INTRODUCTION

Substantial evidence has shown that dopaminergic modulation of glutamate inputs plays a pivotal role in the regulation of attentional processes (Kodama et al, 2002). In particular, the strong influence affected by dopamine on responses mediated by N-methyl-D-aspartate (NMDA) receptors (Clow and Jhamandas, 1989) potentially functions as a filtering device for the extraction of relevant information, either by modifying the signal-to-noise ratio (Cepeda et al, 1992) or by promoting orientation of attentional resources toward significant stimuli (Redgrave et al, 1999). Impairment of filtering mechanisms is conjectured to lead to sensory flooding and sensorimotor gating deficits (McGhie and Chapman, 1961; Braff and Geyer, 1990), which characterize several neuropsychiatric disorders (for a review see Braff et al, 2001), such as bipolar disorder (Saccuzzo and Braff, 1986; Perry et al, 2001), Huntington's disease (Swerdlow et al, 1995), obsessive–compulsive disorder (Swerdlow et al, 1993), Tourette's syndrome (Castellanos et al, 1996), and, particularly, schizophrenia (Braff et al, 1978). The most reliable operational paradigm for the measurement of sensorimotor gating mechanisms is considered the prepulse inhibition (PPI) of the acoustic startle reflex (ASR). Specifically, PPI is the reduction of the startle reflex that occurs when the startling stimulus is preceded by a weak, nonstartling prestimulus (Hoffman and Ison, 1980). Interestingly, both dopaminergic and glutamatergic systems are involved in the neural circuitry of sensorimotor gating and PPI can be disrupted by dopaminergic agonists and NMDA receptor antagonists (for a review see Geyer et al, 2001). Since PPI abolition produced by these compounds on rodents resembles effects present in psychotic patients, and some of these drugs, such as phencyclidine and amphetamine, are also known to induce schizophrenia-like symptoms in humans (Cohen et al, 1962; Bell, 1965; Angrist, 1994; Steinpresis, 1996), the above treatments may serve to model the sensorimotor gating deficits observed in psychosis.

The role of the dopaminergic system in sensorimotor gating has been thoroughly investigated, with evidence indicating D2 receptors, rather than D1, regulate PPI in rats (Peng et al, 1990; Geyer et al, 2001). Indeed, D2 receptors play a key role in schizophrenia, as demonstrated by the fact that antipsychotics predominantly antagonize D2 receptors (Seeman et al, 1975; Kapur and Mamo, 2003). The contribution of D1 appears crucial for the reinforcement of effects mediated by D2 receptors (Geyer et al, 2001). In fact, while D1/D2 agonists such as apomorphine produce dramatic PPI disruption, D2 selective agonists, such as quinpirole, elicit only a mild PPI deficit in Sprague–Dawley rats, and no alterations in Wistar rats (Geyer et al, 1990).

The mechanisms through which NMDA receptor antagonists impair sensorimotor gating are far from being fully elucidated. Although the ability of dizocilpine, the prototypical NMDA receptor antagonist, to increase dopaminergic activity has been well described (Hiramatsu et al, 1989; Rao et al, 1990), many observations confirm that the mechanism through which it produces a robust disruption in PPI is unlikely to be mediated by D2 receptors. A large body of literature shows D2 selective antagonists are unable to reverse PPI disruption induced by NMDA receptor antagonists (Geyer et al, 1990; Johansson et al, 1994). Moreover, although several studies report atypical antipsychotics are effective against deficits caused by NMDA receptor antagonists (Geyer et al, 2001; Bakshi et al, 1994), the assumption remains debated in view of conflicting results on the ability of clozapine and other atypical antipsychotics to antagonize dizocilpine-mediated PPI disruption (Hoffman et al, 1993; Bast et al, 2000).

The role of D1 receptors in the mechanisms accounting for NMDA receptor antagonists has been investigated. In particular, several studies have evidenced the inability of the D1 antagonist SCH 23390 ((R)-(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine) to reverse PPI disruption mediated by dizocilpine and phencyclidine (Bakshi et al, 1994; Wedzony et al, 1994). Moreover, Ralph-Williams et al (2002) demonstrated that dizocilpine induces PPI deficits in D1 and D2 receptor mutant and knockout mice. This result induced the authors of the study to discard hypotheses concerning involvement of both receptors in the NMDA receptor antagonist-mediated impairment of sensorimotor gating. Nonetheless, manifold reports have underlined the activation of D1 receptors as having synergistic effects with dizocilpine, in numerous behavioral tests (Goodwin et al, 1992; Verma and Kulkarni, 1992; Dall'Olio et al, 2000). In light of these premises, the present study was designed to elucidate interactions between D1 and D2 receptors with dizocilpine-mediated disruption of PPI. With this aim in mind, we planned the assessment of the effects on sensorimotor gating of dizocilpine in cotreatment with selective D1 and D2 receptor agonists.

