Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Extensive experimental and clinical evidence indicates that a perturbation of monoaminergic transmission is involved in depressive states (Maes and Meltzer, 1995; Willner, 1995; Ressler and Nemeroff, 1999; Anand and Charney, 2000). Correspondingly, currently used antidepressant agents exert their therapeutic actions via a reinforcement in monoaminergic transmission, although the relative contribution of serotonergic, adrenergic, and dopaminergic mechanisms remains to be clarified (Broekkamp et al., 1995; Burke and Preskorn, 1995; Frazer, 1997; Millan et al., 2000b). Monoamine oxidase inhibitors exert antidepressant actions by preventing the degradation of 5-HT and NE, while mianserin and mirtazapine display antagonist properties at ₂-adrenoceptors (AR) and 5-HT_2C receptors inhibitory to serotonergic, adrenergic, and/or dopaminergic pathways (Burke and Preskorn, 1995; Frazer, 1997; Millan et al., 2000a). Nevertheless, the majority of antidepressant agents reinforce monoaminergic transmission by blocking neuronal uptake of 5-HT and/or NE via actions at SERTs and NETs, respectively (Barker and Blakely, 1995; Frazer, 1997; Sambunaris et al., 1997; Goodnick and Goldstein, 1998; Blakely and Bauman, 2000).

Despite subtle differences among selective 5-HT reuptake inhibitors (SSRIs), such as fluoxetine and the highly selective agent citalopram (Owens et al., 1997; Sánchez and Meier, 1997; Tatsumi et al., 1997; Goodnick and Goldstein, 1998; Popik, 1999), all elevate extracellular levels of 5-HT in corticolimbic structures upon acute and chronic administration (Blier and de Montigny, 1994; Goodnick and Goldstein, 1998; Millan et al., 2000b). On the other hand, the recently introduced morpholine derivative, reboxetine, interacts preferentially with NETs versus SERTs in vitro and markedly increases extracellular levels of NE (Burrows et al., 1998; Riva et al., 1999; Sacchetti et al., 1999). In this respect, reboxetine resembles the first-generation, tricylic agent desipramine (Owens et al., 1997). However, other tricyclic antidepressants, such as clomipramine, interact with both NETs and with SERTs (Burke and Preskorn, 1995; Tatsumi et al., 1997). This dual activity is likewise displayed by the cyclohexanol derivative, venlafaxine, which lacks the undesirable histaminergic, muscarinic, and ₁-AR antagonist properties of clomipramine and other tricyclic agents (Muth et al., 1991; Schweizer et al., 1997; Harvey et al., 2000). Clinical studies have demonstrated the efficacy of venlafaxine in major depressive states, including resistant and, probably, bipolar patients (Amsterdam, 1998; Poirier and Boyer, 1999). It may also possess a rapid onset of activity (Derivan et al., 1995) and be associated with a high remission rate (Ferrier, 1999), although this remains to be confirmed. In addition, venlafaxine improves generalized anxiety disorders (Rickels et al., 2000) and social phobia (Altamura et al., 1999). These observations have triggered a resurgence of interest in antidepressants uniting activity at SERTs and NETs in vivo, with the aim of optimizing clinical efficacy, achieving more rapid action, and enlarging therapeutic reach (Broekkamp et al., 1995; Frazer, 1997; Sambunaris et al., 1997).

Venlafaxine, however, displays only modest affinity for SERTs, and its affinity for NETs is weak (Muth et al., 1991; Owens et al., 1997; Tatsumi et al., 1997; Béïque et al., 1999, 2000). It would thus be of interest to obtain agents of greater potency. In this light, we have generated a series of benzocyclobutane derivatives possessing affinity for both SERTs and NETs. In the present and accompanying papers, the pharmacological profile of one of these, S33005 (Fig. 1), is compared with venlafaxine, citalopram, reboxetine, and clomipramine.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Chemical structure of S33005.

First, we determined their affinities at native, rat, and cloned human (h) SERTs, NETs, and DA transporters (DAT) and evaluated uptake of 5-HT, NE, and DA into cerebral rat synaptosomes. Second, we examined the ability of drugs to prevent in vivo depletion of cerebral pools of 5-HT by parachloroamphetamine (PCA), which exerts its actions following entry into serotonergic terminals via SERTs (Fuller et al., 1991). Third, serotonergic perikarya localized in the dorsal raphe nucleus (DRN) possess both SERTs and inhibitory 5-HT_1A autoreceptors. Blockade of 5-HT uptake therefore suppresses the electrical activity of serotonergic neurones via activation of 5-HT_1A autoreceptors (Blier and de Montigny, 1994; Gartside et al., 1997). Thus, we examined the influence of S33005 and other antidepressants on the firing rate of serotonergic neurones. Similarly, we evaluated their actions at adrenergic neurones in the locus ceruleus (LC), which bear NETs and inhibitory ₂-AR autoreceptors (Scuvee-Moreau and Dresse, 1979; Muth et al., 1991), and at dopaminergic cell bodies in the ventrotegmental area, which possess DATs and inhibitory D₂/D₃ autoreceptors (Millan et al., 2000b). Fourth, the influence of S33005 and the other agents upon extracellular levels of 5-HT, NE, and DA was simultaneously quantified in single dialysis samples of the frontal cortex, dorsal hippocampus, nucleus accumbens, and striatum of freely moving rats. Finally, severalalthough not allclasses of antidepressant diminish the density of cortical 5-HT_2A and -ARs upon long-term treatment (Okada and Tokumitsu, 1994; Newman-Tancredi et al., 1996; Yatham et al., 1999a,b). Thus, the influence of S33005 as compared with venlafaxine upon frontocortical levels of 5-HT_2A and -ARs, as well as dialysis levels of 5-HT, NE, and DA, was evaluated following their administration for 2 or 3 weeks.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. These studies used male Wistar rats of 200 to 250 g (Iffa Credo, L'Arbresles, France) housed in sawdust-lined cages with unrestricted access to standard chow and water. There was a 12-h/12-h light/dark cycle with lights on at 7:30 AM. Laboratory temperature and humidity were 21 ± 0.5°C and 60 ± 5%, respectively. Animals were adapted to laboratory conditions for at least 1 week prior to testing. All animal use procedures conformed to international European ethical standards (86/609-EEC) and the French National Committee (décret 87/848) for the care and use of laboratory animals.

Binding at Native Rat SERTs. Binding affinity was determined by competition with [³H]paroxetine (PerkinElmer Life Sciences, Les Ulis, France). Freshly prepared membranes of rat frontal cortex were homogenized with a Polytron and then centrifuged twice at 20,000g. The pellet was resuspended each time in incubation buffer. Membranes were incubated in triplicate with 2 nM [³H]paroxetine and competing ligand in a final volume of 0.4 ml for 2 h at 25°C. The incubation buffer contained 50 nM Tris-HCl (pH 7.4), 120 nM NaCl, and 5 mM KCl. Nonspecific binding was defined with 10 µM citalopram.

