlai, Golani, Gomez-Marin, Heller, Iraqi, Jaljuli, Karp, Morgan, Nicholson, Pfaff, Richter, Stark, Stiedl, Stodden, Tarantino, Tucci, Valdar, Williams, Wurbel and Benjamini2018; Belzung & Lemoine, Reference Belzung and Lemoine2011). Whereas advancement in genetic and biochemical methods can improve some aspects of validity in animal models, it is still difficult to embody the full complexity of behavioural symptoms that characterize psychopathologies, especially affective disorders in one test or model (Kara & Einat, Reference Kara and Einat2013; Einat, Reference Einat2014). One approach that was suggested to enhance the strength, validity, and reproducibility of animal models was to examine animals in test batteries rather than in single tests (Crawley & Paylor, Reference Crawley and Paylor1997; Crabbe et al., Reference Crabbe, Wahlsten and Dudek1999; Einat, Reference Einat2006). The utilization of test batteries was suggested to result in more predictive models, as they can represent different facets of the pathological conditions and provide better understanding of the complexity of a disorder or treatment effect (van der Staay et al., Reference Van Der Staay, Arndt and Nordquist2009; Flaisher-Grinberg & Einat, Reference Flaisher-Grinberg and Einat2010). This approach resembles the standard methodology in clinical studies where disorders are diagnosed based on a list of symptoms (Sadock et al., Reference Sadock, Sadock and Ruiz2015), and effects of treatments are evaluated based on comprehensive measures that include different domains of the disorder, such as the Hamilton depression scale or the Beck depression inventory (Williams, Reference Williams2001; Wang & Gorenstein, Reference Wang and Gorenstein2013). Similar approaches are also used in humans to study normal (not pathological) mood states such as with the “Positive and Negative Affect Schedule” (Watson et al., Reference Watson, Clark and Tellegen1988) or the Profile of Mood States (Pollock et al., Reference Pollock, Cho, Reker and Volavka1979). Moreover, test batteries are also a practical approach when the number of animals is limited as frequently happens when using mice with targeted mutations and is advantageous in the context of ethical use of animals in research, as it can reduce the total number of animals needed to answer scientific questions (Crawley & Paylor, Reference Crawley and Paylor1997; Crawley, Reference Crawley2000). Altogether, test batteries became a very common tool in the modelling of neuropsychiatric disorders in general and affective disorders in particular. A special emphasis must be given to the testing of intact healthy animals (without interventions), and this is for two main reasons. (1) Intact mice from standard strains are used as background strains for targeted mutations, and significant literature examined the effects of these background strains on the outcomes related to the mutations (Bailey et al., Reference Bailey, Rustay and Crawley2006). As mice with targeted mutations are essential tools in research, it is critical to understand the behaviour of the underlying healthy intact mice strains. (2) Many of the standard screening models in neuropsychopharmacology research are based on examining behaviour in intact mice. For example, tests, such as the forced swim test (FST) and the elevated plus maze (EPM), introduce animals to unique situations where their behaviour is measured with or without additional interventions. The test is based on the responses of intact animals that serve as the reference for any additional interventions. It is, therefore, of critical importance to gain in-depth understanding of the complexity of the behavioural profile of intact animals (Einat et al., Reference Einat, Ezer, Kara and Belzung2018).
Yet, the common practice in the analysis of results from test batteries is to analyse each test separately at the group level. There are a number of problems with this method that were already highlighted in previous work (Stukalin & Einat, Reference Stukalin and Einat2019; Einat et al., Reference Einat, Ezer, Kara and Belzung2018). One concern that is a component of the issue of individual variability is the expectation of consistent behavioural phenotypes of mice across tests. The hidden assumption in test batteries is that model animals will demonstrate a relatively steady phenotype across tests, and that animals that tend to be more depressed-like in one test will also be more depressed-like in another test and vice versa (Einat et al., Reference Einat, Shaldubina, Bersudskey, Belmaker, Soares and Young2007; Castro et al., Reference Castro, Diessler, Varea, Marquez, Larsen, Cordero and Sandi2012). However, the relationship between the behaviour of individual mice across tests is hardly ever assessed, and most studies only explore group effects (Einat et al., Reference Einat, Ezer, Kara and Belzung2018).
