INTRODUCTION

The selective serotonin (5-hydroxytryptamine, 5-HT) reuptake inhibitors (SSRI) have become the most widely prescribed antidepressant drugs because of their relative absence of severe side effects. Their efficacy is comparable to that of tricyclic antidepressants (Tollefson et al, 1994), although the latter may be more effective in severely depressed inpatients (Danish University Antidepressant Group, 1986,1990). The increase in the synaptic concentration of 5-HT that follows reuptake inhibition is limited by a negative feedback involving 5-HT1A autoreceptors (Adell and Artigas, 1991; Artigas et al, 1996). Following repeated treatment, 5-HT1A receptors desensitize (Blier and de Montigny, 1994; Invernizzi et al, 1994; Hervás et al, 2001). This effect results in a reduced efficacy of the negative feedback and therefore in an increase of the extracellular concentration of 5-HT (5-HText) in the forebrain (Bel and Artigas, 1993; Invernizzi et al, 1994; Rutter et al, 1994; Kreiss and Lucki, 1995).

To overcome the above negative feedback and hasten the clinical action of SSRIs, the use of 5-HT1A receptor antagonists, in combination with SSRI, was proposed (Artigas, 1993). In keeping with this hypothesis, selective (eg WAY 100635) and nonselective 5-HT1A receptor antagonists have been shown to potentiate the increase in 5-HText produced by antidepressant drugs in the rat brain (see Artigas et al, 1996 for review). At the clinical level, several open-label (Artigas et al, 1994; Blier and Bergeron, 1995; Bakish et al, 1997) and placebo-controlled trials (Maes et al, 1996; Pérez et al, 1997; Tomé et al, 1997; Zanardi et al, 1997,1998; Bordet et al, 1998; Maes et al, 1999) have shown a faster action and/or greater efficacy of SSRIs when given in combination with the nonselective 5-HT1A receptor antagonist pindolol. However, the efficacy of pindolol in chronically ill or treatment-resistant patients is questionable (Moreno et al, 1997; Tomé et al, 1997; Berman et al, 1999; Pérez et al, 1999).

In vivo, pindolol appears to display a preferential antagonism of pre- vs postsynaptic 5-HT1A receptors in the rat brain (Romero et al (1996), Tada et al (1999), see however Corradetti et al, 1998). This may be related to its preferential occupancy of presynaptic 5-HT1A receptors, as assessed by positron emission tomography scanning in rat and human brains (Hirani et al, 2000; Rabiner et al, 2000; Martinez et al, 2000). However, pindolol displays higher affinity for β-adrenoceptors than for 5-HT1A receptors, which has led to the development of pure 5-HT1A antagonists or compounds with a dual 5-HT reuptake blockade/5-HT1A receptor antagonist action. Yet, given the involvement of hippocampal postsynaptic 5-HT1A receptors in the antidepressant effects of SSRIs (Blier and de Montigny, 1994; Haddjeri et al, 1998) there is some concern that the use of 5-HT1A receptor antagonists not discriminating between pre- and postsynaptic 5-HT1A receptors could cancel the benefits of enhancing presynaptically the 5-HT function. Moreover, recent data suggest that antidepressant drugs increase neurogenesis in the rat brain (Malberg et al, 2000), an effect that could be mediated by the activation of 5-HT1A receptors (Gould, 1999). Therefore, an alternative approach would be the use of drugs with dual 5-HT reuptake blockade/5-HT1A receptor agonist action. These could desensitize presynaptic 5-HT1A receptors by virtue of their action at raphe 5-HT1A autoreceptors. At the same time, they could enhance serotonergic transmission through the activation of postsynaptic 5-HT1A receptors and, unlike selective 5-HT1A receptor agonists, they could keep the tone on other postsynaptic 5-HT receptors because of their ability to inhibit 5-HT reuptake. The present study reports on the in vivo effects of one such compound (VN2222; (±)(1-(benzo[b]thiophen-3-yl)-3-[4-(2-methoxiphenylpiper azin-1-yl]propan-1-ol, Figure 1) (Martínez-Esparza et al, 2001) on the serotonergic system in rat brain using microdialysis and single-unit extracellular recordings.

Figure 1
figure 1

Chemical structure of VN2222, (±)1-(benzo[b]thiophen-3-yl)-3-[4-(2-methoxiphenyl piperazin-1-yl]propan-1-ol.

Full size image

MATERIALS AND METHODS

Animals

Male Wistar rats (Iffa Credo, Lyon, France) weighing 250–320 g at the time of experiments were used. Rats used in chronic treatments had a weight of 175 g at arrival and typically reached 250 g at the time of experiments. Animals were kept in a controlled environment (12 h light–dark cycle and 22±2°C room temperature). Food and water were provided ad libitum before and during the experiments. Animal care followed the European Union regulations (OJ of EC L358/1 18/12/1986).

Drugs and Treatments

5-HT, 8-OH-DPAT [8-hydroxy-2-(di-n-propylamino)tetralin], DOI (1-[2,5-dimethoxy-4-iodophenyl-2-aminopropane]), fluoxetine, and WAY 100635 [N-(2-(4-(2-methoxyphenyl)-1-piperazinyl)ethyl)-N-(2-pyridyl) cyclohexanecarboxamide·3HCl] were from RBI (Natick, MA). p-chlorophenylalanine methyl ester (PCPA) was from Sigma (St Louis, MO). Citalopram HBr was kindly provided by Lundbeck A/S. VN2222 was synthesized by VITA-INVEST (Sant Joan Despí, Barcelona, Spain). Concentrated stock solutions were prepared and aliquots were stored at −80°C. Working solutions were prepared daily by dilution. For local administration of drugs, these were dissolved in the dialysis fluid and applied by reverse dialysis at the stated concentrations (uncorrected for probe recovery). Systemic administration of VN2222 was performed s.c. at the stated doses in 10% Tween 80. 8-OH-DPAT and WAY 100635 were injected s.c. dissolved in saline. An additional experiment in which we examined the effects of the sustained administration of fluoxetine (3 mg/kg day, dissolved in 50% DMSO) and VN2222 (6 mg/kg day, dissolved in DMSO) was carried out in which these drugs were administered with minipumps following described procedures (Hervás et al, 2001). Control rats were treated with vehicle.

