Opposite regulation by typical and atypical anti-psychotics of ERK1/2, CREB and Elk-1 phosphorylation in mouse dorsal striatum

Authors


Address correspondence and reprint requests to Gilberto Fisone, Department of Neuroscience, Karolinska Institutet, Retzius väg 8, S-171 77 Stockholm, Sweden. E-mail: gilberto.fisone@neuro.ki.se

Abstract

The two mitogen-activated protein kinases (MAPKs), extracellular signal-regulated protein kinase 1 and 2 (ERK1/2), are involved in the control of gene expression via phosphorylation and activation of the transcription factors cyclic AMP response element binding protein (CREB) and Elk-1. Here, we have examined the effect of haloperidol and clozapine, two anti-psychotic drugs, and eticlopride, a selective dopamine D2 receptor antagonist, on the state of phosphorylation of ERK1/2, CREB and Elk-1, in the mouse dorsal striatum. Administration of the typical anti-psychotic haloperidol stimulated the phosphorylation of ERK1/2, CREB and Elk-1. Virtually identical results were obtained using eticlopride. In contrast, the atypical anti-psychotic clozapine reduced ERK1/2, CREB and Elk-1 phosphorylation. This opposite regulation was specifically exerted by haloperidol and clozapine on ERK, CREB, and Elk-1 phosphorylation, as both anti-psychotic drugs increased the phosphorylation of the dopamine- and cyclic AMP-regulated phosphoprotein of 32 kDa (DARPP-32) at the cyclic AMP-dependent protein kinase (PKA) site. The activation of CREB and Elk-1 induced by haloperidol appeared to be achieved via different signalling pathways, as inhibition of ERK1/2 activation abolished the stimulation of Elk-1 phosphorylation without affecting CREB phosphorylation. This study shows that haloperidol and clozapine induce distinct patterns of phosphorylation in the dorsal striatum. The results provide a novel biochemical paradigm elucidating the molecular mechanisms underlying the distinct therapeutic actions of typical and atypical anti-psychotic agents.

Abbreviations used
CREB

cyclic AMP response element binding protein

DARPP-32

dopamine and cyclic AMP-regulated phosphoprotein of 32 kDa

ERK1/2

extracellular signal-regulated protein kinase 1 and 2

MAPKs

mitogen-activated protein kinases

MEK

MAPK/ERK kinase

Treatment with anti-psychotic (or neuroleptic) drugs currently represents the most common therapy for schizophrenia. One major limitation in the use of conventional anti-psychotic drugs, such as haloperidol, is that their prolonged administration induces extra-pyramidal side-effects by modifying transmission in the basal ganglia, a group of brain structures involved in the control of voluntary movements. The atypical anti-psychotic, clozapine, has a superior therapeutic efficacy and does not produce extra-pyramidal side-effects. The use of this drug, however, is complicated by the relatively high incidence of agranulocytosis, a condition that requires the immediate suspension of clozapine administration.

A common feature of anti-psychotic drugs is their ability to act as dopamine D2 receptor antagonists. Blockade of dopamine D2 receptors profoundly affects gene expression in the striatum, the major receiving area of the basal ganglia. Administration of haloperidol produces a rapid increase in the expression of various genes including the immediate early gene c-fos (Dragunow et al. 1990; Miller 1990) and the gene for pro-enkephalin (Konradi et al. 1993). Clozapine is much less effective than haloperidol in inducing gene expression in the dorsal striatum (Robertson and Fibiger 1992; MacGibbon et al. 1994). The reasons for this difference are not yet understood and have been attributed to the relatively low affinity of clozapine for dopamine D2 receptors and/or to its ability to interact with different receptors (Meltzer et al. 1989; Guo et al. 1995). Because the dorsolateral striatum has been associated with motor side-effects (Pisa and Schranz 1988), the different patterns of gene induction produced by haloperidol and clozapine in this region has been proposed to be responsible for the lower incidence of extrapyramidal side-effects observed in clozapine-treated patients (Robertson and Fibiger 1992).

