Antipsychotic drugs reverse the AMPA receptor-stimulated release of 5-HT in the medial prefrontal cortex

Authors

  • Mercè Amargós-Bosch,

    1. Department of Neurochemistry and Neuropharmacology, Institut d’ Investigacions Biomèdiques de Barcelona (CSIC), IDIBAPS, Barcelona, Spain
    Search for more papers by this author
  • Albert Adell,

    1. Department of Neurochemistry and Neuropharmacology, Institut d’ Investigacions Biomèdiques de Barcelona (CSIC), IDIBAPS, Barcelona, Spain
    Search for more papers by this author
  • Francesc Artigas

    1. Department of Neurochemistry and Neuropharmacology, Institut d’ Investigacions Biomèdiques de Barcelona (CSIC), IDIBAPS, Barcelona, Spain
    Search for more papers by this author

Address correspondence and reprint requests to Francesc Artigas, PhD; Department of Neurochemistry and Neuropharmacology, Institut d’ Investigacions Biomèdiques de Barcelona (CSIC), IDIBAPS, Rosselló, 161, 6th floor, 08036 Barcelona, Spain. E-mail: fapnqi@iibb.csic.es

Abstract

The prefrontal cortex (PFC) is involved in the pathophysiology of schizophrenia. PFC neuronal activity is modulated by monoaminergic receptors for which antipsychotic drugs display moderate-high affinity, such as 5-HT2A and α1-adrenoceptors. Conversely, PFC pyramidal neurons project to and modulate the activity of raphe serotonergic neurons and serotonin (5-HT) release. Under the working hypothesis that atypical antipsychotic drugs may partly exert their action in PFC, we assessed their action on the in vivo 5-HT release evoked by increasing glutamatergic transmission in rat medial PFC (mPFC). This was achieved by applying S-AMPA in mPFC (reverse dialysis) or by disinhibiting thalamic excitatory afferents to mPFC with bicuculline. The application of haloperidol, chlorpromazine, clozapine and olanzapine in mPFC by reverse dialysis (but not reboxetine or diazepam) reversed the S-AMPA-evoked local 5-HT release. Likewise, the local (in mPFC) or systemic administration of these antipsychotic drugs reversed the increased prefrontal 5-HT release produced by thalamic disinhibition. These effects were shared by the 5-HT2A receptor antagonist M100907 and the α1-adrenoceptor antagonist prazosin. However, raclopride (DA D2 antagonist) had very modest effects. These results suggest that, besides their action in limbic striatum, antipsychotic drugs may attenuate glutamatergic transmission in PFC, possibly by interacting with 5-HT2A and/or α1-adrenoceptors.

Abbreviations used
5-HT

5-hydroxytryptamine or serotonin

AMPA

alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate

CM

centromedial nucleus of the thalamus

DA

dopamine

DOI

1-[2,5-dimethoxy-4-iodophenyl-2-aminopropane]

iGluR

ionotropic glutamate receptors

MD

mediodorsal nucleus of the thalamus

mPFC

medial prefrontal cortex

NMDA

N-methyl-D-aspartate

PFC

prefrontal cortex

The prefrontal cortex (PFC) plays a key role in higher brain functions (Fuster 2001). Many neurochemical, cellular and functional alterations have been reported in the PFC of schizophrenic patients (Weinberger et al. 1994; Andreasen et al. 1997; Bertolino et al. 2000; Lewis and Lieberman 2000; Lewis et al. 2005). In particular, changes in prefrontal GABAergic and glutamatergic transmission have been described (Lewis and Lieberman 2000; Tsai and Coyle 2002; Krystal et al. 2003; Moghaddam 2003), with a significant loss of function of parvalbumin-containing GABAergic interneurons (Lewis et al. 2005).

Non-competitive N-methyl-D-aspartate (NMDA) receptor antagonists are widely accepted as pharmacological models of schizophrenia. The behavioral deficits induced by these agents resemble schizophrenic symptoms, which suggests a glutamatergic hypofunction in schizophrenia. However, neurochemical (Moghaddam et al. 1997) and electrophysiological observations (Suzuki et al. 2002; Jackson et al. 2004) indicate that these agents increase glutamatergic transmission in mPFC, possibly by acting in afferent areas, such as the hippocampus (Jodo et al. 2005).

The activity of projection (pyramidal) neurons in PFC-which make up c. 75% of all neurons in PFC-depends on glutamatergic inputs from cortical and subcortical areas and is locally modulated by GABA interneurons. Main subcortical excitatory inputs arise from the mediodorsal/centromedial nuclei of the thalamus (MD/CM), the hippocampus and the amygdala, which are reciprocally connected with the PFC (Kuroda et al. 1998; Groenewegen and Uylings 2000; Van der Werf et al. 2002). Interestingly, the PFC and the brainstem monoaminergic nuclei (ventral tegmental area, raphe nuclei and locus coeruleus) are also reciprocally connected (Groenewegen and Uylings 2000) and exert a mutual influence. Thus, catecholaminergic and serotonergic axons innervate the PFC and modulate neuronal activity through various inhibitory and excitatory receptors (Araneda and Andrade 1991; Pompeiano et al. 1992, 1994; Pieribone et al. 1994; Aghajanian and Marek 1997, 1999; O’Donnell 2003; Amargós-Bosch et al. 2004; Puig et al. 2005). In turn, the activity of brainstem monoaminergic neurons and transmitter release is modulated by descending inputs from PFC (Aghajanian and Wang 1977; Thierry et al. 1979; Hajós et al. 1998; Jodo et al. 1998; Celada et al. 2001; Martín-Ruiz et al. 2001; Puig et al. 2003).

