Psychopharmacology (2002) 163:26–35DOI 10.1007/s00213-002-1150-y

O R I G I NA L I NVE S T I GA T I ON

Alexandre Urani · Pascal Romieu ·Elodie Portales-Casamar · Fran�ois J. Roman ·Tangui Maurice

The antidepressant-like effect inducedby the sigma1 (s1) receptor agonist igmesine involves modulationof intracellular calcium mobilization

Received: 26 October 2001 / Accepted: 5 May 2002 / Published online: 29 June 20022 Springer-Verlag 2002

Abstract Rationale: Activation of the neuronal sigma1(s1) receptor potentiates calcium mobilization, leading toeffective modulation of postsynaptic responses to neuro-transmitters. At the behavioral level, s1 agonists modulatelearning, response to stress and depression. In particular,the selective s1 agonist igmesine reduced immobility inthe forced swimming test. Objectives and methods: Weinvestigated the effect of modulators of Ca2+ influx andmobilization, administered intracerebroventricularly atdoses ineffective alone, on the igmesine effect. Thetricyclic antidepressant desipramine was also studied forcomparison. Results: The calcium chelator EGTAblocked both igmesine and desipramine-induced decreas-es of immobility duration, indicating the importance ofextracellular Ca2+ influx in the initial action of eachcompound. Both L- and N-type voltage-dependent calci-um channel (VDCC) appeared involved in the s1 agonisteffect. Verapamil, an L-type VDCC antagonist or w-conotoxin GVI, a N-type VDCC antagonist, blockedwhereas (–)-Bay K8644, a L-type VDCC agonist, poten-tiated the igmesine effect. Mobilization of intracellular

Ca2+ stores is involved selectively in the effect mediatedby the s1 receptor, since the membrane permeableintracellular Ca2+ chelator EGTA/AM affected only theigmesine effect. Inositol 1,4,5-trisphosphate (InsP3) re-ceptor-sensitive Ca2+ pools appeared primarily involved,rather than Ca2+/caffeine-sensitive Ca2+ pools. Indeed, theInsP3 receptor positive modulator bradykinin potentiated,whereas the InsP3 receptor antagonist xestospongin Cblocked the igmesine effect. The ryanodine receptoragonist caffeine failed to affect the efficacy of igmesine,whereas the antagonist ryanodine reduced it. Conclu-sions: The s1 receptor-mediated behavioral effect isdependent not only on rapid Ca2+ influx, as observed for aclassical antidepressant, but also on intracellular Ca2+

mobilization.

Keywords s1 Receptor · Igmesine · Ca2+ influx · Ca2+mobilization · Forced swimming test · Depression · Mouse

Introduction

The s1 receptor represents a unique receptor in the brain.The 223-amino acid protein, recently cloned in severalanimal species and in humans, has only one transmem-brane domain and appears unrelated to any other knownmammalian protein (Hanner et al. 1996; Prasad et al.1998). The s1 receptor is widely distributed throughoutthe brain. Its anatomical distribution, elucidated usingautoradiographic techniques with selective radioligands,in situ hybridization or immunohistochemical studies,revealed that the s1 receptor is particularly present withinthe granular layer of the olfactory bulb, the cortex,hippocampus, various hypothalamic nuclei, septum, cau-date putamen, motor nuclei of the hindbrain, andcerebellum (Su et al. 1982; Largent et al. 1986; Gundlachet al. 1986; McLean et al. 1988; Graybiel et al. 1989;Jansen et al. 1991; Okuyama et al. 1995; Bouchard et al.

A. Urani · P. Romieu · E. Portales-Casamar · T. Maurice ())INSERM U. 336, Behavioral Neuropharmacology group,Institut de Biologie, 4, boulevard Henri IV, 34060 Montpellier,Francee-mail: maurice@univ-montp2.frTel.: +33-4-67144270Fax: +33-4-67144251

A. Urani · F.J. RomanDGpartement de Biochimie/Enzymologie, Pfizer GRD-Fresnes, 3/9,rue de la Loge, BP 100, 94265 Fresnes Cedex, France

Present addresses:E. Portales-Casamar, C.R.B.M., CNRS UPR 1086, 1919,Rte de Mende, 34293 Montpellier Cedex 5, France

P. Romieu · T. Maurice, CNRS UMR 5102,University of Montpellier II, c.c. 090, place EugHne Bataillon,34095 Montpellier Cedex 5, France

1996; Alonso et al. 2000; Kitaichi et al. 2000; Zamanilloet al. 2000). At the subcellular level, the s1 receptor isassociated with membranes of the nucleus, endoplasmicreticulum, mitochondria and the plasma membrane,especially at the postsynaptic thickening (Alonso et al.2000).The s1 receptor binds with high affinity a wide variety

of synthetic molecules from different chemical classes.Activation of the receptor results in the modulation ofseveral responses to neurotransmitters. For example, theselective agonist igmesine (JO-1784) (Roman et al. 1990)potentiated the electrophysiological responses to N-methyl-d-aspartate (NMDA) in CA3 hippocampal pyra-midal neurons (Monnet et al. 1992a) or in dopaminergicneurons from the nucleus accumbens (Zhang et al. 1993;Gronier and Debonnel 1999). Igmesine also facilitatedNMDA-evoked [3H]norepinephrine release from rat hip-pocampal slices (Monnet et al. 1992b), potentiated[3H]acetylcholine release from rat hippocampal slices(Junien et al. 1991) and decreased the NMDA-inducedincrease in cGMP levels in rat cerebellar slices (Akunneet al. 2001). In turn, at the behavioral level, the compoundshowed anti-amnesic properties in pharmacological mod-els of amnesia (Earley et al. 1991) or on age-relatedlearning deficits (Maurice et al. 1996b). Igmesine also hadpotent anti-stress activity, as observed against peripheralcolonic motor disturbances (GuG et al. 1992a, 1992b) oron behavioral despair models (Matsuno et al. 1996; Songet al. 1997; Urani et al. 2001). The compound attenuatedseveral physiological responses to the stress-inducedrelease in corticotropin-releasing factor (CRF) (GuG etal. 1994; Song et al. 1997). Igmesine thus appeared as acompound of choice to examine the mechanism sustain-ing the behavioral activation of the s1 receptor.The s1 receptor also binds several endogenous effec-

