Prog. N-h~ d BbL I’syehU 1891. Vol. 15. pp. 205212 02’18!%46/91$0.00 + .50 mntcd In Great mtatn. Au rIghta - 01921FvrgammRcmp1c

ANIMAL MODELS OF ANXIETY AND THE DEVELOPMENT OF NOVEL ANXIOLYTIC DRUGS

DAVID J. SANGRR, GHISLAINE PERRAULT, ELIANB MOREL, DANIELB JOLY and BRANIMIR 2IVKOVIC

Synthdlabo Recherche (L.E.R.S), Bagneux, France

(Final form, October 1990)

Contents

Abstract 1. Introduction 2. Novel Anxiolytics 2.1. o Receptors Ligands 2.2. Serotonergic Mechanisms 3. Summary and Conclusions

References

205 206 207 207 207 210 210

Abstract

Sanger , David J., Ghislaine Perrault, Rliane Morel, Danii?le Joly and Branimir Zivkovic: Animal Models of Anxiety and the Development of Novel Anxiolytic Drugs. Prog. Neuro-Psychopharmacol. h Blol. Psychlat. 1991, 15 : 205-212. -

1.

2.

3.

4.

The behavioural effects of classical anxiolytics such as barbiturates and benzodiazepines have been well characterised. However, recent research has been aimed at the development of novel anxiolytics without problems of sedation, muscle relaxation, amnesia and dependence.

A number of novel w (benzodiazeplne) receptor ligands with anxiolytic properties have been described including alpidem, bretazenil. suriclone and abecarnil. Although these compounds share some behavioural effects with older anxiolytic drugs, such as increasing punished drinking, they also show many differences. Their novel profiles may be related to low intrinsic activity or to selectivity for w receptor subtypes.

The possibility that novel anxiolytics may be found among compounds active at serotonin receptors remains a strong hypothesis. Compounds, which, like buspirone, are active at

5HT1A receptors may be anxiolytic as may be antagonists at %fT2 and 5HT3 receptors. All these ccmpounds have behavioural effects which differ from those of benzodiazepines .

In order more efEectively to screen for and develop novel anxiolytics It will be necessary to refine behavioural models in the light of feedback from the clinic.

Kevwords . Anxiety, punishment, anxiolytics, alpidem. suriclone. abecarnil, bretazenil, busplrone

205

1. Introduction

Until recently the behavioural pharmacology of anxiolytic drugs had probably been one of

the most satisfying areas of study in the discipline. The behavioural effects of these

drugs had been well characterised and generally consistent effects were described under a

variety of experimental conditions. For example, the early demonstrations by Seller and

Seifter (1960) and Cook and Davidson (1973) that benzodiazepines. barbiturates and

meprobamate will produce substantial increases in rates of punished operant responding had

been confirmed in many subsequent experiments (Pollard and Howard, 1990). In addition to

measures of behaviour suppressed by punishment many other behavioural effects of these same

drugs have been demonstrated and there have been a number of scholarly assessments of the

adequacy of these different procedures as animal models for evaluating anti-anxiety

activity (eg Lister, 1990 ; Treit, 1985). However, it seems an indication of how well the

behavioural effects of anxiolytics had been established, that the term animal model of

anxiety was used relatively infrequently and the validity ‘of the different behavioural

tests was subject to relatively little discussion. In terms of the validation criteria for

animal models discussed by Willner (19841 In the context of depression, the pharmacological

validity (ie the extent to which activity in the laboratory tests correlated with or

predicted clinical anxiolytic effects) was high in most cases because the procedures were

developed for their sensitivity to clinically used drugs. Face validity appeared high for

many procedures because the tests seem to involve stressful, fear-inducing or punishing

stimuli, presumably relevant to human anxiety. Construct validity of these behavloural

models, however, has been subject to very little discussion (but see recent discussion by

Lister, 19901.