Classical neuroleptics are believed to exert their therapeutic action by modulating information processing along limbic basal ganglia circuits through the blockade of dopamine (DA) D2 receptors in the ventral striatum (Moore et al. 1999; Grace 2000). The therapeutic effect of these drugs typically occurs after a threshold blockade of c. 80% of DA D2 receptors, which causes extrapyramidal motor symptoms that compromise treatment compliance. However, atypical antipsychotic drugs like clozapine or olanzapine show a greater occupancy of cortical 5-HT2 receptors than of DA D2 striatal receptors at therapeutic doses, as assessed by PET scan (Farde et al. 1992; Kapur et al. 1998), which suggests that these agents partly exert their therapeutic effect through a cortical action. In particular, the presence of antipsychotic-sensitive monoaminergic receptors (e.g. 5-HT1A, 5-HT2A/2C receptors, α1-adrenoceptors, among others) and the role of PFC in behavioral control suggest that antipsychotics may have additional actions in this cortical area.

Here we examined the effect of classical (chlorpromazine and haloperidol) and atypical (clozapine and olanzapine) antipsychotic drugs on the release of 5-HT in mPFC evoked by an increase of glutamatergic transmission, under the working hypothesis that antipsychotics may attenuate the excitatory drive to midbrain and hence, reduce the in vivo terminal 5-HT release. The activity of PFC neurons was enhanced by locally applying S-AMPA and by disinhibiting thalamic afferents to mPFC, a procedure that markedly increases the activity of pyramidal neurons in mPFC (Puig et al. 2003).

Materials and methods

Animals

Male Wistar rats (Iffa Credo, Lyon, France) weighing 280–320 g at the time of the experiments were used. The animals were housed in groups of four per cage until the onset of the experiments and kept under a controlled temperature of 22 ± 2°C and a 12 h lighting cycle (lights on at 07:00). After surgery, rats were housed individually. Food and water were always freely available throughout the experiments. All experimental procedures were in strict compliance with the Spanish legislation and the European Communities Council Directive on ‘Protection of Animals Used in Experimental and Other Scientific Purposes’ of 24 November 1986 (86/609/EEC).

Chemicals

5-HT oxalate, (S)-AMPA (alpha-amino-3-hydroxy-5-methyl-4-isoxazole-4-propionate), bicuculline, chlorpromazine, diazepam, prazosin, reboxetine and raclopride were from Sigma (Tres Cantos, Spain). Haloperidol and clozapine were from Tocris (Bristol, UK). M100907 (R-(+)-alpha-(2,3-dimethoxyphenyl)-1-[4-fluorophenylethyl]-4-piperidinemethanol; Lilly code LY 368675) and olanzapine were from Eli Lilly & Co. Other materials and reagents were from local commercial sources. Drugs were dissolved in the perfusion fluid or water (except clozapine, dissolved in acetic acid, and olanzapine, dissolved in HCl). Concentrated solutions (1 mmol/L; pH adjusted to 6.5–7 with NaHCO3 when necessary) were stored at −80°C and working solutions were prepared daily by dilution in artificial CSF. Concentrations are expressed as free bases. In experiments involving local administration, concentrated drug solutions were diluted with the perfusion fluid and applied by reverse dialysis. Control rats were perfused for the entire experiment with artificial CSF. Vehicles did not significantly affect the 5-HT output in mPFC. The bars in the figures show the period of local drug application (corrected for the void volume of the system). Systemic administration of drugs was carried out s.c. at the stated doses after adjusting the pH to 6–6.5 with NaHCO3. Controls were injected with saline.

The concentrations of antipsychotic drugs and receptor antagonists were taken from previous studies in which PFC neurons were stimulated with the non-selective 5-HT2A receptor agonist 1-[2,5-dimethoxy-4-iodophenyl-2-aminopropane] (DOI) and the, α1-adrenoceptor agonist cirazoline (Martín-Ruiz et al. 2001; Amargós-Bosch et al. 2003, 2004; Bortolozzi et al. 2003; Puig et al. 2003). In order to compare the effects of the antipsychotic drugs on various models of PFC stimulation, the same local concentrations were applied by reverse dialysis. Systemic doses were chosen among various studies showing the effectiveness of antipsychotic drugs in neurochemical or pharmacological models of schizophrenia. The occupancy of various monoaminergic receptors was also considered (Schotte et al. 1993; Chaki et al. 1999).

Despite the in vitro 10−8– 10−7 affinity of antipsychotics for 5-HT2A and α1-adrenoceptors, the use of concentrations in the μmol/L range is required in in vivo microdialysis to significantly affect neurotransmitter receptors or transporters (e.g. Tao et al. 2000; Hervás et al. 2000; Sakai and Crochet 2001; West and Grace 2002). This is due to the fact that effective concentrations at receptors is limited by the low application rate (in the low nmol/h range) and the continuous clearance of applied drug by the CSF. Moreover, a substantial number of post-synaptic receptors in mPFC neurons must be recruited to activate/inhibit the mPFC-raphe circuit and elicit changes in mPFC 5-HT release (Celada et al. 2001; Martín-Ruiz et al. 2001).

Surgery and microdialysis experiments

An updated description of the microdialysis procedures used can be found in Adell and Artigas (1998) and Puig et al. (2003). Briefly, anesthetized rats (sodium pentobarbital, 60 mg/kg i.p.) were stereotaxically implanted with concentric microdialysis probes equipped with a Cuprophan membrane. The probes were perfused at 1.5 μL/min with artificial CSF (125 mmol/L NaCl, 2.5 mmol/L KCl, 1.26 mmol/L CaCl2 and 1.18 mmol/L MgCl2) containing 1 μmol/L citalopram. After 1-h stabilization period, four fractions were collected to obtain basal values before local (reverse dialysis) or systemic administration of drugs. Successive 20-min (30 μL) dialysate samples were collected. At the end of the experiments, rats were killed by an overdose of anesthetic. The placement of the dialysis probes was examined by perfusion of fast green dye and visual inspection of the probe track after cutting the brain at the appropriate levels.