tors, including neuroactive steroids. The neuroactivesteroid/s1 receptor interaction was first demonstratedusing binding studies (Su et al. 1988; Yamada et al. 1994;Maurice et al. 1996a), then reported in physiological(Monnet et al. 1995; Debonnel et al. 1996) and behavioralstudies (Maurice et al. 1997, 1998, 1999; Reddy et al.1998; Urani et al. 1998; Noda et al. 2000). In particular,progesterone acts as a potent s1 receptor antagonist, andphysiological variations in the central levels of the steroiddirectly affect the behavioral efficacy of s1 receptoragonists (Phan et al. 1999; Urani et al. 2001). Forexample, the selective s1 agonist igmesine showed anantidepressant-like effect at 60 mg/kg IP in the forcedswimming test. In mice deprived of circulating steroids byadrenalectomy/castration (AdX/CX), the dose-response ofigmesine was shifted and the drug appeared effective at10 mg/kg (Urani et al. 2001).Recent results demonstrated that activation of the s1

receptor is related to regulation of Ca2+ mobilization. Thes1 agonists increased contractility, beating rate and Ca2+influx in cultured cardiac myocytes (Ela et al. 1994), andincreased intracellular level of inositol 1,4,5-trisphosphate(InsP3) in cultured myocytes (Novakova et al. 1998). Insynaptosomes, the s1/s2 agonist DTG decreased the basal

intracellular free Ca2+ concentration ([Ca2+]i) and blockedthe depolarization-dependent increases in [Ca2+]i (Brentet al. 1997). The s1 ligands reduced the NMDA-inducedincrease in [Ca2+]i by facilitating desensitization of theCa2+ response to NMDA in NG108 cells (Hayashi et al.1995). Moreover, binding studies suggested an importantassociation between the s1 receptor and the InsP3 receptor(Tsao and Su 1996). The s1 receptor may be a sensor/modulator of calcium signaling (Hayashi et al. 2000).Bradykinin, by increasing the formation of InsP3, acti-vates the InsP3 receptor, leading to Ca

2+ mobilizationfrom endoplasmic reticulum pools (Berridge et al. 1993;Thomas et al. 1996). In NG108 cells, s1 agonists such as(+)-pentazocine, PRE-084 or pregnenolone sulfate, en-hanced the bradykinin-induced increase in [Ca2+]i, sug-gesting that the s1 receptor facilitated InsP3-induced Ca2+release (Hayashi et al. 2000). (+)-Pentazocine or preg-nenolone sulfate inhibited, whereas PRE-084 potentiated,KCl-induced increase in [Ca2+]i in a pertussis toxinsensitive manner, suggesting that s1 ligands affected thedepolarization-induced [Ca2+]i influx at the plasmamembrane by a mechanism implying Gi/o proteins(Hayashi et al. 2000). Therefore, the hypothesis wastested whether the s1 receptor-mediated antidepressant-like effect in the forced swimming test is directly relatedto modulation of Ca2+ influx or mobilization fromintracellular pools, in comparison with the mechanisminvolved in the action of classical antidepressants. Theeffect of extracellular and intracellular calcium modula-tors on the behavioral effect of igmesine or desipraminewas thus investigated using direct intracerebroventricularadministration in controls and AdX/CX mice submitted toforced swimming.

Materials and methods

Animals

Animals were male Swiss mice from the breeding center of theFaculty of Pharmacy (Montpellier, France), aged 5–6 weeks andweighing 30€2 g. Animals were housed in groups, with free accessto laboratory food and water, except during behavioral experiments.They were kept in a regulated environment (23€1OC, 40–60%humidity) under a 12-h light/dark cycle (light on at 7:00 a.m.) in aconventional animal facility. Experiments were carried out between2:00 p.m. and 6:00 p.m., in a soundproof and air-regulatedexperimental room, to which mice were habituated at least 30 minbefore each experiment. All animal procedures were conducted inadherence to the European Community Council Directive of 24November 1986 (86-609/EEC) and Decree of 20 October 1987 (87-848).

Drugs

(+)-N-Cyclopropylmethyl-N-methyl-1,4-diphenyl-1-ethyl-but-3-en-1-ylamine hydrochloride (igmesine, CI-1019, JO-1784) was syn-thesized at Pfizer GRD, Fresnes. Desipramine was from Sigma (St-Quentin Fallavier, France). Both drugs were dissolved in distilledwater and injected intraperitoneally (IP). Other drugs were injectedintracerebroventricularly (ICV). Ethyleneglycol-bis(b-aminoethylether)-N,N,N’,N’-tetraacetic acid (EGTA), w-conotoxin GVIA,

27

bradykinin, caffeine, (–)-Bay K8644 and verapamil were fromSigma. Ethyleneglycol-bis(b-aminoethyl ether)-N,N,N’,N’-tetraace-toxymethyl ester (EGTA/AM), ryanodine and xestospongin C werefrom Calbiochem-Novabiochem Corp. (La Jolla, Calif., USA).EGTA, caffeine and ryanodine were dissolved in distilled water. w-Conotoxin GVIA was dissolved in 0.9% NaCl. Xestospongin C wasdissolved in ethanol 0.01% brought up to volume in distilled water.Vehicle for EGTA/AM was DMSO:cremophor:H2O (2:1:17). (–)-Bay K8644 was dissolved in ethanol 5% in distilled water, andverapamil and bradykinin in methanol 2% in distilled water. Doseranges were selected according to previous studies (Laurent et al.1987; Dar 1988; Bourson et al. 1989; Shapira et al. 1990; Basilicoet al. 1992; Smith and Stevens 1995; Whorlow et al. 1996; Suh etal. 1997; Czarnecka et al. 1998; Smith et al.1999; Murzenok et al.2000).

Porsolt forced swimming test

Each mouse was placed individually in a glass cylinder (diameter12 cm, height 24 cm) filled with water up to 12 cm. Watertemperature was maintained at 22–23OC. The animal was forced toswim for 15 min on day 1, and then returned to its home cage. Onday 2, the mouse was placed again into the water and forced toswim for 6 min. The session was videotaped and the duration ofimmobility was measured during the last 5 min. The mouse wasconsidered as immobile when it stopped struggling and remainedfloating, moving only to keep its head above water. Igmesine ordesipramine was administered 30 min before the session on day 2.Ca2+ modulators were administered 10 min before the session.