In electrophysiological experiments, 8-OH-DPAT, fluoxetine, WAY 100635, and VN2222 were administered i.v. through the femoral vein. In experiments assessing the changes in 5-HT1A receptor sensitivity produced by the chronic treatment with VN2222, rats were injected twice daily (12 h apart) for 2 weeks with the drug (2×3 or 2×10 mg/kg s.c.) or vehicle. The same treatment schedule was used to examine the effects of 2-week administration of fluoxetine (3 mg/kg day). Microdialysis experiments were performed on the second day after the last drug injection to avoid residual effects of the drug. The sensitivity of 5-HT1A receptors was also examined using single-unit recordings. These were performed 1 day after microdialysis (3-day washout) to assess the sensitivity of 5-HT1A receptors by both techniques in the same rats. Yet, owing to the greater experimental requirements of electrophysiological vs microdialysis experiments, single-unit recordings could not be performed in all rats. We report on electrophysiological data from two complete groups (rats treated with vehicle and 6 mg/kg day VN2222, n=7 and 5, respectively) and two rats treated with 20 mg/kg day VN2222.

Surgery and Microdialysis Procedures

Microdialysis procedures were performed essentially as described previously (see updated procedure in Adell and Artigas, 1998). In brief, anesthetized rats (pentobarbital 60 mg/kg i.p.) were stereotaxically implanted with I-shaped probes in the dorsal striatum or medial prefrontal cortex. The stereotaxic coordinates (in mm, with respect to bregma and duramater; Paxinos and Watson, 1986) were: prefrontal cortex (AP +3.4, L −0.8, DV −6.0, probe tip: 4 mm), dorsal striatum (AP +0.2, L −3.0, DV −8.0, probe tip 4 mm). Animals were allowed to recover from surgery for 20–24 h and then probes were perfused with artificial CSF (125 mM NaCl, 2.5 mM KCl, 1.26 mM CaCl2, and 1.18 mM MgCl2) pumped at 0.25 μl/min. When required by the experimental paradigm, the dialysis fluid was supplemented with 1 or 10 μM citalopram. Sample collection started 60 min after the beginning of perfusion. Dialysate samples were collected every 20 min (5 μl). Usually 5–6 fractions were collected before drug administration, of which four were used to obtain the individual basal values. At the end of the experiments, rats were killed by an overdose of anesthetic and the placement of the dialysis probes was checked by perfusing Fast Green dye and visual inspection of the probe track after cutting the brain at the appropriate levels.

Biochemical Analyses

The concentration of 5-HT in dialysates was determined using a modification of an HPLC method previously described (Adell and Artigas, 1998). 5-HT was also determined in homogenates of frontal cortex of control rats and of rats treated with PCPA (350 mg/kg, 2 days before). Briefly, after removal of the brains from the skull, pieces of frontal cortex were rapidly dissected out on ice, weighed, and stored frozen (−80°C). Brain tissue was ultrasonically homogenized (15–20 s) in cold perchloric acid (0.4 mol/l) containing 0.1% sodium metabisulfite, 0.01% EDTA, and 0.01% cysteine. After centrifugation (10 min, 10 000 g) aliquots of supernatants were analyzed by HPLC. The composition of HPLC eluant was as follows: 0.15 M NaH2PO4, 1.3 mM octyl sodium sulfate, 0.2 mM EDTA (pH 2.8 adjusted with phosphoric acid), plus 27% methanol. 5-HT was separated on a 3 μm ODS 2 column (7.5×0.46 cm; Beckman, San Ramon, CA) and detected amperometrically with a Hewlett-Packard 1049 detector (oxidation potential +0.6 V). Retention time was 3.5–4 min. 5-HT values were calculated by reference to standard curves run daily.

Single-Unit Recordings

We examined the ability of VN2222 to suppress the firing activity of 5-HT neurons in the dorsal raphe nucleus in untreated rats. This was compared to that of the 5-HT1A receptor agonist 8-OH-DPAT and the SSRI fluoxetine. The inhibitory effects of these agents on serotonergic firing rate were reversed by the administration of low doses of the 5-HT1A receptor antagonist WAY 100635 (5–10 μg/kg i.v.). In chronic experiments, we assessed the sensitivity of 5-HT1A receptors by comparing the ability of 8-OH-DPAT to suppress serotonergic cell firing in groups of rats pretreated with vehicle or VN2222 for 2 weeks after a washout period of 3 days.