In experiments involving the local application of S-AMPA in mPFC, rats were implanted with only one 4-mm probe in this area, at the following coordinates (in mm): AP +3.2, L −0.8, DV −6.0, taken from bregma and duramater (Paxinos and Watson 1986). These microdialysis experiments were conducted in freely moving rats 1 day after implants. After collecting four baseline fractions, S-AMPA was applied in mPFC dissolved in the aCSF used to perfuse the probes (reverse dialysis) for 12 fractions (4 h). Two hours after beginning S-AMPA perfusion (six fractions), the syringe was replaced by one containing S-AMPA plus the test drug (M100907, prazosin, antipsychotics, etc.) and six additional microdialysis fractions were collected.

In the experiments involving the disinhibition of thalamic inputs onto the mPFC, rats were implanted with two microdialysis probes, in mPFC (as above) and in a thalamic area sampling the mediodorsal (MD) and centromedial (CM) nuclei of the thalamus projecting to the mPFC (AP −3.5, L −0.5, DV −6.5; probe tip 1.5 mm). These experiments required the use of anesthetized rats to prevent an excessive behavioral activation produced by bicuculline application in the thalamus. Both the MD and CM nuclei give rise to a dense excitatory input onto mPFC (see third paragraph of Introduction). Previous studies showed that this procedure markedly increases the firing activity of pyramidal neurons in mPFC and doubles the release of 5-HT in this area (Martín-Ruiz et al. 2001; Puig et al. 2003).

On the day after probe implants, rats were anesthetized with chloral hydrate (400 mg/kg i.p.) and supplemental doses of the anesthetic were given when appropriate until the end of the experiments. After collecting baseline dialysate values in mPFC (four fractions), the aCSF used to perfuse the thalamic probes was replaced by one containing 1 mmol/L bicuculline until the end of the experiments (12 more fractions). Two hours after bicuculline application in the CM + MD nuclei, the test drug was applied by reverse dialysis in mPFC or given systemically to examine its effects on prefrontal dialysate 5-HT values.

The concentration of 5-HT in dialysate samples was determined by HPLC (Adell and Artigas 1998) using a Beckman (San Ramon, CA, USA) 3-μm particle size column, and detected with a Hewlett Packard 1049 electrochemical detector set at +0.6 V. Retention time was between 3.5 and 4 min and the limit of detection was typically 1–2 fmol/sample.

Data and statistical analysis

Data (mean ± SEM) are expressed as fmol/fraction (uncorrected for membrane recovery) and are shown in the figures as percentages of basal values, averaged from four pre-drug fractions. Average values of selected time periods were also calculated and shown as bar diagrams. Statistical analysis of drug effects on dialysate 5-HT values has been performed using analysis of variance (anova) for repeated measures with time as repeated factor and drug as independent factor. Statistical significance was set at the 95% confidence level (two tailed).

Results

Local S-AMPA application

Baseline dialysate 5-HT values in the mPFC of freely moving rats were 31 ± 1 fmol/fraction (n = 87). The application of 300 μmol/L S-AMPA in mPFC produced a persistent and stable c. 100% increase in the local 5-HT release (p < 0.0001, time effect; n = 5; Fig. 1). Control rats perfused with aCSF for the whole experiment did not show any alteration of 5-HT levels (n = 5). Although behavioral ratings have not been performed during microdialysis experiments, we noted that the application of S-AMPA in mPFC elicited an overt behavioral activation of the freely moving rats but not seizure activity.

Figure 1.

 The application of (S)-alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate (300 μmol/L) by reverse dialysis in medial prefrontal cortex (mPFC) enhanced the local 5-hydroxytryptamine release (n = 5). The co-perfusion of the classical antipsychotics chlorpromazine (CPZ, n = 5) and haloperidol (HAL 300 μmol/L, n = 5; HAL 100 μmol/L, n = 4) (a) or the atypical antipsychotics clozapine (CLZ, n = 4) and olanzapine (OLZ, n = 4) (b) fully reversed the (S)-alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate-induced elevation in 5-hydroxytryptamine release in mPFC. Bars indicate the period of drug application. See text for statistical analysis.

The application of 300 μmol/L of the classical (chlorpromazine and haloperidol) and atypical (clozapine and olanzapine) antipsychotics in mPFC completely reversed the 5-HT elevation induced by the local S-AMPA application (p < 0.001 for each drug, repeated measures anova; n = 4–5 rats/group) (Fig. 1). This effect was particularly remarkable for haloperidol, which reduced 5-HT values to levels comparable with those produced by the suppression of nerve impulse with tetrodotoxin under the present experimental conditions (e.g. Martín-Ruiz et al. 2001). This concentration of haloperidol had been shown to produce a similar decrease in dialysate 5-HT when administered alone (Amargós-Bosch et al. 2003). A lower haloperidol concentration (100 μmol/L) also reversed the effect of S-AMPA and returned dialysate 5-HT values to baseline (p < 0.001, repeated measures anova; n = 4; Fig. 1). When given alone, this haloperidol concentration reduced maximally dialysate 5-HT to 43 ± 5% of baseline.

