Effects of chronic paroxetine treatment on dialysate serotonin in 5-HT 1B receptor knockout mice

Transcription

1 Journal of Neurochemistry, 2003, 86, doi: /j x Effects of chronic paroxetine treatment on dialysate serotonin in 5-HT 1B receptor knockout mice A. M. Gardier,* D. J. David,*, G. Jego,* C. Przybylski,* C. Jacquot,* S. Durier,* B. Gruwez,* E. Douvier, P. Beauverie, N. Poisson, R. Henà and M. Bourin *Laboratoire de Neuropharmacologie EA3544 MENRT, Faculté de Pharmacie IFR75-ISIT Institut de Signalisation et d Innovation Thérapeutique, Université Paris-Sud, Châtenay-Malabry, France Pharmacie, Centre Hospitalier Spécialisé Paul Guiraud, Villejuif, France àcenter for Neurobiology and Behaviour, Columbia University, New York, New York, USA Laboratoire de Pharmacologie, Faculté de Médecine, Université de Nantes, Nantes, France Abstract The role of serotonin (5-HT) 1B receptors in the mechanism of action of selective serotonin re-uptake inhibitors (SSRI) was studied by using intracerebral in vivo microdialysis in conscious, freely moving wild-type and 5-HT 1B receptor knockout (KO 5-HT 1B ) mice in order to compare the effects of chronic administration of paroxetine via osmotic minipumps (1 mg per kg per day for 14 days) on extracellular 5-HT levels ([5-HT]ext) in the medial prefrontal cortex and ventral hippocampus. Basal [5-HT]ext values in the medial prefrontal cortex and ventral hippocampus, 20 h after removing the minipump, were not altered by chronic paroxetine treatment in both genotypes. On day 15, in the ventral hippocampus, an acute paroxetine challenge (1 mg/kg i.p.) induced a larger increase in [5-HT]ext in saline-pretreated mutant than in wild-type mice. This difference between the two genotypes in the effect of the paroxetine challenge persisted following chronic paroxetine treatment. Conversely, in the medial prefrontal cortex, the paroxetine challenge increased [5-HT]ext similarly in saline-pretreated mice of both genotypes. Such a challenge produced a further increase in cortical [5-HT]ext compared with that in salinepretreated groups of both genotypes, but no differences were found between genotypes following chronic treatment. To avoid the interaction with raphe 5-HT 1A autoreceptors, 1 lm paroxetine was perfused locally through the dialysis probe implanted in the ventral hippocampus; similar increases in hippocampal [5-HT]ext were found in acutely or chronically treated wild-type mice. Systemic administration of the mixed 5-HT 1B/1D receptor antagonist GR (4 mg/kg) in chronically treated wildtype mice potentiated the effect of a paroxetine challenge dose on [5-HT]ext in the ventral hippocampus, whereas systemic administration of the selective 5-HT 1A receptor antagonist WAY did not. By using the zero net flux method of quantitative microdialysis in the medial prefrontal cortex and ventral hippocampus of wild-type and KO 5-HT 1B mice, we found that basal [5-HT]ext and the extraction fraction of 5-HT were similar in the medial prefrontal cortex and ventral hippocampus of both genotypes, suggesting that no compensatory response to the constitutive deletion of the 5-HT 1B receptor involving changes in 5-HT uptake capacity occurred in vivo. As steady-state brain concentrations of paroxetine at day 14 were similar in both genotypes, it is unlikely that differences in the effects of a paroxetine challenge on hippocampal [5-HT]ext are due to alterations of the drug s pharmacokinetic properties in mutants. These data suggest that there are differences between the ventral hippocampus and medial prefrontal cortex in activation of terminal 5-HT 1B autoreceptors and their role in regulating dialysate 5-HT levels. These presynaptic receptors retain their capacity to limit 5-HT release mainly in the ventral hippocampus following chronic paroxetine treatment in mice. Keywords: antidepressant drug, 5-hydroxytryptamine 1B autoreceptor, intracerebral microdialysis, knockout mice, paroxetine, selective serotonin re-uptake inhibitor. J. Neurochem. (2003) 86, Received June 26, 2002; revised manuscript received October 29, 2002; accepted January 26, Address correspondence and reprint requests to A. M. Gardier, Laboratoire de Neuropharmacologie Tour D1, 2ème étage, EA MENRT, Faculté de Pharmacie, Université Paris-Sud, 5 rue J-B. Clément, Châtenay-Malabry cedex, France. alain.gardier@cep.u-psud.fr Abbreviations used: AUC, area under the curve; DMSO, dimethyl sulfoxide; Ed, extraction fraction; 5-HT, serotonin; [5-HT]ext, extracellular serotonin level; KO 5-HT1B, 5-HT1B receptor knockout; SSRI, selective serotonin re-uptake inhibitor. 13

2 14 A. M. Gardier et al. It has been hypothesized for decades that major depressive episodes are associated with hypofunctioning of the central serotonergic system (Meltzer and Lowy 1987). In rodents, numerous intracerebral in vivo microdialysis studies have shown that selective serotonin re-uptake inhibitors (SSRIs) increase synaptic serotonin (5-hydroxytryptamine; 5-HT) availability by selective inhibition of the 5-HT transporter, which blocks 5-HT re-uptake. These experiments have generally demonstrated that a single exposure to SSRIs increases extracellular levels of serotonin ([5-HT]ext) preferentially near the cell bodies and dendrites of serotoninergic neurones in the raphe nuclei rather than in brain regions innervated by serotonergic nerve terminals such as the medial prefrontal cortex (e.g. fluvoxamine, Bel and Artigas 1992; fluoxetine, Malagié et al. 1996). However, the major inconvenience of SSRIs is their long delay of action similar to that of other antidepressant drugs (Nemeroff 1993). Indeed, clinical studies have shown that 2 4 weeks of treatment are required before symptoms of depression begin to be alleviated despite an immediate blockade of 5-HT transporter as demonstrated in vitro (Fuller 1994). This delay corresponds to the time of desensitization of inhibitory 5-HT autoreceptors. By using electrophysiological recordings, several lines of evidence have suggested (1) that the activation of somatodendritic 5-HT 1A autoreceptors located in the dorsal raphe nucleus limits the effects of a single dose of SSRI on the availability of 5-HT in the synapse, and (2) that functional desensitization of somatodendritic 5-HT 1A autoreceptors occurs during long-term treatment with a SSRI, thus allowing an enhancement of 5-HT synaptic transmission (Blier and De Montigny 1983; Le Poul et al. 1997, 2000). These data regarding the somatodendritic 5-HT 1A autoreceptors have been confirmed in intracerebral in vivo microdialysis studies (see Hjorth et al for a review). However, using this technique