Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The psychomotor-stimulating and reinforcing properties of cocaine are believed to depend on the blockade of DA reuptake in brain areas innervated by DA neurons (Koob and Bloom, 1988; Wise, 1996).

In the rat striatum, the cocaine-mediated increase in extracellular DA and the consequent stimulation of postsynaptic DA receptors are primarily associated with the induction of stereotyped behavior (LeMoal and Simon, 1991), whereas in the ventral striatum (nucleus accumbens) the cocaine effect on DA is related primarily to the onset of both the increase of locomotor activity and the reinforcing properties (Maldonado et al., 1993; Ikemoto et al., 1997; Koob and Bloom, 1988; LeMoal and Simon, 1991).

These effects, which are due to the cocaine-induced blockade of the DA transporter (Kuhar et al., 1991), are highly dependent on the rate and amount of DA released from intraneuronal vesicular stores via impulse flow-dependent exocytosis (Florin et al., 1995; White, 1996). Thus, it is possible that the endogenous negative modulation of the firing rate of DAergic neurons may attenuate the rewarding and the reinforcing actions of cocaine. Although in vitro cocaine applied to SNc or VTA DAergic neurons inhibits their firing rate, systemic administration of behaviorally active doses of cocaine produce only a relatively weak inhibitory effect on VTA DA neurons, compared with the stronger enhancement of DA neurotransmission within the nucleus accumbens; the latter effect predominates, and this explains the extremely potent rewarding effects of cocaine (Einhorn et al., 1988). However, the firing rate of DAergic SNc-striatal and VTA-accumbens neurons is under potent trans-synaptic inhibitory control by GABA-releasing neurons that impinge on DAergic neuronal somata (Kalivas, 1993; Wise, 1996).

Interestingly, the BZD-mediated amplification of GABA action (Choi et al., 1981) at GABA_A receptors (Vicini et al., 1987) by FAMs BZDs (Giusti et al., 1993; Costa and Guidotti, 1996) also attenuates cocaine-induced psychomotor stimulant activity in humans and rats, reduces the cocaine-induced increase in DA content in striatal dialysates (Dewey et al., 1997) and reduces cocaine self-administration in rats (Goeders et al., 1993). These data suggest that FAMs be tested in the treatment of craving and other symptoms of cocaine abuse. However, protracted use of this class of drugs is considerably limited by the following untoward effects: 1) in rats they reduce cocaine-induced psychomotor activation and cocaine self-administration in doses that are close to those eliciting sedation and ataxia, and 2) their repeated administration produces tolerance, which eliminates their beneficial effects.

We hypothesized that to reduce cocaine-induced psychomotor activation and cocaine-elicited DA increase in striatal and NAS microdialysates, one could use imidazenil, a potent anxiolytic imidazo-benzodiazepine endowed with partial positive allosteric modulatory activity of GABA action at many GABA_A receptor subtypes, which is virtually devoid of tolerance and dependence liability (Auta et al., 1994; Ghiani et al., 1994; Impagnatiello et al., 1996; Costa and Guidotti, 1996).

The aim of the present study was to compare imidazenil (a PAM) and diazepam (a FAM) in terms of their potency to 1) inhibit the cocaine-induced increase in DA content in the NAS and in striatal dialysates, 2) inhibit the cocaine-induced increase in locomotor activity, 3) potentiate the GABA-elicited inhibition of VTA DAergic neurons firing rates in vitro and 4) elicit tolerance liability to the antagonism of cocaine.

We report here that single doses of either imidazenil or diazepam attenuate the cocaine-induced increase in locomotor activity and in DA content in the dialysate of the NAS and striatum of freely moving rats and that they potentiate GABA-induced inhibition of the firing rate of VTA DA neurons in vitro. Moreover, whereas tolerance to these effects followed protracted treatment with diazepam, protracted treatment with imidazenil (see "Materials and Methods") failed to produce tolerance.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animal housing and surgery. Male Fischer 344 rats (Harlan) were housed individually with a 12/12-hr light/dark cycle; food and water were available ad libitum. Cocaine injections (Sigma Chemical Co., St. Louis MO), behavioral testing and dialysis experiments were performed between 9 A.M. and noon.