Single-unit extracellular recordings were performed as previously described (Sawyer et al, 1985; Celada et al, 1996). Briefly, rats were anesthetized (chloral hydrate 400 mg/kg i.p.) and positioned in a stereotaxic apparatus. Additional doses of chloral hydrate (60 mg/kg) were administered i.v. Body temperature was maintained at 37°C throughout the experiment with a heating pad. All wound margins and points of contact between the animal and the stereotaxic apparatus were infiltrated with lidocaine solution (5%). In order to minimize pulsation, the atlanto-occipital membrane was punctured to release some CSF. For recordings in the dorsal raphe nucleus, a burr hole of approximately 4×4 mm2 was drilled over lambda and the sagital sinus was ligated, cut, and reflected. Single units in the dorsal raphe nucleus were recorded extracellularly with glass micropipettes pulled from 2.0-mm capillary glass (WPI, Saratosa, FL) on a Narishige PE-2 pipette puller (Narishige Scientific Instruments, Tokyo, Japan). Microelectrodes were filled with 2 M NaCl. Typically, impedance was between 4 and 10 MΩ. Descents were carried out along the midline. 5-HT neurons were recorded 5.1–5.8 mm below the brain surface and were identified according to electrophysiological criteria previously described (Wang and Aghajanian, 1977,1982). They exhibited a regular spontaneous firing rate with frequencies of 0.4–2.0 Hz, and 2–5 ms bi- or triphasic extracellular waveform.

Single-unit potentials were amplified with a Neurodata IR283 (Cygnus Technology Inc., Delaware Water Gap, PA), postamplified, and filtered with a Cibertec amplifier (Madrid, Spain), and computed on-line using a DAT 1401plus interface system Spike2 software (Cambridge Electronic Design, Cambridge, UK). Data were also recorded on audiotape for off-line reanalysis if necessary. After recording stable baseline spontaneous activity for at least 5 min, 8-OH-DPAT, fluoxetine, or VN2222 were administered i.v. every 2 min. Only one neuron per rat was recorded.

Data and Statistical Analysis

Microdialysis results are expressed as fmol/fraction (uncorrected for recovery) and shown in figures as percentages of basal values (individual means of four predrug fractions). Statistical analysis of drug effects on dialysate 5-HT was performed using one- or two-way analysis of variance (ANOVA) for repeated measures of raw data with time as repeated factor and dose or pretreatment as independent factor. Student's t-tests have been used where appropriate. Changes in firing rate were quantified by averaging the values in the second minute after drug injection and expressed as percentage of baseline. EC50 and ED50 values were calculated with the GraphPad Prism program (GraphPad software, San Diego, CA). Data are expressed as the mean±SEM. The number of animals in each group is given in figure legends. Statistical significance has been set at the 95% confidence level (two tailed).

RESULTS

In Vivo Effects of the Local and Systemic Administration of VN2222 on Extracellular 5-HT Concentration

Baseline 5-HText values in striatum were 4.2±0.2 fmol/fraction (n=26). The local application of VN2222 (10–300 μM, uncorrected for recovery) significantly increased 5-HText in dorsal striatum. The calculated EC50 value was 59 μM (Figure 2). The systemic administration of VN2222 (1, 3, and 10 mg/kg s.c.) induced a long-lasting and dose-dependent decrease of 5-HText in dorsal striatum reaching a maximum of 50% of baseline at 10 mg/kg (p<0.000001, time factor; p<0.000002, time×dose interaction) (Figure 3a). Average values (fractions 7–13) were 101.2±2.2, 92.7±6.4, 72.6±4.7, and 55.6±2.6% of baseline for controls, 1, 3, and 10 mg/kg VN2222, respectively. In the presence of 1 μM citalopram (to block locally 5-HT reuptake) baseline 5-HT values were 21.0±7.3 fmol/fraction (n=20). In this experimental condition, the s.c. administration of VN2222 induced a greater reduction of 5-HText (Figure 3b), which was also statistically significant (p<0.000002, dose factor; p<0.000001, time factor; p<0.000001, time×dose interaction). The average values (fractions 7–13) were 104.6±3.1, 90.0±2.7, 70.2±2.3, and 47.2±0.9% of baseline for controls, 1, 3, and 10 mg/kg VN2222, respectively. The reduction in 5-HText produced by 10 mg/kg VN2222 was fully counteracted by the administration of 0.3 mg/kg s.c. of the selective 5-HT1A receptor antagonist WAY 100635 (p<0.000001; Figure 3b).

Figure 2
figure 2

Increase of the extracellular 5-HT concentration in rat striatum produced by the local application of VN2222 using reverse dialysis (n=6). Data points are means of the last two fractions at each concentration, expressed as percentage of baseline. For comparison, the effect of the local application of fluoxetine is also shown (adapted from Hervás and Artigas, 1998). The calculated EC50 values for VN2222 and fluoxetine were, respectively, 59 and 31 μM (uncorrected for recovery).

Full size image
Figure 3
figure 3

(a) Effect of the s.c. administration of 1, 3, and 10 mg/kg of VN2222 (arrow) on the extracellular 5-HT concentration in rat striatum (n=5–6 rats/group). (b) Effect of the s.c. administration of 1, 3, and 10 mg/kg of VN2222 (first arrow) on the extracellular 5-HT concentration in rat striatum (n=4–5 rats/group) using a perfusion fluid containing 1 μM of the reuptake inhibitor citalopram. The administration of the 5-HT1A receptor antagonist WAY 100635 (0.3 mg/kg s.c.; second arrow) reversed the inhibition of 5-HT release elicited by VN2222 10 mg/kg s.c. See results for statistical analysis.

Full size image

Single-Unit Recordings in Dorsal Raphe Serotonergic Neurons

The i.v. administration of VN2222 reduced dose dependently the firing rate of identified 5-HT neurons in the dorsal raphe nucleus. Figure 4a shows an integrated firing rate histogram corresponding to the effects of cumulative doses of VN2222 (2–64 μg/kg i.v.). Neuronal activity ceased in most neurons examined at 64–128 μg/kg i.v. VN2222. The calculated ED50 was 14.9 μg/kg i.v. (Figure 4a). The effect of VN2222 was reversed by a low dose of WAY 100635 (5 μg/kg i.v.; Figure 4a). Owing to the high affinity of VN2222 for the 5-HT transporter and 5-HT1A receptors, we also examined the effects of two reference compounds, fluoxetine and 8-OH-DPAT on 5-HT cell firing. Both inhibited the firing rate of dorsal raphe serotonergic neurons, with ED50 values of 790 and 0.76 μg/kg i.v., respectively (Figure 4b).

