Psychopharmacology
Psychopharrnacology (1984) 84:48 - 53
9 Springer-Verlag 1984
The staircase test in mice: A simple and efficient procedure for primary screening of anxiolytic agents J. Simiand, P. E. Keane, and M. Morre Sanofi Recherche, 195 Route d'Espagne, F-31035 Toulouse Cedex, France Abstract. The staircase test consists of placing a naive mouse in an enclosed staircase with five steps and observing the number of steps climbed and rearings made in a 3-min period. All the clinically active anxiolytics tested (chlordiazepoxide, clorazepate, diazepam, lorazepam, meprobamate, phenobarbital) reduce rearing at doses which did not reduce the number of steps climbed. The majority of non-anxiolytic substances tested (haloperidol, chlorpromazine, imipramine, amitriptyline, amphetamine, morphine and carbamazepine) produced a parallel reduction of both behavioural variables. Ethosuximide had no effect on behaviour. The anticonvulsant sodium valproate produced an anxiolytic profile in this test, since it reduced rearing, while increasing step climbing. This result confirms the anxiolytic properties of valproate observed in other behavioural models. Our results indicate that the staircase test in mice is simple, rapid and selective for anxiolytics. The test is well suited for use as a primary screening method. Key words: Rearing - Exploration - Anxiolytics - Benzodiazepines - Neuroleptics - Antidepressants Antiepileptics - Mouse - Screening
The recent discovery of benzodiazepine receptors in the brain (Mohler and Okada 1977; Squires and Braestrup 1977), potential ligands, partial agonists and antagonists (for review, see Usdin et al. 1983) has stimulated the search for anxiolytic drugs which may be more specific than the benzodiazepines (Iversen 1980; Kolata 1982). Ideally, the primary screening of new anxiolytic drugs requires tests which are selective, simple, rapid, economical and reproducible. Unfortunately, the most sensitive and specific methods for determining anxiolytic activity, such as the conflict tests (Geller and Seifter 1960; Vogel et al. 1971), the antagonism of pentylenetetrazol discrimination (Sherman and Lal 1979) or social interaction (File 1980) are poorly adapted for screening purposes. Their exploitation is extremely time consuming, either for animal training procedures (conflict and discrimination) or for the observation of multiple behavioural patterns (social interaction). Furthermore, all these tests use the rat, which is a relatively expensive species for screening studies. For mice, the number of anxiolytic screening tests is limited. Most of them investigate exploratory and locomotor behaviour, with or without punishment (Boissier et al. 1968;
Offprint requests to: J. Simiand
Crawley 1981), or consummatory behaviour after deprivation (Stephens 1973). In tests using locomotor behaviour, a generalized motor effect must be distinguished from an anxiolytic effect (increase in exploratory activity) by a separate motor test (Aron et al. 1971; Crawley 1981). In tests relying upon punishment or food and water deprivation, false positive results may be obtained for drugs which alter processes such as pain threshold or motivation without possessing anxiolytic properties. Since the benzodiazepines in particular increase food and water intake (for review, see Dantzer 1977), there is a need for tests which do not use factors such as electric shock or food deprivation (see File 1980, for discussion). Therefore, it would be important to develop such a test in order to allow a routine, rapid screening of anxiolyfic activity in mice. For this reason, we have investigated the possible adaptation of the staircase test described in rats by Thi6bot et al. (1973, 1976). These authors showed that when a rat is placed for the first time in an enclosed staircase, the anxiolytics specifically reduced rearing activity without reducing (indeed, even increasing) the number of steps climbed. Thi~bot et al. (1973, 1976) suggested that, under these conditions, rearing could be used as an index of anxiety, while the climbing of steps was an index of exploratory and locomotor behaviour. We have modified this test for use in mice, and the results presented here indicate that it is a valuable method for the primary screening of anxiolytics.
Materials and methods
Apparatus The apparatus consists of a white PVC enclosure with a fivestep staircase (Fig. 1). The box is placed in a room with constant lighting, isolated from external noise, and thermostatically controlled (21 + 2 ~C).