in rat forebrain area, it is not clear whether or not presynaptic 5-HT 1B receptors desensitize following chronic SSRI treatment. Terminal 5-HT 1B receptors are located in the CNS on nerve endings of serotonergic neurones (autoreceptors) and possibly on varicosities within the median raphe nucleus (Boschert et al. 1994; Adell et al. 2001) as well as postsynaptically on nonserotonergic neurones (heteroreceptors). The presynaptic receptors are preferentially involved in a local inhibitory control of 5-HT release from the serotonergic nerve endings (Engel et al. 1986; Göthert et al. 1987; Trillat et al. 1997). Intracerebral in vivo microdialysis studies carried out in rats to test for 5-HT 1B receptor function following chronic SSRI treatment have drawn contradictory conclusions regarding desensitization of 5-HT 1B receptors in various brain terminal regions. Several studies found changes in the sensitivity of 5-HT 1B receptors in response to chronic SSRI treatment (Sayer et al. 1999; Dremencov et al. 2000), whereas others did not (Auerbach and Hjorth 1995; Bosker et al. 1995a, 1995b; Hjorth and Auerbach 1999; Cremers et al. 2000). For example, a challenge dose of non-selective 5-HT 1A/1B receptor agonist or antagonist, such as RU24969 or penbutolol respectively, still had an effect on [5-HT]ext in the dorsal hippocampus following chronic SSRI treatment (21 days with fluvoxamine, Bosker et al. 1995a; 14 days with citalopram, Gundlah et al. 1997). These contradictory results may be explained by the poor selectivity of the drugs used to test for 5-HT 1B receptor desensitization; RU24969 and ( )penbutolol (Hjorth and Sharp 1993; Gundlah et al. 1997) are a non-selective 5-HT 1A/1B receptor agonist and antagonist respectively. Regional differences in the response may also explain the heterogenous results (most of the time only one brain region was studied in rats). In addition, the inclusion of a wash-out period before microdialysis sessions during repeated administration of a SSRI seems to be a critical factor in drug challenge studies with 5-HT 1B receptor ligands (Moret and Briley 1996). To understand further the role of 5-HT 1B receptors in the mechanism of action of SSRIs, homozygous mice lacking this serotonin receptor subtype (5-HT 1B receptor knockout; KO 5-HT 1B ) have been generated by homologous recombination (Saudou et al. 1994). By using intracerebral in vivo microdialysis in freely moving KO 5-HT 1B mice, we demonstrated recently that a single administration of a SSRI (paroxetine, fluoxetine) induced a larger increase in hippocampal [5-HT]ext in KO 5-HT 1B than in wild-type mice (Malagié et al. 2001, 2002). Our data thus suggest that this terminal autoreceptor limits the effects of a single administration of a SSRI in the hippocampus of wild-type mice. In the present study, by comparing these wild-type and KO 5-HT 1B genotypes, we examined whether or not terminal 5-HT 1B autoreceptors retain their inhibitory capacity following chronic paroxetine administration. We hypothesized that, if 5-HT 1B presynaptic receptors are desensitized following paroxetine treatment for 14 days (1 mg per kg per day delivered via osmotic minipumps implanted i.p.), then an acute paroxetine challenge (1 mg/kg given 20 h after the removal of the osmotic minipump) should have a comparable effect on [5-HT]ext in paroxetine-treated wild-type and saline-treated KO 5-HT 1B mice. In addition, to test for putative differences in the pharmacokinetic properties of paroxetine between the two genotypes, brain concentrations of paroxetine were measured in the steady state (on day 14, i.e. before removing the osmotic minipumps), as well as 20 h after the removal of minipumps (on day 15, i.e. when the maximal effect of a challenge dose of paroxetine on [5- HT]ext was observed in the ventral hippocampus and medial prefrontal cortex). Previous data have suggested that interaction occurs between the plasma membrane 5-HT transporter and 5-HT 1B receptors (Daws et al. 2000). Basal serotonin dynamics may therefore be altered in projection areas of raphe nuclei serotonergic neurones in mice lacking 5-HT 1B

3 5-HT 1B autoreceptors and chronic antidepression 15 receptors. In a second part of the study, we compared the efficacy of the 5-HT transporter in re-uptaking 5-HT in naive, saline-treated wild-type and KO 5-HT 1B mice. For this purpose, we performed the zero net flux method of quantitative in vivo microdialysis (Parsons and Justice 1994); we quantified basal [5-HT]ext and the extraction fraction of 5-HT (Ed), which provided an in vivo index of 5-HT uptake in the ventral hippocampus and medial prefrontal cortex. Materials and methods Animals The founders of the wild-type and mutant colonies used in the present study were the product of heterozygous matings made at the animal facility of Columbia University. These founders were shipped to France, and their offspring were bred and reared in independent colonies as described previously (Malagié et al. 2001). Wild-type and KO 5-HT 1B mice were obtained from a pure 129/Sv genetic background. Group-housed mice were kept in standard cages on a 12-h light/12-h dark cycle with light onset at hours. Mice had free access to food and water. All procedures used in these studies were performed under the guidelines of the French Ministry of Agriculture for experiments with laboratory animals (law no ) and were approved by the appropriate local committee. Drugs and treatment Groups of 8 16 wild-type or KO 5-HT 1B mice were treated for 14 days with either paroxetine hydrochloride (SmithKline Beecham, Harlow, UK) or vehicle (dimethyl sulfoxide (DMSO) : NaCl 0.9%; 20 : 80 v/v). To allow continuous delivery of paroxetine, the SSRI was administered by means of osmotic minipumps (model 1002; Alzet, Alza Corporation, Palo Alto, CA, USA) that delivered a solution of paroxetine dissolved in vehicle. After filling these pumps so that the mice would receive a paroxetine dose of 1 mg per kg per day for 14 days at a flow rate of 0.25 ll/h, pumps were placed at 37 C for at least 4 h to reach the steady state of diffusion. The low paroxetine dose chosen was similar to that used recently in rats (Malagié et al. 2000) and mice (Malagié et al. 2001). Mice were lightly anaesthetized with ether and minipumps were implanted i.p. under aseptic conditions on day 0. Mice were then housed separately for 14 days. Between and hours on day 14, minipumps were removed under aseptic conditions, microdialysis probes were implanted in wild-type and KO 5-HT 1B mice, then the animals were returned to their home cages. For acute paradigms, naive saline-treated