The S-AMPA-induced elevation of 5-HT release in mPFC could also be reversed by the co-perfusion of the selective 5-HT2A receptor and α1-adrenoceptor antagonists M100907 and prazosin, respectively (Fig. 2). Given the high affinity of the classical antipsychotics for dopamine D2 receptors, we examined the ability of the DA D2/3 receptor antagonist raclopride to reverse the S-AMPA-evoked 5-HT release. Raclopride application in mPFC (100 μmol/L; n = 7) elicited a partial reversal of the effect of S-AMPA which was statistically significant (p < 0.05, repeated measures anova) but of smaller size than that produced by haloperidol or chlorpromazine (Fig. 2).

Figure 2.

 Bar diagram showing the effects of various drugs on the (S)-alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate (S-AMPA)-evoked 5-hydroxytryptamine (5-HT) release in medial prefrontal cortex (mPFC). The black bar shows the effect of the perfusion of S-AMPA alone. The rest of bars show average values of the last three fractions (1 h) of co-perfusion of each drug in combination with S-AMPA using the experimental procedure shown in Fig. 1. In addition to classical (haloperidol, HAL; 100 and 300 μmol/L; chlorpromazine, CPZ 300 μmol/L) and atypical antipsychotics (clozapine, CLZ and olanzapine, OLZ, both at 300 μmol/L) the selective 5-HT2A receptor antagonist M100907 (MDL, 300 μmol/L) and the selective α1-adrenoceptor antagonist prazosin (PRA, 100 μmol/L) respectively, completely reversed the effect of S-AMPA (the data of M100907 and prazosin were taken from Amargós-Bosch et al. 2003). In contrast, the dopamine D2/3 receptor antagonist raclopride (RAC, 100 μmol/L; n = 7) exerted a partial reversal, whereas the anxiolytic drug diazepam (DZP, 10 and 100 μmol/L, n = 4 each) and the antidepressant drug reboxetine (RBX, 50 μmol/L, n = 5) did not attenuate the S-AMPA-evoked 5-HT release. Actually, reboxetine significantly enhanced the S-AMPA-induced elevation in 5-HT release. ap < 0.05 vs. baseline; *p < 0.05 vs. S-AMPA alone.

Contrary to the antipsychotic drugs, neither the anxiolytic drug diazepam (GABAA receptor modulator; 10 and 100 μmol/L, n = 4 each) nor the antidepressant drug reboxetine (noradrenaline reuptake inhibitor; 50 μmol/L, n = 5) counteracted the S-AMPA-induced elevation of 5-HT release when co-perfused in mPFC. Actually, reboxetine significantly enhanced the 5-HT release over S-AMPA alone (p < 0.03, repeated measures anova). Figure 2 shows the summary effects of the antipsychotic drugs, M100907, prazosin, raclopride, diazepam and reboxetine on the S-AMPA-induced elevation of 5-HT release in mPFC.

Figure 3 shows the effect of the local (in mPFC) and systemic administration of classical and atypical antipsychotics, M100907, prazosin and raclopride (selective antagonists of 5-HT2A receptors, α1-adrenoceptors and dopamine D2/3 receptors, respectively) on the basal 5-HT release in mPFC. The local concentrations were as those in Figure 2. Systemic (s.c.) doses were as follows: haloperidol 0.1 and 1 mg/kg, chlorpromazine, clozapine, olanzapine and raclopride, 1 mg/kg, and M100907 and prazosin, 0.3 mg/kg. All drugs, except raclopride, significantly reduced the spontaneous 5-HT release in mPFC compared to baseline (p < 0.05, repeated measures anova). Likewise, the local (but not systemic) administration of M100907 significantly reduced 5-HT release in mPFC. This difference suggests a differential blockade of 5-HT2A receptors by these two administration routes. However, all other compounds (with the exception of raclopride) reduced dialysate 5-HT by both routes.

Figure 3.

 (a) Effect of the local administration of various drugs on the basal 5-hydroxytryptamine (5-HT) release in medial prefrontal cortex. Drugs were applied at varying concentrations, as in Fig. 2. Bars show 1-h average 5-HT values expressed as percentage of baseline. (b) Effect of the systemic administration of various drugs on the basal 5-HT release in medial prefrontal cortex. Doses used were haloperidol (HAL) 0.1 and 1 mg/kg, chlorpromazine (CPZ), clozapine (CZP) and olanzapine (OZP), 1 mg/kg, M100907 (MDL) and prazosin (PRA), 0.3 mg/kg and raclopride (RAC), 1 mg/kg. Data from 4 to 7 rats/group. *p < 0.05 vs. baseline.

Disinhibition of thalamic afferents to mPFC

The baseline dialysate 5-HT value in the mPFC of chloral hydrate anesthetized rats was 28 ± 1 fmol/fraction (n = 148). This value was significantly lower (p < 0.02; Student’s t-test) than that of freely moving rats (31 ± 1 fmol/fraction; n = 87). As previously observed (Martín-Ruiz et al. 2001; Puig et al. 2003), the local application of bicuculline in the CM + MD nuclei of the thalamus induced a persistent elevation of the 5-HT release in mPFC which was very similar to that produced by S-AMPA application (maximal effect 200 ± 10% of baseline; Fig. 4). The concurrent application of haloperidol in mPFC (300 μmol/L, n = 4) completely reversed the 5-HT elevation and reduced dialysate 5-HT to a maximal value of 15 ± 2% of baseline (p < 0.001, repeated measures anova). The application of chlorpromazine in mPFC (300 μmol/L, n = 6) also reversed significantly the increase in 5-HT release produced by thalamic disinhibition and lowered 5-HT values to 77 ± 4% of baseline (p < 0.001, repeated measures anova) (Fig. 4A). Likewise, the application of the atypical antipsychotics clozapine and olanzapine (300 μmol/L each; n = 4 and 6, respectively) significantly reversed the 5-HT increase produced by thalamic disinhibition (p < 0.001 for both agents; repeated measures anova) (Fig. 4B). At these concentrations, the application of the four antipsychotic drugs in the mPFC of chloral hydrate anesthetized rats decreased dialysate 5-HT in the case of haloperidol (maximum 49 ± 3% of baseline; p < 0.0001; n = 3) and clozapine (maximum 61 ± 3% of baseline; p < 0.0001; n = 3) but not of chlorpromazine (maximum 97 ± 12% of baseline, n.s.; n = 7) and olanzapine (maximum 85 ± 12% of baseline; n.s.; n = 4) (data not shown).