Experimental protocol Subjects. Naive male mice (CD1, Charles River) weighing 21 _+ 3 g were used in these studies. The animals where housed in groups of 50 in standard conditions of temperature, lighting and ventilation, with food and water freely available. The mice were kept at least 4 days after their arrival in the laboratory before being used. The day before the test, the animals where randomly divided into groups of 12 mice in plastic cages. All the animals for a single experiment where placed at the same height in the animal house. They were transferred to the laboratory at least 1 h before the start of the test. Each animal was used only once.
49 -I-
I
I
1
__t_
~- 10 cm i f " ( Fig. 1. Experimental apparatus. The staircase is composed of five identical steps 2.5 cm high, 10 cm wide, and 7.5 cm deep. The internal height of the walls was constant along the whole length of the staircase. This modification of the model used for rats by Thi6bot et al. (1973) leads to an equal distribution of rearings at all levels of the staircase
Procedure. The animal was placed singly on the floor of the box with its back to the staircase. The number of steps climbed and the number of rears were counted over a 3-rain period. A step was considered to be climbed only if the mouse had placed all four paws on the step. The number of steps descended was not taken into account, in order to simplify the observations. Preliminary experiments showed that this procedure did not alter the pharmacological profiles observed. After each animal had been tested, the box was rapidly cleaned to eliminate any olfactory cue which might modify the next animal's behaviour. Except where otherwise stated, drugs were administered orally (10 ml/kg) 60 rain before the test to groups of 12 mice. In each experiment, a control group received only the solvent. The treatments were randomised, and the observer was unaware of the treatment given to each group (blind method). All studies were carried out between 8 a.m. and 5 p.m. Drugs Chlordiazepoxide hydrochloride, diazepam, lorazepam, meprobamate, carbamazepine and ethosuximide were used as suspensions in an aqueous solution of 1% carboxymethylcellulose. Dipotassium clorazepate, phenobarbital, sodium valproate, amphetamine sulfate, morphine hydrochloride, chlorpromazine, haloperidol, imipramine and amitriptyline hydrochloride were dissolved in physiological saline. The benzodiazepine receptor antagonist CGS 8216 (Czernick et al. 1982), 2-phenylpyrazole (4,3-c) quinolin-3(5H)-one (Ciba-Geigy, Basel, Switzerland), synthesised in our laboratory, was suspended in 1% carboxymethylcellulose. The doses are expressed in terms of the salt or base where appropriate.
Statistical analysis Statistical significance of differences between control and treated groups with the same drug was ascertained by a combined analysis of variance and an unpaired two-tailed t-test using the Bonferroni method as described by
Wallerstein et al. (1980). The analysis is first based on Levene's test for equal variances. In cases of equality, an ANOVA is carried out and the pooled variance is used for multiple t comparisons (drug groups vs controls). In the case of unequal variances, a one-way analysis of variance is made using the Brown-Forsythe method. The separate variances are then used for the multiple t comparisons. The statistical tests were applied to the real experimental data. Percentages were calculated from the control group of the same experiment. All calculations were carried out on a DIGITAL VAX-I1 computer using statistical programs furnished by the B.M.D.P. (Department of Biomathematics, University of California, Los Angeles). Results
