The 5-HT2A receptor antagonist M100,907 prevents extracellular glutamate rising in response to NMDA receptor blockade in the mPFC

Authors


Address correspondence and reprint requests to Roberto W. Invernizzi, Istituto di Ricerche Farmacologiche ‘Mario Negri’, Via Eritrea 62, 20157 Milano, Italy. E-mail: rinvernizzi@marionegri.it

Abstract

We recently found that intracortical injection of the selective and competitive N-methyl-d-aspartate (NMDA) receptor antagonist 3-(R)-2-carboxypiperazin-4-propyl-1-phosphonic acid (CPP) impaired attentional performance in rats and blockade of 5-hydroxytryptamine (5-HT)2A receptors antagonized this effect. Here, we used the microdialysis technique in conscious rats to study the effect of CPP on extracellular glutamate (GLU) in the medial prefrontal cortex (mPFC) and the regulation of this effect by 5-HT2A receptors. Intraperitoneal injection of 20 mg/kg CPP increased extracellular GLU in the mPFC (201% of basal levels) but had no effect on 5-HT. Intracortical infusion of 100 µm CPP increased extracellular GLU (230% of basal values) and 5-HT (150% of basal values) in the mPFC, whereas 30 µm had no significant effect. The effect of 100 µm CPP on extracellular GLU was abolished by tetrodotoxin, suggesting that neuronal activity is required. Subcutaneous injection of 40 µg/kg M100,907 completely antagonized the effect of 100 µm cpp on extracellular GLU, whereas 10 µg/kg caused only partial attenuation. Likewise, intracortical infusion of 0.1 µm M100,907 completely reversed the increase of extracellular GLU induced by CPP. These findings show that blockade of NMDA receptors in the mPFC is sufficient to increase extracellular GLU locally. The increase of cortical extracellular GLU may contribute to CPP-induced cognitive deficits and blockade of 5-HT2A receptors may provide a molecular mechanism for reversing these deficits caused by dysfunctional glutamatergic transmission in the mPFC.

Abbreviations used
AP7

2-amino-7-phosphonoheptanoic acid

CPP

3-(R)-2-carboxypiperazin-4-propyl-1-phosphonic acid

CSF

cerebrospinal fluid

DA

dopamine

ESPC

excitatory post-synaptic currents

GLU

glutamate

HPLC

high-performance liquid chromatography

5-HT

5-hydroxytryptamine

M100,907

R-(+)-(2,3-dimethoxyphenyl)-1-[2-(4-fluorophenyl)-ethyl]-4-piperidine methanol

mPFC

medial prefrontal cortex

NMDA

N-methyl-d-aspartate

OPA

o-phthaldialdehyde

PCP

phencyclidine

TTX

tetrodotoxin

The non-competitive N-methyl-d-aspartate (NMDA) receptor antagonists phencyclidine (PCP), ketamine and dizocilpine as well as the selective and competitive NMDA receptor antagonist 3-(R)-2-carboxypiperazin-4-propyl-1-phosphonic acid (CPP) impair cognitive-related processes dependent on the prefrontal cortex such as working memory and selective attention (Moghaddam et al. 1997; Moghaddam and Adams 1998; Aura and Riekkinen 1999; Romanides et al. 1999; Higgins et al. 2003). PCP and ketamine increase extracellular glutamate (GLU) in the medial prefrontal cortex (mPFC) (Moghaddam et al. 1997; Moghaddam and Adams 1998; Abekawa et al. 2003) and it has been suggested that the increased availability of GLU at non-NMDA receptors contributes to the motor, neurotoxic and cognitive effects of NMDA receptor antagonists (Olney and Farber 1995; Moghaddam et al. 1997; Moghaddam and Adams 1998; Cartmell et al. 1999). In addition, the finding that PCP or genetic deletion of the NMDA receptor raised extracellular 5-HT in the mPFC (Martin et al. 1998a; Adams and Moghaddam 2001; Miyamoto et al. 2001) suggests that enhanced 5-HT tone in the mPFC may contribute to the effects of NMDA receptor hypofunction.