wild-type and KO 5-HT 1B mice were implanted with a minipump that delivered the vehicle (DMSO : NaCl 0.9%, 20 : 80 v/v) for 14 days, and on day 15 received an acute challenge dose of paroxetine (1 mg/kg i.p.) dissolved in saline in a volume of 10 ml/kg. Control wild-type and KO 5-HT 1B mice received the vehicle (DMSO : NaCl 0.9%, 20 : 80 v/v) for 14 days, and a single i.p. injection of saline on day 15. Dialysis procedure To allow wash-out of paroxetine, the minipumps were removed the day before dialysis sampling, i.e. when dialysis probes were implanted in the mouse brain. Thus, on day 14 concentric dialysis probes were made from cuprophan fibres and constructed as described previously (Trillat et al. 1997). All probes had an active length of 1.6 and 2 mm for the medial prefrontal cortex and ventral hippocampus respectively (outer diameter (OD) 0.30 mm). Animals were anaesthetized with chloral hydrate (400 mg/kg, i.p.) and their osmotic minipumps were removed. Two probes were implanted and cemented in place, one in the right medial prefrontal cortex and the other in the left ventral hippocampus, according to a mouse brain atlas (Franklin and Paxinos 1997) (coordinates: from bregma, medial prefrontal cortex, anterior + 2.0, lateral + 1.2, ventral ) 1.6; ventral hippocampus, anterior ) 2.8, lateral ) 3.0, ventral ) 4.0). The animals were allowed to recover from the surgery overnight. On day 15, approximately 20 h after surgery, the probes were continuously perfused with an artificial CSF (NaCl 147 mm, KCl 3.5 mm, CaCl mm, MgCl mm, NaH 2 PO mm, NaHCO mm; ph 7.4 ± 0.2) at a flow rate of 1.3 ll/min using a CMA/100 pump (Carnegie Medicin, Stockholm, Sweden). Dialysate samples were collected every 15 min in small Eppendorf tubes and were analysed for 5-HT by HPLC (XL-ODS, mm, particle size 3 lm; Beckman, Roissy, France) coupled to an amperometric detector (1049 A; Hewlett-Packard, Les Ulis, France). Four fractions were collected to measure basal values (mean ± SEM) before systemic administration of the drugs. The limit of sensitivity for 5-HT was 0.5 fmol/sample (signal-tonoise ratio 2). The acute challenge dose of paroxetine (1 mg/kg) was then administered. At the end of the experiments, placement of microdialysis probes was verified histologically. To study the effects of a local paroxetine infusion through the dialysis probe located in the ventral hippocampus, wild-type mice received either the vehicle (acute treatment) or paroxetine (chronic treatment, at a dose of 1 mg per kg per day for 14 days) delivered by osmotic minipumps. On day 15, approximately 20 h after the removal of the minipumps, the first eight samples were collected to measure basal [5-HT]ext in the vhpc. We then performed a local paroxetine infusion (1 lm at a flow rate of 1.3 ll/min for 2 h) as the drug challenge, and collected the next eight samples. For interaction studies, either the mixed 5-HT 1B/1D receptor antagonist, GR (4 mg/kg) (N-[4-methoxy-3(4-methylpiperazin-1-yl)phenyl]-2 -methyl-4 -(5-methyl-1,2,4-oxadiazol-3-yl)biphenyl-4-carboxamide) or the selective 5-HT 1A receptor antagonist, WAY (1 mg/kg) (N-[2-[4-2-methoxyphenyl-1-piperazinyl]- ethyl]-n-(2-pyridinyl)cyclohexane carboxamide trihydrochloride) was dissolved in water in a volume of 12.5 ml/kg and administered s.c. 60 min after the i.p. administration of paroxetine or saline (NaCl 0.9%). HPLC analysis of whole-brain paroxetine concentrations Two groups of wild-type mice and two groups of KO 5-HT 1B mice received paroxetine at a dose of 1 mg per kg per day for 14 days via an osmotic minipump implanted in the peritoneum. On day 14, osmotic minipumps were removed from two groups of lightly anaesthetized mice (one wild-type group and one group of knockout mice) and that the remaining two groups (one of each) was killed by cervical dislocation to measure steady-state concentrations of paroxetine in the whole-brain tissue. Paroxetine concentrations were also measured on day 15, approximately 20 h after the removal of the minipumps, i.e. at the time when microdialysis experiments were

4 16 A. M. Gardier et al. Fig. 1 Effects of a challenge dose of paroxetine (1 mg/kg) on extracellular 5-HT levels in the ventral hippocampus of wild-type (WT) and KO 5-HT 1B (KO) mice chronically pretreated with either paroxetine [NaCl 0.9% ( ) or paroxetine (j) as a challenge dose] or vehicle (DMSO : NaCl 0.9%, 20 : 80 v/v) [NaCl 0.9% (m) or paroxetine (d) as a challenge dose]. Chronically treated mice received paroxetine at a dose of 1 mg per kg per day for 14 days delivered via osmotic minipumps. Data are expressed in (a) and (b) as a percentage of baseline levels of 5-HT in wild-type and KO 5-HT 1B mice respectively. Each point represents the mean ± SEM of 8 16 determinations per group. (c) AUC values. Each bar represents the mean ± SEM of [5-HT]ext measured from 0 to 60 min after drug administration (symbolized by an arrow) and expressed as percentage of basal values. ***p < versus corresponding group receiving NaCl 0.9% as a challenge (oneway ANOVA followed by a Fisher protected least significant differences (PLSD) test); #p < 0.05, ##p < 0.01 (two-way ANOVA carried on for mice who received a challenge dose of paroxetine). performed and the effect of a challenge dose of paroxetine on extracellular 5-HT levels was maximal in the ventral hippocampus and medial prefrontal cortex (see Figs 1 and 2). The whole brain was dissected out and paroxetine was extracted from brain homogenates, after adding haloperidol as an internal standard. Fig. 2 Effects of a challenge dose of paroxetine (1 mg/kg) on extracellular 5-HT levels in the medial prefrontal cortex of wild-type (WT) and 5-HT 1B receptor knockout (KO 5-HT 1B ) mice chronically pretreated with either paroxetine [NaCl 0.9% ( ) or paroxetine (j) as a challenge dose] or vehicle (DMSO : NaCl 0.9%, 20 : 80 v/v) [NaCl 0.9% (m) or paroxetine (d) as a challenge dose]. Chronically treated mice received paroxetine at a dose of 1 mg per kg per day for 14 days delivered via osmotic minipumps. Data are expressed in (a) and (b) as a percentage of baseline levels of 5-HT in wild-type and KO 5-HT 1B mice respectively. Each point represents the mean ± SEM of 8 16 determinations per group. (c) AUC values. Each bar represents the mean ± SEM of [5-HT]ext measured from 0 to 60 min after drug administration (symbolized by an arrow) and expressed as percentage of basal values. ***p < versus corresponding group receiving NaCl 0.9% as a challenge (one-way ANOVA followed by a Fisher PLSD test). p < 0.05, p < 0.01 (two-way ANOVA carried on for mice who received a challenge dose of paroxetine).