The activity of 100 untreated mice showed that the number of steps climbed (mean _+ SEM = 24 + 0.9) and the number of rearings (20 + 0.9) were very similar. However, the coefficient of correlation between these two activities, calculated with a group of 40 mice, was not significant (r = 0.26). Thus there does not seem to be any clear relationship between these two behavioural parameters. The two variables were normally distributed. The behaviour of untreated mice was very stable, and we did not observe any statistically significant differences in activity from one experiment to another. All the benzodiazepines tested significantly reduced rearing at doses which either had no effect on, or actually increased step climbing (Fig. 2). The latter variable was in fact markedly increased in a number of cases (diazepam, chlordiazcpoxide and clorazepate), but this increase was rarely statistically significant due to the wide (and reproducible) variability in individual responses. At the highest doses studied (except for clorazepate) there were significant reductions in both the number of rears and the number of steps climbed, presumably due to the sedative and muscle relaxant properties of the drugs. For the benzodiazepines tested, the differential effects on the two variables studied were observed over a wide range of doses. The specific reduction of rearing at low doses could be attributed to the activation of benzodiazepine receptors, since the administration of the receptor antagonist CGS 8216 (which was inactive alone) totally abolished the activity of clorazepate on rearing (Table 1). The other clinically active anxiolytics studied, meprobamate and phenobarbital, produced a marked reduction of rearing and a significant increase in the number of steps climbed (Fig. 3). Again, at high doses the number of steps climbed was reduced as a result of the hypnotic and/or muscle relaxant properties of the drugs. In the case of meprobamate, the dose-range which dissociated the two behavioural parameters was relatively small. Neuroleptics (haloperidol, chlorpromazine) and antidepressants (imipramine, amitriptyline) produced a progressive and simultaneous reduction of both the steps climbed, and rearing (Figs. 3 and 4). A dissociation of the two behavioural patterns was never observed. Of the three antiepileptics studied, ethosuximide had no effect at doses up to 1,024 mg/kg, PO, while carbamazepine produced an identical reduction in both rearing and steps climbed (Fig. 4). In contrast, sodium valproate reduced rearing activity at doses which significantly increased the number of steps climbed (Fig. 5). This was, however, only observed at high doses (512 mg/kg) and over a limited dose range.
50
Diazepam
Chlordiazepoxide 200.
200
q,,
150t
150
\\
\ \ \
100
. _ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .\. . . . .
'~176...... \-- ......... :'< ........ 50 z
16
8
32
4
64 158 rag/kg
16
32 64 rag/kg
Lorazepam
Clorazepate e-
8
200
200. 150.
~D"
150
\
D~
\
\
ZI
100 . . . . . . . . . . . . . . . . . . . . . .
100 ................................. 12k
~ .......
-
50
"~~-L-~""
"'"
50 .
[]. . . .
"1:3- . . . .
.
.
.
,
eeo
0 8
16
32 64 mg/kg
Fig. 2 Effects in mice of chlordiazepoxide, diazepam, clorazepate and lorazepam on the number of rearings and steps climbed in a 3-rain period. Drug effects are shown as mg/kg, PO. All experimental groups contained 12 animals. Percentages are calculated from the same untreated control group. *P < 0.05, **P < 0.01, ***P < 0.001 (analysis of variance, followed by Student's t-test, using the method of Bonferroni
rag/kg
Table 1. Effect ofa benzodiazepine receptor blocker (CGS 8216) on the action of clorazepate in the staircase test in mice Treatment
Doses mg/kg
Rearings mean § SEM
Steps climbed mean • SEM
Saline Clorazepate CGS 8216 Clorazepate § CGS 8216
16 10 16 + 10
16.8 _+1.9 3.3 +_1.0"* 19.8 -t- 2.2
18.5 _+2.4 17.3 • 5.6 24.7 • 2.8
24.3 _+2.2 + +
24.7 _+3.7
Clorazepate (PO) and CGS 8216 (IP) were administered respectively 25 min and 30 min before the test (12 mice per group). ** P < 0.01 compared to saline; + + P < 0.01 compared to clorazepate (A twoway analysis of variance followed by Student's t-test)
Morphine induced a progressive reduction of climbing and rearing behaviour (Fig. 5). At the highest doses (64 and 128 mg/kg), the Straub tail phenomenon was present. Amphetamine also produced a simultaneous reduction of both behavioural parameters (Fig. 5). This was not related to sedation, but rather to an initial immobility when placed in the experimental apparatus. Exploratory behaviour appeared only towards the end of the test period.