The stimulation exerted by haloperidol on c-fos expression appears to be mediated via phosphorylation and activation of the transcription factor cyclic AMP response element binding protein (CREB) (Konradi and Heckers 1995). CREB is phosphorylated on Ser133 in response to stimulation of various protein kinases, including cyclic AMP-dependent protein kinase (PKA) (Gonzales and Montminy 1989), calcium/calmodulin-dependent protein kinases (Sheng et al. 1991) and mitogen activated protein kinases (MAPKs) (Xing et al. 1996). Another transcription factor that may be involved in the haloperidol-mediated stimulation of c-fos expression is Elk-1, a member of the Ets domain transcription factor family (Hipskind et al. 1991; Marais et al. 1993; Treisman 1995). Elk-1 is activated by phosphorylation on Ser383 and Ser389 catalysed by MAPKs (Marais et al. 1993; Hipskind et al. 1994; Gille et al. 1995). In the striatum, glutamate-induced phosphorylation of CREB and Elk-1 appears to be mediated via activation of the two MAPKs, extracellular signal-regulated protein kinase 1 and 2 (ERK1/2) (Sgambato et al. 1998; Vanhoutte et al. 1999; Choe and McGinty 2001).

In this study we have examined the effects, in the dorsal striatum, of low doses of haloperidol, clozapine, as well as eticlopride, a highly selective dopamine D2 receptor antagonist, on the state of phosphorylation of ERK1/2, CREB and Elk-1. In addition, we have determined the effects of the three drugs on the phosphorylation of the dopamine and cyclic AMP-regulated phosphoprotein of 32 kDa, DARPP-32, an important mediator of dopaminergic transmission (Walaas et al. 1983; Hemmings et al. 1984). Finally, we have examined the involvement of ERK1/2 in the haloperidol-induced increase in CREB and Elk-1 phosphorylation.

Materials and methods

Drugs

Eticlopride was purchased from Research Biochemical International (Research Biochemicals Inc., Natick, MA, USA). Haloperidol and clozapine were purchased from Sigma (St Louis, MO, USA). Eticlopride was dissolved in 0.9% saline. Haloperidol and clozapine were dissolved in a solution of 10% acetic acid in saline and the pH was brought to 5.5 with 1.0 m NaOH. SL327 (a gift from Dr James Trzaskos, DuPont Pharmaceuticals, Wilmington, DE, USA) was dissolved in 40% dimethylsulfoxide (DMSO).

Tissue extraction

Male C57BL/6 mice (25–30 g; B & K, Stockholm, Sweden) were injected with drugs or vehicle and killed by decapitation at various times as described. Following decapitation, the heads of the animals were immediately immersed in liquid nitrogen for 6 s. The brains were then removed, the dorsal striata were dissected out within 20 s on an ice-cold surface, sonicated in 750 μL of 1% sodium dodecylsulfate and boiled for 10 min. The effectiveness of this extraction procedure in preventing protein phosphorylation and dephosphorylation, hence ensuring that the level of phosphoproteins measured ex vivo reflects the in vivo situation, has previously been tested (Svenningsson et al. 2000).

Determination of phosphoproteins

Aliquots (5 μL) of the homogenate were used for protein determination using a BCA (bicinchoninic acid) assay kit (Pierce Europe, Oud Beijerland, the Netherlands). Equal amounts of protein (15 or 30 μg for total or phosphoproteins, respectively) from each sample were loaded onto 10% polyacrylamide gels. The proteins were separated by sodium dodecylsulfate-polyacrylamide gel electrophoresis and transferred overnight onto polyvinylidene difluoride membranes (Amersham-Pharmacia Biotech, Uppsala, Sweden) as described (Towbin et al. 1979). Membranes were incubated for 2 h with polyclonal antibodies against phospho[Thr202/Tyr204]-ERK1/2 (Cell Signaling Technology, Beverly, MA, USA) and phospho[Ser133]-CREB (Upstate Biotechnology, Lake Placid, NY, USA) or with monoclonal antibodies against phospho[Ser383]-Elk-1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and phospho[Thr34]-DARPP-32 (Snyder et al. 1992). Antibodies against ERK1/2 (Cell Signaling Technology), CREB (Upstate Biotechnology), Elk-1 (Santa Cruz Biotechnology) and DARPP-32 (Hemmings and Greengard 1986) which were not phosphorylation state specific were used to estimate the total amounts of these proteins in tissue extracts. Antibody binding was revealed using goat anti-rabbit or goat anti-mouse horseradish peroxidase linked IgGs (Pierce Europe, Oud Beijerland, the Netherlands) and the Enhanced Chemiluminescence Plus immunoblotting detection kit (Amersham Pharmacia Biotech, Uppsala, Sweden). Chemiluminescence was detected by autoradiography and quantification of proteins was done by densitometry, using the NIH Image software (version 1.61).