Figure 4.

 The application of 1 mmol/L bicuculline by reverse dialysis in the centromedial and mediodorsal nuclei of the thalamus (CM + MD) increases the 5-hydroxytryptamine (5-HT) release in medial prefrontal cortex (mPFC) of chloral hydrate anesthetized rats (n = 7). The perfusion of 300 μmol/L of the classical (a; HAL, haloperidol, n = 4; CPZ, chlorpromazine, n = 6) or atypical antipsychotics (b; CLZ, clozapine, n = 4; OLZ, olanzapine, n = 6) through the probe in mPFC reversed this effect. Likewise, the local application in mPFC of 300 μmol/L M100907 or 100 μmol/L prazosin (n = 5 each; c) in mPFC reversed the 5-HT elevation induced by the thalamic disinhibition. However, the application of 100 μmol/L raclopride (n = 7) exerted only a partial, although significant attenuation of the effect of thalamic disinhibition on prefrontal 5-HT release. Bars indicate the period of drug application in each area. See text for statistical details.

As previously observed for the local application of S-AMPA (Fig. 2) the local application of M100907 (300 μmol/L, n = 5) and prazosin (100 μmol/L, n = 5) in mPFC also reversed the 5-HT elevation in mPFC induced by the thalamic disinhibition (Fig. 4C). The application of raclopride (100 μmol/L, n = 7) induced a smaller but statistically significant attenuation of the effect of thalamic disinhibition (p < 0.001, repeated measures anova) (Fig. 4C).

We subsequently examined the effect of the systemic administration of classical and atypical antipsychotic drugs on the elevation of 5-HT release induced by thalamic disinhibition. A saline s.c. injection (n = 5) did not alter the effect of thalamic disinhibition on cortical 5-HT release (Fig. 5). However, the s.c. administration of 1 mg/kg of all antipsychotic drugs significantly attenuated the effect of thalamic disinhibition and returned 5-HT values to baseline (p < 0.001 for all drugs, repeated measures anova). A lower haloperidol dose (0.1 mg/kg s.c., n = 4) induced a partial but statistically significant attenuation of the increase in 5-HT release (p < 0.001, repeated measures anova) (Fig. 6).

Figure 5.

 The s.c. administration of vehicle (n = 5; filled circles) did not alter the increase in 5-hydroxytryptamine (5-HT) release produced by disinhibition of thalamic afferents to mPFC. In contrast, the administration of 1 mg/kg of classical (chlorpromazine, n = 7; haloperidol; n = 4; a) and atypical antipsychotics (clozapine, n = 5; olanzapine, n = 4; b) totally reversed the increase in 5-HT release produced by thalamic disinhibition. Likewise, the s.c. administration of the selective 5-HT2A and α1-adrenoceptor antagonists M100907 and prazosin, respectively (0.3 mg/kg; n = 5 each; c) attenuated the effect of thalamic disinhibition on prefrontal 5-HT release. However, the s.c. administration of the selective D2/3 receptor antagonist raclopride (1 mg/kg) did not alter significantly 5-HT release. Arrows show the time of drug injection. See text for statistical analysis.

Figure 6.

 Bar diagram showing the effects of various drugs on the 5-hydroxytryptamine (5-HT) release in medial prefrontal cortex (mPFC) evoked by the application of bicuculline in the mediodorsal and centromedial nuclei of the thalamus. (a) The effects of drugs applied locally in mPFC, as shown in Fig. 4. (b) The effects of systemically administered drugs, as in Fig. 5. Black bars show the average effect of the thalamic disinhibition in the control groups shown in Figs 4 and 5. The rest of bars show average values of the last three fractions (1 h) of administration (local or systemic) of each drug in combination with the thalamic disinhibition following the experimental procedure shown in Figs 4 and 5. All drugs reduced significantly the increase in 5-HT when they were locally applied in mPFC or were systemically administered, except raclopride. This agent exerted a moderate but significant reduction of 5-HT release after its local application but did not reduce 5-HT after systemic administration. Drug concentrations in (a) are as follows: haloperidol, chlorpromazine, clozapine, olanzapine and M100907 (300 μmol/L), prazosin (100 μmol/L) and raclopride (100 μmol/L). Subcutaneous doses in (b) are 1 mg/kg for all antipsychotic drugs (plus 0.1 mg/kg haloperidol), 0.3 mg/kg for M100907 and prazosin and 1 mg/kg for raclopride. Bars show 1-h average 5-HT values expressed as percentage of baseline. ap < 0.05 vs. baseline; *p < 0.05 vs. thalamic disinhibition alone.

The s.c. administration of M100907 (0.3 mg/kg, n = 5) and prazosin (0.3 mg/kg, n = 5) but not raclopride (1 mg/kg, n = 5) significantly reversed the effect of thalamic disinhibition on 5-HT release in mPFC (Fig. 5C). Figure 6 shows the summary effects of the local and systemic administration of antipsychotic drugs and receptor antagonists on the increase of 5-HT release in mPFC produced by thalamic disinhibition.