Discussion Our results indicate that the rearing response in the staircase test is not correlated with the climbing of steps. All clinically active anxiolytics (over a certain range of doses) modified
these two variables (independently), whereas the other psychotropic or neurotropic drugs tested did not dissociate the two behavioural patterns. Indeed the anxiolytics reduced rearing behaviour at doses which produced no significant reduction of, or even increased the number of steps climbed (Figs. 2 and 3). With the exception of valproate, the nonanxiolytic substances tested did not dissociate the two behavioural patterns. The tested neuroleptics, antidepressants, amphetamine, morphine and antiepileptics had identical effects on both step climbing and rearing over the whole range of doses studied (Figs. 3 - 5). However, valproate can be classified as an anxiolytic in the staircase test, since it reduced rearing at doses which increased climbing behaviour (Fig. 5). This confirms the predictive value of the test, as valproate has been found to have an anxiolytic profile in very specific models of anxiety such as conflict tests (Lal et al. 1980; Ohmori et al. 1980) and antagonism of pentylenetetrazol discrimination (Shearman and Lal 1980). Some experiments were carried out in order to determine the reproducibility of the experimental results. Results with chlordiazepoxide, diazepam, valproate and meprobamate were found to vary only slightly from one experiment to another, and the profile for each drug was always identical (results not shown). In screening studies "blind" experimentation does not appear to be necessary, as the criteria for the two behavioural patterns are unambiguous. We have found identical results in open or blind drug trials. The results we obtained using the staircase test in mice are comparable to those observed by Thi6bot et al. (t973, 1976) in rats. The active doses of the anxiolytics, and the behavioural profiles of the drugs are very similar. We have however observed that in mice the increased climbing pro-
51
Meprobamate
Phenobarbital 200
~*
200
/ '\ / /
150
/;Cq
k\
//
150
\\
J
\\ \
\ \
1oo . . . . . . . . . . . . . . . . . . . _~. . . .
50
~-.~..
'~.
""~T
9 E g
3'2
64
'~176 .....\ i
.........
504
',
..~
0J
, 64
158 256 rag/kg
Chlorpromazine
Z Z
, , ~"" 128 256 5i2 rag/kg
Haloperidoi
200
200 _
150
150 t
100 ...............................
Fig. 3
1.......< ...................
50
~..
501
~,,
",, ,,,
i
2
4
8
0.125 0.25
0.5
mg/kg
Imipramine
,,
1 2 mg/kg
Effects of phenobarbital, meprobamate, chlorpromazine and haloperidol on rearing and step-climbing activity of mice in the staircase test. For details see Fig. 2
Amitriptyline
200.
200.
1501
150
1001
100 N
5oi
ol
to \x
e.t~
"-
~oQ
O. 16
32
64
128 256 mg/kg
16
32
128 256 mg/kg
Ethosuximide
Carbamazepine ,tz
"'~
50_ 9o
200
200-
150
150
e~
E E
Z Z
1I
33. 100
100
50
50
0
mt~tllL
1~5 32
64
128 256 rag/kg
Fig. 4
0 128 256 5i2 10'24 mg/kg
Effects of imipramine, amitriptyline, carbamazepine and ethosuximide on rearing and climbing activity of mice in the staircase test. For details see Fig. 2
52 Valproate
200. 150
200
/"
150
\\
/Z]//
Amphetamine
\
.\ / lOO ........... _/_/_.................... _'~........
100_ . . . . . . . . E. . . . . . . . . ]. . . . . . . . ~. . . . . . . . . . . . . . . . . . . .
r~ ,...,
50.
