Introduction

Top Abstract Introduction Methods Results Discussion References

Dopaminergic transmission in the central nervous system plays a role in the control of several important behavioral functions and in sensorimotor integration. Disruption of normal neural processing in dopaminergic areas is associated with Parkinson's disease, Huntington's disease, tardive dyskinesia, schizophrenia and psychomotor stimulant abuse (Scheel-Krüger, 1986; Casey, 1995; Johanson and Schuster, 1995; Kahn and Davis, 1995).

Systemic administration of relatively high doses of indirect- and direct-acting dopamine agonists can produce highly repetitive movements in laboratory rats, mice and cavies. These behavioral stereotypies can involve repetitive quick rears, continuous sniffing accompanied by head bobbing, prolix licking and especially intense biting and gnawing (see reviews by Robbins and Sahakian 1981; Randrup et al., 1988; Cooper and Dourish, 1990). Central dopaminergic neurotransmission is under the regulatory control of GABAergic processing. Stimulation of GABA_A receptors modulates dopaminergic cell firing in several brain areas and is associated with altered dopaminergic effects in vivo (Walters and Pucak, 1996). For example, dopamine-driven orofacial behaviors (e.g., gnawing) can be robustly potentiated by indirect GABA_A agonists (e.g., amino-acetic acid, -vinyl-GABA), postsynaptic GABA_A agonists (e.g., isoguvacine, muscimol, THIP) and benzodiazepine modulators of GABA (e.g., diazepam) (Scheel-Krüger et al., 1978, 1979; Arnt et al., 1979; Delini-Stula, 1979; Hammerstad et al., 1980; Risch et al., 1980; Worms and Lloyd, 1982; Tirelli 1987). The role of GABA_A receptors is further supported by findings that GABA_A antagonists can prevent these effects (Delini-Stula, 1979). In contrast, agonists acting at GABA_B receptors, such as baclofen, do not potentiate gnawing (Delini-Stula 1979; Scheel-Krüger et al., 1979).

Tirelli and Witkin (1995) demonstrated that whereas indirect dopamine agonists, including dopamine releasers (e.g., amphetamines and amfonelic acid) and dopamine uptake inhibitors (e.g., cocaine, mazindol, GBR 12909, GBR 12935), dose-dependently increased gnawing, direct agonists (e.g., apomorphine, quinpirole, SKF 82958) did not facilitate gnawing. This common stereotypic effect engendered by the indirect-acting dopamine agonists is consistent with the common spectrum of other behavioral effects generally observed with these drugs. Indirect dopamine agonists generally increase locomotor activity (Vaugeois et al., 1993; Izenwasser et al., 1994), produce a common constellation of subjective effects (Schuster and Johanson, 1988) as suggested by drug discrimination experiments (Broadbent et al., 1991; Melia and Spealman, 1991; Witkin et al., 1991) and function as reinforcers (Griffiths et al., 1979; Bergman et al., 1989).

Although dopamine uptake blockers share a constellation of behavioral effects (Johanson and Fischman, 1989; Witkin, 1994), recent findings have suggested that differences among these compounds may exist. Although single-site models have also been presented (Dersch et al., 1994; Reith and Selmeci, 1992), ligand binding studies have suggested differences in the interactions of compounds with the dopamine uptake carrier and the potential for multiple binding sites (Madras et al., 1989; Berger et al., 1990; Johnson et al., 1992; Akunne et al., 1994). Structure-activity studies (Meltzer et al., 1994; Newman et al., 1994), investigation of protein chemistry (Vaughn, 1995; Vaughn and Kuhar, 1996) and amino-acid sequence analysis of the cloned transporter (Giros et al., 1994; Kitayama et al., 1992) have provided supporting evidence for the multiplicity of binding domains on the dopamine transporter. Taken together, these observations raise the possibility of observing functional correlates of such mechanistic differences. In fact, several reports have shown significant differences in the behavioral effects of dopamine uptake inhibitors, the most striking of which are the identification of drugs within this class that are devoid of psychomotor stimulant profiles (Rothman, 1990; Witkin and Acri, 1995; Acri et al., 1996).