Binding at hSERTs. Binding affinity was determined by competition with [³H]paroxetine (PerkinElmer Life Sciences). Membranes prepared from HEK293 cells stably expressing recombinant hSERTs were purchased from Receptor Biology (Beltsville, MD) and incubated in triplicate with 2 nM [³H]paroxetine and competing ligand in a final volume of 0.4 ml for 1.5 h at 25°C. The incubation buffer contained 50 mM Tris-HCl (pH 7.4), 120 mM NaCl, and 5 mM KCl. Nonspecific binding was defined with 10 µM citalopram.

Binding at Native Rat NETs. Binding affinity was determined by competition with [³H]nisoxetine (Amersham, Les Ulis, France). Freshly prepared membranes of rat cortex were homogenized with a Polytron and then centrifuged twice at 20,000g. The pellet was resuspended each time in incubation buffer. Membranes were incubated in triplicate with 2 nM [³H]nisoxetine and competing ligand in a final volume of 0.5 ml for 4 h at 4°C. The incubation buffer contained 50 mM Tris-HCl (pH 7.4), 120 mM NaCl, and 5 mM KCl. Nonspecific binding was defined with 10 µM desipramine.

Binding at hNETs. Binding affinity was determined by competition with [³H]nisoxetine (2.0 nM, Amersham). Membranes prepared from Madin-Darby canine kidney cells expressing hNETs were purchased from Receptor Biology and incubated in triplicate with 2 nM [³H]nisoxetine and competing ligand in a final volume of 0.5 ml for 1 h at 4°C. The incubation buffer contained 50 mM Tris-HCl (pH 7.4), 120 mM NaCl, and 5 mM KCl. Nonspecific binding was defined with 10 µM desipramine.

Binding at Native Rat DATs. Binding affinity was determined by competition with [³H]GBR12909 (PerkinElmer Life Sciences). Freshly prepared membranes of rat striata were homogenized with a Polytron and then centrifuged twice at 20,000g. The pellet was resuspended each time in incubation buffer. Membranes were incubated in triplicate with 2 nM [³H]GBR12909 and competing ligand in a final volume of 1 ml for 1 h at 4°C. The incubation buffer contained 50 mM Tris-HCl (pH 7.4), 120 mM NaCl and 4 mM MgCl₂. Nonspecific binding was defined with 10 µM GBR12909.

Binding at hDATs. Binding affinity was determined by competition with [³H]GBR12909 (PerkinElmer Life Sciences). Membranes prepared from CHO cells stably expressing recombinant hDATs were purchased from Receptor Biology and incubated in triplicate with 2 nM [³H]GBR12909 and competing ligand in a final volume of 0.5 ml for 2 h at 4°C. The incubation buffer contained 50 mM Tris-HCl (pH 7.4) and 120 mM NaCl. Nonspecific binding was defined with 10 µM GBR12909.

Data Analysis for Binding Studies. For all of the above protocols, at the end of the incubation period, membranes were filtered through Whatman (Packard, Meriden, CT) GF/B filters pretreated with 0.1% polyethylenimine. Radioactivity retained on the filters was determined by scintillation counting. Binding isotherms were analyzed by nonlinear regression using Prism software (GraphPad Software Inc., San Diego, CA) to determine IC₅₀ values. These were converted to inhibition constants (K_i) by use of the Cheng-Prusoff equation: K_i = IC₅₀ / [(L/K_D)  1], where L is the concentration of ³H-labeled ligand and K_D is its dissociation constant determined in saturation binding experiments. The K_D values were as follows: 0.13 nM for [³H]paroxetine at both native rat SERTs and hSERTs; 1.2 and 2.2 nM for [³H]nisoxetine at native rat NETs and hNETs, respectively; and 1.6 and 1.0 nM for [³H]GBR12909 at native rat DATs and hDATs, respectively.

Interaction with other Binding Sites. The potential interaction of S33005 at diverse (>50) binding sites was evaluated by using standard procedures (see Results for several key sites) detailed elsewhere (Millan et al., 2000a). Inasmuch as S33005 showed negligible (pK_i < 5.0) affinity at all sites examined, these protocols are not further described herein. Data analysis was as described above.

Influence upon [³H]Monoamine Uptake by Rat Brain Synaptosomes. [³H]Monoamine uptake assays were carried out on synaptosomes prepared from rat cortex ([³H]5-HT), rat hypothalamus ([³H]NE), and rat striatum ([³H]DA), essentially as described previously (Janowsky et al., 1986). Synaptosomes were incubated with the radiolabeled neurotransmitter and drug for 15 min at 37°C before rapid filtration. [³H]Monoamine uptake into synaptosomes was determined by liquid scintillation counting.

Influence upon Depletion of Cerebral Pool of 5-HT by PCA. Levels of 5-HT were determined by high-performance liquid chromatography and electrochemical detection as previously described (Millan et al., 2000a) in the frontal cortex and hippocampus of rats 60 min after s.c. administration of S33005, vehicle, or other drugs and 30 min after injection of the vehicle or PCA (5.0 mg/kg, i.p.). Data were analyzed by analysis of variance (ANOVA) followed by Dunnett's test. ID₅₀ values plus 95% confidence limits (CL) were calculated.

Influence upon the Electrical Activity of Serotonergic, Adrenergic, and Dopaminergic Cell Bodies. The influence of S33005 compared with other ligands upon the firing rate of DRN-localized serotonergic perikarya, LC-localized adrenergic perikarya, and ventrotegmental area-localized dopaminergic perikarya was determined by using a procedure described in detail previously (Millan et al., 2000a). Briefly, following anesthesia with chloral hydrate (400 mg/kg, i.p.), rats were placed in a stereotaxic apparatus, and a tungsten microelectrode was lowered into the DRN, LC, or ventrotegmental area. Coordinates were as follows: DRN, AP = 7.8 from bregma, L = 0.0, and H = 5/6.5 from dura; LC, AP = 1.2 from zero, L = 1.2, and H = 5.5/6.5 from dura; and ventrotegmental area, AP = 5.5 from bregma, L = 0.7, and H = 7/8.5 from dura. As detailed elsewhere (Millan et al., 2000a), serotonergic, adrenergic, and dopaminergic neurones in the DRN, LC, and ventrotegmental area, respectively, were recognized by their distinctive waveforms. Following baseline recording over 5 min, vehicle, S33005, or other agents were administered i.v. (in a volume of 0.5 ml/kg) in cumulative doses every 2 to 3 min. After vehicle or drug administration, a further injection of WAY100,635 (0.031 mg/kg, i.v.) or idazoxan (0.063 mg/kg, i.v.) was made for the DRN and LC, respectively. Drug effects were quantified over the 60-s bin corresponding to their time of peak action. Spike2 software (CED, Cambridge, UK) was used for data acquisition and analysis. Data are expressed as percentage of change from baseline firing rate (defined as 0%). Data were analyzed by ANOVA followed by Newman-Keuls test, and ID₅₀ values (95% CL) were calculated.