Serotonergic afferents arising from the raphe nuclei innervate the mPFC (Azmitia and Segal 1978) and inhibit or excite the activity of cortical pyramidal neurones depending on the 5-hydroxytryptamine (5-HT) receptor subtype (Barnes and Sharp 1999). Immunocytochemical, autoradiographic and in situ hybridization studies have found high densities of 5-HT2A receptors and the corresponding mRNA in the mPFC of primates and rats (Pompeiano et al. 1994; Jakab and Goldman-Rakic 1998). 5-HT2A receptors are enriched in pyramidal cells, but are also found in interneurones and, albeit rarely, on cortical axon terminals, forming asymmetric synapses (Willins et al. 1997; Jakab and Goldman-Rakic 1998; Scruggs et al. 2000; Miner et al. 2003). The 5-HT2A receptor agonist and hallucinogenic drug DOI raises extracellular GLU (McGrew et al. 2003; Scruggs et al. 2003), enhancing GLU-induced excitatory post-synaptic currents (EPSC) in the mPFC (Aghajanian and Marek 1997, 1999; Zhai et al. 2003) and cortical c-fos expression (Scruggs et al. 2000). On the other hand, the selective 5-HT2A receptor antagonist R-(+)-(2,3-dimethoxyphenyl)-1-[2-(4-fluorophenyl)-ethyl]-4-piperidine methanol (M100,907) (Kehne et al. 1996) counteracted the effect of NMDA receptor antagonists on the increase in Fos-immunoreactivity (Habara et al. 2001), hyperlocomotion (Maurel-Remy et al. 1995; Martin et al. 1998b; Abekawa et al. 2003), deficits in pre-pulse inhibition of the startle response (Varty et al. 1999) and immobility in the forced swimming test (Corbett et al. 1999). We recently observed that the selective and competitive NMDA receptor antagonist CPP (Lehmann et al. 1987) injected intracortically impaired attentional performance in a 5-CSRT task and that blockade of 5-HT2A receptors with M100,907 antagonized this effect (Carli et al. 2004).

It is not known whether CPP, like the non-competitive NMDA receptor antagonists, can increase cortical extracellular GLU and to what extent serotonergic mechanisms are involved. Therefore, using in vivo microdialysis in the conscious rat we investigated whether CPP injected intraperitoneally (i.p.) or infused intracortically through the probe raised extracellular GLU in the mPFC. We examined the role of 5-HT2A receptors in CPP's effect on GLU in rats pre-treated with the selective 5-HT2A receptor antagonist M100,907 injected subcutaneously (S.C.) or co-infused with CPP through the probe. In view of the poor relationship between the basal extracellular concentration of GLU and neuronal activity (Herrera-Marschitz et al. 1996; Timmerman and Westerink 1997; Del Arco and Mora 2002), in one experiment we used the sodium channel blocker tetrodotoxin (TTX) to study the action potential-dependent effect of CPP on extracellular GLU. Finally, we examined CPP's effect on extracellular 5-HT in the mPFC. Some of this work has already been reported as an abstract (Invernizzi et al. 2003).

Materials and methods

Animal housing

All experiments were conducted in conformity with the institutional guidelines that are in compliance with national (D.L. n. 116, G.U., suppl. 40, 18 Febbraio 1992, Circolare no. 8, G.U., 14 Luglio 1994) and international laws and policies (EEC Council Directive 86/609, OJ L 358,1, 12 December 1987; Guide for the Care & Use of Laboratory Animals, US National Research Council 1996).

Male rats (CD-COBS, Charles River, Calco, Italy) weighing about 250–300 g were housed at constant room temperature (21 ± 1°C) and relative humidity (60 ± 5%) on a 12-h light/dark cycle (light on 07:00 h). They had free access to food and water.

Surgery and microdialysis

Concentric dialysis probes were prepared with AN-69 hollow fibers (310 µm outer diameter; Hospal, Bologna, Italy), essentially as described by Robinson and Whishaw (1988). Because diffusion of 5-HT through AN-69 membrane is markedly delayed (Tao and Hjorth 1992), we used a Cuprophan membrane (216 µm outer diameter, Sorin Biomedica, Italy) for measurements of 5-HT. The exposed membrane was 4 mm long. Stereotaxic co-ordinates for the mPFC (referring to the probe tip) were determined according to the Paxinos and Watson (1986) atlas: AP + 3.7 and L ± 0.7 mm from bregma and V−4.8 from dura surface. About 24 h after surgery, the probes were perfused with artificial cerebrospinal fluid (CSF composition: NaCl 140 mm, CaCl2 1.26 mm, KCl 3 mm, MgCl2 1 mm, Na2HPO4 1.2 mm, glucose 7.2 mm, pH 7.4 with a few drops of 0.6 m NaH2PO4) at 1 µL/min. Samples of dialysate were collected every 20 min and stored at 4°C.

Drugs and reagents

CPP and the d-isomer of the 2-amino-7-phosphonoheptanoic acid (AP7) were purchased from Tocris Cookson Inc. (Ellisville, MO, USA); nicotine and TTX were from Sigma-RBI (Milan, Italy) and M100,907 was a gift from Aventis Pharmaceuticals (Bridgewater NJ, USA). Drugs were infused through the probe or injected s.c. (M100,907) or i.p. (CPP). Control rats were perfused with normal CSF or injected with saline or CSF. All drugs or vehicle were infused during the phase of stable glutamate output defined as three consecutive baseline samples not differing by more than 20%.

Chemical reagents were of analytical grade and were purchased from Merck (Bracco, Milan, Italy) or Sigma-Aldrich (Milan, Italy).

Experimental design

Neuronal origin of GLU

Rats were perfused through the probe with 1 µm TTX to verify the neuronal origin of the basal extracellular GLU. TTX was perfused throughout the 2-h experiment. Other rats were perfused with 60 mm KCl through the probe for 20 min to stimulate exocytotic release. In this case, NaCl concentration was reduced to 83 mm to maintain osmolarity. Another group of rats was given nicotine (100 µm and 1 mm) through the probe to assess the ability of microdialysis to reveal changes in extracellular GLU presumably dependent on neuronal activity. Each rat was perfused with 100 µm nicotine in CSF for 60 min. Thereafter, perfusion medium was manually switched to a solution containing 1 mm nicotine for 60 min.