Figure 4
figure 4

(a) Integrated firing rate histogram showing the effect of cumulative doses of VN2222 (arrows, up to 64 μg/kg i.v.) on the firing rate of a dorsal raphe 5-HT neuron. The suppression of the firing rate produced by VN2222 was reversed by the administration of a low dose of the selective 5-HT1A receptor antagonist WAY 100635 (5 μg/kg i.v.). (b) Dose–response curves of 8-OH-DPAT, VN2222, and fluoxetine to suppress firing rate of 5-HT neurons in the dorsal raphe nucleus. The ED50 values were 0.76 [0.62–0.94], 14.9 [7.1–31.5], and 790 [590–1060] μg/kg i.v., respectively (n=6–8 rats/group; 95% confidence intervals in brackets). (c) Comparison of the effect of VN2222 on the firing rate of dorsal raphe serotonergic neurons in untreated (n=7) or rats depleted of 5-HT by the pretreatment with the 5-HT synthesis inhibitor PCPA (n=5). See results for statistical analysis.

Full size image

To examine whether VN2222 inhibited 5-HT neuronal firing by a direct action on 5-HT1A receptors or indirectly, because of its ability to block 5-HT reuptake and increase 5-HText in the raphe nuclei (as an SSRI), we examined its effects on rats pretreated with the 5-HT synthesis inhibitor PCPA (350 mg/kg, 2 days before). The 5-HT concentration in frontal cortex (wet tissue) was 764±33 pmol/g (n=6) in control rats and 56±8 pmol/g (n=7) in PCPA-treated rats (7% of controls). In the latter group, VN2222 inhibited 5-HT cell firing with an ED50 of 46.9 μg/kg. Two-way ANOVA analysis of the data revealed a significant effect of VN2222 on cell firing (p<0.0001) and a significant VN2222×PCPA interaction (p<0.03) (Figure 4c). One neuron remained unaltered at a VN2222 dose of 64 μg/kg i.v. Omitting this neuron from the calculations resulted in an ED50 value of 27.1 μg/kg i.v. (n=4) and a nonsignificant (p=0.25) VN2222×PCPA interaction.

Effects of VN2222 on Postsynaptic 5-HT1A Receptors

In conditions of blockade of the 5-HT reuptake, the local application of potent 5-HT1A receptor agonists, such as 8-OH-DPAT or BAY×3702, in medial prefrontal cortex reduced 5-HText (Casanovas et al, 1999a). This effect is attributable to an activation of postsynaptic 5-HT1A receptors in this brain area since it was reversed by the coperfusion of WAY 100635 (Casanovas et al, 1999a). We used this experimental paradigm to assess the action of VN2222 at postsynaptic 5-HT1A receptors. Given the 5-HT reuptake blocking properties of VN2222, a maximal concentration of citalopram (10 μM; Hervás et al, 2000) was used to block locally 5-HT reuptake. In this experimental condition, the application of 8-OH-DPAT (100 μM) significantly reduced 5-HText in medial prefrontal cortex compared to control rats, perfused with the dialysis fluid supplemented with 10 μM citalopram (p<0.00001, time effect; p<0.00001, time×treatment interaction). The systemic administration of WAY 100635 0.3 mg/kg s.c. reversed the inhibition of 5-HT release produced by 8-OH-DPAT application (p<0.000001) (Figure 5). In control rats perfused with citalopram 10 μM, WAY 100635 0.3 mg/kg s.c. increased 5-HText above baseline (p<0.01).

Figure 5
figure 5

Effects of the local application of the selective 5-HT1A receptor agonist 8-OH-DPAT (100 μM) and VN2222 (300 μM) on the extracellular 5-HT concentration in medial prefrontal cortex (n=4 and 8, respectively). The 5-HT reduction elicited by 8-OH-DPAT and VN2222 was significantly counteracted by the s.c. administration of WAY 100635 (WAY; 0.3 mg/kg s.c.; arrow). Control rats (n=3) were perfused with the dialysis fluid containing citalopram for the entire experiment and were also injected with WAY 100635, which significantly elevated 5-HText over baseline. The bar shows the period of drug application. See results for statistical analysis.

Full size image

The application of VN2222 (300 μM) reduced 5-HText in medial prefrontal cortex to an extent comparable to that produced by 100 μM 8-OH-DPAT (p<0.0001, time effect; p<0.0002, time treatment interaction). This effect was also attenuated by the s.c. administration of WAY 100635 (p<0.01; Figure 5). In a small group of rats (n=3), the local application of VN2222 counteracted the increase in local 5-HT release induced by the selective stimulation of 5-HT2A receptors in medial prefrontal cortex produced by the 5-HT2 agonist DOI 100 μM (DOI=179±26% of baseline, DOI+VN2222=122±9% of baseline) (data not shown).

Chronic Administration of VN2222: Desensitization of 5-HT1A Autoreceptors

We examined the ability of VN2222 to desensitize 5-HT1A autoreceptors using two different experimental paradigms, in vivo microdialysis and single-unit recordings in the DR.