For several reasons, investigation of the behavioural effects of anxiolytics has become

considerably more difficult in recent years and the emphasis of this research has moved

away from the benzodiazepines. Because benzodiazeplnes have a complex pharmacological

profile including anticonvulsant. muscle-relaxant, sedative and amnesic actions, in

addition to anxiolytic effects, efforts have been made to develop drugs more selectively

anxiolytic without the other pharmacological effects which, while often having therapeutic

utility themselves, may be considered undesirable side effects during the treatment of

anxiety. In addition, there has been much concern about withdrawal syndromes seen after

cessation of chronic benzodiazepine treatment. The evolution of ideas about the nature and

treatment of different forms of anxiety has also led to questions concerning the adequacy

of currently used animal models (Lader, 1989).

Animal model of anxiety 207

2. Novel Anxiolvtics

2.1. w Receotor liqands

The discovery that benzodiazepines bind to specific. high affinity sites in the central

nervous system led to a considerable amount of research on the’nature and function of these

receptors, which have recently been renamed, o receptors (Langer and Arbilla, 1988).

There has also been much effort aimed at the synthesis of novel molecules which. while

binding to the same receptors, would, it was hoped, have more selective pha~cological

profiles. Hany such compounds have been described recently, as reviewed by C?ardner (1988).

In the search for new anxiolytics It was hoped that structurally novel compounds with

affinity for w receptors might be more selectively anxiolytic. such compounds might be

expected to have behavioural effects, such as anti-p~is~ent activity indicating

anxiolytic actions, without the muscle-relaxant, sedative, memory disrupting and tolerance-

and dependence-producing potential of benzodiazepines. A number of such compounds have been

developed for clinical use as anxlolytics and although none has yet entered into widespread

clinical use it is likely that several will do so in the near future. such compounds

include the ~idazopyrid~e, alpidem fzivkovic et al, 19901, the cyclopyrrolone. surfclone

(Julou et al, 1985), the R-carboline, abecarnil (Stephens et al, 1990). and the

imidazodiazepine, bretazenil (martin et al, 1968). Although these compounds do often show

more selective profiles than traditional benzodiazepines, all seeming to have much lower

propensities to produce muscle relaxation, for example, their behavloural effects often

depend on the particular test used to assess potential anxiolytic activity. Table 1 shows

the effects of five o receptor ligands on punished drinking in rats. a procedure widely

used for assessing anxiolytic potential (Vogel et al, 1971). Although all compounds were

active it is clear that the efficacy of the different drugs varied widely. only abecarnil

produced an effect comparable with that of chlordiazepoxide. iiowever. in another punishment

procedure I the four-plate test this time in mice, bretazenfl and abecarnil showed only

weak effects while alpidem and suriclone, two drugs known to have clinical anxiolytic

activity, were inactive (unpublished results).

200 D. 1. Sanger et al.

Table 1

Effects of w Receptor Ligands on Punished Drinking in Rats

Chlordiaxepoxide

Dose - mg/kg

Licks - % Control

3 10 20

239 + 93 739 + 26* 690 + 138**

Sur lclone

Dose - mg/kg

Licks - 0 control

Alpidem

Dose - mg/kg

Licks - % Control

Abecarnil

Dose - mg/kg

Licks - % Control

0.3

207 + 101

10

162 + 32

1.0

200 + 57

1.0 3.0

289 + 55** 270 + 39*

10

375 + 50**

30 60 100

243 f. 63** 257 + 45*” 208 + 30**

3.0 10

685 5 45** 704 + 94**

Bretazenil

Dose - mg/kg

Licks - % Control

1.0 3.0 10

178 + 4 185 + 77 247 + 77*

Each dose was tested in at least 8 rats. Results are expressed as 0 of control because the

control values differed slightly in different experiments: control numbers of licks/min

ranged from 31 + 5 to 56 + 15.

*p < 0.05, **p < 0.01 difference from control.

Such apparently contradictory results may, of course, be explained in terms of species

differences but this is not always the case. For example, in mice, bretazenll, has been

reported to show anxiolytic-like activity In a two-chambered exploration box while having

no effect on activity in the staircase test (Belzung et al, 1989). Similarly, alpidem and

abecarnil while ’ increasing punished drinking in rats did not produce statistically

significant increases in rates of punished operant responding in the same species (Stephens

et al, 1990 : Zivkovlc et al, 1990).