Effect of CPP and AP7

A group of rats received CPP (10 and 20 mg/kg) dissolved in saline or vehicle intraperitoneally. Other rats were given 30, 100 and 300 µm CPP or 1 mm AP7 through the probe. After baseline stabilization, the perfusion fluid was manually switched to a solution containing CPP or AP7 in CSF. Each concentration of CPP or 1 mm AP7 was perfused for 60 min. A separate group of rats was coinfused with 1 µm TTX and 100 µm CPP to assess the contribution of the action potential-dependent effect of CPP on extracellular glutamate. TTX infusion started 20 min before CPP and lasted 140 min.

Effect of M100,907

The 5-HT2A receptor antagonist M100,907 was dissolved in CSF and administered s.c. (10 and 40 µg/kg) or through the probe (0.1 µm) 20 min before 100 µm CPP. The systemic doses used were based on their ability to block the CPP-induced attentional performance deficit (Carli et al. 2004) and DOI-mediated rise in extracellular DA (Gobert and Millan 1999). In vitro relative recovery of M100,907 measured at room temperature through the AN69 probe was 22 ± 3% (n = 3). Based on this result, the concentration of M100,907 in the extracellular fluid surrounding the probe should be about 20 nm.

Extracellular 5-HT in response to CPP

A group of rats was implanted with Cuprophan probes and given 10 and 20 mg/kg CPP or saline intraperitoneally or infused with the drug (30 and 100 µm) through the probe. Dialysate was collected every 20 min and splitted for the simultaneous measurement of 5-HT (15 µL) and glutamate (5 µL).

In control rats, perfused with normal CSF, tubing was disconnected and reconnected to simulate the change of the perfusion medium as in rats given CPP and other treatments through the probe.

Glutamate

The concentrations of GLU in dialysate samples were determined by high-performance liquid chromatography (HPLC) with fluorometric detection after precolumn derivatization with o-phthaldialdehyde (OPA)/β-mercaptoethanol reagent according to Donzanti and Yamamoto (1988). Briefly, the derivatizing reagent was prepared by dissolving 27 mg OPA (Sigma-Aldrich, Milan, Italy) in 1 mL methanol followed by 5 µL of β-mercaptoethanol and 9 mL of 0.1 m sodium tetraborate buffer (pH 9.3). This stock solution was stored for about 5 days at room temperature in a sealed vial darkened with aluminum film. It was diluted 1 : 4 with 0.1 m sodium tetraborate buffer 24 h before use. Twenty-five microlitres of the diluted reagent were added to 5 or 10 µL sample and the reaction was allowed to proceed for 2 min at room temperature before injection onto the chromatograph. GLU was separated with a reverse phase column (HR-80, 80 × 4.6 mm, 3 µm packing; ESA Inc., Chelmsford, MA, USA) protected with a guard column (New Guard RP-18, 7 µm, 15 × 3.2 mm; Perkin Elmer, Norwalk, CT, USA). The mobile phase consisted of 0.1 m Na2HPO4, 0.13 mm Na2EDTA, 28% CH3OH, adjusted to pH 6.4 with 85% H3PO4, pumped at 1 mL/min with a LC-10ADvp pump (Shimadzu, Milan, Italy). GLU was detected by a scanning fluorescence detector (model 470; Waters S.p.A., Milan, Italy) using an excitation wavelength of 330 nm and an emission wavelength of 450 nm. The assay was calibrated daily with 20 pmol/10 µL GLU standard made up in CSF.

5-HT

5-HT was measured as described elsewhere (Invernizzi et al. 1992). Briefly, 5-HT was separated by a reverse phase column (Supelcosil LC18-DB 3 µm, 150 × 4.6 mm; Supelchem, Milan, Italy) and a mobile phase consisting of citric acid 9 mm, sodium acetate trihydrate 48 mm, Na2EDTA 0.1 mm, 100 µL/L triethylamine and 40 mL/L acetonitrile, pumped at 1 mL/min. 5-HT was measured by a Coulochem II electrochemical detector equipped with a 5011 analytical cell (ESA Inc., Chelmstord, MA, USA) at the following potentials (E1 + 50 mV, E2 + 180 mV 5-HT was read as the second electrode output signal.

Histology

At the end of the experiment, rats were deeply anesthetized with chloral hydrate (400 mg/kg i.p.) and killed by decapitation. Their brains were immediately removed and frozen on dry ice. The correct placement of the probes was checked by visual inspection of the probe tracks on 40-µm coronal sections from the mPFC of each rat. Only rats with correct probe placement were considered in the results.