Rats treated with VN2222 (6 and 20 mg/kg day) had a weight gain that did not differ significantly from that of controls (mean values were: controls 74 g, VN2222 6 mg/kg day 83 g, VN2222 20 mg/kg day 85 g). On day 15, rats were implanted with dialysis probes in prefrontal cortex and allowed to recover from anesthesia for ca. 20–24 h, before microdialysis experiments began. Dialysis probes were perfused with 10 μM citalopram to block the 5-HT reuptake in medial prefrontal cortex. After collection of baseline values, rats were administered the selective 5-HT1A receptor agonist 8-OH-DPAT (25 μg/kg s.c.). Baseline 5-HText values did not significantly differ between groups (one-way ANOVA) although there was a tendency towards lower values in VN2222-treated rats: 13.1±1.8 fmol/fraction in controls, 9.4±1.8 fmol/fraction in rats treated with VN2222 6 mg/kg day, and 7.6±1.2 fmol/fraction in those treated with VN2222 20 mg/kg day. One-way ANOVA did not show a significant difference between controls and VN2222-treated rats (F2,20=2.55; p=0.10). The reduction in 5-HText induced by 8-OH-DPAT was significantly less marked in rats pretreated with VN2222 than in controls (Figure 6). Two-way repeated measures ANOVA revealed a significant effect of time (p<0.000001) and of the time×treatment interaction (p<0.002). The injection of 8-OH-DPAT reduced 5-HText maximally to 34±6, 51±8, and 55±7% of baseline for controls, VN2222 6 mg/kg day and VN2222 20 mg/kg day, respectively.

Figure 6
figure 6

Effect of the administration of an 8-OH-DPAT challenge (25 μg/kg s.c.) on the dialysate 5-HT concentration in prefrontal cortex of rats pretreated with vehicle (n=12), 2×3 mg/kg day VN2222 (n=5) and 2×10 mg/kg day VN2222 (n=6). Dialysis probes were perfused with artificial CSF containing 10 μM of the 5-HT reuptake inhibitor citalopram. See results for statistical analysis.

Full size image

After microdialysis experiments, rats were kept in their cages and, on the following day, the effects of the pretreatment with vehicle or 6 mg/kg day VN2222 on 5-HT1A receptor sensitivity were assessed by single-unit recordings of dorsal raphe 5-HT neurons. Once a stable recording was obtained for 5 min, cumulative doses of 8-OH-DPAT were injected (0.25–32 μg/kg i.v.) and individual dose–response curves were constructed (one neuron per rat). The baseline firing rate did not significantly differ among groups (Figure 7). This figure shows also the baseline firing of two neurons of rats treated with 20 mg/kg day VN2222. Figure 8 shows representative firing rate histograms of neurons corresponding to the effects of 8-OH-DPAT in controls and rats pretreated with 6 mg/kg day VN2222. The ED50 of 8-OH-DPAT for the vehicle-treated group was 0.45 μg/kg i.v. The corresponding value in rats treated with 6 mg/kg day VN2222 was 2.34 μg/kg i.v. Two-way ANOVA revealed a significant effect of the pretreatment (VN2222) (p<0.002), dose (8-OH-DPAT) (p<0.0001) and of the dose×pretreatment interaction (p<0.0001) (Figure 8). In the two rats treated with 20 mg/kg day VN2222, 8-OH-DPAT elicited a moderate decrease of the firing rate at 32 μg/kg i.v. in one rat whereas in another rat, firing rate ceased at 4 μg/kg i.v. 8-OH-DPAT.

Figure 7
figure 7

Baseline firing rate of dorsal raphe serotonergic neurons in rats treated with vehicle (n=7), 6 mg/kg day VN2222 (n=5) and 20 mg/kg day VN2222 (n=2). No significant differences were found among groups (one-way ANOVA).

Full size image
Figure 8
figure 8

Integrated firing rate histograms corresponding to the effects of cumulative doses of 8-OH-DPAT (arrows, in μg/kg i.v.) on two representative dorsal raphe 5-HT neurons from rats treated with vehicle (a) and VN2222 (6 mg/kg day). (b) The suppression of serotonergic cell firing induced by 8-OH-DPAT was reversed by a low dose of WAY 100635 (WAY; 5–10 μg/kg i.v.). (c) Dose–response curves of the suppressing effect of 8-OH-DPAT on the firing of dorsal raphe 5-HT neurons in rats treated with vehicle (n=7) and 6 mg/kg day VN2222 (n=5). The calculated ED50 values were 0.45 [0.37–0.54] and 2.34 [1.87–2.92] μg/kg i.v., respectively (95% confidence intervals given in brackets).

Full size image

To further examine the desensitization of 5-HT1A autoreceptors induced by the chronic administration of VN2222, two groups of rats were treated with 6 mg/kg day VN2222 (as above) or vehicle for 14 days and microdialysis experiments were conducted on the 15th day (15 h after last dose). Baseline 5-HT values were 2.6±0.2 fmol/fraction in controls and 3.0±0.3 fmol/fraction in rats treated with VN2222. The administration of a challenge dose of fluoxetine (10 mg/kg i.p.) elevated extracellular 5-HT significantly more in rats pretreated with VN2222 than in controls (p<0.039, group effect; p<0.000001, time effect; p<0.01, time×group interaction; Figure 9).

Figure 9
figure 9

Effect of the administration of a fluoxetine challenge (10 mg/kg i.p., arrow) on the extracellular 5-HT in the medial prefrontal cortex of rats treated with vehicle (open circles; n=7) or VN2222 6 mg/kg day (filled circles; n=8). Two-way repeated measures ANOVA showed a significant effect of the group, time, and time×group interaction (see text for statistical details).