Experimental design

Ca2+ modulators were administered ICV, as previously described(Smith et al. 1995; Maurice et al. 1998; Smith et al. 1999). Briefly,a microsyringe with a 28-gauge stainless-steel needle (Exmire; ItoCorporation, Fuji, Japan) 3 mm long was used for all experiments.Mice were anaesthetized lightly with ether, and the needle insertedunilaterally 1 mm to the right of the midline point equidistant fromeach eye, at an equal distance between the eyes and the ears andperpendicular to the plane of the skull. Drugs or vehicle (5 Sl) weredelivered gradually within approximately 30 s. Mice exhibitednormal behavior within 1 min after injection. The administrationsite was confirmed by injecting Indian ink in preliminary exper-iments and our extensive experience has made possible to injectdrugs within the ventricle with greater than 95% accuracy. Neitherinsertion of the needle nor injection of the vehicle had a significantinfluence on survival or behavioral responses.

Adrenalectomy/castration

Mice were anaesthetized with sodium pentobarbital 2%. Bothadrenal glands were removed through incisions in the back of theanimal, just below the breast ribs. The skin was sutured. Both testeswere ligatured and cut through an incision in the scrotum. Animalsreceived an injection of gentamycin 10 mg/kg IP and recoveredwithin few hours from surgery. After surgery, drinking tap waterwas replaced by saccharose 1%, NaCl 0.9%. Animals were used forbehavioral experiments 6 days after surgery.

Fig. 1 Effects of the extracel-lular Ca2+ chelator EGTA onthe antidepressant-like effect ofigmesine in control mice (A)and in AdX/CX mice (B).EGTA also blocked the effectof desipramine (C). Igmesine(60 mg/kg) or desipramine(30 mg/k g) were injected IP20 min before EGTA (13,40 nmol) administered ICV10 min before the forcedswimming. Duration of immo-bility was recorded during thelast 5 min over a 6-min session.Values are expressed asmean€SEM of the number ofanimals indicated inside eachcolumn. **P<0.01 versus theveh+veh-treated group;##P<0.01 versus the igme-sine+veh-treated group or thedesipramine+veh-treated group(Dunnett’s test)

28

Statistical analysis

Results are expressed as means€SEM. Data were analyzed usingthe Dunnett’s multiple comparisons test after a one-way analysis ofvariance (ANOVA, F values). The criterion for statistical signif-icance was P<0.05.

Results

Effects of extracellular Ca2+ modulators

In intact mice, the s1 receptor agonist igmesine reducedimmobility time in the forced swimming test in a dose-dependent and linear manner. Significant effects wereobserved at doses of 60 and 100 mg/kg IP [F(3,27)=24.05,P<0.001] (Fig. 1A). However, the 100 mg/kg dosage inmice is usually considered as excessive. Toxicologicalstudies demonstrated that mice developed at this dosageabdominal cramps, hypermotility, excitation and tailcatatonia, corresponding to appearance of side-effectsunrelated with primary action at s1 receptors. In addition,the drug precipitated within the abdominal cavity. As aresult, the present study was limited to doses of 30 mg/kg

as the low, sub-effective dose and 60 mg/kg as theeffective dose.Central injection of the extracellular Ca2+ chelator

EGTA at 13–40 nmol/mouse had no effect alone onimmobility duration and failed to affect the response ofmice treated with 30 mg/kg igmesine. However, EGTAsignificantly prevented the effect of 60 mg/kg igmesine ina dose-dependent manner [F(8,66)=9.28, P<0.001](Fig. 1B). In AdX/CX mice, the antidepressant-like effectof 60 mg/kg igmesine was enhanced as compared to intactmice, as previously reported (Urani et al. 2001). EGTAcompletely blocked the enhanced effect of igmesine[F(3,26)=9.67, P<0.001] (Fig. 1C). This observationconfirmed that the s1 receptor-mediated behavioral effectcritically involved extracellular Ca2+ influx. The effect ofdesipramine was also blocked by EGTA, at 13 and40 nmol [F(5,45)=9.70, P<0.001] (Fig. 1D).Igmesine, at 30 mg/kg, did not significantly decrease

the duration of immobility (Fig. 2A). However, co-administration with the L-type VDCC positive modulator(–)-Bay K8644 (0.14, 4.2 nmol) resulted in a significantreduction of immobility duration [F(7,88)=3.32, P<0.001](Fig. 2A). The L-type VDCC antagonist verapamil failedto affect the immobility duration of animals treated with

Fig. 2 Effects of VDCC modu-lators on the antidepressant-likeeffect of igmesine: effect of theL-type VDCC antagonist vera-pamil (A), L-type VDCC ago-nist (–)-Bay K8644 (B), and N-type VDCC antagonist w-cono-toxine (C). Igmesine (30 or60 mg/kg) was injected IP20 min before the VDCC mod-ulators, verapamil (5, 10 nmol),(–)-Bay K8644 (0.14–4.2 nmol),and w-conotoxine (0.1,0.3 nmol), administered ICV10 min before the forcedswimming. Duration of immo-bility was recorded during thelast 5 min over a 6-min session.Values are expressed asmean€SEM of the number ofanimals indicated inside eachcolumn. *P<0.05 and **P<0.01versus the veh+veh-treatedgroup; #P<0.05 and ##P<0.01versus the igmesine+veh-treatedgroup (Dunnett’s test)

29

30 mg/kg igmesine, but it significantly blocked the effectof 60 mg/kg igmesine [F(7,104)=5.43, P<0.001] (Fig. 2B).In addition, the N-type VDCC antagonist w-conotoxinGVIA also blocked the effect of 60 mg/kg igmesine

without affecting the response of mice treated at the lowerdose [F(7,74)=414, P<0.001] (Fig. 2C). The Ca2+ mod-ulator effects were all observed at doses ineffective ontheir own. These results showed that Ca2+ influx through

Fig. 3 Effects of the intracellu-lar Ca2+ chelator EGTA/AM onthe antidepressant-like effect ofigmesine in control mice (A)and in AdX/CX mice (B).EGTA/AM did not affect theeffect of desipramine (C).Igmesine (60 mg/kg) or desi-pramine (30 mg/kg) were in-jected IP 20 min before EGTA(75, 225 nmol) administeredICV 10 min before the forcedswimming. Duration of immo-bility was recorded during thelast 5 min over a 6-min session.Values are expressed asmean€SEM of the number ofanimals indicated inside eachcolumn. *P<0.05 and **P<0.01versus the veh+veh-treatedgroup; ##P<0.01 versus theigmesine+veh-treated group orthe desipramine+veh-treatedgroup (Dunnett’s test)

Fig. 4 Effects of the InsP3 receptor modulators on the antidepres-sant-like effect of igmesine: Effect of bradykinin (A), andxestospongin C (B). Igmesine (30 or 60 mg/kg) was injected IP20 min before bradykinin (0.9–13.4 nmol) or xestospongin C (0.33,3.3 nmol), administered ICV 10 min before the forced swimming.