Statistical analysis

Extracellular levels of GLU and 5-HT, not corrected for in vitro recovery, are expressed as percentages of basal values. Basal values of GLU in different experiments were compared by one-way anova. All time-course data were analyzed by anova for repeated measures with treatment as between factor and time as within factor. Post-hoc comparisons between pre- and post-injection values were made with Dunnett's test, whereas Tukey–Kramer's test was used for comparisons between treatments (i.e. CPP vs. CSF; M100,907 vs. CSF; CSF + CPP vs. M100,907 + CPP). Values missing because of occasional problems in sample collection or analysis were replaced by the mean of the samples immediately before and after. Statistical analyses were done on raw data using the StatView 5.0 for Apple Macintosh computer (SAS Institute Inc., Cary, NC, USA).

Results

Basal concentrations of glutamate: effect of TTX, elevated KCl and nicotine

GLU levels were stable about 2 h after the start of the experiment. Mean basal GLU concentrations in the mPFC of rats implanted with AN-69 probes, obtained by pooling basal values of different experimental groups, were 14.2 ± 0.7 pmol/20 µL (n = 85). No significant differences were found among basal GLU levels in the different experimental groups (F15,69 = 0.6, p > 0.05).

Figure 1 shows that 1 µm TTX had no significant effect on cortical extracellular GLU (F5,5 = 0.8, p > 0.05; one-way anova), whereas the 20-min perfusion with 60 mm KCl increased extracellular GLU by about 2.5-fold (F3,3 = 24.1, p < 0.05; one-way anova). Nicotine significantly increased extracellular GLU (F4,8 = 11.8, p < 0.01). Extracellular GLU reached 284% of basal levels after 1 mm nicotine, whereas 0.1 mm nicotine had no significant effect (132% of basal levels; Fig. 1).

Figure 1.

Effect of the perfusion of 1 µm tetrodotoxin (TTX), 60 mm KCl and 0.1 and 1 mm nicotine though the probe on basal extracellular glutamate (GLU) in the mPFC. Results are mean ± SEM and are expressed as percentage of basal levels. Basal levels of GLU in pmol/60 µL were: TTX, 41.8 ± 6.8 (n = 6); KCl, 35.7 ± 8.4 (n = 4); Nicotine, 33.5 ± 5.7 (n = 5). *p < 0.05 versus baseline levels (Dunnett's test).

Increase of extracellular GLU in response to CPP and effect of TTX

Figure 2(a) shows that intraperitoneal CPP significantly raised extracellular GLU in the mPFC (F2,10 = 4.9, p < 0.05). Post-hoc analysis showed that extracellular GLU in rats given 20 mg/kg CPP was significantly higher than in those receiving saline. Extracellular GLU rose by about 50% during the first hour after the injection, reached 195% at 80 min and was still increased at 2 h (190% of basal levels). A ratio of 10 mg/kg CPP had no significant effects.

Figure 2.

(a) The effect of 10 and 20 mg/kg i.p. CPP or saline on extracellular glutamate (GLU). Arrow indicates the injection of saline or CPP. Results are mean ± SEM. Basal levels of glutamate in pmol/20 µL were: (□) saline, 16.0 ± 1.3 (n = 4); (bsl00084) CPP 10 mg/kg, 12.7 ± 3.8 (n = 5); (○) CPP 20 mg/kg, 13.9 ± 1.8 (n = 4) *p < 0.05 versus saline (Tukey–Kramer's test). (b) The effect of 30 (⋄) and 100 µm (○) CPP on extracellular GLU in rats perfused with normal CSF (□) and the effect of 100 µm CPP in rats co-perfused with 1 µm TTX (bsl00084). TTX perfusion started 20 min before CPP and continued for the rest of the experiment. Each concentration of CPP was perfused through the probe for 1 h (horizontal bar). Results are mean ± SEM. Basal levels of glutamate in pmol/20 µL were: CSF, 17.9 ± 4.6 (n = 5); CPP 30 µm, 18.6 ± 3.7 (n = 4); CPP 100 µm, 13.8 ± 3.1 (n = 9); TTX + CPP, 14.5 ± 1.3 (n = 4). *p < 0.05 versus CSF; #p < 0.05 versus CPP 100 µm (Tukey–Kramer's test).

The infusion of CPP through the probe significantly raised extracellular GLU in the mPFC (F2,15 = 23.9, p < 0.0001) (Fig. 2b). Post-hoc analysis showed that extracellular GLU in rats given 100 µm CPP was significantly higher than in rats infused with CSF (p < 0.05; Tukey–Kramer's test). Extracellular GLU rose by about 225% in response to 100 µm CPP at 60 min. Surprisingly, the maximal increase was reached at 80 min (350%), i.e. 20 min after the end of CPP infusion. Then, extracellular GLU gradually returned to baseline. Infusion of 300 µm CPP did not cause any greater increase in extracellular GLU (n = 2; data not shown). The concentration of 30 µm CPP had no significant effect on extracellular GLU (Fig. 2b, p > 0.05, Tukey–Kramer's test).

The infusion of 1 µm TTX completely prevented the effect of 100 µm CPP on extracellular GLU (Ftreatment1,12 = 29.6, p < 0.001; Fig. 2B).