Immunohistochemistry

Mice were anaesthetized and perfused intracardially via the left ventricle with 100 mL of 4% paraformaldehyde (20 mL/min). The brains were rapidly removed and post fixed in fresh fixative solution at 4°C overnight. Coronal sections (50 μm) were cut using a cryostat (for phospho-CREB) or a vibrotome (for phospho-ERK1/2) and kept at 4°C in 0.1 mm NaF dissolved in phosphate-buffered saline (PBS). For phospho-CREB, free-floating sections were washed for 15 min in PBS/0.3% Triton X-100 and pre-incubated at 20°C for 30 min in PBS/10% goat serum (Vector Laboratories, Peterborough, UK). Sections were rinsed with PBS and then incubated overnight at 4°C in PBS/0.3% BSA/0.3% Triton X-100-containing polyclonal antibody against phospho[Ser133]-CREB (Upstate Biotechnology; diluted 1 : 700). Negative control staining was performed by incubating slices in PBS/0.3% BSA/0.3% Triton X-100 without antibodies. The following day, sections were rinsed in PBS and incubated for 2 h at room temperature with anti-rabbit biotinylated antibodies (Vector Laboratories; diluted 1 : 1000). To bleach endogenous peroxidase, sections were incubated for 15 min in PBS containing 1% H2O2 and 20% methanol, and then rinsed twice in PBS. Slices were then incubated for 60 min in avidin–biotin–peroxidase complex (ABC) solution (final dilution 1 : 100; Vector Laboratories). Immunoreactivity to phospho-CREB was developed by incubating the slices for 2–4 min in a solution containing 0.5 mg/mL of 3,3′-diaminobenzidine (DAB), 3 mg/mL of nickel ammonium sulfate and 0.3 μL/mL of H2O2. At the end of this period, slices were washed twice, mounted onto poly-lysinate slides, dehydrated in ethanol/ethanol-xilene solutions and subjected to light microscopic examination. For phospho-ERK1/2 we used the procedure described by Sgambato et al. (1998).

Results

Regulation of ERK1/2 phosphorylation

Both ERK isoforms, ERK1 (44 kDa) and ERK2 (42 kDa), were detected in the striatum with a large prevalence for ERK2 immunoreactivity (data not shown). Administration of 0.2 mg/kg of haloperidol produced a time-dependent increase in ERK1/2 phosphorylation (Figs 1a and b; open circles) with a maximal effect reached 15 min following injection. The levels of phospho-ERK 1 were still elevated 60 min after haloperidol administration (Fig. 1a, open circles), whereas phospho-ERK2 returned to control levels 30 min after injection (Fig. 1b, open circles). Very similar results were obtained using the specific dopamine D2 receptor antagonist, eticlopride, injected at the dose of 0.5 mg/kg (Figs 1a and b; squares). In contrast, clozapine, injected at the dose of 5 mg/kg, caused a time-dependent decrease in ERK1/2 phosphorylation (Figs 1a and b; closed circles). This effect was particularly evident for phospho-ERK-2, which was markedly reduced at 15 min and almost undetectable at 60 min following clozapine administration (Fig. 1b, closed circles). Phospho-ERK1 was also significantly reduced 60 min after clozapine injection (Fig. 1a, closed circles). The ability of haloperidol to increase ERK1/2 phosphorylation in the dorsal striatum was confirmed by immunocytochemical analysis performed using the same antibodies that had been used for western blotting (Fig. 3e). In the same series of experiments, clozapine slightly decreased ERK1/2 phosphorylation (Fig. 3f).