In that context, this study was designed to explore the relationship between the behaviour of individual mice across tests in some variations of test batteries for depression- and anxiety-like behaviours, in males and females of two strains of mice. We hypothesized that individual animals will demonstrate a consistent behavioural phenotype across tests that are suggested to represent the same (or similar) state. That animals showing a more depressed- and anxiety-like behaviour in one test will also do so in other relevant tests and that animals that demonstrate a more anxious behaviour in one test will also present it in other tests. Specifically, we hypothesized that there will be a correlation among the behaviours in the FST, the tail suspension test (TST), and the sweet solution preference (SSP) test, that there will be a correlation among the behaviours in the open field (OF), the EPM, and the novelty-induced hypophagia tests, that there will be a correlation between the behaviours in the OF and in the amphetamine-induced hyperactivity test, and that there could be a reverse correlation between the behaviours in the amphetamine-induced hyperactivity test and the test related to depression (FST, TST, and SSP). The aims of the study were, therefore, to (1) examine the relationship between the behaviours of ICR and black Swiss mice in a number of dissimilar test batteries for affective- and anxiety-like behaviours and (2) study sex differences in these strains and tests. It was our hypothesis that at least for some of the tests, especially ones representing similar constructs, there will be a significant correlation between the behaviours of individual mice in one test and the behaviour in the other test. We further hypothesized that sex differences will be limited to only a few tests.
Methods
Animals and procedure
ICR (CD-1®) mice (Experiments 1–3; Envigo, Israel) and black Swiss mice (Experiment 4; local colony based on mice from Taconic Labs, USA) were used in four separate experiments. We selected two outbred strains, as outbred strains are more heterogeneous compared with inbred mice and may be a better translational model for the genetically diverse human population (Tuttle et al., Reference Tuttle, Philip, Chesler and Mogil2018). ICR mice were selected because within outbred strains, they are very frequently used as model animals in relevant research related to affective and anxiety disorders (Messiha et al., Reference Messiha, Martin and Bucher1990; Sugimoto et al., Reference Sugimoto, Kajiwara, Hirano, Yamada, Tagawa, Kobayashi, Hotta and Yamada2008; Willner et al., Reference Willner, Moreau, Nielsen, Papp and Sluzewska1996; Sade et al., Reference Sade, Kara, Toker, Bersudsky, Einat and Agam2014). Black Swiss mice were selected, because they were suggested to be a model for a number of behavioural facets of mania (Hannah-Poquette et al., Reference Hannah-Poquette, Anderson, Flaisher-Grinberg, Wang, Meinerding and Einat2011; Ene et al., Reference Ene, Kara and Einat2015; Flaisher-Grinberg & Einat, Reference Flaisher-Grinberg and Einat2010; Kara et al., Reference Kara, Flaisher-Grinberg, Anderson, Agam and Einat2018a). Because of their unique behavioural profile, we expected that these mice would demonstrate consistency in behaviour across specific tests related to affective- and anxiety-like behaviours. The number of animals per experiment was based on experiments in our laboratory where behaviour of mice was examined across tests and without drugs [e.g. (Kazavchinsky et al., Reference Kazavchinsky, Dafna and Einat2019; Shemesh et al., Reference Shemesh, Kara and Einat2018)]. However, Experiment 4 was an exception with only 10 mice per group. This is because supply of black Swiss mice in Israel is very complicated, and we had to rely on our small breeding colony.
All animals were singly housed in transparent Plexiglas cages (36.5 × 14 × 20 cm) with approximately 3 cm wood shavings bedding, cardboard roles, and cotton wool enrichment. Animals were maintained under constant conditions, wherein temperature was set to 22 ± 1 °C, ad libitum access to food and water, and 12-h light/dark cycle. Animals had at least 1-week habituation before the start of experiments and were not disturbed or handled during this period and the experimental period. Single housing was needed because of the SSP test (see in the following for details). Procedures and experiments were performed during the light phase of the light/dark cycle starting 1 h after “lights on” and ending no later than 1 h before “lights off.” All experimental procedures followed the Israeli Ministry of Health directives and were approved by the Tel Aviv-Yaffo Academic College IACUC (protocols MTA-2015-15-3 and MTA-2012-10-3). For all experiments, in each battery, tests were conducted by the same two experimenters, one test per day, in a dedicated testing room, except the SSP test (see in the following for details) that included 48 h of exposure to sweet solution in the home cages.
Experiment 1
Sixty, 8–9 weeks old male ICR mice were tested in four consecutive behavioural tests: spontaneous activity in the OF; EPM; FST; and TST.