Discussion

Psychotic symptoms and cortical glutamatergic transmission

Numerous reports suggest that schizophrenia is associated with an abnormal glutamatergic transmission in PFC (Lewis and Lieberman 2000; Tsai and Coyle 2002; Harrison and Lewis 2003; Krystal et al. 2003; Moghaddam and Krystal 2003). The reduced spine density and synaptic proteins, reduced glutamatergic markers and hypofrontality (Andreasen et al. 1997) suggest a decreased glutamatergic activity. However, hypofrontality appears to be mainly associated with negative symptoms (Potkin et al. 2002) and other studies have reported normal or higher than normal cortical activity in schizophrenic patients, particularly during hallucinations (Catafau et al. 1994; Dierks et al. 1999; Shergill et al. 2000). Likewise, proton magnetic resonance studies reported higher than normal glutamate/glutamine levels in PFC of neuroleptic-naïve schizophrenic patients (Bartha et al. 1997; Théberge et al. 2002).

On the other hand, a reduction of GABAergic markers has been reported in the PFC of schizophrenic patients (see Lewis et al. 2005 for review), which possibly results in a decrease of local inhibitory inputs and increased glutamatergic transmission. In rodents, NMDA receptor antagonists, used as pharmacological models of schizophrenia, increase glutamate outflow (Moghaddam et al. 1997; Ceglia et al. 2004) and pyramidal cell firing in the mPFC (Suzuki et al. 2002; Jackson et al. 2004; Jodo et al. 2005; Kargieman et al. 2006). Likewise, LY-354740, a mGluR2/3 agonist abolished the deleterious effects of ketamine on working memory (Krystal et al. 2005), an effect that may result from a reduction of glutamate release (Moghaddam and Adams 1998). Collectively, these data suggest that psychotic symptoms may be associated with an increased glutamatergic transmission in PFC, yet affective/negative symptoms may represent a different stage of the illness involving distinct neurotransmitter abnormalities.

Experimental models used

In accordance with this view, we tested the effects of conventional and atypical antipsychotics in two experimental conditions evoking an increased glutamatergic tone on mPFC neurons: (a) local activation of AMPA receptors by exogenous S-AMPA application; and (b) thalamic disinhibition. The latter procedure was achieved by applying bicuculline in the CM + MD nuclei, which project densely to mPFC and make synapses with pyramidal neuron spines (Berendse and Groenewegen 1991; Kuroda et al. 1998; Van der Werf et al. 2002). Consistent with this connectivity, MD stimulation increased AMPA-mediated responses in mPFC pyramidal neurons (Pirot et al. 1994). Moreover, thalamic disinhibition increased c-fos expression in mPFC (Erdtsieck-Ernste et al. 1995; Bubser et al. 1998) and dramatically increased the activity of pyramidal neurons in mPFC (Puig et al. 2003). The latter effect was accompanied by an increase in prefrontal 5-HT release that was antagonized by mGluR2/3 agonists and AMPA receptor blockade in mPFC. Likewise, the rise in pyramidal cell firing was abolished by the selective mGluR2/3 agonist LY 379268 (Puig et al. 2003). These observations indicate that thalamic disinhibition enhances AMPA-mediated glutamatergic transmission in mPFC, which subsequently evokes an increase in 5-HT release that follows the increased activity of pyramidal neurons.

Here we employed the extracellular 5-HT concentration in mPFC as an in vivo neurochemical index of the changes induced by AMPA receptor stimulation and antipsychotic drugs in PFC. This experimental approach is based on several observations. First, anatomical, electrophysiological and neurochemical studies indicate the presence of a very close reciprocal relationship between the mPFC and the midbrain raphe nuclei. Hence, the electrical stimulation of the mPFC elicited profound changes in most DR 5-HT neurons and vice-versa (Hajós et al. 1998; Celada et al. 2001; Puig et al. 2005). As a result, the electrical stimulation of the mPFC increased 5-HT release in the DR (Celada et al. 2001). Second, the activation of excitatory (5-HT2A, α1-adrenergic, AMPA) or inhibitory (5-HT1A, μ-opioid, mGluR2/3) receptors in mPFC increased and decreased, respectively, the local 5-HT release and – when examined – the firing of 5-HT neurons and 5-HT release in the DR (Celada et al. 2001; Martín-Ruiz et al. 2001; Amargós-Bosch et al. 2003, 2004; Puig et al. 2003). In particular, the increase of glutamatergic transmission in mPFC produced by disinhibition of the CM + MD nuclei, as well as the local blockade of glutamate reuptake increased 5-HT release in mPFC (Martín-Ruiz et al. 2001; Puig et al. 2003). Third, NMDA receptor antagonists, which increase pyramidal cell firing and glutamate release in mPFC, also increase 5-HT neuron activity (Lejeune et al. 1994) and 5-HT release in mPFC (Martin et al. 1998; Ceglia et al. 2004; Amargós-Bosch et al. 2006; Lópz-Gil et al. 2007) an effect blocked by the application of the AMPA antagonist NBQX in mPFC (Lópz-Gil et al. 2007). Altogether, these observations suggest that the 5-HT release can reliably monitor in vivo changes in excitatory transmission in mPFC.

Notwithstanding these observations supporting the involvement of long loops to midbrain, we cannot exclude a local effect of glutamate or S-AMPA on 5-HT terminals to stimulate 5-HT release. Indeed, S-AMPA increased the local 5-HT release in areas not feeding back directly to the raphe (e.g. striatum; Maione et al. 1997) and presynaptic AMPA receptors seem to modulate glutamate and GABA release in various CNS areas (Satake et al. 2000; Patel et al. 2001; Schenk et al. 2003, 2005). However, since none of the receptors for which antipsychotics exhibit high affinity (in particular 5-HT2A/2C and α1-adrenergic) are present in 5-HT axons, the drug-induced changes in 5-HT release must necessarily involve the blockade of post-synaptic receptors in mPFC neurons.