MATERIALS AND METHODS

Animals

A total of 754 experimentally naïve, male Sprague–Dawley rats from Harlan Laboratories, Italy, weighing between 250 and 300 g, served as subjects. Consistent with the guidelines for housing and behavioral testing presented in the literature on startle measures in rodents (Geyer and Swerdlow, 1998), rats were housed four per cage in a room maintained at a temperature of 22±2°C and a humidity of 60%. Food and water were freely available and animals were held under an artificial 12/12-h light/dark cycle, with lights off from 0800 to 2000. In order to reduce stress during the experiment, each rat was gently handled for 5 min on each of the 7 days prior to behavioral testing. All experimental procedures were approved by the local ethics committee and carried out in strict accordance with the guidelines for experimental animals care (European Economic Community (86/609; DL 27/01/92, number 110)). Besides, all efforts were made to minimize the number of animals used, and sample sizes for every experiment were chosen accordingly, without ever compromising the soundness of statistical analysis.

Drugs

The following drugs were used: SKF 38393 ((±)-1-phenyl-2,3,4,5-tetrahydro-(1H)-3-benzazepine-7,8-diol) hydrobromide, SCH 23390 hydrochloride, haloperidol, clozapine, and dizocilpine maleate. All drugs were purchased from Tocris Cookson, UK. SKF 38393 and SCH 23390 were dissolved in distilled water, while dizocilpine was dissolved in 0.9% saline. Haloperidol was dissolved in 10% acetic acid buffered with NaOH and diluted with saline. Clozapine was dissolved in a single drop of 1 N HCl and diluted with saline. The pH was adjusted to 5 using NaHCO3. All drugs were weighed out as salts and administered in an injection volume of 1 ml/kg.

Apparatus

The apparatus used for detection of startle reflexes (Med Associates, St Albans, USA) consisted of four standard cages placed in sound-attenuated chambers with fan ventilation. Each cage consisted of a plexiglas cylinder of 9 cm diameter, mounted on a piezoelectric accelerometric platform connected to an analogue-digital converter. Background noise and acoustic bursts were conveyed by two separate speakers, each one properly placed so as to produce a variation of sound within 1 dB across the startle cage. Both speakers and startle cages were connected to a main PC, which detected and analyzed all chamber variables with specific software. Before each testing session, acoustic stimuli and mechanical responses were calibrated via specific devices supplied by Med Associates.

Procedure

At 3 days before the experiment, all rats went through a brief base-line startle session. Rats were exposed to a background noise of 70 dB and, after an acclimatization period of 5 min, were presented with a randomized sequence of 12 40-ms bursts of 115 dB, interposed with three trials in which a 82 dB prestimulus preceded the same pulse by 100 ms. Subsequently, treatment groups were established so that the average startle response and %PPI of each group were equivalent in all groups. On the testing day, each rat was placed in a cage for a 5 min acclimatization period with a 70 dB white noise background, which continued for the remainder of the session. Each session consisted of three consecutive sequences of trials (periods). Unlike the first and the third period, during which rats were presented with only five pulse-alone trials of 115 dB, the second period consisted of a pseudorandom sequence of 50 trials, including 12 pulse-alone trials, 30 trials of pulse preceded by 73, 76, or 82 dB prepulses (10 for each level of prepulse loudness), and eight no stimulus trials, where only the background noise was delivered. Intertrial intervals were selected randomly between 10 and 15 s.

Experiment Descriptions

The study was articulated in seven experiments. The first experiment was carried out to verify whether SKF 38393 synergizes with dizocilpine in disrupting PPI. A total of 144 rats were therefore pretreated with either saline or SKF 38393 (10 mg/kg, s.c), and, 5 min post-treatment, they were given injections of saline or a dose of dizocilpine (0.0015, 0.005, 0.015, 0.05, and 0.15 mg/kg, s.c.) At 5 min after this second treatment, animals were subjected to the test session. All groups consisted of 12 rats.

The second experiment (n=96; 12 groups of eight rats each) evaluated the effect of quinpirole (0.3, 0.6 mg/kg, s.c.) in cotreatment with dizocilpine, at three doses (0.0015, 0.015, and 0.15 mg/kg, s.c.). Accordingly, rats were pretreated with either 0.9% saline or quinpirole (0.3, 0.6 mg/kg, s.c.). Immediately after this treatment, animals were injected either with saline or with one of the doses of dizocilpine. At 5 min after this injection, rats went through behavioral testing.

The third experiment (n=126; 18 groups of seven rats each) was directed toward screening effects of SCH 23390 on PPI, both per se and in reversing the PPI disruption induced by the cotreatment with dizocilpine and SKF 38393, as verified in the first experiment. Therefore, rats were injected with SCH 23390, at the doses of 0.05 and 0.1 mg/kg (s.c), or saline. After 10 min, each group of rats was further injected either with saline or SKF 38393 and, 5 min later, with saline or a dose of dizocilpine (0.0015, 0.05, or 0.15 mg/kg). Finally, 20 min after SCH 23390 treatment, all rats went through startle testing.