Microdialysis experiments. Twenty-four hours after probe implantation, rats were connected to a microperfusion pump (Harvard apparatus) with a microdialysis syringe (CMA model, 1-ml vol), and the striatum or the NAS was perfused using artificial cerebrospinal fluid (aCSF) composed of NaCl 145 mM, KCl 2.7 mM, MgCl₂ 1 mM, CaCl₂ 1.2 mM, and Na₂HPO₄ 2 mM, pH 7.4.

Measurement of DA levels in microdialysis samples. The perfusate was assayed for DA content using an HPLC coupled to an ESA coulochem II 5200A detection system with oxidation potential +320 mV at range of 50 nA. The detector was equipped with a high-performance analytical cell (ESA model 5014) tailored for use in the analysis of dialysates. The mobile phase was composed of 0.05 M monobasic sodium phosphate, 0.1 N sodium acetate and 1% methanol and was adjusted to pH = 4.4 with HPLC grade phosphoric acid. The flow rate of the system was 1.0 ml/min. The limit of detection for DA was approximately 3 fmol.

Histological analysis. At the completion of microdialysis experiments, the animals were given an overdose of pentobarbital and perfused through the heart with phosphate-buffered saline followed by 10% formalin/isotonic saline. Brain coronal slices (100 µm) were stained with cresyl violet, and probe placement was determined by light microscopy in the dorsal striatum (not shown) and in the NAS. As an example, in figure 1, we have diagrammatically represented all probe placements in NAS relative to the experiment described in figure 2. These placements were similar in all the reported experiments. Animals in which the probe was located outside the target area were discarded.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Location of the tips of the microdialysis probes in the NAS. Data relative to the experiment described in figure 2. The numbers indicate millimeters rostral to the bregma according to the Atlas of Paxinos and Watson (1986). The microdialysis probes in the shell were localized () to the dorso-medial portion but not to the ventral portion. The figure also indicates the incorrect placements () relative to the animals that were discarded.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   A) Imidazenil (Imid) attenuates the cocaine-induced DA increase in the NAS in a dose-related manner. This inhibition by imidazenil of the cocaine-elicited increase in DA in NAS dialysates is significant at doses of 0.56 (-), 1.2 (-) and 2.5 µmol/kg i.p. (-). Dunnett's post-hoc test: P = .02, .005 and .001 respectively. This test evaluates each dose vs. the vehicle (Veh)-treated group (-); no statistically significant difference was observed for the dose of 0.31 µmol/kg i.p. (-). Each point represents the mean ± S.E.M. of 3 to 5 animals. B) Diazepam (Diaz) (4.4 µmol/kg ip) (-) significantly attenuates cocaine-induced increase in DA content in NAS dialysates when compared with the vehicle (Veh)-treated group (-). t test; P = .02. Statistical analyses were performed by comparing the different AUC. In the experiments of figure 2A and B, basal levels of DA averaged 55 ± 14 fmol/sample.

Locomotor activity. Spontaneous and cocaine-induced locomotor activity in rats was measured using the Photobeam Activity System (San Diego Instruments, Inc.) linked to an IBM 386 computer. A clear Perspex cage (42 cm × 23 cm × 17 cm) containing a small amount of sawdust was surrounded by a metal frame with seven infrared beams and sensors placed 2 in. apart. Total activity (total number of beam breaks) was selected as the measure of locomotion. Visual observation by the operator confirmed that locomotion and not stereotypy was the cause of the behavior. Before the test period, the animals were kept three to a cage in their home cages. Rats were allowed a 15-min habituation period in the activity cages and then were treated intraperitoneally i.p. with vehicle, diazepam or imidazenil. After a 40-min recording period, some rats received an i.p. injection of cocaine, and the locomotor activity was measured for another 40-min period. Locomotor activity was measured between 9 A.M. and noon.

Schedule for long-term diazepam and imidazenil treatment. Diazepam and imidazenil (Hoffmann-La Roche, Nutley, NJ) were suspended in water containing 0.05% Tween 20 (Sigma Chemical Co., St. Louis, MO) and administered in 1-ml volume by oral gavage three times a day (approximately at 9.00 A.M., 2.00 P.M., and 7.00 P.M.) for 14 days at increasing doses. Diazepam was administered in the following regimen: 17.6 µmol/kg, days 1 to 3; 35.2 µmol/kg, days 4 to 6; 52.8 µmol/kg, days 7 to 10; 70.4 µmol/kg, days 11 to 14. This diazepam dose regimen was shown to elicit anticonvulsant tolerance (Auta et al., 1994).