Figure 1.

Regulation of ERK1/2 phosphorylation by haloperidol, clozapine and eticlopride. Mice were treated i.p. with haloperidol (0.2 mg/kg, open circles), clozapine (5 mg/kg, closed circles), or eticlopride (0.5 mg/kg, squares) and killed 15, 30 or 60 min later. Upper panels (a) phospho-ERK1 and (b) phospho-ERK2 were determined by western blotting, using phosphorylation state specific antibodies (see Materials and methods). The amounts of phospho-ERK1/2 are expressed as percentage of those measured in vehicle-treated mice (dotted line). Data represent means ± SEM (n = 7–12). *p < 0.05 and **p < 0.01 versus respective vehicle-treated group, one-way anova followed by Dunnett's test. Lower panels, representative autoradiograms showing phospho-ERK1/2 immunoreactivity in the dorsal striatum 15 min following administration of (c) vehicle (d) haloperidol (0.5 mg/kg) or (e) clozapine (5 mg/kg).

Figure 3.

Immunocytochemical analysis of CREB and ERK1/2 phosphorylation following typical and atypical anti-psychotic treatment. Mice were treated i.p. with vehicle (a, d), haloperidol (0.5 mg/kg; b, e), or clozapine (5 mg/kg; c, f) and killed 15 min later. Brain coronal sections were incubated with antibodies against phospho-CREB (a–c) or phospho-ERK1/2 (d–f). (A) Section of the mouse brain showing the area of the dorsal striatum where high magnification microscopic analyses were performed. The changes in phospho-CREB- and phospho-ERK1/2 immunorectivities induced by the treatments reflect those obtained by western immunoblotting analysis.

Regulation of CREB phosphorylation

As shown in Fig. 2, administration of 0.2 mg/kg of haloperidol stimulated CREB phosphorylation on Ser133. Phospho-CREB was significantly increased at 15 min following haloperidol administration and returned to near basal levels after 30 min (Fig. 2a; open circles). Administration of eticlopride produced a similar increase in CREB phosphorylation, with a maximal effect observed at 15 min following injection (Fig. 2a; squares). Conversely, clozapine produced a rapid and long lasting (60 min) decrease in CREB phosphorylation (Fig. 2a; closed circles). These results were supported by immunocytochemical experiments (Fig. 3a–c).

Figure 2.

Regulation of CREB phosphorylation by haloperidol, clozapine and eticlopride. Mice were treated i.p. with haloperidol (0.2 mg/kg, open circles), clozapine (5 mg/kg, closed circles), or eticlopride (0.5 mg/kg, squares) and killed 15, 30 or 60 min later. Left panel, phospho-CREB was determined by western blotting, using a phosphorylation state specific antibody (see Materials and methods). The amount of phospho-CREB is expressed as percentage of that measured in vehicle treated mice (dotted line). Data represent means ± SEM (n = 8–11). *p < 0.05 and **p < 0.01 versus vehicle-treated group, one-way anova followed by Dunnett's test. Right panels, representative autoradiograms showing phospho-CREB immunoreactivity in the dorsal striatum 15 min following administration of (b) vehicle (c) haloperidol (0.5 mg/kg) or (d) clozapine (5 mg/kg).

Regulation of DARPP-32 phosphorylation

Administration of haloperidol, clozapine or eticlopride stimulated the phosphorylation of DARPP-32 at the PKA site (i.e. Thr34) (Hemmings et al. 1984) (Fig. 4). The effects of clozapine (closed circles) and eticlopride (squares) were still present 60 min after injection, whereas the effect of haloperidol (open circles) lasted for 30 min.

Figure 4.

Regulation of DARPP-32 phosphorylation at Th34 by haloperidol, clozapine and eticlopride. Mice were treated i.p. with haloperidol (0.2 mg/kg, open circles), clozapine (5 mg/kg, closed circles), or eticlopride (0.5 mg/kg, squares) and killed 15, 30 or 60 min later. Left panel, phosphoThr34-DARPP-32 was determined by western blotting, using a phosphorylation state specific antibody (see Materials and methods). The amount of phosphoThr34-DARPP-32 is expressed as percentage of that measured in vehicle-treated mice (dotted line). Data represent means ± SEM (n = 8–11). *p < 0.05 and **p < 0.01 versus vehicle-treated group, one-way anova followed by Dunnett's test. Right panels, representative autoradiograms showing phospho-DARPP-32 immunoreactivity in the dorsal striatum following administration of (b) eticlopride, (c) haloperidol or (d) clozapine.