Influence upon Extracellular Levels of 5-HT, NE, and DA. Quantification of extracellular levels of 5-HT, NE, and DA in single dialysate samples of the frontal cortex, dorsal hippocampus (NE and 5-HT), nucleus accumbens (DA and 5-HT), and striatum (DA and 5-HT) was achieved by using a protocol extensively described previously (Millan et al., 2000a). The guide cannulae were implanted under pentobarbital anesthesia (40.0 mg/kg., i.p.) at the following coordinates 1 week prior to experimentation: frontal cortex, AP = +2.2 from bregma, L = ±0.6, and H = 0.2 from dura; dorsal hippocampus, AP = 3.6 from bregma, L = ±1.2, and H = 2.3 from dura; nucleus accumbens, AP = +0.8 from bregma, L = +0.6, and H = 4.5 from dura; and striatum, AP = +0.5 from bregma, L = 2.8, and H = 3.0 from dura. A cuprophane CMA/11 probe (4 mm in length for the frontal cortex and striatum, 2 mm in length for the hippocampus and nucleus accumbens, and, in each case, 0.24-mm outer diameter; Carnegie Medicine, Stockholm, Sweden) was lowered into position. Three basal samples of 20 min each were taken. Vehicle, S33005, or other drugs were administered s.c., and samples were taken for a further 3 h. NE, 5-HT, and DA levels were quantified by high-performance liquid chromatography followed by coulometric detection as previously described (Millan et al., 2000a). The assay limit of sensitivity was 0.1 to 0.2 pg/sample for 5-HT, NE, and DA in each case. Data were analyzed by ANOVA, with sampling time as the repeated within-subject factor.

Influence upon DRN Firing Rate and Dialysis Levels of 5-HT, NE, and DA in the Frontal Cortex: Chronic Administration. Rats were treated daily with a single injection of either vehicle, S33005 (10.0 mg/kg, s.c.), or venlafaxine (10.0 mg/kg, s.c.) for 14 days. On the 15th day, rats either received an additional injection of the same drug or of vehicle. Thereafter, exactly as described above, extracellular levels of 5-HT, NE, and DA were determined in the frontal cortex of freely moving rats. In a parallel study, following a chronic 14-day treatment, the influence of S33005 or venlafaxine (day 15) upon the electrical activity of serotonergic cell bodies in the DRN was determined.

Influence upon Cortical 5-HT_2A Receptors and -ARs: Chronic Administration. Rats were treated with S33005 (10.0 mg/kg, s.c.), venlafaxine (10.0 mg/kg, s.c.), or vehicle once daily for 2 or 3 weeks. One day following the final injection, as described previously (Newman-Tancredi et al., 1996), using [³H]CGP-12177 as a radioligand, -ARs were examined in homogenates of cortex. Furthermore, using [³H]ketanserin as a radioligand, 5-HT_2A receptors were examined in frontal cortex. B_max and K_D values were determined through conventional procedures (see Newman-Tancredi et al., 1996).

Drugs. Actions of drugs in each of the individual studies described herein were evaluated concurrently. For binding studies, drugs were dissolved in dimethyl sulfoxide (10² M) and dilutions made in the buffer as appropriate. For in vivo studies, drugs were dissolved in sterile water, plus a few drops of lactic acid if necessary, and pH adjusted to as close to normality (>5.0) as possible. Drug salts and sources were as follows. S33005 HCl [()1-(1-dimethylaminomethyl 5-methoxybenzocyclobutan-1-yl) cyclohexanol], WAY100,635 fumarate [(N-{2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl}-N-(2-pyridinyl)cyclohexanecarboxamine], citalopram HCl, idazoxan HCl, reboxetine methane sulfonate, and venlafaxine HCl were all synthesized internally (G. Lavielle and J.-L. Peglion). GBR12935 2HCl [1-[2-(diphenylmethoxy)ethyl]-4-(3-phenylpropyl)-piperazine] was purchased from Sigma/RBI (Natick, MA), and PCA HCl and clomipramine were purchased from Sigma (Saint Quentin-Fallavier, France).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of S33005 as Compared with Its Isomer, S33004, and Racemic S32647. S33005 is the optically pure ()-isomer of racemic (±)-S32647, whereas S33004 is the optically pure (+)-isomer. S32647 potently interacted with native rat SERTs and, less potently, with native rat NETs: pK_i values (K_i in nM) = 8.64 ± 0.02 (2.3) and 6.58 ± 0.09 (263), respectively. Similarly, S33005 displayed marked affinity for these sites: pK_i values (K_i in nM) = 8.71 ± 0.06 (1.9) and 6.75 ± 0.10 (178), respectively. In contrast, S33004 was substantially less active at both SERTs and NETs: pK_i values (K_i in nM) = 6.36 ± 0.05 (437) and 5.12 ± 0.10 (7586), respectively. These observations led us to select S33005 for further study.

Interaction of S33005 as Compared with Reference Antidepressant Agents at Cerebral Rat SERTs, NETs, and DATs. As mentioned above, S33005 possessed pronounced affinity for cerebral rat SERTs and less marked affinity for cerebral rat NETs. A similar pattern of preferential interaction with native SERTs versus NETs was acquired with venlafaxine. However, its affinity at these sites was, respectively, 11- and 6-fold lower than those of S33005 (Fig. 2; Table 1). Clomipramine also resembled S33005 in its dual interaction at SERTs and native NETs, with a clear preference for the former. In distinction to S33005, citalopram revealed pronounced affinity for rat SERTs and negligible affinity for NETs. On the other hand, reboxetine displayed an opposite pattern of preference with superior affinity for native NETs as compared with SERTs. S33005 and all other drugs manifested low affinity for cerebral rat DATs.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Interaction of S33005 as compared with reference antidepressant agents at native cerebral rat SERTs, NETs, and DATs. Isotherms are representative of at least three independent experiments performed in triplicate. All isotherms yielded pseudo-Hill coefficients not significantly different from unity and corresponding to pKi values specified in Table 1.

Interaction of S33005 as Compared with Reference Antidepressant Agents at Cloned hSERTs, hNETs, and hDATs. In analogy to native rat SERTs, S33005 showed high affinity for heterologously expressed hSERTs (Table 1). However, its affinity for hNETs was about 10-fold lower than at native rat NETs. Venlafaxine similarly showed (~8-fold) lower affinity for hNETs than for native NETs. In addition, its affinity for hSERTs was somewhat less pronounced than at native SERTs. In distinction, clomipramine showed more pronounced affinity for hSERTs versus SERTs and slightly inferior affinity for hNETs versus native NETs. Citalopram presented ~3-fold lower affinity at hSERTs versus native SERTs. Its affinity was low at hNETs. Finally, the affinity of reboxetine for hSERTs was less pronounced than its affinity at hNETs. At both native DATs and cloned hDATs, the affinity of S33005 and other ligands for hDAT sites was low.