Extracellular recording of VTA neuron firing rate in brain slices. Brain slices from rats (Fischer 344, 100-200 g) containing the VTA were prepared as previously described (Brodie et al., 1990, 1995; Mueller and Brodie, 1989). Rats received long-term diazepam or imidazenil treatment with the schedule described above.

Data analysis. In the microdialysis experiments, DA recovery averaged 9% to 12% through the microdialysis probe. These data agree with those reported by others (Benveniste and Huttemeier, 1990). Measured levels were expressed as a percentage of the basal release (an average of 3-5 samples collected before drug administration).

Drugs. Imidazenil and diazepam (Hoffmann-LaRoche, Nutley, NJ) solutions were freshly prepared before each microdialysis, locomotor and electrophysiological activity experiment. The drugs were dissolved in a medium of propylene glycol 50% (Fisher Sci, Itasca, IL), polyethylene glycol 11% (Sigma Chemical Co., St. Louis, MO), H₂O 37% and dimethyl sulfoxide, 2% (Sigma). Cocaine hydrochloride (Sigma) was dissolved in isotonic saline for systemic injection. For the acute studies, imidazenil, diazepam and cocaine were always administered i.p.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Both imidazenil and diazepam attenuate the cocaine-induced increase in the DA content of NAS dialysates. The i.p. administration of 45 µmol/kg of cocaine resulted in a 3-fold increase in DA concentration in the dialysates from the NAS. This effect peaked at about 20 min and then began to decline, reaching basal levels in about 60 min (fig. 2). Imidazenil attenuated the cocaine-induced DA increase in a dose-dependent manner (f = 30.1; dF = 1, 13; P > .001; fig. 2A) (F test for the linear portion of the dose-response curve). More specifically, the overall increase in the dialysate DA content produced by cocaine in vehicle-treated rats was decreased by approximately 14%, 33%, 42% and 56% (values were calculated by considering the various AUC) by 0.31, 0.56, 1.2 and 2.5 µmol/kg of imidazenil, respectively (fig. 2A). Imidazenil's inhibition of the cocaine-elicited DA increase in NAS dialysates is significant at doses of 0.56, 1.2 and 2.5 µmol/kg (Dunnett's post-hoc test; P = .02, .005 and .001, respectively). Like to imidazenil, diazepam administered i.p. 40 min before cocaine injection (45 µmol/kg) attenuated the cocaine-induced DA increase in the dialysates of the NAS (fig. 2B). More specifically, a single dose of diazepam (4.4 µmol/kg) reduced the cocaine-induced DA increase in NAS dialysates by about 41% (t = 2.8; dF = 12; P = .02; fig. 2B).

Both imidazenil and diazepam attenuate the cocaine-induced increase in the DA content of striatal dialysates. The i.p. administration of 45 µmol/kg of cocaine resulted in a 2-fold increase in striatal dialysate DA content. This effect reached a peak between 20 and 40 min and then declined toward basal level in about 60 min (fig. 3). Imidazenil administered i.p. to rats 40 min before cocaine injection in a dose (2.5 µmol/kg) that itself failed to alter the DA striatal dialysate content (fig. 3A) reduced the AUC of the cocaine-induced DA increase in the striatal dialysates by almost 50% (t = 3.2; dF = 12; P < .01; fig. 3A). Like imidazenil, a single dose of diazepam (4.4 µmol/kg) administered i.p. to rats 40 min before cocaine injection reduced the cocaine-induced DA increase in the dialysate by almost 50%. The difference in AUC was significant (t = 2.8, dF = 12; P < .02 fig. 3B).

View larger version (13K): [in this window] [in a new window]   Fig. 3.   A) Imidazenil (2.5 µmol/kg i.p.) -) significantly attenuates the cocaine-induced increase in DA content in striatal dialysates when compared with the vehicle-treated group (-) (t test: P < .01). B) Diazepam (4.4 µmol/kg i.p.) (-) significantly attenuates the cocaine-induced increase in DA content in striatal dialysates when compared with the vehicle-treated group (-) (t test, P < .02). Statistical analyses were performed by comparing the different AUC. Each point represents the mean ± S.E.M. of 3 to 5 animals. Basal levels of DA averaged 83 ± 21 fmol/sample.