50
***
o m
8 's 0
0 158
256
5i2
10'24 rag/kg
o15
i
~
~
mg/kg
rX~ 200t z z I
Morphine
150t l
1o . . . . . . . . . t 5o-1 0l
--..... \{:_ . . . . . . . . . . . . . . .
b~-7
i
8
/D,, ~ i
16
~
32
\\
[]-_ J
64
,, ,
128 rag/kg
Fig. 5. Effects of valproate, amphetamine and morphine on rearing and climbing activity of mice in the staircase test. For details see Fig. 2 duced by some anxiolytics is often less pronounced than in rats. This may be related to the different baseline climbing activities in the two species. The mouse has an average control value for this variable of 24 +_ 0.9 steps, which is three-times higher than that observed with rats (about 7 + 1 according to Thi6bot et al. 1973). In both species there is considerable inter-individual variation with regard to the ability of anxiolytics to increase climbing behaviour (results not shown); this explains the frequent lack of statistical significance of this effect. Therefore, the staircase test has the same characteristics in mice as in rats, in contrast to a number of other anxiolytic tests, for which the mouse has proved to be an unsuitable species (File 1980; Vogel et al.
1980). Thi~bot et al. (1973) have suggested that in the staircase test (in rats), rearing could be an index of the anxiety or emotionality of the animal, whereas climbing is more related to exploratory or locomotor activity. However, the theoretical interpretation of behavioural measures in terms of emotional states is not simple. Although rearing and locomotion may be dislocated by anxiolytic and psychotropic drugs (Hughes 1972; Cunha and Masur 1978), some authors consider them both to have a substantial exploratory component (Lilt 1972). In the staircase test, the exact relevance of rearing to anxiety and climbing to exploration is still uncertain, and further analysis of these relationships is necessary. Nevertheless, the procedure seems to be a valuable screening model for anxiolytic drugs. Furthermore, the adaptation of this test to mice may enable an analysis of the two behavioural patterns in a number of inbred strains of mice in which distinct neurochemical differences have been
observed in aminergic (Natali et al. 1980) and benzodiazepine systems (Skolnick et al. 1979). Some strains of mice appear to react differently to stressful situations (Tassin et al. 1980), and a behavioural and biochemical study of these strains may throw some light on the neuronal substrates of drug action and the relevance of rearing and climbing to anxiety and exploration in this model. In conclusion, the staircase test in mice is simple, rapid (five doses of a drug can be studied in 1 day), economical (with respect to apparatus and substances tested), and selective for anxiolytic drugs. The test requires no prior manipulation of animals, such as training, food deprivation or habituation to the experimental environment, and no further investigation as in some other tests (Boissier et al. 1968; Crawley 1981) where anxiolytic activity must be distinguished from general locomotor effects. Since the animals are not deprived or submitted to physical stress such as electric shock, the staircase test is consistent with the ethics of modern animal experimentation.
Acknowledgements. The authors would like to thank MarieChristine Barnouin and Josette Guitard for their excellent technical assistance and Marie Roussel for preparation of the manuscript. Diazepam and chlordiazepoxide were generous gifts from HoffmanLa Roche (France), and imipramine from Ciba-Geigy (France).
References
Aron C, Simon P, Larousse C, Boissier JR (1971) Evaluation of a rapid technique for detecting minor tranquilizers. Neuropharmacology 10: 459 - 469