Interaction of S33005 as Compared with Reference Antidepressant Agents at Additional Sites. S33005 displayed negligible affinity (pK_i < 5.0) for multiple monoaminergic receptors, including diverse sites shown in Table 1 that, as autoreceptors and heteroceptors, modulate corticolimbic serotonergic, adrenergic, and dopaminergic transmission (see Millan et al., 2000b). It also showed negligible (pK_i < 5.0) affinity for histaminergic (H)₁ receptors labeled with [³H]pyralamine (1.0 nM), and for cloned, human, muscarinic h(M)₁, hM₂, hM₃, and hM₄ receptors labeled with [³H]N-methyl-scopolamine (1.0 nM) (not shown). S33005 did not bind (pK_i < 5.0) to either monoamine oxidase A or monoamine oxidase B (not shown). At a broad diversity of other receptors, enzymes, and ion channels (>50 sites), S33005 also showed low affinity (pK_i < 5.0). S33005 was thus highly selective for SERTs and NETs. Like S33005, venlafaxine showed low affinity (pK_i < 5.0) for various monoaminergic receptors and H₁ and hM₁ receptors. The affinity of clomipramine at h5-HT_1B, h5-HT_1D, and 5-HT₃ receptors was negligible, although it showed marked affinity for h5-HT_2A and h5-HT_2C receptors, as well as modest affinity for 5-HT₃ receptors. At native ₁-ARs and cloned h_1A-ARs, its affinity was marked, whereas its affinity for native rat ₂-ARs and cloned h_2A-ARs was low. At hD₂ and hD₃ receptors, the affinity of clomipramine was modest. Clomipramine also displayed high affinity for both H₁ and hM₁ receptors: pK_i values (K_i in nM), 8.0 (10) and 7.5 (31.6), respectively. Citalopram manifested negligible affinity for monoaminergic receptors, with the exception of h5-HT_2C sites for which it showed mild affinity. It showed modest affinity for H₁ receptors: pK_i values (K_i in nM) = 6.3 (501). Finally, reboxetine showed modest affinity for h5-HT_2C sites but negligible affinity for all other sites.

Influence of S33005 as Compared with Reference Antidepressant Agents upon Uptake of 5-HT, NE, and DA into Rat Synaptosomes. In line with its high affinity at SERTs, S33005 potently and concentration dependently inhibited the uptake of [³H]5-HT into cerebral rat synaptosomes (Table 2). It also inhibited [³H]NE uptake at higher concentrations, in line with its lower affinity for NETs. In distinction, only very high concentrations of S33005 modified [³H]DA uptake. A similar pattern of data was acquired for venlafaxine although, in line with its lower affinity than S33005 for native SERTs and NETs, venlafaxine was less potent than S33005 in inhibiting uptake of [³H]5-HT and [³H]NE. Clomipramine also preferentially inhibited [³H]5-HT versus [³H]NE uptake and weakly inhibited [³H]DA uptake. In contrast, in line with its selective interaction with native SERTs versus NETs, citalopram selectively inhibited synaptosomal uptake of [³H]5-HT versus [³H]NE (and [³H]DA). On the other hand, in analogy to its binding profile, reboxetine behaved as a preferential inhibitor of [³H]NE versus [³H]5-HT uptake. It failed to modify [³H]DA uptake.

Influence of S33005 as Compared with Reference Antidepressant Agents upon Depletion of Cerebral Pools of 5-HT by PCA. The administration of PCA (5.0 mg/kg, i.p.) elicited a pronounced reduction of levels of 5-HT in both the frontal cortex and the hippocampus (Fig. 3): frontal cortex, vehicle/vehicle (n = 12), 3.61 ± 0.24 ng/mg of protein versus vehicle/PCA (n = 10), 0.87 ± 0.09, P < 0.001; hippocampus, vehicle/vehicle (n = 12), 4.23 ± 0.22 ng/mg of protein versus vehicle/PCA (n = 10), 0.43 ± 0.08, P < 0.001. Preadministration of S33005 dose dependently, potently, and completely blocked the influence of PCA upon 5-HT levels in both frontal cortex and hippocampus. Administered alone, S33005 did not significantly modify 5-HT levels. For a dose of 10.0 mg/kg, s.c (which abolished the action of PCA): frontal cortex, vehicle/vehicle (n = 12), 3.61 ± 0.24 ng/mg of protein versus S33005/vehicle (n = 10), 3.76 ± 0.21, P > 0.05; hippocampus, vehicle/vehicle (n = 12), 4.23 ± 0.22 ng/mg of protein versus S33005/vehicle (n = 10) 4.54 ± 0.2, P > 0.05. Venlafaxine also inhibited the influence of PCA upon levels of 5-HT in frontal cortex and hippocampus, although it was less potently active than S33005. Citalopram was also highly active, whereas clomipramine less potently attenuated the action of PCA. Reboxetine (10.0 mg/kg, s.c.) did not significantly modify the action of PCA (not shown). Administered alone, venlafaxine, citalopram, clomipramine, and reboxetine did not significantly affect basal levels of 5-HT in either frontal cortex or hippocampus (not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Influence of S33005 as compared with reference antidepressant agents upon the depletion by PCA of 5-HT levels in frontal cortex and hippocampus. Data are means ± S.E.M. They are expressed as percentage blockade of the actions of PCA (5.0 mg/kg, i.p.) with 0 and 100% values equivalent to no and full blockade, respectively. Corresponding ID50 values (95% CL) are indicated in Table 3. ANOVA was as follows. Frontal cortex, S33005, F(5,56) = 62.1, P < 0.001; venlafaxine, F(5,57) = 81.2, P < 0.001; clomipramine, F(4,36) = 74.0, P < 0.001; and citalopram, F(4,28) = 107.7, P < 0.01. Hippocampus, S33005, F(5,50) = 41.1, P < 0.001; venlafaxine, F(5,57) = 53.3, P < 0.001; clomipramine, F(4,36) = 36.9, P < 0.001; and citalopram, F(4,28) = 91.5, P < 0.01. Asterisks indicate significance of drug versus control values in Dunnett's test following ANOVA. *P < 0.05. VEH, vehicle.