Inhibition by imidazenil of the cocaine-induced increase in DA content fails to show tolerance. When rats receiving long-term treatment with increasing doses of diazepam or pharmacologically equieffective doses of imidazenil or vehicle (see "Materials and Methods" for drug schedule) were left drug-free for 18 hr after the last injection, they failed to show differences in basal or cocaine-induced increase in the DA content of NAS or striatal dialysates (data not shown). This result confirms previously reported findings that diazepam and imidazenil were completely eliminated from the rat brain 18 hr after the last injection (Impagnatiello et al., 1996). At this time, a challenge with imidazenil (2.5 µmol/kg) was fully effective in attenuating the cocaine-induced increase in DA content observed both in the NAS (fig. 4A) and in the striatal dialysates (fig. 5A) of freely moving rats receiving long-term treatment with either vehicle or with imidazenil. In contrast, a similar challenge with diazepam (4.4 µmol/kg) failed to be effective in reducing the cocaine-induced increase in DA content in either NAS (fig. 4B) or striatal dialysates (fig. 5B).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   A) Effect of imidazenil challenge on the cocaine-induced increase in DA content of NAS dialysates in rats treated with long-term exposure to vehicle (Veh) or imidazenil (Imid) (see "Materials and Methods"). At 18 hr after the termination of the protracted treatment schedule, the group of animals treated long-term with vehicle was challenged with vehicle (-) or with imidazenil 2.5 µmol/kg i.p. (-). The group treated long-term with imidazenil was challenged with imidazenil 2.5 µmol/kg i.p. (-). Overall F test (f = 15.3; dF = 2,6; P = .004). (-) vs. - P = .003. - vs. - P = .01. Dunnett's post-hoc test for the two individual groups vs. vehicle.   B) Effect of diazepam challenge on the cocaine-induced increase in the DA content of NAS dialysates in rats treated with long-term exposure to vehicle (Veh) or diazepam (Diaz) (see "Materials and Methods"). One group of animals was treated long-term with vehicle and then challenged with vehicle (-) or with diazepam 4.4 µmol/kg i.p. (-). The challenge was carried out 18 hr after termination of the protracted treatment schedule. Another group was treated long-term with diazepam and then challenged with diazepam 4.4 µmol/kg i.p. (-). One time-point from the group of animals chronically treated with diazepam and challenged with diazepam (-) was missing for technical reasons, so the results were analyzed for AUC of three time-points (f = 4.9; dF = 2,5; P = .07). Helmert contrasts for rats chronically treated with diazepam and challenged with diazepam (-) was not significantly different from the control group (-) (t test = .8; P = .45), whereas the rats chronically treated with vehicle and challenged with diazepam (-) showed a significant diminution in cocaine-increased DA levels compared with the two other groups (t = 3.0; P = .03). ANOVA with Helmert contrasts. Statistical analyses were performed by comparing the different AUC. Each point represents the mean ± S.E.M. of 3 to 5 animals. Basal levels of DA averaged 50 ± 12 fmol/sample.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   A) Effect of imidazenil challenge on the cocaine-induced increase in DA content of striatal dialysates 18 hr after termination of the protracted treatment schedule with imidazenil or vehicle (see "Materials and Methods"). One group of animals was treated long-term with vehicle and then challenged with vehicle (-), or with imidazenil 2.5 µmol/kg i.p. (-). Another group was treated long-term with imidazenil and then challenged with imidazenil 2.5 µmol/kg i.p. (-). (-) vs. - P = .006. - vs. - P = .01 Dunnett's post-hoc test for the two individual groups vs. vehicle.   B) Effect of diazepam challenge on the cocaine-induced increase in DA content of striatal dialysates 18 hr after termination of the protracted treatment schedule with diazepam or vehicle (see "Materials and Methods"). One group of animals was treated long-term with vehicle and then challenged with vehicle (-), or with diazepam 4.4 µmol/kg i.p. (-). Another group was treated long-term with diazepam and then challenged with diazepam 4.4 µmol/kg i.p. (-) (f = 3.2; dF = 2,7; P = .10). Rats chronically treated with diazepam and then challenged with diazepam (-) were indistinguishable from rats chronically treated with vehicle and then challenged with vehicle (t = .2; P = .81, not significant), whereas rats chronically treated with vehicle and then challenged with diazepam (-) showed a significant decrease in DA levels (t = 2.5; P = .04) compared with the two other groups (-) and (-). ANOVA with Helmert contrasts. Statistical analyses were performed by comparing the different AUC. Each point represents the mean ± S.E.M. of 3 to 5 animals. Basal levels of DA averaged 101 ± 22 fmol/sample.