Duration of immobility was recorded during the last 5 min over a 6-min session. Values are expressed as mean€SEM of the number ofanimals indicated inside each column. *P<0.05 and **P<0.01versus the veh+veh-treated group; #P<0.05 and ##P<0.01 versusthe igmesine+veh-treated group (Dunnett’s test)

30

either L- or N-type VDCC is involved in the antidepres-sant-like effect induced by the s1 receptor agonistigmesine.

Effects of intracellular Ca2+ modulators

In intact mice, the diminution of immobility durationinduced by 60 mg/kg igmesine was completely blockedby the membrane permeable intracellular Ca2+ chelatorEGTA/AM, when 75–225 nmol was injected centrally[F(7,53)=10.50, P<0.001] (Fig. 3A). EGTA/AM failed toaffect the immobility duration when injected alone orwith 30 mg/kg igmesine (Fig. 3A). The enhanced effectobserved in AdX/CX animals was also completelyblocked by EGTA/AM [F(3,28)=24.97, P<0.001](Fig. 3B). However, EGTA/AM failed to affect thediminution of immobility duration induced by desipra-mine [F(5,59)=20.62, P<0.001] (Fig. 3C). These resultsindicate that the behavioral effect mediated by the s1receptor is selectively dependent on intracellular Ca2+

mobilization.Bradykinin, which is known to increase InsP3 levels,

had no effect alone in the test, but enhanced the effect of

igmesine, which became efficient at 30 mg/kg[F(7,81)=3.41, P<0.001] (Fig. 4A). The InsP3 receptorantagonist xestospongin C failed to affect the immobilityduration when injected alone or with 30 mg/kg igmesine,but it blocked the effect of 60 mg/kg igmesine[F(7,56)=5.44, P<0.001] (Fig. 4B). These results indicatethat mobilization of intracellular Ca2+ from the InsP3receptor-sensitive pools participates in the behavioraleffect mediated by s1 receptors.Caffeine acts as an agonist of Ca2+/caffeine-sensitive

Ca2+ pools, which are closely related to the ryanodinereceptor. The drug, injected in the 13–772 nmol doserange, failed to affect the reduction of immobility inducedby igmesine at either the dose of 30 or 60 mg/kg[F(14,113)=4.24, P<0.001] (Fig. 5A). However, the Ca2+/caffeine-sensitive Ca2+ pool antagonist ryanodine, inef-fective alone or on the response of animals injected with30 mg/kg igmesine, significantly blocked the effect of60 mg/kg igmesine [F(7,100)=5.53, P<0.001] (Fig. 5B).This contradictory effects of ligands of the Ca2+/caffeine-sensitive Ca2+ pools suggest that these Ca2+ pools couldbe involved in the effect of igmesine, although in a lessdirect manner to that observed for InsP3 receptor-sensitivepools.

Fig. 5 Effects of the ryanodinereceptor modulators on the an-tidepressant-like effect ofigmesine: effect of caffeine (A),and ryanodine (B). Igmesine(30 or 60 mg/kg) was injectedIP 20 min before caffeine (13–772 nmol) or ryanodine (5–10 nmol), administered ICV10 min before the forcedswimming. Duration of immo-bility was recorded during thelast 5 min over a 6-min session.Values are expressed asmean€SEM of the number ofanimals indicated inside eachcolumn. *P<0.05 and **P<0.01versus the veh+veh-treatedgroup; ##P<0.01 versus theigmesine+veh-treated group(Dunnett’s test)

31

Discussion

Manipulating Ca2+ fluxes in vivo in particular structuresassociated with behaviorally relevant functions has been ahard challenge. The recent availability of drugs crossingthe blood-brain barrier and ependymocyte periventricularlayer, thus showing correct bioavailability after centralICV administration, has allowed the study of Ca2+

mobilization in vivo. In particular, the involvement ofintracellular Ca2+ in morphine tolerance in mice wasextensively characterized using similar ICV injectionprotocols as used in the present study (Smith and Stevens1995; Smith et al. 1999). The brain areas reached by thedrugs following the ICV injection procedure, includinghippocampal formation, periventricular cortical layersand periaqueductal gray, are involved in the behavioralresponse to stress (Nikulina et al. 1998; Smith et al. 1999;Urani et al. 2001).The present study showed that the mechanism of the

antidepressant-like effects induced by a tricyclic antide-pressant and a selective s1 receptor agonist are highlyCa2+ dependent. However, marked differences could beobserved in the nature of the level of Ca2+ entry necessaryto the pharmacological response. The effect induced bythe tricyclic antidepressant was only dependent onextracellular Ca2+ influx, being blocked after centraladministration of EGTA. The s1 receptor agonist effectwas dependent on both extracellular Ca2+ influx andmobilization of intracellular Ca2+ pools, its antidepressantaction being blocked not only by EGTA but also by theintracellular Ca2+ chelator EGTA/AM.Extracellular Ca2+, through VDCC-dependent Ca2+

influx, is involved in the presynaptic regulation ofneurotransmitter release. Ca2+ channel blockers inhibitthe release of neurotransmitters involved in depression,such as norepinephrine (Ebstein and Daly 1982) andserotonin (Murakami et al. 1978), and impede the efficacyof tricyclic antidepressants, acting by inhibition ofmonoamine transporters. VDCC blockers thus decreasenorepinephrine and/or serotonin release to such an extentthat even after transporter blockade, the synaptic concen-tration in neurotransmitter could not reach sufficientlevels to reverse the depressive state (Srivastava and Nath2000). Our result, showing that chelating the extracellularCa2+ with EGTA impeded the efficacy of desipramine, isthus highly consistent with previous observations andwith the expected Ca2+-dependency of classical trans-porter inhibitors.The Ca2+ dependency of the effects mediated by s1