The present series of experiments was initiated to discriminate among psychomotor stimulants through differential augmentation of their effects by GABA_A agonists. Interactions of the GABA_A agonist THIP with dopamine agonists were evaluated in C57BL/6J mice. With those dopamine agonists for which gnawing was not potentiated by THIP, further detailed studies were performed to achieve potentiation (e.g., dose and route of administration). If potentiation was still not achieved, these dopamine agonists were tested further in the presence of another GABA_A agonist, muscimol, to evaluate the generality of the effect. Based on results, further detailed dose-response experiments were carried out with methylphenidate as a prototype GABA_A-sensitive compound and cocaine as a GABA_A-insensitive compound.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals. A total of 1832 C57Bl/6J male mice (Jackson Laboratories, Bar Harbor, ME), weighing 21 to 32 g, were housed in groups of four in transparent polypropylene cages (19 × 27 × 15 cm high) with sawdust bedding. Mice had free access to commercial rat chow and tap water. The colony room was maintained at an ambient temperature of 22-26°C and was under a 12-h light/dark cycle (lights on at 7:00 A.M.).

General procedure. Mice were first injected either with saline or one of the doses of THIP or muscimol, before receiving an injection of one of the dopamine agonists (i.p. or s.c.). The doses of the dopamine agonists were based on prior observations and selected to produce generally comparable and submaximal levels of gnawing (Tirelli and Witkin, 1995) and the results of pilot experiments. In some cases (e.g., methamphetamine), choice of a comparable submaximal effect was complicated by the shape of the dose-effect curve (Tirelli and Witkin, 1995). When potentiation did not occur initially, dose manipulations were made (see below). Further, full dose-effect experiments with representative compounds (as noted below) were used to fully characterize the manner in which agonist dose affected GABA potentiation. The GABA agonists were administered (s.c.) 15 to 20 min before the dopamine agonist. After administration of the dopamine agonist, the mouse was placed in the test chamber. Because THIP and muscimol do not induce gnawing (Scheel-Krüger et al., 1978; Tirelli, 1987; Tirelli and Witkin, 1995), these compounds were not studied alone. All testing was conducted during the light period of the colony light-dark cycle, between 9:30 A.M. and 5:00 P.M.

Behavioral measurement. Gnawing on corrugated packing paper was the behavioral measure as described previously (Tirelli and Witkin, 1995). Corrugations were 5 mm wide, 2 mm high and separated from each other by 6.5 mm. Mice were tested singly in transparent, acrylic chambers (height 37 cm; area 25 cm × 14 cm), the floor of which was made of a sheet of corrugated paper, with the corrugations facing upward. A paper sheet was used only once. Gnawing was quantified by placing a grid over the corrugations. A score of 1 was given for each grid square in which penetration or tearing was visible. The maximal possible score was 880, but values in this experiment were always less than 615. Testing started immediately after injection and lasted 75 min.

Data analysis. Data are expressed as the mean (± S.E.M.) number of squares in which gnawing traces were visible (as described above). Means from drug dose groups were compared with the values of the respective control mean (dopamine agonist alone) by the a priori Welch-Aspin's test (Marascuilo and Serlin, 1988). This test is derived from the Dunn's test and takes into account deviations from homogeneity of variances. It requires the use of Dunn's test tables for critical values. Statistical significance was declared at a P < .05.

Drugs. All solutions were prepared immediately before injection. THIP hydrobromide (gaboxadol) and muscimol HBr (Research Biochemicals International [RBI], Natick, MA) were dissolved in saline (0.9% NaCl). The following dopamine agonists were dissolved in saline or in distilled water, and slightly heated if necessary: diclofensine HCl (Hoffmann-La Roche, Basel, Switzerland), indatraline HCl (Lundbeck, Copenhagen, Denmark), ()-cocaine HCl (Mallinkrodt/Nuclear, Orlando, FL), nomifensine maleate (RBI), WIN 35,428 (CFT naphthalene sulfonate; Sterling Winthrop, Rensselaer, NY), (+)-amphetamine sulfate, methamphetamine HCl; GBR 12909 diHCl (RBI); GBR 12935 diHCl (RBI). Amfonelic acid (RBI) was dissolved in a minimum of dilute NaOH and mazindol (RBI) in minimal HCl and made up to volume with distilled water. Drug doses are in terms of the drug forms just listed. THIP and muscimol were administered s.c. and the dopamine agonists were given by i.p. injection in a volume of 0.01 ml/g b.wt.

	Results

Top Abstract Introduction Methods Results Discussion References

None of the postsynaptic dopamine agonists studied, apomorphine (D₁/D₂), pergolide (D₂/D₁), RU24213 (D₂) or SKF 38393 (D₁), induced gnawing when preceded by injections of either THIP or muscimol (table 1). This lack of potentiation was observed despite an additional period of testing of 75 min. Direct observation of the mice indicated that only rare occurrences of spontaneous or exploratory gnawing and biting (that did not pierce the paper) occurred.