Effects of combined administration of haloperidol and clozapine on ERK1/2, CREB, and DARPP-32 phosphorylation

Previous studies indicated that clozapine pre-treatment reduces the ability of haloperidol to induce catalepsy and stimulate c-fos expression in the dorsal striatum (Young et al. 1999). As both ERK1/2 and CREB have been linked to c-fos expression (Konradi and Heckers 1995; Sgambato et al. 1998; Vanhoutte et al. 1999; Choe and McGinty 2001), we examined the interaction between haloperidol and clozapine on the regulation of these two phosphoproteins. We found that, when administered 5 min before haloperidol, clozapine strongly reduced the ability of the typical anti-psychotic to stimulate ERK1/2 (Fig. 5a) and CREB (Fig. 5b) phosphorylation. In contrast, clozapine did not affect the increase in phosphorylation produced by haloperidol on Thr34 of DARPP-32, the effects of the two drugs being similar and not additive (Fig. 5c).

Figure 5.

Effect of clozapine on the haloperidol-induced increases in ERK1/2, CREB and DARPP-32 phosphorylation. Mice were treated i.p. with clozapine (5 mg/kg) 5 min before receiving vehicle or haloperidol (0.2 mg/kg) and killed 15 min later. (a) phospho-ERK1/2 (b) phospho-CREB, and (c) phospho-DARPP-32 were determined by western blotting, using phosphorylation state specific antibodies (see Materials and methods). Upper panels, representative autoradiograms. Lower panels, the amounts of phosphoproteins are expressed as percentage of those measured in vehicle-treated mice. Data represent means ± SEM (n = 7–17). *p < 0.05 and **p < 0.01 versus vehicle-treated group, one-way anova followed by Newman-Keuls test. †p < 0.05 interaction betweeen clozapine and haloperidol treatments, two-way anova.

Effect of SL327 on ERK1/2, Elk-1, CREB, and DARPP-32 phosphorylation

Our data showed that haloperidol increased, whereas clozapine decreased, the state of phosphorylation of ERK1/2 and CREB (Figs 1–3). We therefore examined the effects of the two neuroleptic drugs on the phosphorylation of Elk-1, a transcription factor directly activated by ERK1/2. We found that systemic administration of clozapine reduced the phosphorylation of Elk-1 (Fig. 6). Conversely, haloperidol produced a significant increase in Elk-1 phosphorylation (Fig. 7b). As the increases in ERK1/2, CREB and Elk-1 phosphorylation produced by haloperidol occurred within the same time frame, we checked the requirement of ERK1/2 activation for the haloperidol-induced increase in CREB and Elk-1 phosphorylation. Pre-treatment with 100 mg/kg (i.p.) of the MAPK/ERK kinase (MEK) inhibitor, SL327 (Selcher et al. 1999), reduced both basal and haloperidol-stimulated levels of phosphorylated ERK1/2 (Fig. 7a). SL327 also prevented the increase in Elk-1 phosphorylation caused by haloperidol (Fig. 7b). In contrast, the MEK inhibitor did not modify either basal or haloperidol-stimulated CREB phosphorylation (Fig. 7c). SL327 also failed to modify the increase in DARPP-32 phosphorylation induced by haloperidol at the PKA site (Fig. 7d), thereby confirming its specificity for MEK.

Figure 6.

Regulation of Elk-1 phosphorylation by clozapine. Mice were treated i.p. with clozapine (5 mg/kg, closed circles) and killed 15, 30 or 60 min later. Phospho-Elk-1 was determined by western blotting, using a phosphorylation state specific antibody (see Materials and methods). Upper panel, representative autoradiogram. Lower panel, the amount of phospho-Elk 1 is expressed as percentage of that measured in vehicle-treated mice (dotted line). Data represent means ± SEM (n = 5–8). *p < 0.05 versus vehicle-treated group, one-way anova followed by Dunnett's test.