receptor agonists may involve completely differentmechanisms. The present results illustrated dual depen-dency upon both extracellular Ca2+ influx and intracel-lular Ca2+ mobilization. The igmesine effect wasexamined on both control and AdX/CX animals, for theexperiments using the Ca2+ chelators EGTA and EGTA/AM. Removing the sources of peripheral steroids, theadrenal and gonadal glands, potentiates the antidepres-sant-like effect of s1 agonists, as observed previously(Urani et al. 2001). The s1 receptor represents a key target

for the rapid, non-genomic modulatory effects of neu-roactive steroids (Su et al. 1988; Monnet et al. 1995; forreview, see Maurice et al. 1999). These steroids includeprogesterone, pregnenolone, dehydroepiandrosterone, andtheir sulfate esters, but also testosterone, which is notsynthesized centrally. Progesterone behaved as a potents1 receptor antagonist. Other steroids such as testoster-one, or metabolites such as dihydrotestosterone, couldshare this action. Their decreased levels in AdX/CX micemay explain the potentiation of s1 receptor-mediatedbehavioral effects (Phan et al. 1999; Urani et al. 2001). Inthe present study, both EGTA and EGTA/AM ICVinjections completely blocked the augmented effect inAdX/CX mice as well as the effect observed in controlanimals, thus demonstrating the specificity of the Ca2+

dependency of the s1 receptor-mediated effect.The involvement of s1 receptor in modulating extra-

cellular Ca2+ influx through a direct effect on VDCC hasalready been suggested in previous studies. First, s1receptor agonists increased contractility, beating rate andCa2+ influx in cultured cardiac myocytes, putativelythrough increased production of InsP3 (Ela et al. 1994;Novakova et al. 1998). Second, DTG decreased the basalintracellular free Ca2+ concentration ([Ca2+]i) in synapto-some preparations and blocked the depolarization-depen-dent increases in [Ca2+]i (Brent et al. 1997). Finally, PRE-084 potentiated, whereas both (+)-pentazocine and theneuroactive steroid pregnenolone sulfate inhibited, thedepolarization-induced increase in cytosolic Ca2+ con-centration in NG108 cells, through specific activation ofthe s1 receptor (Hayashi et al. 2000). In the present study,involvement of both L-type and N-type VDCC in theantidepressant-like effect of igmesine was observed. Onone hand, the L-type VDCC opener (–)-Bay K8644potentiated the igmesine-induced effect, whereas the L-type VDCC blocker verapamil blocked it. On the otherhand, the N-type VDCC blocker w-conotoxin GVIAblocked the effect of the s1 receptor agonist. It is atpresent unclear whether pre- or postsynaptic VDCC areinvolved in the regulation of the s1 receptor-mediatedeffect. The s1 receptor has been observed mainly in thepostsynaptic neurons (Alonso et al. 2000), however, s1receptor agonists may exert their modulation of dopaminerelease from striatal neurons though a presynaptic mech-anism (Gonzalez-Alvear and Werling 1994, 1995; Chakiet al. 1998). Thus, two hypotheses could be proposed.First, presynaptic VDCC could be involved in a similarmanner as proposed for classical antidepressants, i.e. inthe release of the neurotransmitter involved in theresponse to behavioral despair, namely norepinephrineor serotonin. VDCC modulators would thus affect theintensity of the neurotransmission upstream to the neu-romodulatory effect induced by the s1 receptor, within thepostsynaptic neuron. In particular, each VDCC blocker,verapamil or w-conotoxin GVIA, may decrease theintensity of neurotransmitter release, and thus of theactivation of postsynaptic receptors to a level that renderssubsequent activation of the s1 receptor ineffective.Second, postsynaptic VDCC may be activated secondary

32

to the activation of s1 receptor. The s1 receptor, locatedon the endoplasmic reticulum membranes (Alonso et al.2000; Hayashi et al. 2000), is translocated to the vicinityof the plasmic membranes after activation (Morin-Surunet al. 1999; Hayashi and Su 2001). On the plasmamembrane, the s1 receptor interacts and modulates theactivity of VDCC, as observed in previous studies (Brentet al. 1997; Hayashi et al. 2000). In terms of behavioralconsequences, it was previously observed that selectives1 receptor agonists could attenuate the learning impair-ments induced in mice by acute systemic administrationof nimodipine (Maurice et al. 1995). The present studyalso brings a pertinent example of the interaction of s1receptor with VDCC activation, showing that both L-typeand N-type VDCC are involved in the s1 receptor-mediated antidepressant-like effect in mice. The exactmechanism, and particularly the localization of the VDCCinvolved in this effect, has, however, to be determined.Simpler, for example in vitro, models have to be used forsuch characterization.Neurons have two separate intracellular Ca2+ mobili-

zation pathways. One pool mobilizes Ca2+ from theendoplasmic reticulum through InsP3-receptor activation.The other involves a Ca2+ channel closely resembling theryanodine receptor purified from skeletal muscle thatmobilizes the Ca2+ channel upon stimulation by intracel-lular Ca2+ or caffeine. Ryanodine acts as a competitiveantagonist on this channel (for review, see Sutko et al.1997). Previous studies, particularly by Su and colleagues(Tsao and Su 1996; Hayashi et al. 2000; Hayashi and Su2001), showed that the s1 receptor located on endoplas-mic reticulum membranes interacted with InsP3 receptors,but not with Ca2+/caffeine-sensitive Ca2+ pools (Tsao andSu 1996). On NG108 cells, the s1 agonists PRE-084, (+)-pentazocine and pregnenolone sulfate potentiated thebradykinin-induced increase in [Ca2+]i (Hayashi et al.2000). The mechanism of this potentiation involved atrimeric complex formed between the s1 receptor, thecytoskeletal adaptor protein ankyrin B and the InsP3receptor type 3 (Hayashi and Su 2001). The s1 agonistsprovoke the dissociation of ankyrin B from the InsP3receptor, whereas s1 antagonists, NE-100 or the steroidprogesterone, dissociated the s1 receptor alone, ankyrin Bremaining associated with the InsP3 receptor. An excel-lent correlation was found for several ligands betweentheir efficacies in potentiating the bradykinin-inducedincrease in [Ca2+]i and in dissociating ankyrin B from theInsP3 receptor, suggesting that this simple but originalmechanism could constitute the molecular basis of the s1receptor-mediated neuromodulatory action (Hayashi andSu 2001). Results from the present study are in agree-ment, since the igmesine antidepressant-like effect was:(i) potentiated by bradykinin and (ii) inhibited by theInsP3 receptor antagonist xestospongin C. These obser-vations suggest that a direct effect of the s1 receptoragonist on the InsP3 receptor-sensitive Ca

2+ pools, asdescribed in the relevant study on NG108 cells, underliesits antidepressant-like effect.