In another experiment we examined the effect of the competitive NMDA receptor antagonist AP7 on extracellular GLU. We found that the infusion of 1 mm AP7 for 60 min almost doubled extracellular GLU levels (basal 30.8 ± 7.2, AP7 61.8 ± 13.8 pmol/h; n = 8; F1,7 = 7.2, p < 0.05).

Blockade of 5-HT2A receptors prevents the effect of CPP on extracellular glutamate

As shown in Fig. 3(a), extracellular levels of GLU in the mPFC of rats infused with 100 µm CPP were significantly higher than after normal CSF (p < 0.05; Tukey–Kramer's test). Subcutaneous and intracortical injection of M100,907 prevented the rise of extracellular GLU in response to CPP (Fig. 3a). anova indicated a significant effect of treatment (F3,18 = 7.8, p < 0.01), time (F5,90 = 2.4, p < 0.05) but not treatment by time interaction (F3,18 = 0.7, p > 0.05). Post-hoc analysis showed that 40 µg/kg M100,907 completely prevented the increase of extracellular GLU induced by CPP (p < 0.05 vs. CPP alone; Tukey–Kramer's test). Likewise, the co-perfusion of 0.1 µm M100,907 abolished the effect of 100 µm CPP on extracellular GLU (p < 0.05 vs. CPP alone; Tukey–Kramer's test). M100,907 10 µg/kg also attenuated the increase of extracellular GLU induced by CPP but the effect was not significant (p > 0.05 vs. CPP alone; Tukey–Kramer's test).

Figure 3.

The 5-HT2A receptor antagonist M100,907 prevents CPP increasing extracellular glutamate (GLU) in the mPFC. (a) Rats were injected subcutaneously with CSF, 10 or 40 µg/kg M100,907, 20 min before the perfusion of (⋄) 100 µm CPP (horizontal bar) through the probe. One group of rats was co-perfused with 0.1 µm M100,907 and 100 µm CPP through the probe. M100,907 started 20 min before CPP and continued for the rest of the experiment. (b) The effect of (bsl00066) 40 µg/kg s.c. M100,907 (arrow) or (bsl00001) 0.1 µm M100,907 perfused through the probe (horizontal bar) on extracellular GLU. Results are mean ± SEM. Basal levels of GLU in pmol/20 µL were: VEH + CPP, 14.8 ± 3.5 (n = 5); (○) 10 µg/kg M100,907 + CPP, 13.9 ± 2.1 (n = 6); (bsl00084) 40 µg/kg M100,907 + CPP, 18.6 ± 4.9 (n = 6); (□) 0.1 µm M100,907 + CPP, 13.5 ± 2.7 (n = 6); 40 µg/kg M100,907 (bsl00066), 12.9 ± 2.1 (n = 6); 0.1 µm M100,907, 12.3 ± 1.0 (n = 6). *p < 0.05 versus VEH + CPP (Tukey–Kramer's test).

Figure 3(b) shows that by itself M100,907 (40 µg/kg s.c.) or 0.1 µm through the probe had no effect on extracellular GLU (Ftreatment1,10 = 0.01, p > 0.05; Ftime7,70 = 1.8, p > 0.05; Ftreatment × time 7,70 = 1.7, p > 0.05).

Effect of CPP on extracellular 5-HT

Figure 4 shows that extracellular levels of 5-HT in the mPFC of rats given intraperitoneal CPP were significantly higher than in control rats (Ftreatment2,10 = 9.4, p < 0.01; Ftime6,60 = 1.5, p > 0.05; Ftreatment × time 12,60 = 1, p > 0.05). Then 10 mg/kg CPP increased extracellular 5-HT by 140% of basal levels at 40 min, whereas 20 mg/kg had no significant effect.

Figure 4.

Extracellular 5-HT in the mPFC of rats given 10 (bsl00084) and 20 mg/kg (□) CPP or saline (arrow; a) or 30 and 100 µm CPP infused through the probe for 1 h (horizontal bar; b). Results are mean ± SEM. Basal levels of 5-HT (fmol/20 µL) were: (○) Saline (n = 4) 4.0 ± 0.6; CPP 10 mg/kg (n = 5) 5.4 ± 0.6; CPP 20 mg/kg (n = 5) 3.2 ± 0.5; CSF (n = 4), 4.8 ± 0.4; (bsl00084) CPP 30 µm (n = 5), 4.9 ± 0.6; (□) CPP 100 µm (n = 5), 4.7 ± 0.3. *p < 0.05 versus saline or CSF (Tukey–Kramer's test).

The infusion of CPP through the probe significantly increased extracellular 5-HT (Ftreatment2,12 = 5.1, p < 0.05; Ftime6,72 = 1.6, p > 0.05; Ftreatment × time12,72 = 6.7, p < 0.0001). Extracellular 5-HT in rats infused with 100 µm CPP reached about 150% of baseline levels at 60 min and had fallen back at 100 min (Fig. 4b). Extracellular levels of 5-HT in rats infused with 30 µm CPP were slighltly lower than in those infused with CSF but the difference was not significant.