Figure 7.

Effect of SL327 on the haloperidol-induced increases in ERK1/2, Elk-1, CREB and DARPP-32 phosphorylation. Mice were treated i.p. with SL327 (100 mg/kg) 45 min before receiving vehicle or haloperidol (0.5 mg/kg) and killed 15 min later. (a) phospho-ERK1/2 (b) phospho-Elk-1 (c) phospho-CREB, and (d) phospho-DARPP-32 were determined by western blotting, using phosphorylation state specific antibodies (see Materials and methods). Upper panels, representative autoradiograms. Lower panels, the amounts of phosphoproteins are expressed as percentage of those measured in vehicle-treated mice. Data represent means ± SEM (n = 4–10). *p < 0.05, **p < 0.01 and ***p < 0.001 versus vehicle-treated group, one-way anova followed by Newman-Keuls test. †p < 0.05 interaction between SL327 and haloperidol treatments, two-way anova.

Total levels of ERK1/2, CREB, Elk-1 and DARPP-32

Total levels of ERK1/2, CREB, Elk-1 and DARPP-32 were measured in the same tissue extracts used for the determination of phosphorylated proteins and were not modified by any of the treatments described in the study (data not shown).

Discussion

The present study demonstrates that, in the dorsal striatum, signal transduction pathways linked to gene transcription are differently regulated by standard clinical doses of haloperidol and clozapine. Haloperidol increases the phosphorylation of ERK1/2, CREB and Elk-1, whereas clozapine decreases the levels of all three phosphorylated proteins. In contrast, both anti-psychotic drugs stimulate DARPP-32 phosphorylation at the PKA site.

The increases in protein phosphorylation produced by haloperidol are identical to those observed following administration of eticlopride, a selective dopamine D2 receptor antagonist. Dopamine D2 receptors are negatively coupled to adenylyl cyclase and cyclic AMP production (Stoof and Kebabian 1981). Thus, the ability of haloperidol and eticlopride to increase CREB and DARPP-32 phosphorylation is most likely due to the removal of the inhibition exerted by dopamine D2 receptors on PKA, as previously suggested (Konradi and Heckers 1995; Svenningsson et al. 2000). Blockade of the D2 receptor-mediated activation of calcineurin also is likely to contribute to the increase in DARPP-32 phosphorylation observed with these drugs (Nishi et al. 1997).

It was previously reported that the dopamine D2 receptor agonist, quinpirole, stimulated ERK1/2 phosphorylation in brain slices (Yan et al. 1999). A similar effect was produced by quinpirole in the striatum, following lesion of the nigrostriatal pathway with 6-hydroxydopamine (Cai et al. 2000). However, in a recent study, Gerfen et al. (2002) found that the stimulatory effect exerted by quinpirole on ERK1/2 phosphorylation is mostly restricted to cholinergic interneurons, which, in the striatum, represent only a small fraction of the total neuronal population. Moreover, the same authors observed an increase in the state of phosphorylation of ERK1/2 in the dorsal striatum following systemic administration of eticlopride. Our results support these findings and show that, in the mouse, the typical neuroleptic haloperidol produces a similar activation of the MAPK pathway. In the dorsal striatum, administration of quinpirole has been shown to prevent the increase in ERK1/2 phosphorylation produced by stimulation of glutamatergic cortical areas (Gerfen et al. 2002). Thus, it is possible that haloperidol exerts its effect by blocking an inhibitory tone normally exerted by dopamine D2 receptors on glutamate-induced ERK1/2 phosphorylation.