Effect of antipsychotic drugs

Antipsychotic drugs reversed the increase in 5-HT output evoked by the increase in AMPA-mediated transmission in mPFC. This effect (1) was common to classical and atypical drugs; (2) was observed after local (in mPFC) and systemic drug administration; and (3) was consistently found in the two experimental models used (S-AMPA application and thalamic disinhibition). Interestingly, neither diazepam nor reboxetine reversed the effect of S-AMPA on 5-HT release, which suggests that this effect may be specific for antipsychotic drugs. Reboxetine was used instead of the more widely used 5-HT reuptake inhibitors since these would interfere with the 5-HT measure.

These observations add to previous studies indicating that antipsychotic drugs can also reverse the increase in mPFC 5-HT release produced by the stimulation of 5-HT2A and α1-adrenoceptors in mPFC (Amargós-Bosch et al. 2003; Bortolozzi et al. 2003). Interestingly, the physiological or pharmacological stimulation of these receptors (5-HT2A, α1-adrenoceptors, AMPA receptors) in PFC pyramidal neurons enhances their activity (Araneda and Andrade 1991; Aghajanian and Marek 1997; Marek and Aghajanian 1999; Puig et al. 2003, 2005) and the local 5-HT release (Martín-Ruiz et al. 2001; Amargós-Bosch et al. 2003; Bortolozzi et al. 2003). Thus, the present findings suggest that antipsychotic drugs may act in part by attenuating an excessive cortical excitatory neurotransmission.

However, the mechanisms underlying this effect are unclear. Indeed, the interpretation of the present results is complicated by the complex pharmacological profile of antipsychotic drugs, which show affinity for cholinergic, catecholaminergic and serotonergic receptors (Meltzer 1995, 1999; Bymaster et al. 1996; Arnt and Skarsfeldt 1998). Some of these receptors are expressed in PFC in moderate to high density, such as 5-HT1A, 5-HT2A, 5-HT2C, α1-adrenoceptors and D2 receptors (see Introduction). Other receptors, such as the 5-HT7 receptor, for which clozapine has high affinity (Ruat et al. 1993; Shen et al. 1993), is less expressed in cortex and does not seem to play a major role in the depolarization of PFC pyramidal neurons in adult rat brain (Beique et al. 2004). Hence, some of the conclusions raised in the present study are based on the analogies between the action of antipsychotic drugs and that of selective antagonists for such monoaminergic receptors.

The effect of antipsychotic drugs on AMPA-stimulated 5-HT release cannot be accounted for by a direct competition of antipsychotic drugs at iGluRs (Bymaster et al. 1996; Arnt and Skarsfeldt 1998) and possibly results from changes in the activity of mPFC neurons which are later translated into 5-HT changes. Projection (pyramidal) neurons, which make up 80% of the cortical neuronal population, integrate a large number of excitatory, inhibitory and modulatory signals and express most monoaminergic receptors (see Introduction) for which antipsychotics show high affinity. Interestingly, the four antipsychotic drugs tested reduced 5-HT in non-stimulated conditions, suggesting that 5-HT release is tonically controlled by mPFC activity. Moreover, this suggests that the antipsychotic-induced reversal of the effects of AMPA receptor over-stimulation is due to opposite effects on PFC neuronal activity. It is possible that the blockade of certain monoaminergic receptors by antipsychotic drugs modifies the overall glutamatergic tone in mPFC and consequently, the local 5-HT release. The more moderate effect of the antipsychotic drugs on dialysate 5-HT in basal conditions in the anesthetized rats (when compared with the cortical over-stimulation produced by S-AMPA or thalamic disinhibition) suggests that the effect of antipsychotic drugs depends to some extent on the degree of cortical activation. Further studies should examine this point.

Role of monoaminergic receptors

M100907 and prazosin cancelled the effect of S-AMPA and thalamic disinhibition on dialysate 5-HT, which supports the involvement of 5-HT2A and/or α1-adrenoceptors in the action of the antipsychotic drugs tested. In contrast, raclopride (DA D2/3 antagonist) was partly or totally ineffective after its local or systemic administration, respectively. The systemic dose used (1 mg/kg) produces a total occupancy of DA D2 receptors in rodent brain (Millan et al. 1998; Assiéet al. 2006) which suggests that higher doses would have been ineffective as well. Indeed, the excitatory effect of DA on PFC pyramidal neurons was insensitive to the D2/3 receptor antagonist (−)sulpiride (Ceci et al. 1999), in agreement with the more prominent role of D1 receptors to mediate the effect of dopamine on cortical transmission (Gonzalez-Islas and Hablitz 2003; O’Donnell 2003). This suggests that D2 receptor blockade alone does not play a major role in the observed effects despite full occupancy of D2 receptors by conventional antipsychotics at the doses used (Schotte et al. 1993). This view is not contradictory with the major role played by DA D2 receptors in antipsychotic drug action. Indeed, all antipsychotic drugs show moderate-high affinity for DA D2 receptors (Bymaster et al. 1996; Arnt and Skarsfeldt 1998). Yet, the therapeutic action of DA D2 receptor blockade is presumably exerted in subcortical regions, mostly in the nucleus accumbens and anatomically related structures (Moore et al. 1999; Grace 2000).