Influence of S33005 as Compared with Reference Antidepressant Agents upon the Electrical Activity of Monoaminergic Cell Bodies. S33005 potently, dose dependently, and completely inhibited the firing rate of serotonergic perikarya localized in the DRN of anaesthetized rats (Figs. 4 and 5). This action was abolished by the selective 5-HT_1A receptor antagonist, WAY100,635. Expressed relative to baseline values (0%), S33005 (0.125 mg/kg, i.v.) + vehicle = 100.0 ± 0.0% versus S33005 + WAY100,635 (0.031 mg/kg, i.v.) = 1.1 ± 7.6%, P < 0.001. In line with previous work (Millan et al., 2000b), WAY100,635 did not modify the electrical activity of serotonergic neurones alone (not shown). Over a higher dose range, S33005 dose dependently reduced the firing rate of adrenergic neurones of the LC, an action abolished by the ₂-AR antagonist idazoxan. S33005 (2.0 mg/kg, i.v.) + vehicle = 98.8 ± 0.3% versus S33005 + idazoxan (0.063 mg/kg, i.v.) = 13.3 ± 5.2%, P < 0.001. Administered alone, idazoxan elicited a significant increase in firing rate. Idazoxan (0.063) = +42.0 ± 14.0% versus vehicle = +1.3 ± 2.2%, P < 0.05. In contrast to serotonergic and adrenergic neurones, S33005 exerted little influence upon dopaminergic neurones of the ventrotegmental area. A slight tendency toward an increase in firing rate was seen at modest doses, which transformed into a tendency for inhibition at high doses, but neither effect attained statistical significance. In addition, S33005 did not modify the firing pattern (regular or burst) of dopaminergic neurones (not shown). In contrast, as a positive internal control, the selective DA reuptake inhibitor, GBR12935, dose dependently inhibited firing of ventrotegmental area dopaminergic neurones: ID₅₀ = 0.8 (0.6-1.1) mg/kg, i.v. Its actions were blocked by the D₂/D₃ antagonist, haloperidol (0.016 mg/kg, i.v.). Vehicle + GBR12935 (4.0 mg/kg, i.v.) = 83.4 ± 12.5 versus haloperidol + GBR12935 = 2.2 ± 2.2%, P < 0.001. 

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Dose-dependent influence of S33005 upon the electrical activity (firing rate) of representative serotonergic, adrenergic, and dopaminergic neurones localized in the DRN, LC, and ventrotegmental area (VTA), respectively, of anesthetized rats. S33005 was administered in cumulative doses at 3-min intervals. For the DRN, this was followed by administration of the selective 5-HT1A receptor antagonist, WAY100,635, and for the LC, by administration of the selective 2-adrenoceptor antagonist, idazoxan. Both WAY100,635 and idazoxan rapidly and completely reversed the inhibitory influence of S33005 upon DRN and LC neurones, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Influence of S33005 as compared with reference antidepressant agents upon the electrical activity (firing rate) of serotonergic, adrenergic, and dopaminergic cell bodies localized in the DRN, LC, and VTA, respectively, of anesthetized rats. Data are means ± S.E.M. of firing rate expressed relative to basal pretreatment values (defined as 0%). n  5 per value. Corresponding ID50 values (95% CL) are indicated in Table 3. ANOVA was as follows. S33005: DRN, F(6,24) = 38.8, P < 0.001; LC, F(6,30) = 72.1, P < 0.001; and VTA, F(5,25) = 2.3, P > 0.05. Venlafaxine: DRN, F(7,35) = 68.9, P < 0.001; LC, F(5,25) = 57.2, P < 0.001; and VTA, F(6,36) = 2.5, P > 0.05. Clomipramine: DRN, F(5,30) = 46.8, P < 0.001; LC, F(5,25) = 5.3, P < 0.01; and VTA, F(4,20) = 25.5, P < 0.001. Citalopram: DRN, F(5,25) = 113.9, P < 0.001; LC, F(5,25) = 4.1, P < 0.01; and VTA, F(6,30) = 6.2, P < 0.001. Reboxetine: DRN, F(6,24) = 6.6, P < 0.001; LC, F(5,25) = 57.2, P < 0.001; and VTA, F(5,30) = 2.3, P > 0.05. Asterisks indicate significance of drug to control values in Newman-Keuls test following ANOVA. *P < 0.05. VEH, vehicle.

Influence of S33005 as Compared with Reference Antidepressant Agents upon Extracellular Levels of 5-HT, NE, and DA in Dialysates of the Frontal Cortex of Freely Moving Rats. S33005 elicited a rapid, sustained, and dose-dependent elevation in extracellular levels of 5-HT in frontal cortex of freely moving rats (Figs. 6 through 8; Table 3). In the same dialysis samples, levels of NE and DA were likewise augmented. Area under the curve (AUC) analysis revealed that the most pronounced influence of S33005 was, in fact, upon extracellular levels of NE, with those of 5-HT and DA displaying a less marked increase. A similar pattern of data was acquired with both venlafaxine and clomipramine, although they were less potent than S33005. In distinction, citalopram preferentially elevated levels of 5-HT versus NE and DA and only evoked a slight rise in levels of NE and DA even at the highest dose tested. On the contrary, reboxetine potently increased levels of NE and, less markedly, DA without significantly modifying levels of 5-HT. Surprisingly, and for reasons remaining to be clarified, the induction of NE levels by reboxetine was less pronounced at the highest dose evaluated (40.0 mg/kg, s.c.). Upon i.p. and p.o. administration, S33005 also dose dependently (0.63-40.0 mg/kg, in each case), significantly (P < 0.01, in each case), and markedly elevated dialysate levels of 5-HT, NE, and DA in frontal cortex (not shown).

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Influence of S33005 as compared with venlafaxine upon extracellular levels of 5-HT, NE, and DA in dialysates of the frontal cortex of freely moving rats. Data are means ± S.E.M. of 5-HT, NE, and DA levels expressed relative to basal pretreatment values (defined as 100%). These were, representatively (S33005, 10.0 mg/kg, s.c.), 0.79 ± 0.06, 1.49 ± 0.05, and 1.14 ± 0.06 pg/20 µl of dialysate for 5-HT, NE, and DA, respectively. n  5 per value. ANOVA was as follows. 5-HT: S33005, 0.01, F(1,8) = 0.1, P > 0.05; 0.04, F(1,9) = 15.2, P < 0.01; 0.16, F(1,8) = 23.6, P < 0.01; 0.63, F(1,8) = 10.6, P < 0.05; 2.5, F(1,8) = 27.8, P < 0.01; 10.0, F(1,11) = 47.3, P < 0.01; and 40.0, F(1,9) = 100.4, P < 0.01. Venlafaxine, 0.16, F(1,8) = 0.6, P > 0.05; 0.63, F(1,8) = 11.3, P < 0.01; 2.5, F(1,8) = 60.0, P < 0.01; 10.0, F(1,11) = 35.4, P < 0.01; and 40.0, F(1,9) = 125.6, P < 0.01. NE: S33005, 0.01, F(1,8) = 0.3, P > 0.05; 0.04, F(1,9) = 0.3, P > 0.05; 0.16, F(1,8) = 21.3, P < 0.01; 0.63, F(1,8) = 12.5, P < 0.01; 2.5, F(1,9) = 179.9, P < 0.01; 10.0, F(1,10) = 55.5, P < 0.01; and 40.0, F(1,8) = 41.3, P < 0.01. Venlafaxine, 0.16, F(1,8) = 1.7, P > 0.05; 0.63, F(1,8) = 51.4, P < 0.01; 2.5, F(1,8) = 39.6, P < 0.01; 10.0, F(1,11) = 62.8, P < 0.01; and 40.0, F(1,9) = 198.4, P < 0.01. DA: S33005, 0.01, F(1,8) = 0.1, P > 0.05; 0.04, F(1,9) = 0.1, P > 0.05; 0.16, F(1,8) = 7.1, P < 0.05; 0.63, F(1,8) = 10.1, P < 0.05; 2.5, F(1,9) = 16.7, P < 0.01; 10.0, F(1,11) = 31.6, P < 0.01; and 40.0, F(1,8) = 26.3, P < 0.01. Venlafaxine, 0.16, F(1,8) = 0.1, P > 0.05; 0.63, F(1,8) = 3.8, P > 0.05; 2.5, F(1,8) = 30.0, P < 0.01; 10.0, F(1,11) = 101.8, P < 0.01; and 40.0, F(1,9) = 135.5, P < 0.01. Asterisks indicate significance of drug versus vehicle values. *P < 0.05.