The involvement of the Ca2+/caffeine-sensitive Ca2+

pools appeared less clear. Caffeine itself failed to affectthe antidepressant-like effect induced by igmesine, withabsence of either potentiation or inhibition, suggestingthat the Ca2+/caffeine-sensitive Ca2+ pools are notinvolved in the s1 receptor-mediated effect. However,ryanodine significantly inhibited the igmesine effect. Thenature of the response to ryanodine is difficult to interpretand may reflect either the complex effects of ryanodineon the conductance and gating of its channels, as observedon cardiac muscle preparation (Sutko et al. 1997) or thecomplex mobilization of Ca2+/caffeine-sensitive Ca2+

pools and interaction of intracellular pools. Moreover,the use of caffeine to activate the ryanodine receptor iscomplicated by the observation that it is an inhibitor ofthe InsP3 receptor (Parker and Ivorra 1991). Furtherstudies on more adequate preparations are necessary toclarify the particular effect of ryanodine in our paradigm.The present study brought evidence that disruptions in

extracellular and intracellular Ca2+ directly affected theantidepressant efficacy of a s1 receptor agonist. Ourresults indicate that extracellular Ca2+ through voltage-sensitive Ca2+ channels and mobilization from InsP3receptor-sensitive Ca2+ pools are involved with the s1receptor-mediated effect. In addition, the effect of aclassical antidepressant was observed to depend solelyupon involvement of extracellular Ca2+ through voltage-sensitive Ca2+ channels. These findings agree with therecent reports by Hayashi and colleagues (Hayashi et al.2000; Hayashi and Su 2001), suggesting that the cellularrole of the s1 receptor could be to regulate the efficacy ofintracellular Ca2+ mobilization.

Acknowledgements The authors acknowledge Drs Michel Vignes(CNRS UMR 5102, Montpellier, France) and Tsung-Ping Su (IRP,NIDA, NIH, Besthesda, Md., USA) for valuable advice during thestudy. This work was supported by Pfizer GRD, Fresnes.

References

Akunne HC, Zoski KT, Whetzel SZ, Cordon JJ, Brandon RM,Roman F, Pugsley TA (2001) Neuropharmacological profile ofa selective sigma ligand, igmesine: a potential antidepressant.Neuropharmacology 41:138–149

Alonso G, Phan V, Guillemain I, Saunier M, Legrand A, Anoal M,Maurice T (2000) Immunocytochemical localization of the s1receptor in the adult rat central nervous system. Neuroscience97:155–170

Basilico L, Parolaro D, Rubino T, Gori E, Giagnoni G (1992)Influence of w-conotoxin on morphine analgesia and with-drawal syndrome in rats. Eur J Pharmacol 218:75–81

Berridge MJ (1993) Inositol trisphosphate and calcium signaling.Nature 361:315–325

Bouchard P, Roman F, Junien JL, Quirion R (1996) Autoradio-graphic evidence for the modulation of in vivo sigma receptorlabeling by neuropeptide Y and calcitonin gene-related peptidein the mouse brain. J Pharmacol Exp Ther 276:223–230

Bourson A, Moser PC, Gower AJ, Mir AK (1989) Central andperipheral effects of the dihydropyridine calcium channelactivator BAY K 8644 in the rat. Eur J Pharmacol 160:339–347

Brent PJ, Herd L, Saunders H, Sim AT, Dunkley PR (1997) Proteinphosphorylation and calcium uptake into rat forebrain synap-

33

tosomes: modulation by the sigma ligand, 1,3-ditolylguanidine.J Neurochem 68:2201–2211

Chaki S, Okuyama S, Ogawa S, Tomisawa K (1998) Regulation ofNMDA-induced [3H]dopamine release from rat hippocampalslices through sigma-1 binding sites. Neurochem Int 33:29–34

Czarnecka E, Tymczyszyn W, Pietrzak B (1998) Effect ofnifedipine and verapamil on hypotensive action of clonidinein rabbits. Pol J Pharmacol 50:193–201

Dar MS (1988) The biphasic effects of centrally and peripherallyadministered caffeine on ethanol-induced motor incoordinationin mice. J Pharm Pharmacol 40:482–487

Debonnel G, Bergeron R, de Montigny C (1996) Potentiation bydehydroepiandrosterone of the neuronal response to N-methyl-d-aspartate in the CA3 region of the rat dorsal hippocampus: aneffect mediated via s (sigma) receptors. J Endocrinol 150:S33–42

Earley B, Burke M, Leonard BE, Gouret CJ, Junien JL (1991)Evidence for an anti-amnesic effect of JO 1784 in the rat: apotent and selective ligand for the sigma receptor. Brain Res546:282–286

Ebstein RP, Daly JW (1982) Release of norepinephrine anddopamine from brain vesicular preparations: effects of calciumantagonists. Cell Mol Neurobiol 2:205–213

Ela C, Barg J, Vogel Z, Hasin Y, Eilam Y (1994) Sigma receptorligands modulate contractility, Ca++ influx and beating rate incultured cardiac myocytes. J Pharmacol Exp Ther 269:1300–1309

Gonzalez-Alvear GM, Werling LL (1994) regulation of [3H]dopa-mine release from rat striatal slices by sigma ligands. J Phar-macol Exp Ther 271:212–219

Gonzalez-Alvear GM, Werling LL (1995) Sigma1 receptors in ratstriatum regulate NMDA-stimulated [3H]dopamine release viaa presynaptic mechanism. Eur J Pharmacol 294:713–719

Graybiel AM, Besson MJ, Weber E (1989) Neuroleptic-sensitivebinding sites in the nigrostriatal system: evidence for differen-tial distribution of sigma sites in the substantia nigra, parscompacta of the cat. J Neurosci 9:326–338

Gronier B, Debonnel G (1999) Involvement of sigma receptors inthe modulation of the glutamatergic/NMDA neurotransmissionin the dopaminergic systems. Eur J Pharmacol 368:183–196