Doses of indirect agonists that did not induce gnawing when given alone were studied in conjunction with increasing doses of GABA_A agonists. Dose-effect functions for dopamine agonists in which GABA agonist potentiation of gnawing was observed are shown in figure 1. THIP (4, 6 and 8 mg/kg) readily potentiated gnawing produced by the indirect dopamine agonists acting via release ((+)-amphetamine, methamphetamine), uptake inhibition (methylphenidate, indatraline, mazindol, diclofensine, nomifensine, GBR12935) or both mechanisms (amfonelic acid). THIP was effective at 6 and 8 mg/kg for all nine agonists presented in figure 1, whereas the lower dose (4 mg/kg) facilitated gnawing triggered by (+)-amphetamine, amfonelic acid, mazindol and nomifensine. GABA facilitation of gnawing was also achieved under different control levels of gnawing. For example, methylphenidate produced a gnawing score of 24 at 10 mg/kg and 110 at 15 mg/kg. Nonetheless, as with the 10 mg/kg dose shown in figure 1, THIP also significantly augments the gnawing induced by 15 mg/kg methylphenidate (score = 100), producing a maximum gnawing score of 290 (fig. 2). Figure 2 also shows more clearly that the augmentation was bitonic. The reduced potentiation or frank depression (with muscimol) is associated with profound sedative effects of THIP or muscimol when given alone. For example, in the horizontal wire test of ataxia, both THIP and muscimol produce dose-dependent behavioral toxicity with significant effects starting at 5 mg/kg for THIP and 1.5 mg/kg for muscimol (data not shown).

View larger version (55K): [in this window] [in a new window]   Fig. 1.   Potentiation of indirect dopamine agonist-induced gnawing by the GABAA agonist THIP (s.c.) in male C57BL/6J mice. The indirect dopamine agonists act via release (methylphenidate, (+)-amphetamine, methamphetamine), uptake inhibition (indatraline, mazindol, diclofensine, nomifensine, GBR12935) or both mechanisms (amfonelic acid). The number in parentheses represents the dose of dopamine agonist which was injected 15 to 20 min after THIP. At higher doses, all these dopaminergic compounds induced intensive gnawing (Tirelli and Witkin, 1995). Dopamine agonists were given intraperitoneally. The number of mice per group was constant within each experiment. *P<.05, **P<.01 compared to saline control.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Dose-dependent augmentation of methylphenidate-induced gnawing by s.c. administration of the GABAA agonists THIP or muscimol. Other details are as in figure 1.

In contrast, as depicted in figure 3, the effects of the dopamine uptake blockers WIN 35428, bupropion, GBR 12909 and cocaine were not potentiated by THIP; gnawing was even decreased in cocaine. To determine if higher levels of gnawing produced by these compounds could be augmented, each was evaluated at a higher dose. Although the increase in dose of the dopamine agonists was effective in increasing basal levels of gnawing (right portions of each panel in fig. 3), THIP was still ineffective in increasing gnawing to higher values. This finding was substantiated and extended in additional experiments in which the route of cocaine administration and the dose (and hence the basal level of gnawing) were manipulated (fig. 4). Across a broader range of doses of THIP, gnawing induced by either 15 or 30 mg/kg cocaine, given by either the i.p. or s.c. routes, was not augmented by THIP. Higher doses of THIP resulted in a decrease in the gnawing response relative to cocaine alone. The lack of responsiveness of cocaine to THIP is further emphasized by contrast with methlyphenidate. Although both drugs dose-dependently increase gnawing, the dose-response curve for methylphenidate was shifted to the left by THIP, whereas that of cocaine was shifted downward by THIP (fig. 5).

View larger version (38K): [in this window] [in a new window]   Fig. 3.   The GABAA agonist THIP (s.c.) does not facilitate gnawing induced by several dopamine agonists in male C57BL/6J mice. *denotes a significant decrease as compared with the control value after a priori Welch-Aspin tests taken at a P level of .05. Other details are as described in the legend of figure 1.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Lack of potentiation of cocaine by a range of doses of THIP (s.c.). Cocaine was given at either 15 (left panel) or 30 mg/kg (right panel) by either the s.c. (open circles) or i.p. routes (closed circles). Values with * or * * are significantly different from the control value (SAL + dopamine agonist) as found by the a priori Welch-Aspin test at a P level either of .05 or .01, respectively. Values from the s.c. experiments with + or ++ are significantly different from the comparable dose values studied under the i.p. route, respectively. Other details are as described in the legend of figure 1.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Dose-effect functions for methylphenidate and cocaine when given alone and in the presence of 6 mg/kg THIP. Values with * or * * are significantly different from the control value (saline + 5 mg/kg methylphenidate or cocaine) as found by the a priori Welch-Aspin test at a P level either of .05 or .01, respectively. + or ++ denote significant differences within each dose comparison. Other details are as described in the legend of figure 1.