View larger version (33K): [in this window] [in a new window]   Fig. 7.   Influence of citalopram, reboxetine, and clomipramine upon extracellular levels of 5-HT, NE, and DA in dialysates of the frontal cortex of freely moving rats. Data are means ± S.E.M. of 5-HT, NE, and DA levels expressed relative to basal pretreatment values (defined as 100%). For representative basal levels, see legend to Fig. 6. ANOVA was as follows. 5-HT: citalopram, 0.16, F(1,9) = 4.2, P > 0.05; 0.63, F(1,9) = 62.9, P < 0.01; 2.5, F(1,9) = 69.4, P < 0.01; 10.0, F(1,8) = 140.4, P < 0.01; and 40.0, F(1,9) = 107.3, P < 0.01. Clomipramine, 0.63, F(1,8) = 0.1, P > 0.05; 2.5, F(1,8) = 16.3, P < 0.01; 10.0, F(1,10) = 52.0, P < 0.01; and 40.0, F(1,9) = 62.4, P < 0.01. Reboxetine, 0.0025, F(1,8) = 1.3, P > 0.05; 0.04, F(1,8) = 0.1, P > 0.05; 0.63, F(1,8) = 0.1, P > 0.05; 10.0, F(1,12) = 0.3, P > 0.05; and 40.0, F(1,9) = 0.2, P > 0.05. NE: citalopram, 0.16, F(1,9) = 0.8, P > 0.05; 0.63, F(1,9) = 3.8, P > 0.05; 2.5, F(1,9) = 3.3, P > 0.05; 10.0, F(1,8) = 41.2, P < 0.01; and 40.0, F(1,10) = 19.3, P < 0.01. Clomipramine, 0.63, F(1,8) = 1.4, P > 0.05; 2.5, F(1,8) = 64.5, P < 0.01; 10.0, F(1,9) = 76.5, P < 0.01 and 40.0, F(1,9) = 80.5, P < 0.01; Reboxetine, 0.0025, F(1,8) = 3.3, P > 0.05; 0.04, F(1,8) = 88.1, P < 0.01; 0.63, F(1,9) = 44.7, P < 0.01; 10.0, F(1,12) = 46.1, P < 0.01; and 40.0, F(1,9) = 10.9, P < 0.01. DA: citalopram, 0.16, F(1,9) = 0.3, P > 0.05; 0.63, F(1,9) = 1.7, P > 0.05; 2.5, F(1,9) = 0.8, P > 0.05; 10.0, F(1,8) = 1.5, P > 0.05; and 40.0, F(1,10) = 30.5, P < 0.01. Clomipramine, 0.63, F(1,8) = 1.0, P > 0.05; 2.5, F(1,8) = 13.5, P < 0.01; 10.0, F(1,9) = 39.3, P < 0.01; and 40.0, F(1,9) = 120.9, P < 0.01; Reboxetine, 0.0025, F(1,8) = 1.2, P > 0.05; 0.04, F(1,8) = 19.1, P < 0.01; 0.63, F(1,9) = 112.8, P < 0.01; 10.0, F(1,12) = 25.7, P < 0.01; and 40.0, F(1,9) = 155.3, P < 0.01. Asterisks indicate significance of drug versus vehicle values. *P < 0.05.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   Relative potency of S33005 as compared with reference antidepressant agents upon extracellular levels of 5-HT, NE, and DA in dialysates of the frontal cortex of freely moving rats. Data (means ± S.E.M.) are area under the curve analyses based on individual time courses presented in Figs. 6 to 8. Corresponding minimal effective doses (P < 0.05 to vehicle) are indicated in Table 3. VEH, vehicle.

Influence of S33005 as Compared with Reference Antidepressant Agents upon Extracellular Levels of 5-HT, NE, and DA in Dialysates of the Dorsal Hippocampus, Nucleus Accumbens, and Striatum of Freely Moving Rats. S33005 elicited a pronounced and sustained elevation in levels of 5-HT and NE in the dorsal hippocampus (Fig. 9). For 5-HT AUC analysis, vehicle = 99.2 ± 2.5% versus S33005 = 207.8 ± 6.1%. For NE AUC analysis, vehicle = 97.8 ± 2.5% versus S33005 = 231.7 ± 8.1%. Venlafaxine similarly increased extracellular levels of 5-HT and NE in this structure. For 5-HT AUC analysis, vehicle = 99.2 ± 2.5% versus venlafaxine = 205.2 ± 10.5%. For NE AUC analysis, vehicle = 97.8 ± 2.5% versus venlafaxine = 204.1 ± 7.0%. Clomipramine also enhanced levels of both 5-HT and NE (not shown). For 5-HT, F(1,16) = 30.8, P < 0.01; and AUC analysis, vehicle = 99.2 ± 2.5% versus clomipramine = 165.3 ± 5.9%. For NE, F(1,16) = 34.4, P < 0.01; and AUC analysis, vehicle = 97.8 ± 2.5% versus clomipramine = 189.2 ± 7.4%. In contrast, citalopram selectively elevated levels of 5-HT as compared with NE (not shown). For 5-HT, F(1,15) = 87.0, P < 0.01; and AUC analysis, vehicle = 99.2 ± 2.5% versus citalopram = 193.0 ± 9.0%. For NE, F(1,15) = 0.7, P > 0.05; and AUC analysis, vehicle = 97.8 ± 2.5% versus citalopram = 102.1 ± 3.5%. On the other hand, reboxetine selectively increased levels of NE as compared with 5-HT. For 5-HT, F(1,16) = 0.2, P > 0.05; and AUC analysis, vehicle = 99.2 ± 2.5% versus reboxetine = 101.5 ± 3.0%. For NE, F(1,16) = 32.6, P < 0.01; and AUC analysis, vehicle = 97.8 ± 2.5% versus reboxetine = 215.0 ± 8.3%.