In order to explain the different pharmacological profiles of such compounds two

hypotheses have been considered. Such compounds may, in some cases, act as partial agonists

Animal model of anxiety a09

at o receptors thus showl@g pharmacological effects which are produced by benzodlazepines

at low levels of receptor occupation without showing effects normally associated with high

levels of occupation. The pharmacological effects of bretazenil are, in general, consistent

with this hypothesis although the differential sensitivity of different tests of punis~ent

or exploratory activity remains to be explained. Thus bretazenil has been reported to

produce pharmacological activity only at relatively high levels of receptor occupancy

(Potier et al, 1988) and has also been shown to antagonise certain effects of

benzodiazeplnes and other agonists (Martin et al, 1988. Sanger, 19871.

This hypothesis of partial agonist activity at a single receptor site does not, however,

provide an adequate explanation for the pharmacological activity of sane other compounds,

Including alpidem and abecarnll. Both of these compounds seem to produce very little muscle

relaxation but, unlike bretazenil. they do decrease rates of spontaneous and conditioned

behaviour , although at doses higher than those giving rise to ~tic~~ls~t or

antipunishment effects. In studies of the discriminative stimulus properties of these

drugs, also, results are inconsistent with partlal agonism. Thus alpidem did not substitute

for chlordlazepoxide but did substitute for the hypnotic imidazopyridine, zolpidem

(Zivkovic et al, 1990) a result different from that obtained with bretazenil and other

compound% believed to act a% partial agonists (Sanger et al, 1987). Abecarnil was

generalised in rats trained to discriminate a benzodiaxepfne but in rats trained with

abecarnil the stimulus control did not generalise to benzodiazepines although it did to

zolpidem (Rndrews and Stephens, 1989). Such results seem only to be explicable in terms of

selectivity for certain receptor subtypes. Indeed, alpidem has been shown to be selective

for the o1 and “3

subtypes &anger et al, 1990). However, the functional

significance of u3 sites is presently unknown and as the hypnotic, zolpidem, also has

selectivity Ear the o1 site, It Is clear that activity at this site is not only related

to anxiolytic eEfects. To explain the anxiolytic profile of alpidem it has been suggested

that both receptor selectivity and a low level of intrinsic activity are involved (Zivkovic

et al, 1990).

2.2. Serotonerqic mechanisms

For some years serotonin has been associated with the neural mechanisms underlying

anxiety and the actions of anxiolytic drugs (Iversen, 1984). The recent finding that the

novel anxiolytic drug, buspirone, has aEfinity for the SHTIA subtype of serotonin

receptor (Glaser and Traber, 1983) has greatly intensified the search for new anxiolytics

acting through serotonerqic synapses. Hany new compounds have recently been described as

Potential anxlolytics, including agonist% at 5HTll( receptors feg ipsapirone - Traber et

al, 1984) and antagonists at 5HT2 feg ritanserin - Meert and Janssen, 1989) and 5?lT3

(eg ondansetron - Jones et al, 1988) receptors.

fn behavioural tests in the laboratory the activities of these compounds often difEer

significantly from those of benzodiazepine anxiolytics. For example. in tests of punished

210 D. J. Sanger et al.

responding in rodents a,nd primates, compounds acting at 5HT receptors rarely give rise to

large and consistent increases in response rate similar to those seen with benzodiazepines

(Dourish, 1987). However, In pigeons, buspirone, ipsapirone and other compounds active at

5HT1* receptors, produced large increases in punished responding and smaller increases

were produced by ritanserin and ketanserin (Gleeson et al, 1989). Antagonists at 5HT3

receptors have not been shown to increase rates of punished behavior although they have

been reported to produce changes in patterns of exploratory behaviour similar to those

associated with benzodiazeplnes (Jones et al, 1988). Similar effects are not found in all

laboratories, however (Pile and Johnston, 1989). It is also notable that the effects of

buspirone and related drugs in tests such as the exploration of elevated mazes differ

greatly from laboratory to laboratory (Soderpalm et al, 1989; Moser, 1989). Presumably,

such apparently contradictory results depend on experimental details. The critical

methodological variables remain to be identified, however.