The possibility that GABA_A agonists other than THIP would augment the gnawing response of the THIP-insensitive compounds was investigated with muscimol (fig. 6). Pretreatment with muscimol at 2 mg/kg potentiated gnawing induced by WIN 35428 at 2.5 mg/kg and bupropion at 30 mg/kg. Additionally, gnawing was potentiated in mice receiving 40 mg/kg bupropion after 1 mg/kg muscimol. At that dose of bupropion, 3 mg/kg muscimol reduced gnawing. The potentiations observed with WIN 35428 and bupropion were restricted to a single dose of muscimol and were generally weak compared with the data presented in figure 1. Neither GBR 12909 nor cocaine were potentiated by muscimol regardless of the dose of either the dopamine agonist or the dose of muscimol. Note the large variability in gnawing in these experiments. These findings were substantiated in experiments in which either 15 or 30 mg/kg cocaine, given either by the i.p. or s.c. routes, was not potentiated by muscimol (fig. 7). In fact, higher doses of muscimol decreased the cocaine-induced gnawing (fig. 7) as they did methylphenidate-induced gnawing (fig. 2).

View larger version (37K): [in this window] [in a new window]   Fig. 6.   Effects of the GABAA agonist muscimol (s.c.) on gnawing induced by dopamine agonists in male C57BL/6J mice. Other details are as described in the legend of figure 1.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Lack of potentiation of cocaine by a range of doses of muscimol (s.c.). Cocaine was given at either 15 (left panel) or 30 mg/kg (right panel) by either the s.c. (open circles) or i.p. routes (closed circles). Other details are as in figures 1 and 4.

In control experiments, methylphenidate, GBR12935, diclofensine and mazindol were also tested in the presence of muscimol (fig. 8). In contrast to the compounds shown in figure 6, muscimol facilitated gnawing at all doses tested. Moreover, the augmentation in the gnawing response to these dopamine agonists was greater than that observed with WIN 35428 or bupropion. Note, however, the tendency for the highest dose of muscimol to produce a smaller degree of potentiation than the lower doses of muscimol for all compounds except methylphenidate. The maximal potentiation of gnawing produced by muscimol was generally greater than that observed with THIP for these compounds (compare figs. 1 and 8). Nonetheless, when a broader range of doses of the two agonists is explored, this difference may not be as marked (cf. fig. 2).

View larger version (43K): [in this window] [in a new window]   Fig. 8.   The GABAA agonist muscimol (s.c.) potentiates dopamine agonist-induced gnawing in male C57BL/6J mice. Dopamine agonists were given i.p. with the exception of methylphenidate (s.c.). Values with * or ** are significantly different from the control value (SAL + dopamine agonist) as found by the a priori Welch-Aspin test at a P level either of .05 or .01, respectively. Other details are as described in the legend of figure 1.

The D₁ antagonist, SCH 23390, was more potent in blocking cocaine-induced gnawing than methylphenidate-induced gnawing (Tirelli and Witkin, 1995). Because these two compounds were also differentiated based on GABA potentiation in the present study, we used SCH 23390 to differentiate among the structural analogs, GBR 12909 and GBR 12935. SCH 23390 given 30 min before those drugs (both at 40 mg/kg i.p.) dose-dependently attenuated gnawing with equivalent potency [ED₅₀ = 0.02 mg/kg (95% CL: 0.004-0.12) for GBR 12909 and 0.02 mg/kg (95% CL: 0.01-0.04) for GBR 12935].

	Discussion

Top Abstract Introduction Methods Results Discussion References

There were three primary findings in the present study. First, indirect- but not direct-acting dopamine agonists induce gnawing in the present experimental configuration. These differences have been reported and discussed previously (Tirelli and Witkin, 1995). Second, the gnawing induced by some of the indirect-acting dopamine agonists was potentiated in mice in which GABA_A receptors were stimulated by either THIP or muscimol. The GABA_A potentiation of gnawing triggered by indirect dopamine agonists reported here confirms and extends previous observations (see references cited in the introduction and below). These findings can now be extended to several indirect dopamine agonists, the dopamine releasers methamphetamine and amfonelic acid, the dopamine uptake inhibitors indatraline, nomifensine, GBR 12909 and GBR 12935 and the catecholamine uptake inhibitors mazindol and diclofensine. The third finding of note is the differential sensitivity of the indirect agonists to the effects of THIP and muscimol.