The fourth experiment (n=104) was designed to assess the ability of haloperidol (0.1, 0.5 mg/kg, i.p.) to reverse the effect induced by the cotreatment of dizocilpine and SKF 38393. The schedule of the experiment followed the same used in the third experiment, allowing a period of 60 min before startle testing for haloperidol. All groups consisted of 12 rats, with the exception of the ones that received the antipsychotic in cotreatment with dizocilpine and SKF 38393, which comprised 16 rats each.

The fifth experiment (n=108, nine groups of 12 rats each) was performed to verify whether the same treatment with haloperidol used in the fourth experiment was effective on the PPI disruption induced by quinpirole (0.3, 0.6 mg/kg, s.c.), administered 5 min before startle evaluation.

The sixth experiment (n=104) was aimed at verifying the effects of clozapine (5, 10 mg/kg, i.p.) on the combination with SKF 38393 and dizocilpine. The experimental procedure was identical to the one described in the third and fourth experiment, with a time interval of 40 min between the administration of clozapine and startle testing. All groups consisted of 12 rats, with the exception of the ones that were administered clozapine together with the combination of dizocilpine and SKF 38393, which comprised 16 rats each.

In the seventh experiment, we verified the effects of the same doses of clozapine used in the sixth experiment against a treatment with only dizocilpine (0.15 mg/kg, s.c.). Therefore, 72 rats (six groups of 12 rats each) were administered with either saline or one of the doses of clozapine (5, 10 mg/kg, i.p.) and, after 35 min, were treated with saline or dizocilpine (0.15 mg/kg), in order to be tested 5 min later.

Data Analysis

For each animal, the mean startle amplitudes for the first and the second halves of the second period of the session (blocks, six pulse-alone trials each) were analyzed with a three- or four-way analysis of variance (ANOVA), with pre-pretreatment (where present), pretreatment, and treatment as between-subjects factors and blocks as repeated measures. The %PPI was calculated with the following formula: 100−((mean startle amplitude for prepulse+pulse trials/mean startle amplitude for pulse-alone trials) × 100) and analyzed in multifactor ANOVAs (with specific design and comparisons noted below for each experiment) with the different combinations of injections for pretreatment and treatment as between-subjects factors and trial types as repeated measures. Post hoc analyses were performed using Tukey's test. Alpha was set at 0.05.

RESULTS

Throughout the study no-stimulus trials data were found negligible in comparison with other startle values; therefore, they will not be presented here.

Effects of SKF 38393 Pretreatment in Cotreatment with Dizocilpine

Tested in the experiment were the effects of dizocilpine treatment at five doses (0.0015, 0.005, 0.015, 0.05, and 0.15 mg/kg, s.c.) after saline or SKF 38393 (10 mg/kg, s.c.) pretreatment. Startle magnitudes were evaluated using a three-way ANOVA (pretreatment and treatment being between-subjects factors, and blocks being repeated measures). ANOVA exhibited a significant block effect (F(1,132)=25.06, P<0.001). Moreover, as shown in Table 1, neither pretreatment with SKF 38393 nor treatment with dizocilpine significantly altered baseline startle (pretreatment: F(1,132)=0.03; NS; treatment: F(5,132)=1.16; NS, ANOVA), although the doses of 0.05 and 0.15 mg/kg dizocilpine slightly increased startle amplitude. Finally, ANOVA found no pretreatment × treatment interaction (F(5,132)=0.05; NS).

Table 1 Mean Startle Amplitudes of Treatment Groups in the First Experiment
Full size table

PPI was evaluated by a three-way, fixed factors, repeated-measures ANOVA, with pretreatment, treatment, and trial types as factors. As indicated in Figure 1, statistical evaluation remarkably attested a significant interaction between pretreatment and treatment (F(5,132)=8.60; P<0.0001, ANOVA). In detail, post hoc analysis proved that dizocilpine significantly disrupted PPI in comparison with saline at the doses of 0.05 and 0.15 mg/kg (P<0.0001 in comparison with saline), while SKF 38393 failed to elicit any alteration of PPI. Furthermore, SKF 38393 dramatically potentiated the disruptive effect of dizocilpine in the ‘dose window’ between 0.005 and 0.05 mg/kg at all prepulse levels (P<0.01 for all groups, in comparison with controls), while no such effect was observed for higher doses (presumably due to a ‘floor’ effect). As expected, significant differences were additionally found between different prepulse levels (F(2,264)=171.92; P<0.0001, ANOVA).

Figure 1
figure 1

Effects of the synergy of dizocilpine and SKF 38393 (10 mg/kg). Pretreatments and treatment are indicated over and below the braces, respectively. All dizocilpine doses are given in mg/kg. Values represent mean±SEM. for each treatment. Prepulses are indicated by the intensity corresponding to decibels above background noise. Sal, saline; SKF, SKF 38393; diz, dizocilpine. °°°P<0.001 compared with the respective sal+diz group; ***P<0.001 compared to SKF+sal group; +++P<0.001 compared to sal+sal group.