View larger version (24K): [in this window] [in a new window]   Fig. 9.   Influence of S33005 as compared with venlafaxine upon extracellular levels of 5-HT, NE, and DA in dialysates of the dorsal hippocampus, nucleus accumbens, and striatum of freely moving rats. Data are means ± S.E.M. of 5-HT, NE, and DA levels expressed relative to basal pretreatment values (defined as 100%). These were, representatively (S33005, 10 mg/kg, s.c.), 0.68 ± 0.05 and 0.83 ± 0.04 pg/20 µl of dialysate for 5-HT and NE, respectively, in dorsal hippocampus; 0.63 ± 0.04 and 4.8 ± 0.5 pg/20 µl of dialysate for 5-HT and DA, respectively, in nucleus accumbens; and 0.49 ± 0.03 and 16.8 ± 1.5 pg/20 µl of dialysate per values for 5-HT and DA, respectively, in the striatum. n  5 per value. ANOVA was as follows. Dorsal hippocampus, S33005, 5-HT, F(1,15) = 266.2, P < 0.01 and NE, F(1,15) = 123.3, P < 0.01; and venlafaxine, 5-HT, F(1,16) = 24.6, P < 0.01; and NE, F(1,16) = 61.6, P < 0.01. Nucleus accumbens, S33005, 5-HT, F(1,9) = 486.1, P < 0.01 and DA, F(1,9) = 0.5, P > 0.05; and venlafaxine, 5-HT, F(1,10) = 52.6, P < 0.01 and DA, F(1,10) = 0.1, P > 0.05. Striatum, S33005, 5-HT, F(1,10) = 137.6, P < 0.01 and DA, F(1,11) = 5.2, P < 0.05; and venlafaxine, 5-HT, F(1,9) = 107.7, P < 0.01 and DA, F(1,11) = 3.5, P > 0.05. Asterisks indicate significance of drug versus vehicle values. *P < 0.05.

Influence of Chronic Administration of S33005 as Compared with Venlafaxine upon Extracellular Levels of 5-HT, NE, and DA in Frontal Cortex and the Electrical Activity of DRN Serotonergic Neurones. Following administration of S33005 for 2 weeks, there was no significant modification in basal dialysate levels of DA in frontal cortex (Fig. 10). Levels of 5-HT showed a (nonsignificant) tendency for an increase. Furthermore, there was a significant elevation in basal levels of NE. Expressed relative to basal levels, an additional injection of S33005 at the dose used for chronic treatment (10.0 mg/kg, s.c.) evoked an elevation in frontocortical dialysate levels of 5-HT, NE, and DA similar to that seen in rats treated chronically with vehicle (Fig. 11). A comparable pattern of data was obtained for venlafaxine inasmuch as the influence of its acute administration upon 5-HT, NE, and DA levels in frontal cortex was not significantly modified by chronic administration (not shown). Nevertheless, in distinction to S33005, venlafaxine did not show an increase in basal levels of NE following chronic administration (Fig. 10).

View larger version (17K): [in this window] [in a new window]   Fig. 10.   Influence of chronic (2 weeks) administration of S33005 (10.0 mg/kg, s.c.) and venlafaxine (10.0 mg/kg, s.c.) upon extracellular levels of 5-HT, NE, and DA in dialysates of the frontal cortex. Data are means ± S.E.M. n  5 per value. The asterisk indicates a significant (P < 0.05) difference of chronic S33005 to chronic vehicle values in Student's t test.

View larger version (14K): [in this window] [in a new window]   Fig. 11.   Influence of chronic (2 weeks) administration of S33005 (10.0 mg/kg, s.c.) upon modulation of extracellular levels of 5-HT, NE, and DA in dialysates of the frontal cortex by additional, acute administration of S33005 (10.0 mg/kg, s.c.). Dialysate levels are expressed as a percentage of basal preinjection values, which were defined as 100% (see Fig. 10). Data are means ± S.E.M. n  5 per value. There was no significant difference in ANOVA (P > 0.05) for influence of S33005 (acute) in rats chronically treated with S33005 as compared with vehicle (VEH).

Influence Upon Cortical Levels of 5-HT_2A Receptors and -ARs. Following chronic (2 weeks) administration of S33005, relative to vehicle treatment (B_max, 22.9 ± 1.3 fmol/mg of tissue = 100.0 ± 5.5%), there was a significant (P < 0.05) reduction in the density (B_max) of 5-HT_2A receptors in frontal cortex (85.0 ± 2.5%) in the absence of an alteration of affinity (vehicle, K_D = 1.64 ± 0.12 nM versus S33005, K_D = 1.39 ± 0.08 nM) (Fig. 12). Venlafaxine tended to decrease the density of 5-HT_2A receptors, although this effect did not attain statistical significance (92.2 ± 3.7%, P > 0.05); the K_D was also not affected (1.64 ± 0.14 nM). At 3 weeks, relative to vehicle (B_max, 28.0 ± 1.4 fmol/mg of tissue = 100.0 ± 5.0%), both S33005 and venlafaxine yielded a significant decrease in B_max: 79.6 ± 2.8 and 77.8 ± 2.8%, respectively, P < 0.01 in each case. K_D values were unaffected: vehicle = 2.27 ± 0.11 nM, S33005 = 2.03 ± 0.19 nM, and venlafaxine = 2.18 ± 0.21 nM. Compared with vehicle-treated controls (B_max, 5.58 ± 0.27 fmol/mg of tissue = 100.0 ± 4.6%), following 2 weeks of treatment, S33005 tended to decrease levels of -ARs in cortex (86.6 ± 4.2%), but this effect just failed to reach statistical significance (P = 0.056). Venlafaxine, in distinction, exerted little influence upon -ARs (97.7 ± 5.3%). Following 3 weeks of administration, S33005 and venlafaxine did not significantly modify -AR density or affinity (not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 12.   Scatchard analysis of the influence of chronic (2 weeks) administration of S33005 upon the density of 5-HT2A receptors in frontal cortex. A, from representative experiments each performed in triplicate. For further details, see Results. B, data are means ± S.E.M. n  5 per group. Asterisks indicate significance of differences to vehicle in Student's two-tailed t test. *P < 0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Receptor Binding Profile. S33005 possessed pronounced affinity for native rat and cloned human SERTs, yielding pK_i values similar to citalopram and clomipramine and superior to venlafaxine and reboxetine (Muth et al., 1991; Owens et al., 1997; Tatsumi et al., 1997; Béïque et al., 1999; Riva et al., 1999). Citalopram was ~3-fold less potent at human versus rat SERTs. This observation, obtained with [³H]paroxetine, may be compared with ratios of 2- and 10-fold for binding studies performed with [³H]citalopram at rat SERTs as compared with [³H]citalopram and [³H]5-HT, respectively, at hSERTs (Owens et al., 1997). On the contrary, clomipramine shows ~10-fold higher affinity at human versus rat SERTs (Table 1; Owens et al., 1997). Such contrasting affinities (Barker and Blakely, 1995; Sur et al., 1998) emphasize the importance of determining drug affinities in both species.