Experiments 2 and 3
Forty, 8–9 weeks old ICR mice, 20 females, and 20 males were tested in seven consecutive behavioural tests: spontaneous activity in an OF; defensive marble burying; SSP; EPM; novelty-induced hypophagia; FST, and amphetamine-induced hyperactivity. Experiment 3 was intended to be a replication of Experiment 2, but, unfortunately, due to technical error, the order of tests was altered in Experiment 3 compared with Experiment 2. As the order of tests can influence the behavioural outcomes, we could not consider these experiments as replications but as two separate experiments.
Experiment 4
Ten male and 10 female, 8–10 week old black Swiss mice were tested in three consecutive behavioural tests: spontaneous activity in the OF; EPM; and FST.
Behavioural tests
The tests used in this study are all prototypic tests that are frequently used in the research related to the exploration of the underlying basis of affective and anxiety disorders and in the attempts to develop better treatments. Although, in general, these tests are not ideal for many reasons, they are heavily used. For example, a PubMed search for “EPM” during the last 5 years yields 2550 results, and similar searches for the other tests show hundreds to thousands of results, suggesting that many researchers are using them.
Specific tests that were used in each of the experiments are detailed in Table 1. All tests were conducted as detailed elsewhere (Ene et al., Reference Ene, Kara and Einat2015, Reference Ene, Kara, Barak, Reshef Ben-Mordechai and Einat2016) with minor modifications: (1) in the marble burying test, marbles were placed on top of a 2 cm layer of wood chips bedding (and not 5 cm as described). (2) In the novelty-induced hypophagia test, animals had food restriction for 24 h and not 12–14 h as described earlier. In brief: (1) Spontaneous activity – the spontaneous activity test is a standard method to evaluate mice behaviour in the context of neuropsychiatric testing. Each mouse was individually placed in the centre of a small OF (Plexiglas box 38.5 × 38.5 cm with 35 cm walls) where behaviour was digitally recorded from above for 30 min. Recording and scoring were done automatically using a specialized software (Viewer; BioBserve Bonn, Germany). Track length (measured in cm) and time in the centre (15 × 15 cm area) were recorded and analysed. At the end of session, the mouse was returned to its home cage and the arena cleaned with alcohol solution before the start of the next session. (2) Sweet Solution Preference (SSP) – the SSP is a non-intrusive test that can model aspects of reward seeking behaviour (Willner, Reference Willner1997; Flaisher-Grinberg et al., Reference Flaisher-Grinberg, Overgaard and Einat2009). The SSP was conducted in the home cages of the mice. A bottle of 1% saccharin solution was presented in the cage for 48 h, on top of the regular water bottle and food. Bottles (saccharin solution and water) were weighed at the start of the experiment and every 24 h thereafter, resulting in three measures, time 0, time 24 h, and time 48 h. SSP was computed by dividing the amount of consumed saccharin solution by the amount of consumption of all liquids (saccharin solution + water). After 48 h, the saccharin bottles were removed. (3) EPM – the EPM is a standard model for anxiety-like behaviour (Lister, Reference Lister1987). The EPM apparatus was a black Plexiglas plus shaped maze with two enclosed arms with walls and two open arms without walls. The maze was elevated 50 cm above the ground. The size of the arms was 35 cm long and 5 cm wide, with a centre area of 5 × 5 cm. The walls of the closed arms were 15 cm high, and the open arms had a 1 cm lip. Each mouse was placed in the centre of the maze for a 5 min session and was recorded from above using the Viewer software (BioBserve Bonn, Germany). Time and frequency of visits to each arm were analysed by the software. At the end of session, each mouse was returned to its home cage and the maze cleaned with alcohol solution before the start of the next session. (4) Marble burying test – the marble burying test is associated with anxiety-like behaviour (Nicolas et al., Reference Nicolas, Kolb and Prinssen2006; Ene et al., Reference Ene, Kara, Barak, Reshef Ben-Mordechai and Einat2016). Each mouse was singly placed in a small cage with 5 cm layer of saw dust bedding and with 25 glass marbles placed in close contact in the middle of the cage. Mice were left with the marbles for a 30-min session after which they were placed back in their home cages, and the number of marbles that was more than two-thirds covered with bedding was counted. (5) Novelty induced hypophagia test – the hypophagia test examines an additional facet of anxiety or anhedonia (Deacon, Reference Deacon2011). Following 24 h with limited access to food, mice were placed individually in a plastic transparent cylinder with a 17 cm diameter and 20 cm height. Shredded sweet popcorn was spread across the cylinder’s