Full size image

Effects of Quinpirole Pretreatment in Combination with Dizocilpine

In the second experiment, three different dizocilpine doses, ranging from 0.0015 to 0.15 mg/kg (s.c.), were administered following injection of the D2-receptor agonist quinpirole (0.3 and 0.6 mg/kg, s.c.) or saline. A three-way ANOVA, with pretreatment and treatment as independent variables and blocks as repeated measures, was performed to scrutinize startle amplitudes. Table 2 shows that, although no fully significant effect was observed for pretreatment, a definite statistical trend was exhibited for quinpirole to increase startle amplitude (F(2,84)=1.69; P<0.1). As in the first experiment, dizocilpine weakly, but not significantly, increased ASR at the highest doses (F(3,84)=0.47; NS). Finally, a significant difference between block values revealed an habituation effect for startle magnitudes (F(1,84)=12.02; P<0.001).

Table 2 Mean Startle Amplitudes of Treatment Groups in the Second Experiment
Full size table

A three-way ANOVA designed identically as in the previous experiment was used to analyze PPI levels. As expected, ANOVA revealed a significant effect for pretreatment (F(2,84)=12.30; P<0.0001, ANOVA) and post hoc revealed that both at the dose of 0.3 mg/kg (P<0.01) or at the dose of 0.6 mg/kg (P<0.001) quinpirole is able to disrupt PPI. In addition, dizocilpine was shown to disrupt PPI at the dose of 0.15 mg/kg ((F(3,84)=93.73; P<0.0001, ANOVA); P<0.0005 between 0.15 mg/kg dizocilpine and saline, Tukey) (Figure 2). Remarkably, no significant effect was detected for the interaction pretreatment × treatment (F(6,84)=1.58; NS, ANOVA), showing no synergistic effect between quinpirole and dizocilpine on PPI parameters. Eventually, a significant difference between prepulse intensities was also found (F(2,168)=70.76; P<0.0001).

Figure 2
figure 2

(a and b) Effects of the cotreatment of dizocilpine (0.0015, 0.015, and 0.15 mg/kg) and quinpirole (0.3 and 0.6 mg/kg). All quinpirole and dizocilpine doses are given in mg/kg. For all groups n=8. Values represent mean±SEM for each treatment. Prepulses are indicated by the intensity corresponding to decibels above background noise. Sal, saline; diz, dizocilpine; quin, quinpirole; **P<0.01, ***P<0.001 compared to groups receiving saline pre-treatment; °°°P<0.001 compared to groups receiving saline treatment. For further details, see text.

Full size image

Effects of SCH Pretreatment vs SKF and Dizocilpine

The third experiment evaluated the capability of SCH 23390 (0.05, 0.01 mg/kg, s.c.), a D1 receptor antagonist, in reversing the PPI disruption induced by a cotreatment with SKF (10 mg/kg, s.c.) and dizocilpine (0.015, 0.15 mg/kg, s.c.). Startle amplitudes were analyzed by a four-way ANOVA with the three series of treatments as independent variables and blocks as repeated measures. ANOVA showed neither SCH (F(2,110)=0.04; NS) nor SKF (F(1,110)=0.01; NS) were able to significantly alter startle amplitudes. Besides, startle magnitude increase induced by dizocilpine, as shown in Table 3, did not appear statistically significant (F(2,110)=3.03; NS). Instead, ANOVA showed a significant effect for blocks (F(1,110)=12.12; P<0.001]. A further four-way ANOVA, with the same independent variables and with PPI levels as repeated measures, served to test PPI values. ANOVA revealed either in SCH groups (F(2,110)=10.92; P<0.0001) or in SKF groups (F(1,110)=6.44; P<0.05] a significant alteration of PPI. As expected, dizocilpine significantly reduced PPI (F(2,110)=289.49; P<0.0001). A significant difference between prepulse levels was also found (F(2,220)=280.93; P<0.0001). As in the first experiment, a significant PPI reduction was observed for the groups treated with SKF 38393 and dizocilpine in comparison with controls treated with saline and the D1 agonist (Figure 3). Moreover, SCH 23390, at both doses antagonized the disruption induced by the cotreatment with SKF 38393 and 0.015 mg/kg dizocilpine. As expected, no antagonism of SCH 23390 was observed against the cotreatment with SKF 38393 and the dose of dizocilpine effective per se. Remarkably, Tukey's test revealed a statistical trend for the dose of 0.1 mg/kg SCH 23390 to attenuate the disruptive effect of 0.15 mg/kg dizocilpine+saline (P<0.1).