Attenuation by imidazenil of cocaine-induced locomotor activity fails to show tolerance. In separate groups of animals without implantation of the microdialysis probe, we also studied the effects of diazepam and imidazenil on spontaneous and cocaine-induced locomotor activity.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   A) Effect of imidazenil challenge on the cocaine-induced increase in locomotor activity 18 hr after the termination of the protracted treatment schedule with imidazenil or vehicle (see "Materials and Methods"). Imidazenil challenge (2.5 µmol/kg i.p.) failed to change spontaneous locomotor activity (t = 0.6; dF = 4; N.S.) but significantly attenuated the cocaine-induced increase in locomotor activity in both long-term vehicle-treated animals (t = 7.8; dF = 4; * P < .001; n = 3) and long-term imidazenil-treated animals (t = 7.4; dF = 6; ** P = .001; n = 5). Note that 18 hr after termination of protracted imidazenil treatment, neither the spontaneous nor the cocaine-stimulated locomotor activity differed from that of long-term vehicle-treated animals. B) Effect of diazepam challenge on the cocaine-induced increase in locomotor activity 18 hr after termination of the protracted treatment schedule with diazepam or vehicle (see "Materials and Methods"). Diazepam challenge (4.4 µmol/kg i.p.) significantly attenuated the spontaneous (t = 3.0; dF = 4;  P < .05; n = 3) and the cocaine-induced increase in locomotor activity in animals treated long-term with vehicle (t = 6.4, dF = 4; * P < .003; n = 3) but not in animals treated long-term with diazepam (t = 2.4; dF = 4; # P = .08; n = 3) (ANOVA followed by t test). Note that 18 hr after termination of protracted diazepam treatment, neither the spontaneous nor the cocaine-stimulated locomotor activity differed from that of long-term vehicle-treated animals. Results similar to those reported in this figure were obtained in two other experiments.

Potentiation by imidazenil of GABA inhibition of VTA DAergic neuron firing is devoid of tolerance. After long-term diazepam or imidazenil treatment (see "Materials and Methods"), the ability of either of these BDZ recognition site ligands to potentiate the GABA inhibition of VTA DAergic neuron firing was tested electrophysiologically. Extracellular single-unit recordings were obtained from DAergic VTA neurons in brain slices prepared from rats treated long-term with diazepam or imidazenil according to the protocol in "Materials and Methods." Figure 7 illustrates that in control rats, the reduction in firing rate elicited by 10 µM GABA is greatly potentiated by an addition of 5 µM diazepam. In contrast, as shown in the upper part of figure 8, after long-term treatment with diazepam, the GABA-elicited reduction in DA neuronal firing rate is only marginally increased in the VTA.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Diazepam potentiated GABA inhibition of VTA neurons in brain slices from control rats. Each vertical bar represents the average firing rate over a 5-sec interval; the horizontal bar represents the duration of application of GABA (10 µM). The top portion of the figure illustrates the effect of 10 µM GABA before the administration of diazepam; in this cell, 10 µM GABA produced a 11.5% decrease in firing rate. The bottom portion of the figure shows the effect of that same concentration of GABA on the same neuron during administration of 5 µM diazepam. In the presence of 5 µM diazepam, 10 µM GABA produced a 68% decrease in firing rate.

View larger version (25K): [in this window] [in a new window]   Fig. 8.   GABA inhibition is not potentiated by diazepam in long-term diazepam-treated rats. VTA neurons in slices from six rats were tested with various concentrations of GABA in the presence of vehicle and two concentrations of diazepam. The spontaneous firing rate of VTA neurons was decreased by GABA in a concentration-dependent manner. In the top portion of the figure are firing-rate histograms recorded from the same VTA neuron in vehicle (top left) or in 5 µM diazepam (top right). The horizontal bar represents the duration of application of GABA (10 µM). In vehicle, 10 µM GABA produced an 11.0% decrease in firing rate, and in diazepam, 10 µM GABA produced an 18% decrease in firing rate. The bottom portion of this figure illustrates the responses of the population of cells tested. To present the mean data concisely, we subtracted the response (percent inhibition) to GABA in the absence of diazepam from the GABA response of the same neuron in the presence of diazepam. The means of these differences were plotted as a function of GABA concentration for cells from rats treated chronically with diazepam. A value of zero represents no change in GABA potency produced by diazepam. No statistically significant effect of either 1 µM or 5 µM doses of diazepam on GABA potency was observed (two-way ANOVA; P > .05, n = 6).