Full size image

In an additional experiment we examined the effect of the sustained administration of VN2222 (6 mg/kg day) and fluoxetine (3 mg/kg day) on extracellular 5-HT in prefrontal cortex using minipumps. Baseline 5-HT values were collected on the 14th day after implant, with the minipumps on board. The 5-HT values in the control groups for fluoxetine and for VN2222 did not differ significantly and were pooled to perform subsequent statistical analyses. Fluoxetine pretreatment increased dialysate 5-HT by 79% (7.5±0.7 vs 4.2±0.3 fmol/fraction; p<0.00001) whereas VN2222 elicited a more moderate but still significant 33% increase (5.6±0.5 vs 4.2±0.3 fmol/fraction; p<0.015) (Table 1).

Table 1 Dialysate 5-HT Levels in the Prefrontal Cortex of Rats Treated with Fluoxetine 3 mg/kg day and VN2222 6 mg/kg day with Minipumps for 2 Weeks
Full size table

DISCUSSION

VN2222 is a new serotonergic agent with putative antidepressant properties. In agreement with its in vitro receptor profile (Martínez-Esparza et al, 2001; Table 2), the in vivo pharmacological properties of VN2222 are consistent with a dual action at the 5-HT transporter and 5-HT1A receptors. Thus, it inhibits 5-HT reuptake in vivo with a potency slightly lower than that of fluoxetine. On the other hand, its neurochemical and electrophysiological actions in rat brain are consistent with those of a pre- and postsynaptic 5-HT1A receptor agonist, in the dorsal raphe nucleus and the medial prefrontal cortex, respectively.

Table 2 Summary of the In Vitro Affinities of VN2222
Full size table

VN2222 increased 5-HText in the striatum when applied locally by reverse dialysis in the same concentration range as fluoxetine. The ratio between the respective EC50 values is ∼2 (59 μM for VN2222, 28–31 μM for fluoxetine; Hervás and Artigas, 1998; Hervás et al, 2000). However, its systemic administration reduced 5-HText in the same brain area. The striatum was chosen because (together with prefrontal cortex) it is one of the brain areas where the 5-HT release is more sensitive to the activation of 5-HT1A autoreceptors by direct and indirect (eg SSRI) agonists (Kreiss and Lucki, 1994; Casanovas et al, 1997; Romero and Artigas, 1997; Casanovas et al, 2000). The discrepancy between the effects of the local (5-HT increase) and systemic administration (5-HT decrease) of VN2222 is likely accounted for by its agonist effect at raphe 5-HT1A autoreceptors. This action would reduce 5-HText to an extent larger than the potential increase produced by systemic reuptake blockade. Indeed, the systemic administration of 5-HT uptake inhibitors activates a 5-HT1A autoreceptor-mediated negative feedback that offsets the increase in forebrain 5-HText produced by reuptake blockade (Adell and Artigas, 1991; Artigas et al, 1996). Hence, the systemic administration of SSRIs reduces 5-HText when the 5-HT reuptake is locally blocked, an action that illustrates their 5-HT release-reducing properties (Rutter and Auerbach, 1993; Hjorth and Auerbach, 1994a; Romero and Artigas, 1997). This effect is because of the activation of raphe 5-HT1A autoreceptors since it is antagonized by the local (in raphe) or systemic administration of WAY 100635 and other nonselective 5-HT1A antagonists (Hjorth and Auerbach, 1994a; Romero and Artigas, 1997). For SSRIs, the balance between the two opposite factors controlling forebrain 5-HText (inhibition of 5-HT reuptake vs activation of 5-HT1A autoreceptors) is favorable to the former, whereas in the case of VN2222, the latter action predominates. Thus, high doses of SSRIs increase 5-HText after single treatment whereas VN2222 reduces 5-HText.

In the presence of 1 μM citalopram in the dialysis fluid, VN2222 reduced 5-HText slightly more than in standard dialysis conditions and this effect was counteracted by WAY 100635. This supports the exclusive involvement of 5-HT1A receptors in the reduction of 5-HT release induced by VN2222. The dissimilar effects of SSRIs and VN2222 in the absence of citalopram, together with the fact that VN2222 reduced 5-HText in both experimental conditions, suggest that VN2222 has direct agonist properties at 5-HT1A autoreceptors. This view is supported by the results of single-unit recording experiments, in which VN2222 suppressed serotonergic cell firing and this effect was reversed by WAY 100635. The difference in the i.v. and s.c. doses of WAY 100635 used in electrophysiogical and microdialysis experiments of the present study is consistent with previous data showing antagonism of 5-HT1A receptor-mediated effects (Forster et al, 1995) and reflects differences in the bioavailability by both routes. Thus, maximal effects of i.v. WAY 100635 take place in 1–2 min whereas they occur at 40–60 min by the s.c. route. An additional reason to use low i.v. doses of WAY 100635 was the observation that in some instances, higher doses suppressed serotonergic cell firing, in agreement with previous observations (Martin et al, 1999). Although VN2222 shows an appreciable affinity for α1-adrenoceptors ( Table 2) its suppressing effects on 5-HT release and cell firing do not appear to be mediated by a blockade of the α1-adrenoceptor-mediated activation of 5-HT neurons, as these effects are fully counteracted by WAY 100635.

The affinity of VN2222 for 5-HT1A receptors is lower than that of 8-OH-DPAT (Martínez-Esparza et al, 2001). In agreement, the calculated ED50 of VN2222 to suppress 5-HT cell firing (15 μg/kg i.v.) is greater than that of 8-OH-DPAT (0.76 μg/kg i.v.) but comparable or lower than that of other 5-HT1A receptor agonists, such as buspirone (11 μg/kg i.v.; VanderMaelen et al, 1986), ipsapirone (30–125 μg/kg i.v.; Basse-Tomusc and Rebec, 1986; Cox et al, 1993), or flesinoxan (21–108 μg/kg i.v.; Gobert et al, 1995; Hadrava et al, 1995).