53 Boissier JR, Simon P, Aron C (1968) A new method for rapid screening of minor tranquillizers in :mice. Eur J Pharmacol 4:145-151 Crawley JN (1981) Neuropharmacologic specificity of a simple animal model for the behavioral actions of benzodiazepines. Pharmacol Biochem Behav 15: 695 - 699 Cunha JM, Masur J (1978) Evaluation of psychotropic drugs with a modified open field test. Pharmacology 16:259-267 Czernik AJ, Petrack B, Kalinsky H J, Psyclaoyos S, Cash WD, Tsai C, Rinehart RK, Granat FR, Lovell RA, Brundish DE, Wade R (1982) CGS 8216: Receptor binding characteristics of a potent benzodiazepine antagonist. Life Sci 30: 363 - 372 Dantzer R (1977) Behavioral effects of benzodiazepines: a review. Biobehav Rev 1 : 7 1 - 8 6 File S (1980) The use of social interactions as a method for detecting anxiolytic activity of chlordiazepoxide-like drugs. J Nenrosci Methods 2:219-238 Geller I, Seifter J (1960) The effects of meprobamate, barbiturates, d-amphetamine and promazine on experimentally induced conflict in the rat. Psychopharmacologia ]Z:482-492 Hughes RN C1972) Chlordiazepoxide modified exploration in rats. Psychopharmacologia 24: 4 6 2 - 469 Iversen LL (1980) The present status of benzodiazepines in psychopharmacology. Arzneim Forsch 30: 9 0 7 - 909 Kolata G (1982) New valiums and anti-valiums on the horizon. Science 216: 6 0 4 - 605 Lal H, Shearman GT, Fielding S, Dunn R, Kruse M, Theurer K (1980) Evidence that GABA mechanisms mediate the anxiolytic action of benzodiazepines: A study with valproic acid. Neuropharmacology 19 : 785-- 789 Lilt J (1973) The analysis of habituation. Acta Neurobiol Exp 33 : 771 - 789 Mohler H, Okada T (1977) Benzodiazepine receptor: Demonstration in the central nervous system. Science 198 : 849- 851 Natali JP, Mc Rae-Degueurce A, Keane PE, Debilly G, Pujol JF (1980) Genetic studies of daily variations of first-step enzymes of monoamine metabolism in the brain of inbred strains of mice and hybrids. II. Daily variations of tyrosine hydroxylase activity in the locus coeruleus. Brain Res 191:205-213 Ohmori K, Yamanami S, Shuto K, Marumo H (1980) Effect of sodium dipropylacetate on conflict behavior in rats. Jpn J Pharmacol 30: 9 2 5 - 927
Sherman G, Lal H (1979) Discriminative stimulus properties of pentylenetetrazol and Bemegride: Some generalization and antagonism tests. Psychopharmacology 64: 315 - 319 Sherman G, Lal H (1980) Generalization and antagonism studies with convulsant, gabaergic and anticonvulsant drugs in rats trained to discriminate pentylenetetrazol from saline. Neuropharmacology 19 :473 - 479 Skolnick P, Syapin P, Paugh BA, Paul S (1979) Reduction in benzodiazepine receptors associated with Purkinje cell degeneration in nervous mutant mice. Nature 277:397-399 Squires RF, Braestrup C (1977) Benzodiazepine receptors in rat brain. Nature 266: 732- 734 Stephens RJ (1973) The influence of mild stress on food consumption in untrained mice and the effects of drugs. Br J Pharmacol 47:146 P Tassin JP, Herve D, Blanc G, Glowinski J (1980) Differential effects of a two-minute open field session on dopamine utilization in the frontal cortices of BALB/C and C57BL/6 mice. Neurosci Lett 17:67-71 Thi6bot MH, Soubri~ P, Simon P, Boissier JR (1973) Dissociation de deux composantes du comportement chez le rat sous l'effet de psychotropes. Application i l'~tude des anxiolytiques. Psychopharmacologia 31 :77 - 90 Thi6bot MH, Soubri6 P, Simon P, Boissier JR (1976) Specificit6 d'action des tranquillisants mineurs dans le test de l'escalier. Relation entre ces effets et leurs propri6t6s anxiolytiques. J Pharmacol (Paris) 7: 87-102 Usdin E, Skolnick P, Tallman JF, Greenblatt D, Paul S (1982) Pharmacology of benzodiazepines. Mac Millan Press Ltd, London Vogel JR, Beer B, Clody DE (1971) A simple and reliable conflict procedure for testing anti-anxietyagents. Psychopharmacologia 21:1-7 Vogel RA, Frye GD, Wilson JH, Kuhn CM, Koepke KM, Mailman RB, Mueller RA, Breese GR (1980) Attenuation of the effects of punishment by ethanol: Comparisons with chlordiazepoxide. Psychopharmacology 71:123 - 129 Wallerstein S, Zucker CL, Fleiss JL (1980) Some statistical methods useful in research. Circ Res 47:1 - 9
Received December 19, 1983; Final version May 21, 1984