Although the potentiation of dopamine-driven gnawing is clear, the question of whether GABA_A agonists can increase the maximal efficacy of dopaminergic agents is not as straightforward. Although several compounds produced maximal gnawing scores in the presence of THIP or muscimol which were equivalent to their maximal effects when given alone (see Tirelli and Witkin, 1995) (e.g., methlyphenidate, GBR 12935), the GABA-enhanced effects of methamphetamine were ~6-fold higher than maximal effects of methamphetamine alone (300 vs. 50). Unlike methamphetamine, there is little difference in maximal effect with the GABA_A agonist THIP on board (fig. 5). Systematic efficacy comparisons, however, would require full dose-effect functions for all of the drugs and is beyond the scope of this study.

The lack of gnawing induced by direct agonists as well as the inability of THIP or muscimol to augment the gnawing response induced by either apomorphine or cocaine is inconsistent with previous observations. Prior reports have demonstrated gnawing facilitation of apomorphine- or cocaine-induced gnawing by muscimol, although much less intensively than methylphenidate in mice (Scheel-Krüger et al., 1978; Arnt et al., 1979). It is plausible that the difference in these results from the present findings are caused by differences in environmental testing conditions on which the intensity and occurrence of stereotyped gnawing and oral behaviors critically depend (Ljundberg and Ungerstedt, 1977; Benus et al., 1991; Eilam et al., 1992). This issue has been discussed previously (Tirelli and Witkin, 1995).

The present results demonstrate that stereotyped behavioral effects of dopamine uptake inhibitors can be differentiated by GABA_A agonists. Whereas effects of structurally diverse inhibitors of dopamine uptake were readily augmented by THIP or muscimol, effects of bupropion, WIN 35428, GBR 12909 and cocaine were not robustly enhanced (table 2). The lack of GABA-potentiated gnawing with these dopamine agonists could not be attributed to the dose studied, the effect of the dopamine agonists alone, the maximal effects of the agonists or the range of doses of the GABA agonists tested. The doses of the uptake blockers that were not potentiated produced effects alone that were of the same order of magnitude as those that were potentiated (compare figs. 1 and 2). Moreover, the ineffective compounds were studied further at higher doses which achieved greater, but still submaximal, amounts of gnawing when given alone (figs. 2, 4, 5 and 7). Because the GABA agonists were given across a wide range of doses, producing minimal augmentation at the lowest dose and submaximal augmentation at the highest dose, the lack of sufficient GABA receptor activation is also unlikely to account for the differences in dopamine agonist augmentations.

The maximal effects achieved by the dopamine agonists alone were also not related to the lack of GABA enhancement. For example, when given alone across a wide range of doses, methamphetamine produced significant but minimal amounts of gnawing (maximal score <100). Amfonelic acid, on the other hand, produced greater maximal increases in gnawing (~400) (Tirelli and Witkin, 1995). Nevertheless, THIP enhanced gnawing of both methamphetamine and amfonelic acid. Of the compounds that were not enhanced by THIP, a range of maximal effects have been demonstrated under the present testing conditions. Maximal gnawing scores for WIN 35428, bupropion, GBR 12909 and cocaine were approximately 500, 150, 200 and 350, respectively (Tirelli and Witkin, 1995).

Pharmacological or structural differences among the dopamine uptake inhibitors also appear unrelated to the differential GABA augmentation observed. Thus, although some of the compounds nonselectively block uptake of monoamines (e.g., cocaine, mazindol), others, notably GBR 12909 and GBR 12935, are selective inhibitors of dopamine uptake (Van der Zee et al., 1980). Nonetheless, given that dopamine-related gnawing also depends on several other systems (Cools, 1977; Mogenson, 1987; Scheel-Krüger, 1986), the possibility that these nondopaminergic actions of the dopamine agonists might account for differential GABA modulation cannot be dismissed entirely. Differences in chemical structure, in a broad sense, also did not account for differences in GABA augmentation. Notably, the diphenyl-substituted piperazine analogs, GBR 12909 and GBR 12935, although both producing gnawing when given alone, were differentiated by GABA agonists.