floor. Each mouse was placed in the arena for 2 min. In case the mouse ate from the popcorn, it was left in the arena for an additional 2 min (from the first bite). In case the mouse did not eat during the initial 2-min session, it was placed back in its home cage for a few minutes and then reintroduced into the arena, again for 2 min. If the mouse did not eat during the second exposure, the procedure was repeated for a third time. The latency to eat and the total duration of the eating activity was recorded. At the end of the experiment, mice were placed in their home cage and were once again provided with free access to food. (6) FST – the FST is one of the most frequently used tests for the screening of antidepressant activity (Porsolt et al., Reference Porsolt, Bertin and Jalfre1977; Kara et al., Reference Kara, Stukalin and Einat2018b; Kazavchinsky et al., Reference Kazavchinsky, Dafna and Einat2019) Mice were placed individually in a transparent Plexiglas cylinder with a 45 cm height and a 20 cm diameter. The cylinder was filled with water at a temperature of 23–24 °C and a depth of 18 cm. Each mouse was placed in the water for a 6 min session during which behaviour was digitally recorded and scored for the duration of active (swim/struggle) and passive (floating or immobility) behaviours using a specialized software (FST, BioBserve, Bonn, Germany). Immobility time during the last 4 min of each session served as the main measure of analysis. At the end of each session, mice were taken out of the water and placed in their home cages. Water in the cylinders was replaced every three trials. (7) TST – the TST is a standard test used for screening antidepressant activity (Steru et al., Reference Steru, Chermat, Thierry and Simon1985; Kara et al., Reference Kara, Karpel, Toker, Agam, Belmaker and Einat2014) and was recently demonstrated to have relatively strong external validity (Stukalin et al., Reference Stukalin, Lan and Einat2020). Mice were suspended 50 cm above the floor with an adhesive tape placed approximately 1 cm from the tip of the tail for a 6-min session. Session was digitally recorded, and immobility time during the last 4 min of the test was scored manually from recordings. Mice were considered immobile only when they hung passively and completely motionless. (8) Amphetamine-induced hyperactivity – amphetamine-induced hyperactivity is frequently used to model manic-like behaviour (Kara et al., Reference Kara, Karpel, Toker, Agam, Belmaker and Einat2014), although its validity in that context was recently questioned (Lan & Einat, Reference Lan and Einat2019). Mice were placed in a small OF (Plexiglas box 38.5 × 38.5 cm with 35 cm walls) and their behaviour tracked as described earlier for the spontaneous activity test. Each session consisted of 60 min and amphetamine (1.0 mg/kg dissolved in saline to 10 ml/kg volume) was injected after the first 30 min, thereby allowing comparison of undragged and drugged behaviour. At the end of session, the mouse was returned to its home cage, and the arena was cleaned with alcohol solution before the next session.
Table 1. Specific behavioural tests utilised in the different experiments
Statistical analysis
Statistical analyses were performed using Statistica 13.0 software (Dell, Tulsa, OK). Pearson’s correlation (two-tailed) was used to evaluate relationship of individual responses across and within tests. When both females and males were used for an experiment, we performed an overall ANOVA for sex effect across all main behavioural measures (Stukalin & Einat, Reference Stukalin and Einat2019). Power analysis for correlations was conducted using an online calculator at https://www.masc.org.au/stats/PowerCalculator/PowerCorrelation.
Results
The main aim of the study was to examine the relationship between the behaviours of ICR and black Swiss mice in a number of dissimilar test batteries for affective- and anxiety-like behaviours for intact outbred mice (without interventions). Accordingly, the results are centred on correlations rather than group outcomes per-ce, and group outcomes are presented in Supplementary Table 1.
Experiment 1 – male ICR mice
We found correlations within tests, between the behaviours of individual mice within the OF (distance and time in centre) and between the behaviours of individual mice in the EPM (open/closed time ratio and distance, open/closed time ratio and open/closed entries ratio, and open/closed entries ratio and distance). Moreover, there were correlations across tests between the distance in the open field and distance in the EPM (Fig. 1 A), and a reverse correlation between distance in the open field and immobility in the FST (Fig. 1B). No other significant correlations were demonstrated (Table 2) including no correlation between behaviour in the FST and behaviour in the TST. Power for all correlations was at the range of 0.80–0.85. These data suggest consistency of activity levels across tests but possibly not affective-like state.