Discussion

Both intraperitoneal injection and intracortical infusion of the selective NMDA receptor antagonist CPP raised extracellular GLU in the mPFC of awake rats and the blockade of 5-HT2A receptors prevented the effect of locally infused CPP. The systemic dose and the intracortical concentration of CPP that significantly increased extracellular GLU fit well with those used in vitro and in vivo to selectively block NMDA receptors (Chapman et al. 1987; Lehmann et al. 1987; Del Arco and Mora 2002). In addition, intracortical infusion of the competitive NMDA receptors antagonist AP7 (Perkins et al. 1982) also raised extracellular GLU (see Results). These results indicate that selective blockade of NMDA receptors in the mPFC is sufficient to increase extracellular GLU in this brain region. The involvement of NMDA receptors in the mechanism by which CPP increases extracellular GLU is further supported by the finding that systemic administration of non-competitive NMDA receptor antagonists such as PCP and ketamine increase extracellular GLU (Moghaddam et al. 1997; Adams and Moghaddam 1998; Moghaddam and Adams 1998; Abekawa et al. 2003).

Electrophysiological studies have shown that systemic but not locally applied PCP increased the activity of cortical neurones (Suzuki et al. 2002; Jodo et al. 2003). Also, systemic but not local infusion of ketamine through the probe increased extracellular GLU in the mPFC (Lorrain et al. 2003). These results suggested that the effects of systemic ketamine and PCP were due to the blockade of NMDA receptors outside the mPFC. However, an increase, decrease, or no effect have all been reported with different doses of ketamine (Moghaddam et al. 1997). Thus, further studies are needed to clarify the role of cortical NMDA receptors in the effect of ketamine on extracellular GLU.

The ability of the microdialysis technique to reveal changes in neuronal release of GLU is debated (Herrera-Marschitz et al. 1996; Timmerman and Westerink 1997). Basal extracellular levels of GLU do not reflect neuronal release but mainly depend on the activity of the cystine-GLU exchanger (Baker et al. 2002). However, the stimulation of exocytotic release induced by increased K+ in the perfusion medium or electrical stimulation of glutamatergic pathways raised GLU in the dialysate (Timmerman and Westerink 1997). Accordingly, we found that basal levels of GLU were largely insensitive to TTX, whereas K+-evoked release increased dialysate GLU. In addition, we found that the infusion of nicotine through the probe increased extracellular GLU. The activation of pre-synaptic acetylcholine heteroreceptors, with concentrations of nicotine similar to those used in the present study, stimulate the neuronal release of GLU (Gioanni et al. 1999; Lambe et al. 2003; Reno et al. 2004) and other neurotransmitters (Wonnacott 1997). This finding suggests that under the present experimental conditions, stimulation of the activity-dependent release of GLU raised GLU levels in the dialysate. Thus, it is conceivable that the CPP-induced increase of extracellular GLU reflects an increase of neuronal release. This interpretation is supported by the finding that TTX completely abolished the increase of extracellular GLU induced by CPP. Likewise, the increase of extracellular levels of striatal GLU induced by NMDA antagonists or stimulation of the corticostriatal glutamatergic pathway was prevented by locally perfused TTX (Dijk et al. 1995; Liu and Moghaddam 1995; Lada et al. 1998).

The increase of GLU reached its peak 20 min after the end of CPP infusion. This delayed effect (Figs 2 and 3) is probably related to the use of probes made with AN-69 membrane as it was not seen with the Cuprophan membrane (data not shown).

The mechanism by which CPP increased extracellular GLU is unclear. NMDA receptors are preferentially expressed by pyramidal neurones but are also present on non-pyramidal cells in the rat cortex (Aoki et al. 1994; DeBiasi et al. 1996; Conti et al. 1997). Thus, NMDA receptors expressed by both neuronal populations may have contributed to CPP effects on extracellular GLU. The location of some NMDA receptors on pre-synaptic terminals forming asymetric synapses means they are presumably autoreceptors (Conti et al. 1997). However, the stimulation of NMDA autoreceptors facilitated GLU release and synaptic currents whereas their blockade had no effect or even reduced GLU release (Connick and Stone 1988; Martin et al. 1991; Berretta and Jones 1996). Thus, it is unlikely that the blockade of autoreceptors contributed to the effect of CPP on extracellular GLU. This suggests that the increase of extracellular GLU by CPP was probably due to the control exerted by cortical NMDA receptors on other neurotransmitter systems which in turn enhance extracellular GLU.

Blockade of NMDA receptors with PCP and MK-801 reduced extracellular GABA in the mPFC (Yonezawa et al. 1998) and there is evidence that GLU release is inhibited by GABA (Pende et al. 1993). In addition, PCP markedly reduced bursting activity of cortical pyramidal neurones (Shi and Zhang 2003) and it has been suggested that the response of GABAergic interneurones depends on the firing pattern of pyramidal cells (Thomson 2000; Shi and Zhang 2003). Thus, the effect of CPP on extracellular GLU may be mediated by direct or indirect suppression of cortical GABAergic transmission, which in turn enhances the release of GLU from afferents to the mPFC. Studies in cortical slices and synaptosomes showing that stimulation of pre-synaptic GABAB heteroreceptors inhibited cortical GLU release are consistent with this interpretation (Potashner 1979; Pende et al. 1993; Perkinton and Sihra 1998).