Accumulating evidence points to the existence of differences in the manner in which indirect dopamine agonists interact with the dopamine uptake carrier (Andersen et al., 1987; Madras et al., 1989; Johnson et al., 1992; Akunne et al., 1994). Such differences may constitute a mechanism by which differential behavioral effects of dopamine uptake blockers may be initiated. Pristupa et al. (1994), for example, presented evidence suggesting that GBR 12935 and WIN 35428 do not interact with a common functional form or state of the dopamine transporter. Differences in the interactions of cocaine and GBR 12935 with the dopamine uptake carrier have also been suggested (Berger et al., 1990). By use of other methods, Meiergerd and Schenk (1994) observed that mazindol and nomifensine may interact with a kinetically activated dopamine transporter differently from cocaine and GBR 12909. Congruently with these observations at a molecular level, the present work on interactions of GABA agonists uncovered differences throughout these compounds; gnawing induced by GBR 12935, nomifensine and mazindol but not by WIN 35428, GBR 12909 or cocaine was potentiated by GABA agonists.

Other studies have also suggested that functional interactions of GBR 12909 and cocaine with the dopamine transporter may differ (Rothman et al., 1989, 1991). However, those functional differences did not agree with the differentiation among compounds in the present study in which both cocaine and GBR 12909 were insensitive to GABA augmentation. Although many preclinical comparisons of GBR 12909 and cocaine have demonstrated their similar behavioral effects (Melia and Spealman, 1991; Witkin et al., 1991; Tirelli and Witkin, 1995), behavioral differences have also been reported (Elmer et al., 1996). GBR 12909, along with bupropion and mazindol, have also been suggested to be non-euphoria-producing compounds, in marked contrast to cocaine (Rothman, 1990; Rothman and Glowa, 1995).

GABAergic potentiation of dopamine-triggered orofacial behaviors, and the GABAergic attenuation of dopamine agonist-induced hyperkinesia, result from increased GABA activity mainly within the nigral, pallidal and/or subthalamic areas (see overviews by Cools, 1977; Mogenson, 1987; Scheel-Kruger, 1986; Gerfen, 1992). Differential actions of the dopamine uptake inhibitors in discrete brain areas (Arnt et al., 1987; Karoum et al., 1994) may be a potential source for the differential GABA-potentiation observed here. The manner in which DA and GABA interact depends on brain area (Waszcak and Walters, 1983; Chiodo and Berger, 1986; Walters and Pucak, 1996). In addition, differences in sensitivity of neuronal populations to GABA (Beauregard and Ferron, 1991) could result in differential reciprocal control of critical dopaminergic pathways. Existing data also suggest either differences in the uptake carrier or the uptake process in different regions of the brain (Hadfield and Nugent, 1983; Izenwasser et al., 1990; Lew et al., 1991; Elsworth et al., 1993). Regionally specific effects might contribute to the differential behavioral effects observed here in the presence of GABA agonists.

Another potential determinant of the differences in the abilities of GABA_A agonists to potentiate gnawing is the induction of behaviors incompatible with gnawing. Previous observations with dopamine agonists alone have suggested that competing behavioral responses may impede the gnawing response under the experimental conditions studied here (Tirelli and Witkin, 1995). Similarly, the addition of THIP or muscimol could have induced behavioral sequences or toxicities that retarded the expression of gnawing in the present study.

Muscimol generally exerted more intense potentiation than observed with THIP. Although differences between these agonists may be caused by the higher affinity of muscimol for the GABA_A site, THIP is considered one of the most selective GABA_A agonists (Fonnum, 1987; Lloyd and Morselli, 1987). Muscimol, but not THIP, was also capable of producing small, yet restricted increases in gnawing engendered by WIN 35428 and bupropion. This is the first report of qualitative differences in the behavioral effects of these drugs of which we are aware. Muscimol and THIP are both isoxazole derivatives, with THIP representing the likely conformation of muscimol that binds to the receptor (Krogsgaard-Larsen et al., 1994). Nonetheless, qualitative differences in the interactions of these compounds with GABA_A receptors may exist. For example, whereas GABA and dihydromuscimol demonstrate similar efficacy in electrophysiological studies of Xenopus oocytes across GABA_A receptor isoforms composed of different alpha subunits, THIP displayed large differences in efficacy that depended on the subunit composition of the receptor (Ebert et al., 1994). Given the differential central localization of GABA_A receptor subtypes (Lüddens et al., 1995), the potential for GABA agonists to produce different functional effects may also be possible (Mereu et al., 1992).