Table 3 Mean Startle Amplitudes of Treatment Groups in the Third Experiment
Full size table
Figure 3
figure 3

(a and b) Representation of the effects of SCH 23390 on the synergy SKF+diz. Prepretreatments and pretreatments+treatments are indicated over and below the braces, respectively. All dizocilpine doses are given in mg/kg. For all groups n=7. Values represent mean±SEM for each treatment. Prepulses are indicated by the intensity corresponding to decibels above background noise. Sal, saline; diz, dizocilpine; SCH, SCH 23390; SKF, SKF 23390. In graph 5 (a) ++P<0.01, +++P<0.001 compared to sal+SKF+sal group; ##P<0.01 compared to sal+SKF+diz 0.015; ***P<0.001 compared to SCH 0.05+SKF+diz 0.15; °°°P<0.001 compared to SCH 0.1+SKF+diz 0.15. In graph (b) +++P<0.001 compared to sal+sal+sal group; ***P<0.001 compared to SCH 0.05+sal+diz 0.15; °°°P<0.001 compared to SCH 0.1+sal+diz 0.15. For further details see text.

Full size image

Effects of Haloperidol Pretreatment vs SKF and Dizocilpine

In the fourth experiment haloperidol was intraperitoneally administered at two different doses (0.1 and 0.5 mg/kg) prior to cotreatment with SKF (10 mg/kg, s.c.) and 0.015 mg/kg (s.c.) dizocilpine. Startle amplitudes were evaluated with three-way ANOVA using the design of the previous experiments. ANOVA showed haloperidol able to significantly reduce startle magnitudes in comparison with saline (F(2,74)=17.98; P<0.001, ANOVA) and post hoc analysis assessed a significant effect both for the dose of 0.1 mg/kg (P<0.05) and for the dose of 0.5 mg/kg (P<0.001) of haloperidol (see Table 4). A significant effect for blocks was also found (F(1,74)=10.53; P<0.01, ANOVA). No effects were observed either for treatment (F(1,74)=0.08; NS, ANOVA) or for interaction pretreatment × treatment (F(2,74)=0.00; NS, ANOVA).

Table 4 Mean Startle Amplitudes of Treatment Groups in the Fourth Experiment
Full size table

Three-way ANOVA also provided analysis of PPI values. As expected, haloperidol was not able to alter PPI (F(2,74)=0.48; NS) at any given dose, while a synergistic effect between SKF and dizocilpine was confirmed (F(1,74)=48.12; P<0.0001, ANOVA) (see Figure 5); besides, ANOVA showed a significant effect between prepulse levels (F(2,148)=138.77; P<0.0001). As indicated in Figure 4, haloperidol interestingly appeared unable to reverse disruption patterns mediated by the SKF–dizocilpine cotreatment (F(2,74)=0.09; NS, ANOVA).

Figure 5
figure 5

Effects of haloperidol pretreatment vs quinpirole. All haloperidol and quinpirole doses are given in mg/kg. Values represent mean±SEM for each treatment. Prepulses are indicated by the intensity corresponding to decibels above background noise. For all groups n=12. Sal, saline; hal, haloperidol; quin, quinpirole; **P<0.01 in comparison with sal+sal group; °P<0.05, °°P<0.01 in comparison with sal+quin 0.3 group; +P<0.05, ++P<0.01 in comparison with sal+quin 0.6 group. For further details, see text.

Full size image
Figure 4
figure 4

Representation of the effects of haloperidol on the synergy SKF+diz. Prepretreatments and pretreatments+treatments are indicated over and below the braces, respectively. All haloperidol doses are given in mg/kg. Values represent mean±SEM for each treatment. N=16 for hal 0.1+SKF+diz and hal 0.5+SKF+diz groups; for all remaining groups n=12. Prepulses are indicated by the intensity corresponding to decibels above background noise. Sal, saline; diz, dizocilpine; hal, haloperidol; SKF, SKF 23390. ***P<0.001 compared to sal+sal treatment groups. For further details see text.

Full size image

Effects of Haloperidol Pretreatment vs Quinpirole

In the fifth experiment haloperidol was intraperitoneally administered at two different doses (0.1 and 0.5 mg/kg) prior to treatment with quinpirole (0.3, 0.6 mg/kg, s.c.). Startle amplitudes were evaluated with three-way ANOVA using the design of the previous experiment. ANOVA showed a significant effect for pretreatment (F(2,99)=35.24; P<0.0001, ANOVA) and post hoc analysis highlighted a significant reduction of startle magnitudes induced both by the dose of 0.1 mg/kg (P<0.001) and by the dose of 0.5 mg/kg (P<0.001) of haloperidol (see Table 5). A significant effect for blocks was also found (F(1,99)=5.20; P<0.05, ANOVA). Instead, no effect were observed for the interaction pretreatment × treatment (F(4,99)=0.49; NS). Three-way ANOVA also provided analysis of PPI values. A significant pretreatment × treatment was revealed (F(4,99)=2.92; P<0.05) and post hoc analysis showed that while haloperidol is unable to alter PPI per se at any given dose, it totally prevents PPI disruption induced by both doses of quinpirole (see Figure 5). Moreover, ANOVA also showed a significant effect between prepulse levels (F(2,198)=210.27; P<0.0001).