3. Sununarv and Conclusions

The behavioural pharmacology of anxiolytic drugs has entered a new and perhaps more

difficult stage. Many compounds are currently in preclinical and clinical development and

much emphasis continues to be placed on behavioural methods as predictors of clinical

anxiolytic activity. However, as it is hoped that newer drugs will have more specific

anti-anxiety activity than drugs currently being used, it seems clear that their

behavioural profiles will differ from those associated with benzodiazepines and

barbiturates. Much recent research has conCirmed that behavioural procedures such as

punished responding and enhanced exploration in novel environments are sometimes

insensitive to novel compounds which may produce effects different from those associated

with benzodiazepines. Future research efforts therefore, will need to identify the

behavioural tests with the highest validity for predicting activity in the clinic. As yet.

only busplrone has entered into widespread clinical use as a new and different anxiolytic.

However, in the next few years a variety of other new anxiolytics will probably become

available. It will then be possible for laboratory scientists to evaluate their results in

the light of clinical experience.

References

ANDRRUS, J.S. and STRPHRNS, D.N. (1989) Discriminative stimulus properiies of the anxiolytic beta-carboline ZK 112119. Sot. Neurosci. Abs B:633

BELZUNG. C., HISSLIN, R. and VOGEL, E. (1989) Behavioral ePfects of the benzodiazepine receptor partial agonist RO 16-6028 in mice. Psychopharmacology 97: 388-391.

Animal model of anxiety 211

FILE, S.E. and JOHNSTON, A.L. (1989) Lack of effects of 5HT3 receptor antagonists in the social interaction and elevated plus-maze tests of anxiety in the rat. Psychopharmacology 99: 248-251.

COOK. L. and DAVIDSON, A.B. (1973) Effects of behaviorally active drugs in a conflict- punishment procedure in rats. In: The Benzodlazepines. , 5’. Garattini, B. Mussini and R.O. Randall teds) pp. 327-345. Raven Press, New York.

DOURISH, C.T. (1987) Brain 5-mlA IWCeptOrS and anxiety. In : Brain 5-HTlA Receptors. C.T. Dourish, S. Ahlenius and P.H. Hutson teds), pp. 261-277, Ellis-Ho&wood, Chichester.

GARDNER, C.R. (1988) Pharmacological profiles in vlvo of benzodiazepine receptor ligands. Drug Devel. Res. u:l-28.

GRLLRR, I. and SEIFTER, J. (1960) The effects of meprobamate, d-amphetamine, and prcxnazine on experimentally induced conflict in the rat. Psychophamacologla 1: 482-492.

GLASRR, T. and TRABRR, J. (1983) Buspirone : action on serotonin receptors in calf hlppocampus. Eur. J. Phamacol. a: 137-138

GLRBSON. S., AHLRRS, S-T., MANSBACH, R.S.. FOUST, J.H. and BARRETT, J.R. (1989) Behavioral studies with anxiolytic drugs. VI. Effects on punished responding of drugs interacting with serotonin receptors. J. Phamacol. Bxp. Ther. 250: 809-817.

IVRRSEN, S.D. (1984) 5-HT and anxiety. Neuropharmacology 23: 1553-1560.

JONES, B.J.. COSTALL, B.. DOMRNBY, A.M., KELLY, M-E., NAYLOR, R-J.. OAKLEY, N.R. and TYRRS, M.B. (1988) The potential anxiolytic activity of GR38032F. a 5-HT3-receptor antagonist. Brit. J. Pharmacol. 93 : 985-993

JULOU, L., BLANCHARD, J.C. and DREYFUS, J.F. (1985) Pharmacological and clinical studies of cyclopyrrolones : zopiclone and suriclone. Pharmacol. Biochem. Behav. 23: 653-659.

LADER, M. (1989) Drug treatment of anxiety: Implications for animal models. Behav. Pharmacol. 1: 95-100

LANGER. S-2. and ARBILLA, S. (1988) Limitations of the benzodiazepine receptor nomenclature: a proposal for a pharmacological classification as omega receptor subtypes. Fund. Clin. Phamacol. 2: 159-170.