Dopamine (DA) and 5-HT may have excitatory and inhibitory effects on GLU release, depending on the receptor subtype (Yamamoto and Davy 1992; Srkalovic et al. 1994; Dijk et al. 1995; Maura and Raiteri 1996; Exposito et al. 1999; Abekawa et al. 2000; Scruggs et al. 2003). CPP and the structural analogue AP-5 infused into the mPFC raised extracellular DA locally (Del Arco and Mora 1999; Feenstra et al. 2002), suggesting that activation of dopaminergic receptors may have contributed to the effect of CPP. However, results are contrasting regarding the control exerted by D2 receptors on extracellular GLU (Yamamoto and Davy 1992; Exposito et al. 1999), whereas the stimulation of D1 receptors reduced extracellular GLU in the mPFC (Abekawa et al. 2000).

Our results clearly show that subcutaneous injection of the 5-HT2A receptor antagonist M100,907 completely prevented the increase of extracellular GLU induced by CPP. The effect of CPP on GLU may therefore depend on the stimulation of 5-HT2A receptors in the mPFC as a consequence of the increased tone of endogenous 5-HT. The present study did not directly address this issue which, however, is supported by two main findings. First, the infusion of CPP through the probe raised extracellular 5-HT in the mPFC. This effect was observed with the same concentration of CPP (100 µm) that increased extracellular GLU, whereas 30 µm CPP had no effect on extracellular GLU and 5-HT. These results are consistent with previous microdialysis studies in which blockade of NMDA receptors by PCP and MK-801 or genetic deletion of the receptor increased extracellular 5-HT in the mPFC and/or other regions of the rat brain (Whitton et al. 1992; Yan et al. 1997; Martin et al. 1998a; Adams and Moghaddam 2001; Miyamoto et al. 2001). Second, the 5-HT2A receptor agonist DOI increased extracellular GLU (McGrew et al. 2003; Scruggs et al. 2003) and the frequency of the GLU-dependent late component of evoked EPSC in slices of the mPFC (Aghajanian and Marek 1999). Thus CPP, by releasing 5-HT, may activate 5-HT2A receptors in the mPFC to enhance GLU release. However, this interpretation hardly accounts for the effect of systemic CPP that at 10 mg/kg increased extracellular 5-HT but had no effect on GLU, whereas at 20 mg/kg it increased GLU but had no effect on 5-HT. Thus, it appears that systemic CPP may increase extracellular GLU independently of its effect on cortical 5-HT.

Differences in the effects of systemic and intracortical CPP on 5-HT are not surprising because NMDA receptors are widespread in the brain and may exert opposite effects on cortical 5-HT release depending on their localization in the raphe nuclei or cortex (Tao and Auerbach 1996). Interestingly, the hypermotility induced by systemic MK-801 was not affected by 5-HT depletion, whereas its reversal by M100,907 was prevented by the depletion of endogenous 5-HT (Martin et al. 1998b). Thus, it cannot be excluded that the blockade of 5-HT2A receptors unmasks the effect of endogenous serotonin on other receptor subtypes, such as 5-HT1A and 5-HT3 receptors that inhibit GLU release in the rat brain (Srkalovic et al. 1994; Dijk et al. 1995; Maura and Raiteri 1996). 5-HT1A receptors are particularly interesting in this respect as they are co-expressed with 5-HT2A receptors in cortical pyramidal neurons (Martin-Ruiz et al. 2001) and exert opposite effects on the excitability of these cells (Araneda and Andrade 1991), on the release of 5-HT in the mPFC (Martin-Ruiz et al. 2001) and on attention, as assessed by a 5-CSRT task in rats (Winstanley et al. 2003).

The infusion of 0.1 µm M100,907 through the probe prevented the increase in GLU induced by CPP. This finding indicates that blockade of 5-HT2A receptors in the mPFC plays a major role in controlling the effect of CPP on extracellular GLU. It may appear surprising that the CPP-induced raise of extracellular GLU was antagonized here using 0.1 µm M100,907, as higher concentrations (100–300 µm) were used in previous studies to block cortical 5-HT2A receptors (Martin-Ruiz et al. 2001; Amargos-Bosch et al. 2003). However, brain extracellular concentrations of the drug in rats given 5 mg/kg M100,907, a dose 125 times that used in the present study, were about 1 µm (Scott and Heath 1998). Taking into account the in vitro recovery of M100,907 (about 20%, present paper), its concentration in the extracellular fluid should be about 20 nm. This is proportional to the concentration found in rats given 5 mg/kg M100,907 (Scott and Heath 1998) and well above the Ki of the drug (about 1 nm) for 5-HT2A receptors (Kehne et al. 1996; Sorbera et al. 1998).