5HT1A receptor agonists and 5-HT reuptake blockers inhibit directly and indirectly, respectively, the firing activity of serotonergic neurons (Scuvée-Moreau and Dresse, 1979; Quinaux et al, 1982; Sprouse and Aghajanian, 1986; Blier and de Montigny, 1987). Since VN2222 appears to display both activities in vivo, either could theoretically contribute to its firing-suppressant effects. However, the inhibition of 5-HT reuptake seems to play a minor role in this effect, since (a) the ED50 of VN2222 to suppress 5-HT cell firing is 53 times lower than that of fluoxetine, whereas its EC50 to inhibit locally 5-HT reuptake is double that of fluoxetine and (b) VN2222 inhibited the serotonergic cell firing in rats depleted of 5-HT by pretreatment with the 5-HT synthesis inhibitor PCPA. This drug abolishes the inhibitory activity of indirect 5-HT1A receptor agonists, including 5-HT reuptake blockers and cocaine (Trulson and Crisp, 1986; Cunningham and Lakoski, 1990). There was some difference between the ED50 values in both situations but this was small and mostly because of the influence of one neuron that deviated from the average sensitivity and showed no inhibition at 64 μg/kg VN2222, which may be because of adaptive mechanisms in some 5-HT neurons after irreversible inhibition of tryptophan hydroxylase by PCPA.

The local application of 8-OH-DPAT and VN2222 in medial prefrontal cortex markedly reduced basal 5-HText. These experiments were carried out in the presence of a maximal concentration of citalopram (10 μM; Hervás et al, 2000) to mask completely the 5-HT uptake inhibition elicited by VN2222. It was hypothetized that in this experimental condition we would observe the putative 5-HT1A receptor agonist action of VN2222, because the 5-HT reuptake was completely blocked by citalopram. In agreement with previous observations (Casanovas et al, 1999a), WAY 100635 antagonized the reduction induced by both agents, which supports the involvement of postsynaptic 5-HT1A receptors in these effects. Interestingly, WAY 100635 administration increased 5-HText above baseline in control rats perfused with 10 μM citalopram. This suggests that the 5-HT elevation produced by this citalopram concentration was sufficient to activate postsynaptic 5-HT1A receptors and self-attenuate 5-HT release. In the absence of citalopram, WAY 100635 administration does not increase 5-HText in this and other forebrain areas of unanesthetized rats (Invernizzi et al, 1997; Romero and Artigas, 1997; Hervás et al, 2000; Romero et al, unpublished observations).

The reduction in 5-HText in medial prefrontal cortex elicited by the local application of 5-HT1A receptor agonists is because of the activation of postsynaptic 5-HT1A receptors (Casanovas et al, 1999a; Celada et al, 2001). Pyramidal neurons in prefrontal cortex project to and control the activity of ascending 5-HT neurons in the dorsal raphe nucleus through the activation of 5-HT1A and 5-HT2A receptors (Sesack et al, 1989; Hajós et al, 1998; Peyron et al, 1998; Celada et al, 2001; Martín-Ruiz et al, 2001). Thus, the activation of 5-HT1A receptors in pyramidal neurons results in neuronal hyperpolarization (Araneda and Andrade, 1991) and a reduction of the excitatory input onto ascending 5-HT neurons (Celada et al, 2001), which decreases firing-dependent 5-HT release. Thus, in common with selective 5-HT1A agonists like 8-OH-DPAT or BAY×3702, VN2222 displays agonist properties at postsynaptic 5-HT1A receptors. Moreover, VN2222 counteracted the 5-HT2A receptor-mediated elevation in 5-HText induced by the local application of DOI, as observed previously for the potent 5-HT1A receptor agonist BAY×3702 (Martín-Ruiz et al, 2001). The agonism of VN2222 at cortical postsynaptic 5-HT1A receptors may be important for its antidepressant effects in animal models of depression (Martínez-Esparza et al, 2001) because 5-HT1A receptor agonists show anxiolytic/antidepressant properties (Lucki et al, 1994; de Vry, 1995). In addition, electrophysiological data support that the activation of hippocampal 5-HT1A receptors may be important for the antidepressant action (Blier and de Montigny, 1994; Haddjeri et al, 1998). However, the present data cannot establish whether VN2222 acts as a full or partial agonist in vivo at cortical 5-HT1A receptors, yet the similar reduction of 5-HT release elicited by VN2222 and 8-OH-DPAT suggests a substantial agonist effect of the former. Moreover, in hippocampal membranes, VN2222 inhibits the forskolin-induced cAMP formation with a potency lower than that of 8-OH-DPAT (Martínez-Esparza et al, 2001).

The repeated administration of VN2222 resulted in a functional desensitization of 5-HT1A autoreceptors, as assessed by in vivo microdialysis and single-unit recordings of dorsal raphe 5-HT neurons. The five-fold shift of the sensitivity of 5-HT1A receptors produced by 6 mg/kg day VN2222 is greater than that produced by maximal doses of SSRIs and selective 5-HT1A receptor agonists (Haddjeri et al, 1999; Le Poul et al, 1995,1999,2000) and suggests a very efficient desensitization of 5-HT1A autoreceptors controlling serotonergic cell firing. Interestingly, the baseline firing rate of rats treated chronically with VN2222 was higher, although not significantly different from controls, which suggests that 5-HT neurons had—at least—recovered their normal firing rate, in agreement with previous observations with other antidepressant drugs (Blier and de Montigny, 1994).