View larger version (26K): [in this window] [in a new window]   Fig. 9.   GABA inhibition is potentiated by imidazenil in long-term imidazenil-treated rats. VTA neurons in slices from five rats were tested with various concentrations of GABA in the presence of vehicle and two concentrations of imidazenil. The spontaneous firing rate of VTA neurons was decreased by GABA in a concentration-dependent manner. In the top portion of the figure are firing-rate histograms recorded from the same VTA neuron in vehicle (top left) or in 5 µM imidazenil (top right). The horizontal bar represents the duration of application of GABA (10 µM). In vehicle, 10 µM GABA produced a 13% decrease in firing rate, and in imidazenil, 10 µM GABA produced a 63% decrease in firing rate. The bottom portion of this figure illustrates the responses of the population of cells tested. To present the mean data concisely, we subtracted the response (percent inhibition) to GABA in the absence of imidazenil from the GABA response of the same neuron in the presence of imidazenil. The means of these differences are plotted as a function of GABA concentration for cells from rats treated chronically with imidazenil. A value of zero represents no change in GABA potency produced by imidazenil. The potency of GABA was significantly increased by imidazenil (two-way ANOVA; P < .001, n = 5). Both concentrations of imidazenil (1 µM and 5 µM) significantly increased the potency of GABA (Student-Newman-Keuls; P < .05).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The following structural requirements define the susceptibility of GABA_A receptor subtypes to the allosterically elicited amplification of GABA action by anxiolytic BZD: 1) the presence of at least one among ₁, ₂, ₃ and ₅ GABA_A receptor subunits; 2) the presence of the S or L variants of ₂ subunits; 3) the presence of one or more  subunits and 4) the absence of the ₆ receptor subunit (Puia et al.1991; Dcic et al., 1995; MacDonald and Olsen, 1994; Sieghart, 1995). In the framework of these GABA_A receptor structural requirements, the chemical structure of anxiolytic BZDs confers particular profiles of their pharmacological action. In the introduction, we defined the different activity profiles of PAM and FAM anxiolytic BZD (Giusti et al., 1993; Costa and Guidotti, 1996). It is important to differentiate PAM BZD from another class of BZDs that maximize the amplification of GABA action in selective GABA_A receptor subtypes, including a specific  subunit. An example of this class of SAM is zolpidem, which by selectively amplifying GABA action in receptors, including the ₁ subunit, facilitates the onset of sleep. Such a differentiation is important, because unlike PAMs, the pharmacological action of SAMs includes tolerance and dependence liability when they are given in high doses. In contrast, PAMs, even when given in doses that are multiples of those that are maximally active, fail to elicit tolerance and dependence liability (Costa and Guidotti, 1996).

From the standpoint of the therapeutic use of BZDs, the lack of tolerance and dependence liability confers a unique characteristic on PAMs (Impagnatiello et al., 1996; Costa and Guidotti, 1996). Despite claims to the contrary, there are only a few known BZDs that, on the basis of experimentation in rats and monkeys, can be considered PAM BZDs devoid of tolerance and dependence liability (Costa and Guidotti, 1996). One of them is imidazenil, a PAM that fails to be self-administered in rats and monkeys and therefore is presumably devoid of dependence liability. In rats trained to recognize FAMs, imidazenil in doses 250 times greater than the pharmacologically active doses fails to generalize to classic BZDs with FAM activity (Paronis et al., 1997).

In the present experiments with freely moving rats, we found that the effect of a single imidazenil dose on the cocaine-elicited increase in DA content in NAS dialysate was to inhibit the action of cocaine in a dose-dependent manner (fig. 2A). Similarly, the cocaine-elicited increase in DA content in NAS dialysates can be inhibited by appropriate doses of diazepam (fig. 2B). Also a single dose of either drug can inhibit the increase in DA content elicited by cocaine in striatal dialysates (fig. 3, A and B).