Fig. 1. (A) Correlation between distance travelled in the open field and distance travelled in the EPM in Experiment 1 with male ICR mice and (B) reverse correlation between distance travelled in the open field and immobility time in the FST in Experiment 1 with male ICR mice.
Table 2. Correlations within and across tests, Experiment 1
OF Dist = distance travelled in the open field; OF time centre = time spent in the centre area of the open field; EPM time ratio = time in open arms divided by time in closed arms of the EPM; EPM Entries ratio = # of entries to open arms divided by # of entries to closed arms of the EPM; EPM Dist = distance travelled in the EPM; FST immobility = immobility (floating) time in the FST; TST immobility = immobility (hanging) time in the TST. Bold cells indicate significant correlations.
Experiment 2
To evaluate a possible general effect of sex, we performed an overall ANOVA for sex effect across all main behavioural measures (Stukalin & Einat, Reference Stukalin and Einat2019), including (1) open field distance; (2) marbles buried; (3) SSP day 1; (4) EPM open/closed time ratio; (5) novelty-induced hypophagia (time); (6) FST immobility time; (7) spontaneous activity before amphetamine; and (8) amphetamine-induced hyperactivity. Considering that we found no statistical difference between the sexes [F (8,31) = 0.8, p = 0.6] and no differences in the homogeneity of variance (Levene’s test, data not shown), we pooled the data for females and males.
Significant correlations were found within tests in the SSP (day 1 and day 2, Fig. 2A, r = 0.71, p < 0.001) and in the amphetamine test (distance before amphetamine and distance after amphetamine, Fig. 2B, r = 0.48, p = 0.002). Interestingly, there was no within-test correlation in the EPM between the open/closed time ratio and open/closed number of entries ratio (r = 0.21, p = 0.19). Across tests, correlations (Table 3) were shown only between EPM time ratio and activity in the open field: (1) open field activity (r = 0.35, p = 0.03) and (2) activity before amphetamine (r = 0.49, p < 0.001). Power for all correlations was at the range of 0.80–0.92.
Fig. 2. (A) Correlation between SSP in day 1 and day 2 of the test in Experiment 2 with female and male ICR mice and (B) correlation between activity before and after amphetamine injection in Experiment 2 with female and male ICR mice.
Table 3. Correlations within and across tests, Experiment 2
OF Dist = distance travelled in the open field; EPM time ratio = time in open arms divided by time in closed arms of the EPM; EPM Entries ratio = # of entries to open arms divided by # of entries to closed arms of the EPM; Marbles = number of marbles in the defensive marble burying test; Hyponeophagia = latency to eat in the hyponeophagia test; FST immobility = immobility (floating) time in the FST; Amph before = distance travelled in the open field before injection of amphetamine; Amph after = distance travelled in the open field after the injection of amphetamine. Bold cells indicate significant correlations.
Experiment 3
To evaluate a possible general effect of sex, we performed an overall ANOVA for sex effect across all main behavioural measures (Stukalin & Einat, Reference Stukalin and Einat2019), including (1) open field distance; (2) marbles buried; (3) SSP day 1; (4) EPM open/closed time ratio; (5) novelty-induced hypophagia (time); (6) FST immobility time; (7) spontaneous activity before amphetamine; and (8) amphetamine-induced hyperactivity. Considering that we found no statistical difference between the sexes [F (8,31) = 1.42, p = 0.23] and no differences in the homogeneity of variance except in the SSP measure (Levene’s test, data not shown), we pooled the data for females and males.
As shown in Table 4, significant correlations were found within tests in the EPM (Fig. 3A; open/closed time ratio and open/closed number of entries ratio: r = 0.88, p < 0.001) and in the amphetamine test (Fig. 3B; distance before amphetamine and distance after amphetamine: r = 0.45, p = 0.004). A non-significant trend was also observed for the correlation in the SSP between day 1 and day 2 of the test (r = 0.27, p = 0.09). Across tests correlation (Table 4) was shown only between EPM entries ratio and immobility in the FST (r = 0.35, p = 0.03). Power for most of the correlations in this experiment was high, ranging between 0.9 and 0.99, except the correlation between EPM entries ratio and FST at 0.51, and correlation between activity before and after amphetamine at 0.52.