5 5-HT 1B autoreceptors and chronic antidepression 17 Paroxetine was quantified by HPLC with UV detection (214 nm) according to an analytical method similar to that previously described for fluoxetine (Malagié et al. 1996), with minor modifications. Essentially, the chromatographic column was a C8, mm internal diameter column with particle size 5 lm (Kromasil, Les Ulis, France), the elution solvent was a mixture of phosphate buffer 0.05 M (55 volumes) and acetonitrile (45 volumes), ph 3.5, and the pump flow rate was maintained at 1.3 ml/min. The limit of detection of the paroxetine assay was around 20 nmol/l using approximately 300 mg tissue (signal-to-noise ratio 2). Mean ± SEM paroxetine concentrations were calculated for each group. Zero net flux method and data analysis We performed the zero net flux method of quantitative microdialysis in wild-type and KO 5-HT 1B mice. The extraction fraction of 5-HT (Ed, slope of the regression line) provides an index of 5-HT uptake in vivo (Justice 1993; Parsons and Justice 1994). Different 5-HT concentrations (C in ; 0, 5.0, 10 and 20 nm) were perfused for 30 min through the dialysis probe in the medial prefrontal cortex and ventral hippocampus. A linear equation was constructed from dialysate 5-HT concentrations (C out ) collected during 5-HT perfusion in the two brain areas studied for each animal (as described by Shippenberg et al. 2000). The net change in 5-HT (C in ) C out ) was regressed against C in. Data analysis and statistics The basal value of [5-HT]ext was calculated from the first four samples collected. All subsequent samples, as well as the area under the curve (AUC) values calculated as the amount of 5-HT outflow collected during the 0 60 min after treatment from the ventral hippocampus and medial prefrontal cortex, were expressed as percentage of basal values. Statistical analyses were performed using the computer software StatView 4.02 (Abacus Concepts, Inc, Berkely, CA, USA). For each brain structure, a three-way ANOVA on AUC values was performed, with the chronic drug treatment (saline, paroxetine 1 mg per kg per day), the mice genotype (wild type or KO 5-HT 1B ) and the single challenge injection performed on day 15 (saline, paroxetine 1 mg/kg) as main factors. The significance level was set at p < For the zero net flux experiment one-way ANOVA was used to assess the effects of genotype on [5-HT]ext and E d in each brain region of wild-type and KO 5-HT 1B mice. Results Effects of chronic administration of paroxetine on basal values of dialysate 5-HT in the ventral hippocampus and medial prefrontal cortex of wild-type and KO 5-HT 1B mice The basal [5-HT]ext in the two brains regions and both strains was measured 20 h after the removal of the osmotic minipumps (Table 1). In the ventral hippocampus, the chronic paroxetine treatment did not modify the basal [5-HT]ext (F 1,92 ¼ 0.49, p ¼ 0.48) and this level did not differ significantly between the two genotypes (F 1,92 ¼ 0.08, p ¼ 0.78). Similarly, in the medial prefrontal cortex, basal [5-HT]ext did not differ significantly between the two genotypes (F 1,80 ¼ 0.71, p ¼ 0.40) or between the two treated groups (F 1,80 ¼ 0.07, p ¼ 0.79). Effects of an acute paroxetine challenge on dialysate 5-HT in the ventral hippocampus of wild-type and KO 5-HT 1B after treatment with paroxetine for 14 days In the ventral hippocampus (Fig. 1), ANOVA of AUC values revealed significant main effects of genotype (F 1,96 ¼ 4.76, p < 0.05) and acute paroxetine challenge (F 1,96 ¼ 128.6, p < 0.001), but no significant effect of chronic treatment (F 1,96 ¼ 0.06, p ¼ 0.8). Thus, a single systemic challenge dose of paroxetine (1 mg/kg, i.p.) significantly increased [5-HT]ext in the two genotypes studied. However, paroxetine induced a larger increase in [5-HT]ext in KO 5-HT 1B than in wild-type mice, but this increase was not modified by chronic treatment with paroxetine for 14 days. Effects of an acute paroxetine challenge on dialysate 5-HT in the medial prefrontal cortex of wild-type and KO 5-HT 1B after treatment with paroxetine for 14 days In the medial prefrontal cortex (Fig. 2), ANOVA on AUC values revealed significant main effects of chronic treatment (F 1,49 ¼ 10.0, p < 0.01) and acute paroxetine challenge (F 1,76 ¼ 54.7, p < 0.001), but no significant genotype factor (F 1,49 ¼ 0.59, p ¼ 0.44). Thus, a single systemic challenge Table 1 Effects of chronic paroxetine treatment on basal values of extracellular 5-HT levels in the ventral hippocampus and medial prefrontal cortex of wild-type and KO 5-HT 1B mice 5-HT (fmol per 20 ll) Wild type KO 5-HT 1B Ventral hippocampus Vehicle (DMSO : NaCl 0.9%, 20 : 80 v/v) 4.2 ± 0.4 (n ¼ 22) 4.3 ± 0.3 (n ¼ 22) Paroxetine (1 mg per kg per day for 14 days) 4.0 ± 0.3 (n ¼ 26) 4.1 ± 0.3 (n ¼ 26) Medial prefrontal cortex Vehicle (DMSO : NaCl 0.9%, 20 : 80 v/v) 2.8 ± 0.3 (n ¼ 19) 3.0 ± 0.3 (n ¼ 20) Paroxetine (1 mg per kg per day for 14 days) 3.2 ± 0.3 (n ¼ 21) 3.1 ± 0.3 (n ¼ 23) Values are mean ± SEM. Paroxetine was delivered 14 days by osmotic minipumps implanted in the peritoneum. Exposure to paroxetine was stopped by the removal of the minipump 20 h before the beginning of the microdialysis experiment.