Prevention of PCA-Induced 5-HT Depletion. Following access to serotonergic terminals via SERTs, PCA triggers 5-HT release (Fuller et al., 1991). Correspondingly, S33005, venlafaxine, citalopram, and clomipramine suppressed depletion of cerebral pools of 5-HT by PCA, in analogy to their attenuation of its behavioral actions (accompanying paper), whereas reboxetine was inactive. The high potency of S33005 versus venlafaxine in this model underpins in vitro observations discussed above.

Electrical Activity of Monoaminergic Neurones. Reflecting its marked activity at SERTs, S33005 potently and WAY100,635-reversibly inhibited the firing rate of serotonergic perikarya, and, in line with its lower activity at NETs, higher doses of S33005 idazoxan-reversibly inhibited the electrical activity of adrenergic cell bodies. In correspondence with binding studies discussed above, such actions were expressed less potently by venlafaxine (Muth et al., 1991; Gartside et al., 1997; Béïque et al., 1999). Clomipramine likewise inhibited electrical activity of the DRN but only weakly affected the LC, probably since its actions at sites other than NETs influence the activity of adrenergic cell bodies (Scuvee-Moreau and Dresse, 1979; Gartside et al., 1997). This combined inhibition of serotonergic and adrenergic neurones by S33005 contrasted with their selective inhibition by citalopram and reboxetine, respectively (Popik, 1999; Riva et al., 1999). Interestingly, the potency of S33005 as compared with citalopram and reboxetine for inhibition of serotonergic and adrenergic neurones, respectively, was pronounced when compared with their relative affinities for SERTs (S33005 versus citalopram) and NETs (S33005 versus reboxetine). This particularly high activity of S33005 in vivo, underpinned by dialysis studies considered below, may reflect pharmacokinetic factors or, possibly, the existence of specific cerebral isoforms of SERTs and NETs for which S33005 has higher affinity than revealed by binding studies (see below). Notably, ventrotegmental area-localized dopaminergic cell bodies were not inhibited by S33005, consistent with its low activity at DATs. Indeed, although a modest inhibitory influence of fluoxetine upon dopaminergic neurones was suggested to underlie (rare) cases of akathisia (Prisco and Esposito, 1995), we have not found fluoxetine or citalopram to inhibit dopaminergic neurones in our studies (Fig. 5; Millan et al., 2000b). Rather, the latter slightly excited dopaminergic perikarya at high doses, an observation also acquired with clomipramine and reboxetine and justifying further characterization.

Influence upon Extracellular Levels of Monoamines. In support of its potent interaction at SERTs, S33005 markedly and dose dependently elevated dialysate levels of 5-HT in frontal cortex, hippocampus, and nucleus accumbens, structures implicated in the etiology and treatment of depressive states. Comparable, dose-dependent actions of venlafaxine and citalopram were observed, extending previous studies (Béïque et al., 1999, 2000; Popik, 1999; Millan et al., 2000b). Clomipramine acted similarly, whereas reboxetine (a weak ligand of SERTs) was ineffective and, in amplification of a single-dose study by Sacchetti et al. (1999), reboxetine dose dependently elevated frontocortical levels of NE. Most striking, however, was the pronounced influence of S33005, venlafaxine, and clomipramine upon dialysate levels of NE despite their preferential interaction with SERTs versus NETs in vitro. The reasons underlying this marked influence upon adrenergic transmission in vivo (Béïque et al., 1999, 2000; Dawson et al., 1999) require clarification. However, as alluded to above, S33005, venlafaxine, and clomipramine may potently interact with a distinctive population of NETs controlling NE levels in corticolimbic structures (Hughes and Stanford, 1998; Béïque et al., 1999, 2000; Kitayama et al., 1999). This may also be pertinent to the mild influence of high doses of citalopram (and other SSRIs) upon NE levels, an action mediated independently of serotonergic mechanisms (Millan et al., 2000b).

Chronic Administration. Chronic administration of tricyclics reduces cortical levels of -ARs, but the significance of this change to the evolution of antidepressant efficacy has been challenged since it occurs rapidly (1-3 days) (Okada and Tokumitsu, 1994; Newman-Tancredi et al., 1996; Anand and Charney, 2000). Moreover, several classes of antidepressant, including SSRIs, do not reliably down-regulate -ARs (Okada and Tokumitsu, 1994; Millan et al., 1997). Although venlafaxine desensitized pineal -ARs, it failed to reduce central -ARs (Muth et al., 1986; Schweizer et al., 1997; Nalepa et al., 1998), an observation confirmed herein, and extended to S33005. In contrast to -ARs, down-regulation of cortical 5-HT_2A receptors gradually increases over several (2-3) weeks of administration, paralleling clinical progression of therapeutic activity (Maes and Meltzer, 1995; Newman-Tancredi et al., 1996). Furthermore, antidepressant treatment of depressed patients likewise decreases levels of 5-HT_2A receptors (Yatham et al., 1999a,b). Thus, it is important that S33005 and venlafaxine reduced the density of 5-HT_2A receptors in frontal cortex. Interestingly, S33005, but not venlafaxine, significantly diminished 5-HT_2A receptor density at 2 weeks. Adrenergic mechanisms have been implicated in the influence of antidepressants upon 5-HT_2A sites (Gravel and de Montigny, 1987; Yatham et al., 1999a,b), so this difference may reflect the elevation in extracellular NE levels observed with S33005 but not venlafaxine at this time. Although this rapid reduction in 5-HT_2A receptor density with S33005 provides an interesting distinction to venlafaxine, and hints at a more rapid onset of action, it should be regarded as a "marker" rather than as a mechanism of antidepressant action. Indeed, the relationship between activity at 5-HT_2A receptors and depressive states is still under experimental and clinical exploration, and decreased transmission at 5-HT_2A sites is unlikely to be sufficient for expression of antidepressant properties (Yatham et al., 1999a,b; Meyer et al., 2001).

Summary and Conclusions. In conclusion, the novel cyclobutane derivative, S33005, potently interacts with SERTs and, less markedly, with NETs. In freely moving rats, S33005 markedly elevates extracellular levels of 5-HT and NE throughout corticolimbic structures, as well as DA levels in frontal cortex. S33005 may, thus, be differentiated from the SSRI, citalopram, and from the NE reuptake inhibitor, reboxetine. Although S33005 resembles venlafaxine, it displays higher potency. This preferential SERT > NET profile also bears comparison with clomipramine, but S33005 is devoid of the latter's affinity for adrenergic, histaminergic, and muscarinic receptors. In line with these observations, S33005 displays robust activity in behavioral paradigms suggestive of antidepressant activity (accompanying paper).