Table 5 Mean Startle Amplitudes of Treatment Groups in the Fifth Experiment
Full size table

Effects of Clozapine Pretreatment vs SKF and Dizocilpine

In this experiment was tested the ability of clozapine, administered at the doses of 5 mg/kg (i.p.) and 10 mg/kg (i.p.), to reverse PPI disruption induced by cotreatment with SKF (10 mg/kg, s.c.) and dizocilpine (0.015 mg/kg, s.c.). Startle magnitudes were studied in a three-way ANOVA, which showed clozapine capable, at the dose of 10 mg/kg, of significantly reducing startle amplitudes ((F(2,74)=36.01; P<0.001, ANOVA); P<0.001, Tukey) (Table 6), while no effect was observed for treatment (F(1,74)=0.12; NS) or for interaction between pretreatment and treatment (F(2,74)=0.32; NS) (Table 6). Moreover, ANOVA revealed a block effect (F(1,74)=10.96; P<0.01). PPI values analysis revealed not only that clozapine was unable to alter PPI per se at any given dose (F(2,74)=0.97; NS] but was also ineffective in reversing PPI disruption induced by the cotreatment with the D1 agonist and the NMDA antagonist (F(2,74)=0.26; NS, ANOVA) (Figure 6). Instead, statistical analysis confirmed a significant effect for the treatment (F(1,74)=63.73; P<0.0001, ANOVA) and for prepulse levels (F(2,148)=376.89; P<0.0001) (Figure 6).

Table 6 Mean Startle Amplitudes of Treatment Groups in the Sixth Experiment
Full size table
Figure 6
figure 6

Representation of the effects of clozapine on the synergy SKF+diz. Pre-pretreatments and pretreatments+treatments are indicated over and below the braces, respectively. All clozapine and dizocilpine doses are given in mg/kg. Values represent mean±SEM for each treatment. N=16 for clo 5+SKF+diz and clol 10+SKF+diz groups; for all remaining groups n=12. Prepulses are indicated by the intensity corresponding to decibels above background noise. Sal, saline; clo, clozapine; SKF, SKF 38393; diz, dizocilpine. ***P<0.001 compared to sal+sal treatment groups. For further details see text.

Full size image

Effects of Clozapine Pretreatment vs Dizocilpine

The last experiment tested the ability of clozapine, administered at the doses of 5 mg/kg (i.p.) and 10 mg/kg (i.p.), to reverse PPI disruption induced by dizocilpine (0.15 mg/kg, s.c.). A three-way ANOVA was performed to study startle magnitudes; as in the previous experiment, clozapine is able to alter startle amplitudes (F(2,66)=36.22; P<0.001) and post hoc analysis confirmed a significant reduction of startle values both for the dose of 5 mg/kg (P<0.05) or for the dose of 10 mg/kg (P<0.001) of clozapine (Table 7). No effect was observed for treatment (F(1,66)=3.60; NS) or for interaction between pretreatment and treatment (F(2,66)=0.04; NS). Moreover, ANOVA revealed a block effect (F(1,66)=10.04; P<0.001]. PPI values analysis showed a significant interaction pretreatment × treatment (F(2,66)=8.10; P<0.001] and post hoc analysis, revealed that both doses of clozapine are able to partially prevent PPI-disruption induced by dizocilpine (see Figure 7). Besides, as expected, post hoc analysis also showed that both doses of clozapine are intrinsically unable to alter PPI. ANOVA also pointed out a significant difference between PPI levels (F(2,132)=172.99; P<0.0001).

Table 7 Mean Startle Amplitudes of Treatment Groups in the Last Experiment
Full size table
Figure 7
figure 7

Effects of clozapine pretreatment vs dizocilpine. All clozapine doses are given in mg/kg. Values represent mean±SEM for each treatment. Prepulses are indicated by the intensity corresponding to decibels above background noise. For all groups n=12. Sal, saline; clo, clozapine; diz, dizocilpine. ***P<0.001 compared to respective control groups; °°P<0.01, °°°P<0.001 compared to sal+diz group. For further details see text.

Full size image

DISCUSSION

Consistent with previous observations (Peng et al, 1990; Wan et al, 1996), the present study revealed that both SKF 38393 and SCH 23390 are unable per se to alter either PPI or startle magnitude at any prepulse intensity, suggesting that neither the activation nor the blockade of D1 receptors modifies sensorimotor gating and startle reflex. However, SKF 38393 was able to magnify dramatically the disruptive effect of dizocilpine on sensorimotor gating. Indeed, while the minimal threshold dose of dizocilpine to trigger a disruptive effect on PPI in the presence of SKF 38393 was 0.005 mg/kg, the NMDA receptor antagonist alone exhibited a comparable effect on the same parameter when administered in a dosage 10 times higher, in accordance with previous studies (Hoffman et al, 1993). Although the present design lacks an isobolographic analysis, and is therefore incapable of detecting pharmacological synergism sensu stricto in its most conclusive form (Berenbaum, 1981, 1989), our results provide clear evidence that dizocilpine powerfully potentiates the action of the D1 agonist. While this effect was observed at doses of dizocilpine between 0.005 and 0.05 mg/kg, no significant difference was detected at higher administrations of dizocilpine, plausibly because of a ‘floor effect’, as mentioned in the result section. Interestingly, the effect of the combination was prevented by pretreatment with the D1 antagonist SCH 23390, at both given doses, indicating that the observed effect is mediated by D1 receptors.