6 18 A. M. Gardier et al. dose of paroxetine (1 mg/kg, i.p.) induced a significant increase in [5-HT]ext in both saline- and paroxetinepretreated wild-type and KO 5-HT 1B mice. In addition, paroxetine induced a larger increase in [5-HT]ext in mice that had been treated chronically with paroxetine for 14 days than in naive saline-treated mice, but this increase was similar in wild-type and KO 5-HT 1B mice. Whole-brain paroxetine concentrations The brain tissue concentrations of paroxetine were measured following chronic administration for 14 days (Table 2). On day 14, brain paroxetine concentrations were not statistically different in wild-type and KO 5-HT 1B mice (p ¼ 0.28, Student unpaired t-test). On day 15, 20 h after the removal of osmotic minipumps, i.e. when the effect of a paroxetine challenge on [5-HT]ext was maximal in forebrain regions, paroxetine concentrations were below the detection limit of the paroxetine assay in the whole brain of both genotypes. Basal 5-HT dynamics in the ventral hippocampus and medial prefrontal cortex of naive, non-treated wild-type and KO 5-HT 1B mice Figure 3 shows the zero net flux plot of the change in perfusate 5-HT concentration (C in ) C out ) as a function of perfusate 5-HT concentration (C in ) in the ventral hippocampus of wild-type and KO 5-HT 1B mice. The concentration at which C in ) C out equals zero corresponds to equilibrium conditions and provides an estimate of the actual [5-HT]ext (Justice 1993; Parsons and Justice 1994). The slope of the regression line is equal to the E d whereas the y-intercept corresponds to 5-HT levels that would be obtained in a conventional dialysis experiment. [5-HT]ext and E d expressed as absolute values are shown in Fig. 3 (inset). A summary of the results obtained is shown in Table 3. Basal [5-HT]ext did not differ between the two genotypes (F 1,7 ¼ 0.019, p > 0.89). No differences between genotypes in dialysate 5-HT levels (y-intercept; p > 0.36) or Table 2 Whole-brain paroxetine concentrations in wild-type and KO 5-HT 1B mice following chronic treatment with the drug Paroxetine (nmol/l) Wild type KO 5-HT 1B Day ± ± 8.8 Day 15 ND ND Values are mean ± SEM (four to five mice). To test for putative differences in the pharmacokinetic properties of paroxetine between the two genotypes, brain concentrations of paroxetine were measured at the steady state (on day 14, i.e. before removing the osmotic minipumps), and 20 h after the removal of the minipumps (on day 15, i.e. when the effect of the challenge dose was maximal). Mice were killed by cervical dislocation. ND, not detectable. Fig. 3 Zero net flux plot of basal 5-HT dynamics in the ventral hippocampus of wild-type (WT) and KO 5-HT 1B (KO) mice. Plots show the mean ± SEM gain or loss of 5-HT (C in ) C out ) as a function of C in (0, 5, 10, 20 nm 5-HT) and the average linear regression of the data in the two genotypes. The C in at which C in ) C out ¼ 0 equals the [5- HT]ext and the slope of linear regression corresponds to the extracellular fraction of the probe (E d ). The y-intercept corresponds to dialysate levels that would be obtained in a conventional dialysis experiment. No differences were observed between genotypes for E d. Inset: Mean ± SEM E d and [5-HT]ext calculated for each genotype. These variables were not significantly different between the two genotypes studied. Table 3 Basal 5-HT dynamics in wild-type and KO 5-HT 1B mice Medial prefrontal cortex Ventral hippocampus Wild type Basal [5-HT]ext (nm) 1.14 ± ± 0.38 Ed 0.49 ± ± 0.06 Dialysate (y-intercept) (nm) 0.56 ± ± 0.26 KO 5-HT 1B Basal [5-HT]ext (nm) 1.44 ± 0.31 nm 1.66 ± 0.40 Ed 0.45 ± ± 0.11 Dialysate (y-intercept) (nm) 0.50 ± ± 0.03 Values are mean ± SEM (n ¼ 7 8). Values for [5-HT]ext, E d and dialysate levels were obtained for each animal by constructing a linear equation of dialysate 5-HT concentrations obtained during local perfusion of various concentrations of 5-HT. There were no significant differences between genotypes. E d (F 1,7 ¼ 0.638, p > 0.45) were observed in the ventral hippocampus. Figure 4 shows the zero net flux plot of basal 5-HT dynamics in the medial prefrontal cortex as a function of genotype. No differences were observed between the genotypes in basal [5-HT]ext (F 1,14 ¼ 1.62, p > 0.22) or dialysate 5-HT levels (y-intercept; F 1,13 ¼ 0.126, p > 0.73) or for E d (F 1,14 ¼ 0.18, p > 0.68) in the medial prefrontal cortex.

7 5-HT 1B autoreceptors and chronic antidepression 19 Table 4 Comparison of the effect of a local paroxetine challenge on extracellular 5-HT levels in the ventral hippocampus in wild-type mice following acute or a chronic antidepressant treatment [5-HT]ext (fmol per 20 ll) Acute Chronic Basal 4.52 ± 0.24 (n ¼ 6) 4.38 ± 0.30 (n ¼ 7) Local paroxetine 9.91 ± 0.72*** (n ¼ 6) 8.54 ± 0.63*** (n ¼ 7) infusion (1 lm) into the ventral hippocampus Fig. 4 Zero net flux plot of changes in perfusate 5-HT concentration (C in ) C out ) as a function of perfusate 5-HT concentration (C in ) in the medial prefrontal cortex of wild-type (WT) and KO 5-HT 1B (KO) mice. The plots show the mean ± SEM gain or loss of 5-HT (C in ) C out ) as a function of C in (0, 5, 10, 20 nm 5-HT) and the average linear regression of the data in the two genotypes. No differences were observed between genotypes for E d. Inset: Mean ± SEM E d and [5-HT]ext values calculated for each genotype. These variables were not significantly different between the two genotypes studied. Effects of local paroxetine challenge (1 lm) on dialysate 5-HT in the ventral hippocampus of wild-type mice after chronic paroxetine treatment Systemic administration of paroxetine as a challenge may lead to an indirect activation of somatodendritic 5-HT 1A receptors located in the dorsal raphe nucleus and terminal 5-HT 1B autoreceptors located in the ventral hippocampus. By using a local paroxetine infusion through the dialysis probe implanted in the ventral hippocampus of wild-type mice receiving chronic paroxetine treatment, we should observe the involvement of terminal 5-HT 1B autoreceptors only, thus avoiding the interaction with raphe 5-HT 1A autoreceptors. An intrahippocampal paroxetine infusion led to an increase in [5-HT]ext in the ventral hippocampus in acute and chronic experiments (Table 4). Local paroxetine infusion produced a 1.9- to 2.2-fold increase in hippocampal [5-HT]ext in chronically and acutely treated wild-type mice respectively (p < versus baseline). There were no statistically significant differences in basal [5-HT]ext concentrations after paroxetine infusion between acute or chronic paradigms in the hippocampus of wild-type mice. Effects of co-administration of