Table 4. Correlations within and across tests, Experiment 3
OF Dist = distance travelled in the open field; EPM time ratio = time in open arms divided by time in closed arms of the EPM; EPM Entries ratio = # of entries to open arms divided by # of entries to closed arms of the EPM; Marbles = number of marbles in the defensive marble burying test; Hyponeophagia = latency to eat in the hyponeophagia test; FST immobility = immobility (floating) time in the FST; Amph before = distance travelled in the open field before injection of amphetamine; Amph after = distance travelled in the open field after the injection of amphetamine. Bold cells indicate significant correlations.
Fig. 3. (A) Correlation between EPM open/closed time ratio and open/closed number of entries ratio in Experiment 3 with female and male ICR mice and (B) correlation between activity before and after amphetamine injection in Experiment 3 with female and male ICR mice.
Experiment 4
To evaluate a possible general effect of sex, we performed an overall ANOVA for sex effect across all main behavioural measures (Stukalin & Einat, Reference Stukalin and Einat2019), including (1) open field distance; (2) open field centre time; (3) FST immobility; and (4) EPM open/closed time ratio. Considering that we found no statistical difference between the sexes [F (4,15) = 1.15, p = 0.37] and no differences in the homogeneity of variance (Levene’s test, data not shown), we pooled the data for females and males. A significant correlation was found within the EPM test (Fig. 4; open/closed time ratio and open/closed number of entries ratio: r = 0.72, p < 0.001). No correlations were identified across tests (Table 5). Power for all correlations was at the range of 0.80–0.99.
Fig. 4. Correlation between EPM open/closed time ratio and open/closed number of entries ratio in Experiment 4 with female and male black Swiss mice.
Table 5. Correlations within and across tests, Experiment 4
OF Dist = distance travelled in the open field; OF Centre = time in the centre area of the open field; EPM time ratio = time in open arms divided by time in closed arms of the EPM; EPM Entries ratio = # of entries to open arms divided by # of entries to closed arms of the EPM; EPM total entries = number of entries to all arms of the EPM; FST immobility = immobility (floating) time in the FST. Bold cells indicate significant correlations.
Discussion
Within the attempts to develop better and more comprehensive animal models for neuropsychiatric disorders, one approach suggests the development and utilization of test batteries that will include separate tests for the different behavioural facets of the disorder that is being modelled (Bailey et al., Reference Bailey, Rustay and Crawley2006). A background hypothesis for this approach is that a good model for a neuropsychiatric disorder will respond in a similar way across dissimilar tests for the different facets of the disease (Einat, Reference Einat2014). For example, one would expect that a good model for depression would demonstrate increased immobility in the FST, increased immobility in the TST, anhedonia in the SSP test, and similar changes in other tests related to depression. Increased attention to individual variability would further suggest that there should be some relationship between the behaviours of the individual animal in the different tests (Einat et al., Reference Einat, Ezer, Kara and Belzung2018). To explore this idea, this study followed individual mice exposed to a battery of behavioural tests related to affect and anxiety. We hypothesized that individual animals will demonstrate a consistent behavioural phenotype across tests, and that animals showing a more depressed- and anxiety-like behaviour in one test will also do so in other relevant tests. In contrast to our expectations, such consistent phenotypes were not demonstrated, as there were hardly any correlations between the behaviours of individual mice across tests, and even these few correlations can easily be attributed to statistical error due to multiple comparisons (Stukalin & Einat, Reference Stukalin and Einat2019). No clear phenotypes is demonstrated in any of the strains (ICR and black Swiss), in both females and males, and in a number of variations in batteries. The only consistent correlations that were identified were within tests in some of the tests that include more than one measure.
The current findings, therefore, appear to cast some doubt on the benefits of using test batteries. Moreover, one cannot but wonder if these findings raise even broader questions regarding the validity of the tests themselves to model components of the disorders. The lack of correlation between the behaviour of individual mice across tests that are proposed to model different but close facets of a disorder presents a significant problem in the value of these tests. For example, both the FST and the TST are very similar screening tests for antidepressant activity and are based on a similar rationale related to the development of despair in depressed patients (Porsolt et al., Reference Porsolt, Bertin and Jalfre1978; Steru et al., Reference Steru, Chermat, Thierry and Simon1985). Yet, no correlation between these behaviours in the two tests was found (Table 2). Interestingly, two recent meta-analysis studies, one regarding the FST (Kara et al., Reference Kara, Stukalin and Einat2018b) and the other regarding the TST (Stukalin et al., Reference Stukalin, Lan and Einat2020), may indicate that these tests are very different. The FST was found to have validity in qualitatively predicting antidepressant-like effects, whereas the TST was found to also quantitatively show effects of dose across different experiments. Similarly to expected correlation between the behaviour in the FST and TST, both the measure of “time in the centre” of the open field and “open/closed time” in the EPM are considered to reflect the balance between anxiety and exploratory drive (Cryan & Holmes, Reference Cryan and Holmes2005), but no correlation was shown between them (Tables 2 and 5). It is, however, possible that the size of the arena used as open field was not large enough to detect anxiety-like behaviour, and testing in larger arenas would have resulted in different outcomes.