Taken together, these data provide new and intriguing evidence on the interactions between NMDA and D1 receptors on sensorimotor gating, and align with other evidence with regard to the ability of D1 agonists to enhance behavioral effects of NMDA receptor antagonists, such as locomotor activity (Goodwin et al, 1992; Svensson et al, 1992; Martin et al, 1994) and stereotypic responses (Verma and Kulkarni, 1992). Our finding is also consistent with the observation of Dall'Olio et al (2000) that the administration of dizocilpine prolonged SKF 38393-induced grooming in rats. Interactions between dopaminergic system and NMDA receptors appear highly complex and variable, according to the different brain regions, the type and topography of receptors, and the role of other neurotransmitters (Konradi et al, 2002). The body of evidence, however, indicates that one of the main areas where convergence between glutamatergic and dopaminergic fibers plays a critical role in sensorimotor gating is the core of the nucleus accumbens (NAcc) (Wan and Swerdlow, 1996). In this respect, it is worth noting that, while dizocilpine is unable to suppress PPI if infused into the NAcc, it can do so if injected into the areas that project the main glutamate inputs to this brain region, namely the lateral amygdala, the prefrontal cortex, and the dorsal hippocampus (Phillipson and Griffiths, 1985; Bakshi and Geyer, 1998). Evidence of NMDA and dopamine receptor interactions in NAcc is supported by several experimental and anatomical data (Sesack and Pickel, 1990, 1992; Youngren et al, 1993; Taber et al, 1996). On the other hand, the direction of the dopaminergic modulation of NAcc medium spiny neuron firing is known to depend on the glutamatergic tone by descending glutamate projections (Gonon and Sundstrom, 1996). Several biochemical studies indicate that low doses of NMDA receptor antagonists increase the release of glutamate (Liu and Moghaddam, 1995; Moghaddam et al, 1997), in all probability by preventing glutamate from driving GABAergic inhibitory neurons (Farber, 2003). It has been suggested that this increase may activate glutamatergic neurotransmission at non-NMDA receptors, like AMPA/kainate receptors (Hauber and Andersen, 1993; Olney and Farber, 1995; Bubser et al, 1995; Moghaddam et al, 1997; Deutsch et al, 2002). Following this evidence, a possible interpretation of our data may be based on a synergic activity of D1 receptors on the outcomes of a general activation of NAcc mediated by non-NMDA receptors. This interpretation is in line with findings the other groups, who demonstrated that simultaneous activation of D1 receptors and blockage of glutamate reuptake or activation of AMPA receptors in the NAcc enhances locomotor activity (Pierce et al, 1996; Kim et al, 2001). Alternatively, other areas, where dopamine is known to play a key modulatory role on glutamatergic fibers, such as prefrontal cortex, may be responsible for the effects observed in the present study. Our results, while not allowing to substantiate clearly the mechanism of action of the interactions between D1 and NMDA receptors in the circuit subserving sensorimotor gating, stimulate further biochemical and functional investigations in this direction, to better delineate the function of D1 receptors in the NMDA hypofunctional state.

In the third experiment, while 0.05 mg/kg SCH 23390 proved unable to prevent dizocilpine-mediated PPI disruption, the same compound significantly blunted the same effect at the dose of 0.1 mg/kg. In apparent contrast with these last results, Wedzony et al (1994) found that 0.1 mg/kg SCH 23390 failed to antagonize the PPI disruption mediated by dizocilpine. Nonetheless, it should be noted that the doses of NMDA receptor antagonist used in that study were higher than in our experiment. Furthermore, since SCH 23390 is known to antagonize serotonin receptors (Hicks et al, 1984; Briggs et al, 1991; Bischoff et al, 1986; Millan et al, 2001) and dizocilpine is further known to potentiate serotonergic functions (Dall'Olio et al, 1999), we assume that the observed intrinsic effect of SCH 23390 might depend on the blockade of serotonergic system. In fact, the role of the latter in the modulation of dizocilpine-mediated PPI abolition has been investigated by several researchers, providing controversial results. While Varty and Higgins (1995) reported that both metergoline, a 5HT1/2 antagonist, and ketanserin, a 5HT2 antagonist, are able to attenuate the effects of dizocilpine in PPI, Zhang et al (1997) found no actions of serotonergic antagonists on the same paradigm. The evaluation of this issue was beyond the aims of the present study, but the observation of this phenomenon fosters further investigation on the role of serotonin on the synergy of D1 agonists and dizocilpine. In view of this, it is of interest that Dall'Olio et al (2000) demonstrated that the blockade of serotonergic system hampers the dizocilpine-induced potentiation of responses elicited by the stimulation of D1 receptors.