The ex vivo ED50 values of clozapine for 5-HT2 and α1-adrenoceptor occupancy in rat brain are 1.3 and 0.58 mg/kg s.c., respectively, whereas the corresponding values for haloperidol are 2.6 and 0.4 mg/kg s.c. (Schotte et al. 1993). Similar occupancies have been reported elsewhere (Chaki et al. 1999). Therefore, it is conceivable that both compounds produce a substantial occupancy of α1-adrenoceptors at 1 mg/kg, whereas clozapine can additionally occupy 5-HT2 receptors. However, it is unlikely that 0.1 mg/kg haloperidol occupy a substantial proportion of 5-HT2A or α1-adrenoceptors, which makes it difficult to interpret the reversal produced by this haloperidol dose. It is noteworthy that 1 mg/kg s.c. clozapine is amongst the lowest range of doses of clozapine proven effective in different pharmacological or behavioral models.

5-HT2A receptors are abundantly expressed by pyramidal neurons in mPFC (Santana et al. 2004) and mediate the excitatory actions of 5-HT in vivo (Amargós-Bosch et al. 2004; Puig et al. 2005) and in vitro (Araneda and Andrade 1991; Aghajanian and Marek 1999). Similarly, α1-adrenoceptors are densely expressed in PFC (Pieribone et al. 1994; Day et al. 1997; Domyancic and Morilak 1997), and their stimulation also depolarized pyramidal neurons and increased excitatory post-synaptic currents (Araneda and Andrade 1991; Marek and Aghajanian 1999). In close parallelism, the local application of DOI (5-HT2A/2C agonist) or cirazoline (α1-adrenoceptor agonist) in mPFC increased 5-HT release (Martín-Ruiz et al. 2001; Amargós-Bosch et al. 2003). Interestingly, all these effects (increase in excitatory post-synaptic currents and in 5-HT release) were blocked by AMPA antagonists and mGluR2/3 agonists. Conversely, the stimulatory effect of S-AMPA on 5-HT release was blocked by M100907 and prazosin (Amargós-Bosch et al. 2003; this study). Furthermore, some behavioral effects of NMDA receptor antagonists are blocked by 5-HT2A receptor antagonists (e.g. Varty et al. 1999; Mirjana et al. 2004). Altogether, these observations suggest a complex interplay between AMPA-mediated transmission, on the one hand, and 5-HT2A and α1-adrenoceptors on the other which may account for the effects of antipsychotic drugs observed in the present study. The similarity of the effects of prazosin and M100907 can possibly be accounted for by the common intracellular pathways activated by 5-HT2A and α1-adrenoceptors (both are coupled to PLC), the very high co-expression of the respective mRNAs in PFC neurons and the heterodimerization of these receptors in artificial cell systems (N. Santana et al., unpublished observations).

The reversal of the AMPA- and thalamic-induced increase in 5-HT release exerted by M100907 and prazosin seems to argue against the specificity of the observed effect, since none is an antipsychotic drug on its own. However, prazosin addition enhanced the antipsychotic effect of raclopride in rats (Wadenberg et al. 2000) and both M100907 and prazosin have been reported to block behavioral effects of hallucinogenic compounds such as DOI or non-competitive NMDA receptor antagonists in rats (Schreiber et al. 1995; Dursun and Handley 1996; Varty et al. 1999; Mirjana et al. 2004). Hence, blockade of 5-HT2A receptors and α1-adrenoceptors in PFC may add to blockade of DA D2 receptors in limbic striatum to account for the full therapeutic effect of antipsychotics. Both receptors have been claimed as targets for antipsychotic drugs (Meltzer 1999; Svensson 2003) and classical and atypical drugs share high affinity for α1-adrenoceptors (Bymaster et al. 1996; Arnt and Skarsfeldt 1998). However, the latter pharmacological activity has been disregarded as potentially therapeutic because of the cardiovascular side effects produced by peripheral blockade of α1-adrenoceptors.

Conclusions and limitations of the study

One limitation of the present study is that the neurochemical measure used does not allow to fully elucidate the cellular and network elements involved and thus, further electrophysiological work is required. Current studies indicate that both clozapine and haloperidol are able to reverse the increase in cell firing experienced by mPFC pyramidal cells after the administration of phencyclidine (L. Kargieman et al., unpublished resutls). Overall, these observations suggest that antipsychotic drugs may reverse the increase in mPFC glutamatergic transmission produced in experimental models of schizoprehnia, perhaps through the blockade of post-synaptic 5-HT2A and/or α1-adrenoceptors. This effect would attenuate excitatory transmission in mPFC and subsequently reduce local 5-HT release. One possible exception to the parallel changes in cortical activity and local 5-HT release may be that of risperidone, which increased 5-HT release in mPFC by itself (Hertel et al. 1996). The reasons for this dissimilar behavior of risperidone with the antipsychotic drugs used herein are unknown, but might be related to the higher affinity (10--00-fold difference) of risperidone for α2-adrenoceptors (Arnt and Skarsfeldt 1998) which are involved in the modulation of 5-HT release in PFC (Bel and Artigas 1996).

In summary, both classical and atypical antipsychotics counteract the increase in 5-HT release produced by exogenous (S-AMPA application) and endogenous (thalamic disinhibition) increases in prefrontal glutamatergic transmission. This effect possibly involved the blockade of α1-adrenergic and/or 5-HT2A receptors, for which these drugs display high affinity. Since pyramidal neurons in PFC project to ventral striatum, an attenuation of prefrontal excitatory inputs onto nucleus accumbens neurons might add to the blockade of DA D2 receptors in this area, which is considered to underlie antipsychotic action (Moore et al. 1999; Grace 2000).

Acknowledgements

Work supported by grant SAF 2004-05525. Support from the Generalitat de Catalunya (2005SGR00758s) also acknowledged. Support from the Spanish Ministry of Health, Instituto de Salud Carlos III, Red de Enfermedades Mentales (REM-TAP Network) is also acknowledged. MAB was recipient of a predoctoral fellowship from IDIBAPS. We thank Leticia Campa for skilful technical assistance and the pharmaceutical companies for drug supply.

Ancillary