GuG M, Junien JL, Del Rio C, Bueno L (1992a) Neuropeptide Yand sigma ligand (JO 1784) suppress stress-induced colonicmotor disturbances in rats through sigma and cholecystokininreceptors. J Pharmacol Exp Ther 261:850–855

GuG M, Yoneda M, Monnikes H, Junien JL, Tache Y (1992b)Central neuropeptide Y and the sigma ligand, JO 1784, reversecorticotropin-releasing factor-induced inhibition of gastric acidsecretion in rats. Br J Pharmacol 107:642–647

GuG M, Gleizes-Escala C, Del Rio-Lacheze C, Junien JL, Bueno L(1994) Reversal of CRF- and dopamine-induced stimulation ofcolonic motility by CCK and igmesine (JO 1784) in the rat. BrJ Pharmacol 111:930–934

Gundlach AL, Largent BL, Snyder SH (1986) Autoradiographiclocalization of sigma receptor binding sites in guinea pig andrat central nervous system with (+)3H-3-(3-hydroxyphenyl)-N-(1-propyl)piperidine. J Neurosci 6:1757–1770

Hanner M, Moebius FF, Flandorfer A, Knaus HG, Striessnig J,Kempner E, Glossmann H (1996) Purification, molecularcloning, and expression of the mammalian sigma1-binding site.Proc Natl Acad Sci USA 93:8072–8077

Hayashi T, Su TP (2001) Regulating ankyrin dynamics: roles ofsigma-1 receptors. Proc Natl Acad Sci USA 98:491–496

Hayashi T, Kagaya A, Takebayashi M, Shimizu M, Uchitomi Y,Motohashi N, Yamawaki S (1995) Modulation by sigmaligands of intracellular free Ca2+ mobilization by N-methyl-d-aspartate in primary culture of rat frontal cortical neurons.J Pharmacol Exp Ther 275:207–214

Hayashi T, Maurice T, Su TP (2000) Ca2+ signaling via sigma1-receptors: novel regulatory mechanism affecting intracellularCa2+ concentration. J Pharmacol Exp Ther 293:788–798

Jansen KL, Faull RL, Dragunow M, Leslie RA (1991) Distributionof excitatory and inhibitory amino acid, sigma, monoamine,

catecholamine, acetylcholine, opioid, neurotensin, substance P,adenosine and neuropeptide Y receptors in human motor andsomatosensory cortex. Brain Res 566:225–238

Junien JL, Roman FJ, Brunelle G, Pascaud X (1991) JO1784, anovel sigma ligand, potentiates [3H]acetylcholine release fromrat hippocampal slices. Eur J Pharmacol 200:343–345

Kitaichi K, Chabot J, Moebius FF, Flandorfer A, Glossmann H,Quirion R (2000) Expression of the purported sigma1 (s1)receptor in the mammalian brain and its possible relevance indeficits induced by antagonism of the NMDA receptor complexas revealed using an antisense strategy. J Chem Neuroanat20:375–387

Largent BL, Gundlach AL, Snyder SH (1986) Pharmacological andautoradiographic discrimination of sigma and phencyclidinereceptor binding sites in brain with (+)-[3H]SKF 10,047, (+)-[3H]-3-[3-hydroxyphenyl]-N-(1-propyl)piperidine and [3H]-1-[1-(2-thienyl)cyclohexyl]piperidine. J Pharmacol Exp Ther238:739–748

Laurent S, Girerd X, Tsoukaris-Kupfer D, Legrand M, Huchet-Brisac AM, Schmitt H (1987) Opposite central cardiovasculareffects of nifedipine and Bay K8644 in anesthetized rats.Hypertension 9:132–138

Matsuno K, Kobayashi T, Tanaka MK, Mita S (1996) Sigma 1receptor subtype is involved in the relief of behavioral despairin the mouse forced swimming test. Eur J Pharmacol 312:267–271

Maurice T, Su TP, Parish DW, Privat A (1995) Prevention ofnimodipine-induced impairment of learning by the selective sligand PRE-084. J Neural Transm 102:1–18

Maurice T, Roman FJ, Privat A (1996a) Modulation by neuros-teroids of the in vivo (+)-[3H]SKF-10,047 binding to sigma1receptors in the mouse forebrain. J Neurosci Res 46:734–743

Maurice T, Roman FJ, Su TP, Privat A (1996b) Beneficial effectsof sigma agonists on the age-related learning impairment in thesenescence-accelerated mouse (SAM). Brain Res 733:219–230

Maurice T, Junien JL, Privat A (1997) Dehydroepiandrosteronesulfate attenuates dizocilpine-induced learning impairment inmice via sigma1 receptors. Behav Brain Res 83:159–164

Maurice T, Su TP, Privat A (1998) Sigma1 (s1) receptor agonistsand neurosteroids attenuate b25–35-amyloid peptide-inducedamnesia in mice through a common mechanism. Neuroscience83:413–428

Maurice T, Phan VL, Urani A, Kamei H, Noda Y, Naneshima T(1999) Neuroactive steroids as endogenous effectors for thesigma1 (s1) receptor: pharmacological evidence and therapeuticopportunities. Jpn J Pharmacol 81:125–155

McLean S, Weber E (1988) Autoradiographic visualization ofhaloperidol-sensitive sigma receptors in guinea-pig brain.Neuroscience 25:259–269

Monnet FP, Debonnel G, de Montigny C (1992a) In vivoelectrophysiological evidence for a selective modulation of N-methyl-d-aspartate-induced neuronal activation in rat CA3dorsal hippocampus by sigma ligands. J Pharmacol Exp Ther261:123–130

Monnet FP, Blier P, Debonnel G, de Montigny C (1992b)Modulation by sigma ligands of N-methyl-d-aspartate-induced[3H]noradrenaline release in the rat hippocampus: G-proteindependency. Naunyn Schmiedeberg’s Arch Pharmacol 346:32–39

Monnet FP, MahG V, Robel P, Baulieu EE (1995) Neurosteroids,via sigma receptors, modulate the [3H]norepinephrine releaseevoked by N-methyl-d-aspartate in the rat hippocampus. ProcNatl Acad Sci USA 92:3774–3778

Morin-Surun MP, Collin T, Denavit-SaubiG M, Baulieu EE,Monnet FP (1999) Intracellular s1 receptor modulates phos-pholipase C and protein kinase C activation in the brain stem.Proc Natl Acad Sci USA 96:8196–8199