5-HT2A receptors are expressed, albeit in small amounts, by axons and terminals of the mPFC (Willins et al. 1997; Jakab and Goldman-Rakic 1998; Miner et al. 2003). Glutamatergic afferents to the mPFC arise from different regions of the brain including thalamus, hippocampus, amygdala and other cortical regions (Krettek and Price 1977; Groenewegen 1988; Jay et al. 1992; Kuroda et al. 1995; Bacon et al. 1996; Pirot et al. 1996). Thus, it is conceivable that the reversal of the CPP-induced increase of extracellular GLU by M100,907 is mediated by pre-synaptic 5-HT2A heteroreceptors on glutamatergic axon terminals. Thus, the DOI-induced increase of EPSC in the mPFC slices and c-fos expression in the somatosensory cortex depends on the integrity of thalamo-cortical afferents (Aghajanian and Marek 1997, 1999; Scruggs et al. 2000; Marek et al. 2001). However, this interpretation is challenged by in vivo findings that DOI-induced activation of pyramidal neurones in the mPFC of anesthetized rats did not depend on thalamo-cortical afferents (Puig et al. 2003). Thus, the effect of M100,907 on the increase of GLU induced by CPP could be mediated by 5-HT2A receptors on GLU afferents arising from other brain regions or, most probably, from those present in large amounts on cortical pyramidal neurons (Jakab and Goldman-Rakic 1998; Miner et al. 2003). Electrophysiological studies in slices of the mPFC have shown that the major effect of 5-HT is on pyramidal cells (Aghajanian and Marek 1997) and M100,907 prevented PCP-induced blockade of NMDA responses in pyramidal neurons (Wang and Liang 1998). 5-HT2A receptors are also present on GABAergic interneurones (Willins et al. 1997; Jakab and Goldman-Rakic 1998). However, the local infusion of DOI raised extracellular GABA levels and c-fos expression (Abi-Saab et al. 1999) while M100,907 inhibited K+-evoked [3H]GABA release in rat cortical slices (Cozzi and Nichols 1996). Thus, a direct effect of M100,907 on GABAergic interneurones is unlikely to be involved in the reversal of CPP's effect on extracellular GLU.

In contrast with the present findings, Adams and Moghaddam (2001) found that M100,907 did not antagonize the increase of cortical extracellular GLU induced by PCP. This discrepancy may be due to the fact that we used different NMDA receptor antagonists (CPP/AP7 vs. PCP) and route of administration (intracortical vs. intraperitoneal). PCP has considerable affinity for 5-HT2 and sigma receptors, ionic channels other than NMDA receptors and DA and 5-HT transporters (Nabeshima et al. 1988; Wong et al. 1988; Javitt and Zukin 1991; Maurice et al. 1991; Kapur and Seeman 2002), whereas CPP is far more selective (Lehmann et al. 1987). Thus, besides NMDA receptor blockade, the interaction with these mechanisms may contribute to the effect of PCP on cortical GLU. We did not examine whether M100,907 antagonized the effect of systemic CPP on extracellular GLU so it cannot be excluded that different routes of administration of PCP and CPP might account for the discrepancy between the present results and those by Adams and Moghaddam (2001).

Behavioral deficits induced by NMDA receptor antagonists have been shown to be due to increased extracellular GLU in the mPFC (Moghaddam et al. 1997; Moghaddam and Adams 1998). Thus, the modulatory role of 5-HT2A receptors on GLU availability may partly explain why the 5-HT2A receptor antagonist M100,907 prevents some of the behavioral effects of the NMDA receptor antagonists (Maurel-Remy et al. 1995; Martin et al. 1998b; Corbett et al. 1999; Varty et al. 1999; Abekawa et al. 2003; Higgins et al. 2003). We have shown that 10–50 ng/µL CPP injected into the mPFC caused an attentional deficit and increased the anticipatory responses in rats performing a 5-CSRT task, and these effects were reversed by 10–40 µg/kg M100,907 (Carli et al. 2004). In line with these results, a recent study showed that M100,907 attenuated MK-801-induced anticipatory responding in a 5-CSRT task in rats (Higgins et al. 2003). The fact that the same doses of M100,907 that prevent deficits in attentional performance (Carli et al. 2004) attenuate the effect of CPP on extracellular GLU (present study) suggests that the stimulation of GLU release may play a role in the attentional performance deficits caused by CPP. Further studies assessing the effects of drugs that suppress GLU release, such as the agonists of the group II metabotropic glutamate receptors (Moghaddam and Adams 1998; Cartmell and Schoepp 2000), on CPP-induced attentional deficits and GLU release would be needed to confirm this hypothesis.

In conclusion, the present study found that blockade of 5-HT2A receptors prevented the increase in extracellular GLU induced by intracortical infusion of the NMDA receptor antagonist CPP. This suggests that blockade of 5-HT2A receptors may provide a molecular mechanism for reversing behavioral deficits caused by dysfunctional glutamatergic transmission in the mPFC.

Acknowledgements

This study was supported in part by a grant from MIUR-FIRB (RBAU01ZS5C). We are grateful to Dr S. Kongsamut from Aventis Pharmaceuticals (NJ, USA) for the generous gift of M100,907 and to J. Baggott for language editing.

Ancillary