LANGRR, S-2.. ARBILLA, S., TAN, S., LLOYD, K-G.. GEORGE, P., ALLEW, J. and WICK, A. (1990) Selectivity for omega receptor subtypes as a strategy for the development of anxiolytic drugs. Pharmacopsychiatry 23 (Suppl. III): 103-134.

LISTeR, R.G. (1990) Ethologically-based animal models of anxiety disorders. Pharmac. Ther. 46: 321-340.

MARTIN, J-R., PIRRI, L., BONETTI. B.P., SCHAFFNSR. R.. BURKARD, W.P., CUMIN , R. and HABFELY, W.E. (1988) Ro 16-6028 : A novel anxiolytic acting as a partial agonist at the benzodiazeplne receptor. Pharmacopsychiatry 21: 360-362.

MEERT, T.F. and JANSSEN. P.A.J. (1989) Psychopharmacology of ritanserin : comparison with chlordlazepoxide. Drug Dev. Res. Is: 119-144.

MOSER, P.C. (1989) An evaluation of the elevated plus-maze test using the novel anxiolytic buspirone. Psychopharmacology 99: 48-53.

POLLARD, G.T. and HOWARD, J.L. (1990) Effects of drugs on punished behavior : pre-clinical test for anxiolytics. Phamac. Ther. 45: 403-424.

PGTIER. MC., PRADO DE CARVALHO, L., VENAULT, P., CHAPGUTHIER, 0. and ROSSIER. J. (1988) Demonstration of the partial agonist profiles of Ro 16-6028 and RO 17-1812 in mice in viva. Eur. J. Pharmacol. 156: 169-172.

212 D. J. Sanger et al.

SANGBR, 0.~. (1987) Purther investigation of the stimulus properties of chlordiazepoxide and zolpidem. Agonism and aptagonism by two novel benzodiazeplnes. Psychopharmacology 365-368.

SANGRR, 0-J.. PRRRAULT, G., MOREL, E., JOLY, 0. and ZIVKOVIC, 8. (1987) The behavioral

93:

profile of zolpldcm, a novel hypnotic drug of lmidazopyridine structure. Physlol. Behav. 1: 235-240.

SODRRPALM, B., HJORTH, S. and RNGEL, J.A. (1989) Effects of 5HTlA receptor agonists and L-5-HTP In Montgomery’s conflict test. Pharmac. Biochem. Behav. 2: 259-265.

STRPHRNS, D-N., SCHWEIDRR, H.H., KEHK, U., AWDREWS, J.S. RRTTIG, K.J., TIJRSKI, L., SCHMIRCHBN, R., TIJRNRR, J-0.. JENSEN, L.H., PETERSEN. B-N.. HONORB, T. and BONDO-HANSEN, J. (1990) Abecarnil, a metabolically stable, anxioselective 8-carboline acting at benzodiazepine receptors. J. Pharmacol. Exp. Ther. 253: 334-343.

TRABER, J., DAVIES, H.A., DOMPRRT, W.V., GLASER, T., SCHWRMAN. T. and SBIDEL, P.R. (1984) Brain serotonln receptors as a target for the putative anxiolytic TVKQ 7821. Brain Res. Bull. 12: 741-744.

TRBIT, 0. (1985) Animal models for the study of anti-anxiety agents : A review. Neurosci. Biobehav. Rev. 2: 203-222.

VOGEL, J., BEER, 8. and CLODY. 0. (1971) A simple and rellable,conflict procedure for testing antianxiety agents. Psychopharmacologia 21: 1-7.

WILLNKR, P. (1984) The validity of animal models of depression. Psychopharmacology 83: 1-16

ZIVKOVIC, B., MOREL, E., JOLY. 0.. PERRAULT, G., SANGER, D.J. and LLOYD, K.G. (1990) Pharmacological and behavioral profile of alpidem as an anxiolytic. Pharmacopsychiatry 23 (suppl. III): 108-113.

Inquiries and reprint requests should be addressed to:

Dr. D.J. Sanger Synthilabo Recherche (L.E.R.S.) 31 ave P.V. Couturier 92220 Bagneux France