Murakami H, Kaji E, Segawa T (1978) Influences of verapamil, X-537A, A-23187 and adenosine 3’,5’-cyclic monophosphate onrelease of 5-hydroxytryptamine from rat brain slices. JpnJ Pharmacol 28:589–596

34

Murzenok PP, Huang BS, Leenen FH (2000) Sympathoinhibitionby central and peripheral infusion of nifedipine in sponta-neously hypertensive rats. Hypertension 35:631–636

Nikulina EM, Marchand JE, Kream RM, Miczek KA (1998)Behavioral sensitization to cocaine after a brief social stress isaccompanied by changes in fos expression in the murinebrainstem. Brain Res 810:200–210

Noda Y, Kamei H, Kamei Y, Nagai T, Nishida M, Nabeshima T(2000) Neurosteroids ameliorate conditioned fear stress: anassociation with sigma receptors. Neuropsychopharmacology23:276–284

Novakova M, Ela C, Bowen WD, Hasin Y, Eilam Y (1998) Highlyselective sigma receptor ligands elevate inositol 1,4,5-trispho-sphate production in rat cardiac myocytes. Eur J Pharmacol353:315–317

Okuyama S, Chaki S, Yae T, Nakazato A, Muramatsu M (1995)Autoradiographic characterization of binding sites for [3H]NE-100 in guinea pig brain. Life Sci 57:PL333–337

Parker I, Ivorra I (1991) Caffeine inhibits inositol trisphosphate-mediated liberation of intracellular calcium in Xenopus oocytes.J Physiol 433:229–240

Phan VL, Su TP, Privat A, Maurice T (1999) Modulation ofsteroidal levels by adrenalectomy/castration and inhibition ofneurosteroid synthesis enzymes affect s1 (sigma1) receptor-mediated behaviour in mice. Eur J Neurosci 11:2385–2396

Prasad PD, Li HW, Fei YJ, Ganapathy ME, Fujita T, Plumley LH,Yang-Feng TL, Leibach FH, Ganapathy V (1998) Exon-intronstructure, analysis of promoter region, and chromosomallocalization of the human type 1 sigma receptor gene.J Neurochem 70:443–451

Reddy DS, Kaur G, Kulkarni SK (1998) Sigma (s1) receptormediated antidepressant-like effects of neurosteroids in thePorsolt forced swim test. Neuroreport 9:3069–3073

Roman FJ, Pascaud X, Martin B, Vauche D, Junien JL (1990) JO1784, a potent and selective ligand for rat and mouse brainsigma-sites. J Pharm Pharmacol 42:439-440

Shapira S, Adeyemo OM, Feuerstein G (1990) Integrated auto-nomic and behavioral responses to L/N Ca2(+)-channel blockeromega-conotoxin in conscious rats. Am J Physiol 259:R427–R438

Smith FL, Stevens DL (1995) Calcium modulation of morphineanalgesia: role of calcium channels and intracellular poolcalcium. J Pharmacol Exp Ther 272:290–299

Smith FL, Dombrovsky DS, Dewey WL (1999) Involvement ofintracellular calcium in morphine tolerance in mice. PharmacolBiochem Behav 62:381–388

Song C, Earley B, Leonard BE (1997) Effect of chronic pretreat-ment with the sigma ligand JO 1784 on CRF-induced changesin behaviour, neurotransmitter and immunological function inthe rat. Neuropsychobiology 35:200–204

Srivastava SK, Nath C (2000) The differential effects of calciumchannel blockers in the behavioral despair test in mice.Pharmacol Res 42: 293–297

Su TP (1982) Evidence for sigma opioid receptor: binding of[3H]SKF-10047 to etorphine-inaccessible sites in guinea-pigbrain. J Pharmacol Exp Ther 223:284–290

Su TP, London ED, Jaffe JH (1988) Steroid binding at sigmareceptors suggests a link between endocrine, nervous, andimmune systems. Science 240:219–221

Suh HW, Song DK, Choi SR, Huh SO, Kim YH (1997) Differentialeffect of w-conotoxin GVIA, nimodipine, calmidazolium andKN-62 injected intrathecally on the antinociception induced byb-endorphin, morphine and [d-Ala2,N-MePhe4,Gly-ol5]-en-kephalin administered intracerebroventricularly in the mouse.J Pharmacol Exp Ther 282:961–966

Sutko JL, Airey JA, Welch W, Ruest L (1997) The pharmacologyof ryanodine and related compounds. Pharmacol Rev 49:53–98

Thomas AP, Bird GS, Hajnoczky G, Robb-Gaspers LD, Putney JWJr (1996) Spatial and temporal aspects of cellular calciumsignaling. FASEB J 10:1505–1517

Tsao LI, Su TP(1996) IP3 receptor antagonist heparin uncompet-itively inhibits [3H](+)-SKF-10047 binding to sigma receptors.Eur J Pharmacol 311:R1–2

Urani A, Privat A, Maurice T (1998) The modulation byneurosteroids of the scopolamine-induced learning impairmentin mice involves an interaction with sigma1 (s1) receptors.Brain Res 799:64–77

Urani A, Roman FJ, Phan VL, Su, TP, Maurice T (2001) Theantidepressant-like effect induced by s1-receptor agonists andneuroactive steroids in mice submitted to the forced swimmingtest. J Pharmacol Exp Ther 298:1269–1279

Whorlow SL, Loiaccono RE, Angus JA, Wright CE (1996)Distribution of N-type Ca2+ channel binding sites in rabbitbrain following central administration of w-conotoxin GVIA.Eur J Pharmacol 315:11–18

Yamada M, Nishigami T, Nakasho K, Nishimoto Y, Miyaji H(1994) Relationship between sigma-like site and progesterone-binding site of adult male rat liver microsomes. Hepatology20:1271–1280

Zamanillo D, Andreu F, Ovalle S, Perez MP, Romero G, Farre AJ,Guitart X (2000) Up-regulation of sigma1 receptor mRNA in ratbrain by a putative atypical antipsychotic and sigma receptorligand. Neurosci Lett 282:169–172

Zhang J, Chiodo LA, Wettstein JG, Junien JL, Freeman AS (1993)Repeated administration of Sigma ligands alters the populationactivity of rat midbrain dopaminergic neurons. Synapse13:223–230

35