Unlike the results across tests, the results within tests suggest that animals do show consistent behaviour when repeatedly presented with the same task or two related tasks within the same test. In that way, there was a significant correlation between SSP in day 1 and day 2 of the tests in both Experiment 2 (r = 0.71, p < 0.001) and Experiment 3 (r = 0.37, p = 0.02). Similarly, there was a significant correlation between the “time ratio” measure and the “distance” measure in the EPM (Table 2) or between the “time ratio” measure and the “number of entries ratio” measure in the EPM (Tables 2–5). Finally, there was a significant correlation between activity levels within one session in the open field before and after an amphetamine injection (Tables 3 and 4). These results suggest that for each separate test, it is possible to expect consistent response. In that context, a recent study demonstrated consistency in the behaviour of mice across repeated exposures to the FST (Kazavchinsky et al., Reference Kazavchinsky, Dafna and Einat2019).
Whereas the results regarding the behaviour in batteries may appear to be somewhat disappointing, it is important to remember that the mice tested in the current study were intact animals that did not experience any manipulation or treatment before testing. It is, therefore, possible that when we expose model animals to external interventions before testing, there would be a stronger correlation between the results in the different tests. It is, therefore, possible that if mice are exposed to manipulations that are related to the induction of depression-like behaviour such as stress, the mice that are more susceptible to the intervention will show consistent depression-like behaviour across tests, whereas the opposite will occur in animals that are more resilient. Similarly, drug treatment may also increase consistency of individual results across tests, as the major factor will be the response of the individual animals to the drug. Indeed, in a pilot study, we conducted with ICR mice, testing the individual variability of response to chronic oral lithium treatment, animals were exposed to the open field and the SSP tests and showed significant correlation between the individual behaviour in the two tests where mice that were more active in the open field showed higher reward seeking behaviour in the SSP (r = 0.4, p = 0.03). Similarly, in a recent published paper from our group, we examined the effects of a number of light and photoperiod manipulations on behaviours related to depression and anxiety in the diurnal fat sand rats (Psammomys obesus) (Bilu et al., Reference Bilu, Einat, Tal-Krivisky, Mizrahi, Vishnevskia-Dai, Agam and Kronfeld-Schor2019). In that study, we report the effects of interventions of behaviours at the group level but re-analysing the data at the individual levels show that after interventions, there are correlations both within tests and across tests. Within tests, there is a correlation between the first and second “time to sink” measures in the modified FST (r = 0.64, p < 0.001) and between EPM “open time” and “open/closed time ratio” measures (r = 0.87, p < 0.001). Moreover, across tests, there is a correlation between “time to sink” in the FST and “open time” in the EPM (r = 0.37, p = 0.01), suggesting that the manipulations resulted in the development of individual phenotypes in the sand rats. It is, therefore, possible to suggest that although, in intact animals, the individual behaviour across tests does not reflect a coherent disease-relevant phenotype, when animals are exposed to interventions, the individual properties of the animal are exposed, either in the context of susceptibility and resilience to pathological-like interventions or in the context of response to drugs. Because the main objectives of using animal models are in the context of exploring underlying pathology or response to treatment, it is highly important at this point to design the appropriate studies that will directly evaluate the behaviour of animals in test batteries after some of the commonly used manipulations.
In summary, the current findings demonstrate within test but not across tests constancy of behaviour in intact, healthy mice, and additional work is now suggested to further explore the utility of test batteries in modelling affective- and anxiety-like behaviours.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/neu.2020.20
Authors’ contribution
All authors were involved in the conception and design of the study. LK and SD performed the experiments under the supervision of HE. All authors took part in the analysis of the data. LK and SD wrote together the initial draft of the MS. HE wrote the final for submission.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflict of interest statement
None.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional guides on the care and use of laboratory animals.