paroxetine with GR or WAY on dialysate 5-HT in the ventral hippocampus of wild-type mice after chronic paroxetine treatment In the ventral hippocampus in wild-type mice chronically treated with paroxetine (Fig. 5), ANOVA on AUC values revealed significant effects of an acute paroxetine challenge combined with a 5-HT autoreceptor antagonist (F 1,21 ¼ 6.54, Values are mean + SEM. Wild-type mice received either vehicle (DMSO : NaCl 0.9%, 20 : 80 v/v) (acute treatment) or paroxetine (chronic treatment, at a dose of 1 mg per kg per day) for 14 days delivered via osmotic minipumps. We used a local paroxetine infusion (1 lm at a flow rate of 1.3 ll/min lasting for 2 h) through the dialysis probe located into the ventral hippocampus 20 hours after the removal of the minipumps. ***p < versus basal value for the acute or chronic experiment (one-way ANOVA followed by Fisher PLSD test). p < 0.01). A single systemic challenge dose of paroxetine (1 mg/kg, i.p.) followed by systemic administration of the mixed 5-HT 1B/1D receptor antagonist GR (4 mg/kg, s.c.) significantly increased [5-HT]ext in the ventral hippocampus of wild-type mice. The AUC value in this group of mice was significantly higher than that found in wild-type mice receiving either paroxetine alone or paroxetine plus WAY , a selective 5-HT 1A receptor antagonist. Discussion By using intracerebral in vivo microdialysis, we recently showed the administration of a single dose of a SSRI (paroxetine, Malagié et al. 2001; fluoxetine, Malagié et al. 2002) induced a smaller increase in [5-HT]ext mainly in the ventral hippocampus of wild-type than in KO 5-HT 1B mice. These results led us to hypothesize that a suboptimal increase in [5-HT]ext at serotonergic forebrain nerve terminals in wild-type mice might be caused by indirect activation of 5-HT 1B autoreceptors by the excess of endogenous 5-HT that follows the selective blockade of the 5-HT transporter by a SSRI. Evidence in favour of this hypothesis came from the fact that the mixed 5-HT 1B/1D receptor antagonist GR potentiated the effect of a single administration of paroxetine on [5-HT]ext in the ventral hippocampus of wild-type mice (Malagié et al. 2001). In mice lacking 5-HT 1B receptors, this negative feedback control of the amount of intrasynaptic 5-HT is missing, and 5-HT 1B autoreceptors do not limit the efficacy of SSRIs. Furthermore, in wild-type mice, repeated treatment with a SSRI may desensitize terminal 5-HT 1B autoreceptors, such this negative feedback mechanism may no longer be functional. An acute

8 20 A. M. Gardier et al. (a) (b) Prx + GR Prx + WAY Prx + NaCl Fig. 5 Effects of autoreceptor blockade on paroxetine-induced increase in dialysate 5-HT in the ventral hippocampus of wild-type mice chronically treated with paroxetine (1 mg per kg per day) for 14 days. Mice received a paroxetine challenge (1 mg/kg, i.p.; first arrow), then either saline or the mixed 5-HT 1B/1D receptor antagonist GR (4 mg/kg, s.c.; second arrow) or the selective 5-HT 1A receptor antagonist WAY (1 mg/kg, s.c.). (a) Mean ± SEM [5- HT]ext levels expressed as percentage of baseline following treatment with paroxetine (Prx) + NaCl (n ¼ 6), Prx + WAY (n ¼ 12) or Prx + GR (n ¼ 6) respectively. Basal values were 5.01 ± 0.24 fmol per 20 ll in Prx + NaCl group, 5.25 ± 0.18 fmol per 20 ll in Prx + WAY group and 4.52 ± 0.32 fmol per 20 ll in Prx + GR group in the ventral hippocampus of wild-type mice. (b) Mean ± SEM AUC values calculated for the 5-HT outflow collected for min after paroxetine administration expressed as percentage of basal values. **p < 0.01 versus group chronically treated with paroxetine for 14 days, then challenged with paroxetine (1 mg/kg i.p.) on day 15 (one-way ANOVA followed by a Fisher PLSD test). challenge dose of paroxetine would then be able to further increase [5-HT]ext at serotonergic nerve terminals (Bel and Artigas 1993), i.e. may produce an effect on hippocampal [5- HT]ext in paroxetine-treated wild-type mice similar to that measured in saline-treated KO 5-HT 1B mice; this was not observed in the present study in the ventral hippocampus. Basal [5-HT]ext levels on day 15 were not significantly different in the brain regions of the four different groups of mice studied [wild-type or KO 5-HT 1B, chronically treated or not (naive) with paroxetine 1 mg per kg per day for 14 days]. These results are consistent with some microdialysis studies performed in rats, for example with citalopram (Auerbach and Hjorth 1995; Invernizzi et al. 1995; Cremers et al. 2000) or fluvoxamine (Bosker et al. 1995a, 1995b). By contrast, Bel and Artigas (1993), using osmotic minipumps to deliver fluvoxamine for 14 days, found a threefold to sixfold increase in basal [5-HT]ext in the medial prefrontal cortex. However, this effect was observed when the pumps were still implanted in the peritoneum of the animals, i.e. when they were still exposed to the drug. Numerous neurochemical studies with SSRIs have been performed in rats (Bosker et al. 1995a, 1995b; Pineyro and Blier 1996; Gobbi et al. 1997; Hjorth and Auerbach 1999; Cremers et al. 2000) after choosing an appropriate wash-out period, as was used here with paroxetine. We believe that this wash-out period is important to rule out effects caused by the presence of residual drug such that only the consequences of the effect of an acute challenge dose of paroxetine on dialysate [5-HT]ext are observed. On day 14, we found no differences in steady-state brain concentrations of paroxetine between the two genotypes. On day 15, brain tissue paroxetine concentrations were measured when the effect of the paroxetine challenge on dialysate [5-HT]ext was maximal in the two brain regions studied. Indeed, 20 h after removal of the minipumps, brain paroxetine concentrations had dropped below the detection limit of the paroxetine assay (20 nmol/l) in both genotypes, confirming that the wash-out period was sufficient to allow the elimination most of the paroxetine from the brain. These data are fully compatible with the short plasma elimination half-life of paroxetine in rodents (approximately 4 h in rats; Caccia et al. 1993). Taken together with our microdialysis data, these results suggest that differences found in [5-HT]ext between wild-type and KO 5-HT 1B mice are unlikely explained by differences in brain concentrations of paroxetine. Unfortunately, the lack of sensitivity of the paroxetine assay did not allow obtain data for mouse hippocampus and cortex, but only for the whole brain of these animals. It is not known whether 20 nm is a pharmacologically active concentration of the drug in these mice. It is possible that a partial blockade