No interaction between dizocilpine and D2 receptor activation was found, in accordance with previous observations in PPI (Keith et al, 1991) and with other behavioral studies (Ferré et al, 1994; Dall'Olio et al, 1996), where D2 agonists elicited either no potentiation or even counteracted some of the effects of NMDA receptor antagonists. Furthermore, haloperidol was shown unable to reverse the PPI disruption induced by the cotreatment with SKF 38393 and dizocilpine at the same doses, which were effective in preventing quinpirole-mediated PPI attenuation. Although the lack of effect of either quinpirole or haloperidol in our study might be attributable to the choice of the doses or of the route of administration, our results likely indicate that D2 receptors are not involved in the mechanism of potentiation between D1 receptors and the blockade of NMDA receptors.

Remarkably, dizocilpine did not significantly alter startle amplitude at any given dose either per se or in synergy with the D1 agonist, although in both cases it increased startle amplitude when administered at 0.05 and 0.15 mg/kg. The evidence with regard to the effects of dizocilpine on ASR appears contradictory in the literature: in fact, while Mansbach and Geyer (1989) showed moderate (0.01–0.3 mg/kg), but not high doses (0.5–1 mg/kg) caused an overall increase in ASR amplitude, a study by Hoffman et al (1993), performed with a different protocol, showed dizocilpine (0.1 mg/kg) failed to alter the baseline startle amplitude; finally, Al-Amin and Schwarzkopf (1996) suggested a biphasic effect of the same compound on this parameter in Fischer rats, finding the dose of 0.05 mg/kg augmented the baseline startle, while higher doses reduced it. Taken together, the variations between these results show that the role of NMDA receptors in the neurobiological circuitry accounting for startle responses might be highly complex and possibly related to the strain of the animals, as well as other paradigm variables, like the loudness level of pulses and the intertrial time.

Rats treated with quinpirole showed an augmentation of startle reflex, although not fully significant, while haloperidol and clozapine induced a significant reduction of startle reflex. In parallel, acute administrations of dizocilpine did not significantly alter startle amplitude at any given dose or per se in synergy with D1 and D2 agonists, although it produced an increase of startle amplitude when administered at 0.05 and 0.15 mg/kg. Our data on dopaminergic agents are consistent with other studies (Peng et al, 1990; Davis 1980; Davis and Aghajanian, 1976) and suggests involvement of D2 receptors in the neural circuitry accounting for ASR magnitude.

Although our protocol is not designed to assess habituation specifically, significant differences between the two blocks of the stimuli sequence were found throughout the study, in agreement with previous reports about the inability of dopaminergic agonists (Davis and Aghajanian, 1976; Davis, 1980) to impair habituation processes. The fourth and sixth experiments showed PPI disruption caused by the association of dizocilpine and SKF 38393 cannot be attenuated either by haloperidol or by clozapine, at any given dose. In fact, although both antipsychotics are known to antagonize D1 receptors, the given dose of D1 agonist was arguably too high for any efficient competition to the D1 site to occur. On the other hand, it is to stress the point that, provided the observed effect is the result of a magnification of the effect of dizocilpine, PPI disruption induced by the latter is not reversed by haloperidol, as mentioned in the introduction. The ability of clozapine to reverse PPI disruption provoked by dizocilpine is highly questioned in literature, since different studies have obtained contrasting results (Bakshi et al, 1994; Hoffman et al, 1993). Interestingly, in the present study we have shown that clozapine is able to attenuate, but not fully antagonize, dizocilpine-mediated PPI disruption in our experimental setting. In this respect, the abolition of sensorimotor gating mediated by the synergy of SKF 38393 and dizocilpine would appear to differ from the one induced by the NMDA receptor antagonist alone, indicating that the contemporary activation of D1 agonist and blockade of NMDA receptor either potentiates chiefly the disruptive effects of dizocilpine, which are not under control of clozapine or involves also different, hitherto unknown mechanisms of action.

As mentioned in the introduction, D1 receptors are synergic with D2 receptors, even if the neurobiological links of this relationship remain unclear. Our study demonstrates D1 receptors also serve the same function for PPI disruption mediated by dizocilpine. Further studies are required to understand whether this observation can be extended to other NMDA receptor antagonists, such as phencyclidine and ketamine. Whatever the mechanism, we assume that D1 receptors play a key role as enhancers of the action of the main receptors involved in PPI disruption. Since the blockade of NMDA receptors is regarded as a putative model of negative symptoms in schizophrenia, further research is required to assess the clinical value of our observation in the management and treatment of psychotic patients.