Likewise, in microdialysis experiments, the 8-OH-DPAT challenge reduced 5-HT release significantly less in rats pretreated with VN2222, further supporting the view that this agent desensitized 5-HT1A autoreceptors. In microdialysis experiments, only one dose of 8-OH-DPAT was used. Therefore, we could not determine whether the difference between controls and VN2222-treated rats was because of a change in potency or in maximal effect. However, the electrophysiological data suggest a change in potency. Similarly, a challenge dose of fluoxetine was able to induce a greater increase in 5-HText in rats pretreated with VN2222 than in controls, an observation that also supports the view that VN2222 effectively desensitized 5-HT1A autoreceptors after repeated treatment for 2 weeks.

The increase in 5-HText produced by minipump administration of fluoxetine 3 mg/kg day is in agreement with previous data from this laboratory (Hervás et al, 2000), although the effect size was slightly lower in the present experiments because of higher baseline 5-HText values in controls. VN2222 produced a more moderate increase in 5-HText which was, however, statistically significant (Table 1). Again, since acute treatment with this daily dose reduced 5-HText, this is an additional argument in favor of an effective desensitization of 5-HT1A receptors by VN2222.

Taken together, electrophysiological and microdialysis experiments suggest that a 2-week VN2222 treatment markedly reduced the effectiveness of 5-HT1A receptors controlling presynaptic serotonergic function. Moreover, (a) postsynaptic 5-HT1A receptors in cortico-limbic areas are involved in the effects of systemically administered 8-OH-DPAT (Ceci et al, 1994; Romero et al, 1994; Artigas et al, 1998; Hajós et al, 1999; Celada et al, 2001) and (b) these do not desensitize after prolonged treatments (Blier and de Montigny, 1994; Dong et al, 1997; Le Poul et al, 2000). Hence, it is possible that the actual sensitivity of raphe 5-HT1A autoreceptors after repeated VN2222 treatment is actually lower than that determined in the present experiments using the systemic 8-OH-DPAT challenge.

The present data are in keeping with previous work showing that prolonged administration of 5-HT1A receptor agonists and SSRIs desensitizes dorsal raphe 5-HT1A autoreceptors (Blier and de Montigny, 1987,1994; Hensler et al, 1991; Bohmaker et al, 1993; Invernizzi et al, 1994; Le Poul et al, 1995,1999,2000; Dong et al, 1997; Casanovas et al, 1999b; Hervás et al, 2000). However, some microdialysis studies have failed to observe such effects (Sharp et al, 1993; Hjorth and Auerbach, 1994b; Auerbach and Hjorth, 1995; Bosker et al, 1995; Invernizzi et al, 1995). The origin of these discrepancies is unclear and may involve differences in drug half-lives of the agents used, route of administration (eg repeated injections vs minipumps) and regional effects, since most of the latter studies were conducted in the hippocampus, mostly innervated by median raphe 5-HT neurons. Our own microdialysis data, obtained in prefrontal cortex, are consistent with single-unit recordings in the dorsal raphe indicating a desensitization of 5-HT1A receptors. More importantly, some microdialysis and single-unit data were obtained in the same animals, which reinforces the association between both measures. It should be mentioned that the present observations have been obtained with a relative low dose of VN2222 (6 mg/kg day). This is equivalent to 3 mg/kg day fluoxetine in terms of in vivo reuptake blockade and lower than the standard 10 mg/kg day dose, which may appear too high compared with the standard clinical regimen (20 mg/day). It is thus possible that a more effective desensitization and/or greater 5-HText increments could have been obtained after repeated treatment with larger doses of fluoxetine and VN2222.

In summary, the present results show that VN2222 inhibits the 5-HT reuptake and is a direct agonist at pre- and postsynaptic 5-HT1A receptors in vivo. SSRIs possess significant clinical antidepressant activity and experimental data support the involvement of postsynaptic 5-HT1A receptors in their clinical action. VN2222 shares both pharmacological activities in vivo, and in particular, it behaves as a cortical 5-HT1A receptor agonist and effectively desensitizes 5-HT1A autoreceptors. The therapeutic lag of SSRIs can be attributable to the 5-HT1A-autoreceptor-mediated impairment of serotonergic activity that follows reuptake blockade in the raphe (presynaptic component) and to the absence of a sufficient tone on postsynaptic 5-HT receptors (postsynaptic component). As 5-HT1A receptor desensitization progresses, the tone on postsynaptic receptors increases, which results in a therapeutic action. VN2222 shares with SSRIs the property of suppressing serotonergic activity (presynaptic component). Such a reduction may appear contradictory with the desired enhancement of the tone on postsynaptic 5-HT receptors demanded of antidepressant agents, which could potentially result in a greater latency to onset of action. However, recent evidence suggests that antidepressant drugs enhance neurogenesis in the rat hippocampus (Malberg et al, 2000), an effect possibly mediated by the activation of postsynaptic 5-HT1A receptors (Gould, 1999). Moreover, blockade of 5-HT2A neurotransmission by M100907 enhanced the antidepressant effects of fluoxetine in the DRL 72-S reinforcement schedule without concurrently increasing 5-HText (Marek et al, 2001). Since 5-HT2A colocalize with 5-HT1A receptors in pyramidal neurons of medial prefrontal cortex (Martín-Ruiz et al, 2001) and both mediate opposing effect on pyramidal cell excitability (Araneda and Andrade, 1991) these behavioral results suggest that the superior antidepressant effect of fluoxetine+M100907 is because of an enhanced 5-HT1A–mediated transmission. VN2222 may activate postsynaptic 5-HT1A receptors during a time when SSRIs exert little or no activation of such receptors, an action that may result in a more rapid and/or effective antidepressant action.