of the 5-HT transporter still occurs in vivo at this low brain concentration in mice as paroxetine is able to specifically inhibit [ 3 H]5-HT uptake into rat brain synaptosomes with a Ki of 1.1 nm in vitro (Thomas et al. 1987). When an acute challenge dose of paroxetine (1 mg/kg) was administered 20 h after cessation of this chronic treatment, different effects on dialysate [5-HT]ext were observed in the two genotypes and in the two brain areas studied. In the ventral hippocampus, as expected (Malagié et al. 2001), a challenge dose of paroxetine induced a larger increase in [5-HT]ext in vehicle-treated KO 5-HT 1B mice than in vehicle-treated wild-type mice. Such a potentiated response measured in mutants compared with wild-type mice was still observed following chronic treatment with paroxetine. The 5-HT outflow in the ventral hippocampus of wildtype mice did not reach that measured in KO 5-HT 1B mice. These results suggest that terminal 5-HT 1B autoreceptors retain their capacity to limit the paroxetine-induced increase

9 5-HT 1B autoreceptors and chronic antidepression 21 in [5-HT]ext in the ventral hippocampus following chronic paroxetine treatment in wild-type mice. However, interpretation of these data is complicated by the fact that regional patterns of compensation involving somatodendritic 5-HT 1A receptors in KO 5-HT 1B mice have been found by some authors, but not by others. Various pharmacological tests have been used. Administration of the selective 5-HT 1A receptor agonist R-8-hydroxy-2(di-n-propylamino)tetralin (R-8-OH-DPAT) evoked a significantly diminished response on dialysate 5-HT levels in the ventral hippocampus (but not in the striatum) of these mutants, suggesting the potential desensitization of 5-HT 1A receptors in the median raphe nucleus (Knobelman et al. 2001). By measuring peripheral responses (control of body temperature and heart rate), no compensation involving 5-HT 1A autoreceptors was found in KO 5-HT 1B mice (Bouwknecht et al. 2002), whereas we reported an adaptive central thermoregulatory process involving this 5-HT 1A autoreceptor in these mutants following administration of 8-OH-DPAT (Gardier et al. 2001). In the present study, we confirmed the results obtained by authors using the microdialysis technique in rats with various SSRIs (Auerbach and Hjorth 1995; Bosker et al. 1995a, 1995b; Cremers et al. 2000); no functional desensitization of terminal 5-HT 1B autoreceptors occurred in the ventral hippocampus following treatment with paroxetine (1 mg per kg per day) for 14 days. Such a low-dose regimen ensures the selectivity of the serotonergic response on the 5-HT transporter and was found to be active mainly on hippocampal dialysate 5-HT after a single administration in mice (Malagié et al. 2001). Furthermore, a single 0.5-mg/kg s.c. dose of paroxetine increased hippocampal 5-HT levels about twofold in guinea-pigs (Cremers et al. 2001). A 0.8- mg/kg dose of paroxetine did not increase dialysate 5-HT levels in the frontal cortex of anaesthetized rats (Gartside et al. 1995). Such a low dose causes a significant blockade of 5-HT re-uptake, in that it increases dialysate 5-HT in the dorsal raphe nucleus and causes a complete cessation of 5-HT neuronal activity in this region. A higher dose (2.4 mg/kg) caused an increase (to 200% of baseline) in dialysate 5-HT in the frontal cortex (Gartside et al. 1995). It is conceivable that either a longer duration of paroxetine treatment, or repeated administration of a higher paroxetine dose, would amplify the changes we observed. However, our treatment schedule was already efficacious as we observed a potentiation of the effect of paroxetine in the medial prefrontal cortex of both genotypes at a dose of 1 mg per kg per day. In this latter brain region, this potentiation was not likely to involve terminal 5-HT 1B autoreceptors as it occurred in both genotypes after treatment with paroxetine for 14 days; a challenge dose of paroxetine induced similar increases in [5-HT]ext in saline-treated wild-type and salinetreated KO 5-HT 1B mice. In addition, a further increase in cortical [5-HT]ext occurred in both paroxetine-treated wildtype and paroxetine-treated KO 5-HT 1B mice. These results suggest that no functional desensitization of terminal 5-HT 1B autoreceptors occurred in the medial prefrontal cortex. Regarding the effect of paroxetine in the medial prefrontal cortex, there is a discrepancy between the results of our present study and a previous one (Malagié et al. 2001), in which a paroxetine dose of 1 mg/kg increased cortical [5- HT]ext in KO 5-HT 1B, but not in wild-type mice. We believe that this increase was randomly obtained for two main reasons. First, two-way ANOVA of AUC values (percentage of basal 5-HT) revealed no significant genotype factor in our previous work (Malagié et al. 2001). In the present study we again found no significant genotype factor in the medial prefrontal cortex, and significant main effects of a paroxetine challenge, which was probably due to the chronic paroxetine pretreatment. Second, to confirm the absence of differences in the paroxetine response between the two genotypes in the medial prefrontal cortex, we examined the effects of a single lower paroxetine dose (0.5 mg/kg) on cortical [5-HT]ext in 129/Sv wild-type and KO 5-HT 1B mice. We found no differences between the two genotypes, as increases in [5-HT]ext (AUC percentage) induced in the medial prefrontal cortex by such an acute paroxetine challenge were ± 12% (n ¼ 8) and ± 10.8% (n ¼ 8) in wildtype and KO 5-HT 1B mice respectively. It is therefore likely that the potentiation of the paroxetine-induced increase in [5-HT]ext (relative to wild-type mice) occurred in the ventral hippocampus, but not in the frontal cortex of KO 5-HT 1B mice. Paroxetine induced moderate increases in cortical [5-HT]ext in both wild-type and KO 5-HT 1B mice, these effects being lower than those we observed in the hippocampus. We previously found, by administering a single systemic dose of SSRI, that terminal 5-HT 1B autoreceptors limit the effects of SSRI preferentially in the ventral hippocampus compared with the medial prefrontal cortex (Malagié et al. 2001, 2002, for paroxetine and fluoxetine respectively). Our results suggest that, in the medial prefrontal cortex, a specific trait shared by wild-type and KO 5-HT 1B mice was similarly modified, e.g. desensitization of an autoreceptor different from the 5-HT 1B subtype, which normally limits the efficacy of SSRIs in this particular brain region of both genotypes. The ventral hippocampus and the medial prefrontal cortex are preferentially innervated by the median and dorsal raphe nuclei respectively (Jacobs and Azmitia 1992; Hervas et al. 1998, 2000). Somatodendritic 5-HT 1A autoreceptors exert a relatively greater negative feedback control over the dorsal raphe nucleus (and changes in cortical [5-HT]ext) than in the median raphe nucleus (and changes in hippocampal [5-HT]ext) (Malagié et al. 1996; Gardier et al. 1996; Hjorth et al. 2000). It is therefore likely that the well demonstrated desensitization of somatodendritic 5-HT 1A autoreceptors located in the dorsal raphe nucleus occurred in the median prefrontal cortex of paroxetinetreated wild-type mice as well as in KO 5-HT 1B mice. This