In addition to the activation of ERK1/2, haloperidol produces a simultaneous increase in the phosphorylation of CREB and Elk-1. ERK1/2 are known to phosphorylate CREB via stimulation of the p90RSK (p90 ribosomal 6S kinase) (Xing et al. 1996). In addition, ERK1/2 directly phosphorylate Elk-1 (Marais et al. 1993; Gille et al. 1995). In agreement with these observations, it was reported that, in the striatum, activation of ERK1/2 was required for the increases in CREB and Elk-1 phosphorylation produced by glutamate (Sgambato et al. 1998; Vanhoutte et al. 1999; Choe and McGinty 2001). Our results show that SL327, a drug that prevents ERK1/2 activation by blocking MEK-catalysed phosphorylation, abolished the haloperidol-induced increase in Elk-1 phosphorylation. In contrast, SL327 did not affect the increase in CREB phosphorylation produced by haloperidol. Thus, it appears that, in the dorsal striatum, distinct signalling pathways are implicated in the haloperidol-induced stimulation of Elk-1 and CREB phosphorylation. Haloperidol increases Elk-1 phosphorylation by activation of ERK1/2. In contrast, the increase in CREB phosphorylation produced by the typical anti-psychotic is independent of the MAPK cascade and most likely due to disinhibition of the cyclic AMP pathway and activation of PKA (Konradi and Heckers 1995; Leveque et al. 2000).

Administration of clozapine stimulates the phosphorylation of Thr34 of DARPP-32, suggesting that, similarly to haloperidol and eticlopride, clozapine regulates DARPP-32 phosphorylation by acting as a dopamine D2 receptor antagonist, via disinhibition of PKA. The idea that, in the dorsal striatum, clozapine and haloperidol share at least part of their mechanisms of action (i.e. both act as dopamine D2 receptor antagonists) is supported by the observation that their stimulatory effects on DARPP-32 phosphorylation are not additive.

Clozapine produced a sustained decrease in the state of phosphorylation of ERK1/2, CREB and Elk-1. This effect may depend on the ability of clozapine to interfere with multiple receptors, thereby affecting other signalling pathways involved in the phosphorylation of these proteins (Brunello et al. 1995). For instance, it has been proposed that the ability of PKA to stimulate CREB phosphorylation depends on secondary activation of NMDA receptors and calcium channels (Rajadhyaksha et al. 1999; Leveque et al. 2000). Increased calcium concentration is also critically involved in ERK1/2 phosphorylation (Grewal et al. 1999). Thus, the decreases in ERK1/2, CREB and Elk-1 phosphorylation produced by clozapine in the dorsal striatum may depend on the ability of this drug to reduce intracellular calcium. Interestingly, clozapine has been shown to inhibit voltage-dependent calcium channels (Park et al. 2001). Future studies will be necessary to assess the involvement of these specific changes in the ability of clozapine to reduce ERK1/2, CREB and Elk-1 phosphorylation.

The negative control exerted on the MAPK cascade and on transcription factors interacting with regulatory elements (i.e. cyclic AMP response element and serum response element) on the c-fos promoter, may account for the lower efficacy shown by clozapine in inducing immediate early genes in the dorsal striatum (Deutsch et al. 1992; Nguyen et al. 1992; Robertson and Fibiger 1992). We also found that clozapine abolished the increases in ERK1/2 and CREB phosphorylation induced by haloperidol, thereby providing a plausible explanation for the ability of clozapine to prevent haloperidol-induced c-fos expression and catalepsy (Young et al. 1999).

In conclusion, this study demonstrates that, in the dorsal striatum, administration of haloperidol results in the activation of ERK1/2, which in turn is responsible for the phosphorylation and activation of Elk-1. Furthermore, haloperidol stimulates the phosphorylation of CREB via an ERK1/2-independent mechanism, most likely involving activation of PKA. In contrast, the atypical anti-psychotic, clozapine, reduces ERK1/2, CREB and Elk-1 phosphorylation. These distinct patterns of protein phosphorylation induced by haloperidol and clozapine may be related to the low incidence of extrapyramidal side-effects observed in clozapine-treated schizophrenic patients and represents a potential tool to characterize the profile of novel anti-psychotic drugs.

Acknowledgements

We thank L. Collin, V. Fontaine, R. Rimondini, A. Oliveira and S. Plantman for technical assistance. This work was funded by Swedish Research Council Grant 13482 (to GF), the Foundation Blanceflor Boncompagni-Ludovisi, née Bildt (to LP), Mariano Scippacercola Foundation (to AU) and by US Public Health Service Grants MH 40899 and DA 10044 (to PG).

Ancillary