A comparison between diazepam and imidazenil indicates that imidazenil (2.5 µmol/kg) is more potent than diazepam (4.4 µmol/kg) in inhibiting the cocaine-induced DA increase in NAS dialysates (fig. 2, A and B). In naive (figs. 2 and 3) or long-term vehicle-treated rats (figs. 4 and 6), a single dose of diazepam that reduces the cocaine-elicited increase in locomotor activity and the cocaine-elicited increase in the NAS and in striatal DA microdialysates, also effectively lowers spontaneous locomotor activity. Thus it has been difficult to determine whether the anticocaine effect of diazepam or of other FAMs (see figs. 2, 3 and 6 and Goeders et al., 1993) is specific or simply reflects the consequences of reduced spontaneous motor activity. However, we show that imidazenil, a drug that also acts at GABA_A receptors but fails to produce sedative and ataxic effects (Giusti et al., 1993), reduces the cocaine-elicited increase in extracellular DA in NAS and the striatum and in locomotor activity. Thus the results with imidazenil indicate that potentiation of GABA_A receptor function with PAMs reduces cocaine-induced DA increases in NAS and striatal microdialysates and the cocaine-induced psychomotor activation independently of the sedative action that is absent in this class of drugs (Costa and Guidotti, 1996).

We then determined whether the protracted administration of diazepam or imidazenil for 14 days at increasing doses (as reported in "Materials and Methods") remained active when a challenge with imidazenil or diazepam was given 18 hr after the last injection of protracted treatment with each of two BZDs (Auta et al., 1994; Impagnatiello et al., 1996, see "Results"). The data in figures 4B and 5B show that a challenge of 4.4 µmol/kg of diazepam loses its ability to reduce the cocaine-elicited DA increase in NAS and striatal dialysates. In contrast, imidazenil maintains its ability to inhibit the cocaine-elicited DA increase in NAS (fig. 4A) and striatal (fig. 5A) dialysates. Thus tolerance eventually undermines the ability of diazepam to inhibit the cocaine-induced increase in DA in brain dialysates and locomotor activity, whereas tolerance to imidazenil does not develop. Moreover, we demonstrate that direct application of either imidazenil or diazepam to VTA brain slices potentiated GABA-elicited inhibition of VTA-DA cell firing. When the rats used for such VTA preparations received equipotent treatment schedules (see "Materials and Methods") with either imidazenil or diazepam for 14 days, we found that DAergic VTA neurons from rats receiving diazepam showed tolerance to the enhancement of GABA-elicited inhibition of VTA neuronal firing (fig. 8), whereas rats receiving imidazenil did not show tolerance (fig. 9). Thus, after long-term treatment with increasing doses (see "Materials and Methods"), imidazenil fails to show tolerance liability to the potentiation of GABA-elicited inhibition of DAergic neuronal activity in the VTA. Thus the action of imidazenil on cocaine-induced DA increases in NAS and striatal microdialysates and on locomotor activity may support the conclusion that imidazenil might be useful in the clinical management of the psychomotor stimulant effects and reinforcing properties of cocaine, because tolerance does not develop to its action to prevent the effects of cocaine.

However, one could suggest that the reduced increase in the DA content in the dialysates might have a greater postsynaptic effect in stimulating motor activity in the presence of imidazenil. To test this, we determined whether the motor stimulation elicited by cocaine in rats receiving protracted treatment with imidazenil or diazepam is still inhibited when challenged 18 hr after the last injection of protracted treatment with the BZDs with the standard dose of imidazenil or diazepam. The data reported in figure 6, A and B, show that the ability of the challenge dose of diazepam to reduce the increase in motor activity elicited by cocaine is subject to the development of tolerance; in contrast, the efficacy of a challenge dose of imidazenil is virtually intact in rats receiving protracted treatment with either vehicle or imidazenil.

Thus, because imidazenil continues to inhibit efficaciously the increased motor activity elicited by cocaine, the hypothesis that imidazenil may increase the efficacy of the smaller amounts of DA that are increased in the NAS and striatum dialysates after cocaine treatment does not appear to be tenable.

In conclusion, the present experiments suggest that treatment with imidazenil might be tested in the self-administration of cocaine in rats and monkeys to evaluate its ability to reduce cocaine craving in patients dependent on cocaine. We have initiated these studies on the action of imidazenil in the self-administration of cocaine in rats. So far, imidazenil appears to reduce cocaine self-administration in rats without evidence of tolerance liability.