10 22 A. M. Gardier et al. effect thus led to a potentiation of paroxetine-induced increases in [5-HT]ext in the medial prefrontal cortex of both genotypes. In the present study, we tested this hypothesis in the hippocampus, by studying the effect of WAY , a selective 5-HT 1A receptor antagonist, on paroxetine-induced increases in [5-HT]ext. To our knowledge, only one study (Gundlah et al. 1997) has used microdialysis in conscious rats chronically treated with a SSRI and obtained results suggesting that there may be differences between two nerve terminal brain regions in the regulation of 5-HT release. Gundlah and co-workers (1997) studied the effects of a challenge dose of WAY (0.3 mg/kg, s.c.), and did not to find a desensitization of somatodendritic 5-HT 1A autoreceptors in the frontal cortex (compared with the dorsal hippocampus) in rats chronically treated with citalopram. It is therefore not surprising that, under our experimental conditions, administration of WAY (1 mg/kg, s.c.) did not potentiate paroxetineinduced increases in [5-HT]ext in wild-type mice following chronic paroxetine treatment. This lack of effect could not be attributable to down-regulation of somatodendritic 5-HT 1A autoreceptors as the ventral hippocampus is not very sensitive to this selective 5-HT 1A receptor antagonist in rats acutely treated with a SSRI (Malagié et al. 1996; Hervas et al. 1998; Hervas et al. 2000). As systemic paroxetine administration may also activate somatodendritic 5-HT 1A receptors, comparison of the effects of a local 1-lM paroxetine infusion by reverse dialysis between wild-type and KO 5-HT 1B mice might help to clarify the role of terminal 5-HT 1B autoreceptors in the mechanism of action of SSRI following acute and chronic treatment. Here, local paroxetine infusion doubled hippocampal [5-HT]ext in both acutely and chronically treated wild-type mice. This result suggests that terminal 5-HT 1B autoreceptors in the ventral hippocampus are not desensitized following chronic paroxetine treatment. Similar results have been found in the rat frontal cortex and dorsal hippocampus with 1 lm citalopram perfused for 14 days (Auerbach and Hjorth 1995). The increased hippocampal (in KO 5-HT 1B mice) and cortical (in both genotypes) 5-HT output may indicate a reduced efficacy of the 5-HT transporter in KO 5-HT 1B mice. Similar effects have already been reported in rat brain regions following repeated administration of SSRI (Pineyro et al. 1994). However, a higher density of 5-HT re-uptake sites (specifically labelled by [ 3 H]citalopram), and therefore a higher re-uptake capacity, was found in the dorsal raphe nucleus of KO 5-HT 1B mice (Evrard et al. 1999). To test this hypothesis in vivo, we used the zero net flux method of quantitative microdialysis and compared the efficacy of the 5-HT transporter in naive, saline-treated wild-type and KO 5-HT 1B mice. The 5-HT transporter has been found to be up- or down-regulated in KO 5-HT 1B mice depending on the brain region studied (Ase et al. 2001). For example, the density of the 5-HT transporter was increased in the ventral hippocampus of these mutants. This latter change was correlated with a 5-HT hyperinnervation in this brain region of mutants, which may partly explain our microdialysis data obtained under pharmacological conditions in KO 5-HT 1B mice following administration of paroxetine. Our results do not support this hypothesis because the functional status of the 5-HT transporter was not altered in vivo in KO 5-HT 1B mice, either in the ventral hippocampus or in the medial prefrontal cortex. In the two brain areas studied, we found that basal [5-HT]ext and Ed did not differ between wild-type and KO 5-HT 1B mice. These results suggest that constitutive deletion of the 5-HT 1B receptor is not associated with alteration of 5-HT uptake. In summary, the present findings obtained by use of pharmacological and null mutation approaches in mice indicate that dialysate [5-HT]ext in the ventral hippocampus is mainly under the control of terminal 5-HT 1B autoreceptors and that there is no functional desensitization of 5-HT 1B receptors in the ventral hippocampus following repeated treatment with a low dose of paroxetine; increases in [5-HT]ext induced by a challenge dose of paroxetine did not reach the levels of those measured in saline-treated KO 5-HT 1B mice. We observed identical potentiated responses in the ventral hippocampus of saline-treated and paroxetine-treated KO 5-HT 1B mice. In the medial prefrontal cortex, a challenge dose of paroxetine induced similar increases in [5-HT]ext in saline-treated wild-type and saline-treated KO 5-HT 1B mice. Furthermore, a further increase in cortical [5-HT]ext was measured in both paroxetine-treated wild-type and paroxetine-treated KO 5-HT 1B mice. Our results reinforce the hypothesis recently developed by Hjorth et al. (2000) suggesting an accessory role of nerve terminal 5-HT 1B over somatodendritic 5-HT 1A autoreceptors in restraining increases in [5-HT]ext induced by SSRIs. References Adell A., Celada P. and Artigas F. (2001) The role of 5-HT 1B receptors in the regulation of serotonin cell firing and release in the rat brain. J. Neurochem. 79, Ase A. R., Reader T. A., Hen R., Riad M. and Descarries L. (2001) Regional changes in density of serotonin transporter in the brain of 5-HT 1A and 5-HT 1B knock-out mice, and of serotonin innervation in the 5-HT1B knockout. J. Neurochem. 78, Auerbach S. B. and Hjorth S. (1995) Effect of chronic administration of the selective serotonin (5-HT) uptake inhibitor citalopram on extracellular 5-HT and apparent autoreceptor sensitivity in rat forebrain in vivo. Naunyn Schmiedebergs Arch. Pharmacol. 352, Bel N. and Artigas F. (1992) Fluvoxamine preferentially increases extracellular 5-hydroxytryptamine in the raphe nuclei: an in vivo microdialysis study. Eur. J. Pharmacol. 229, Bel N. and Artigas F. (1993) Chronic treatment with fluvoxamine increases extracellular serotonin in frontal cortex but not in raphe nuclei. Synapse 15,