Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Cocaine inhibits the reuptake of serotonin, norepinephrine, and dopamine by binding to their respective transporters and acts as a local anesthetic by binding to sodium channels. Despite this multiplicity of actions, most of the behavioral effects of cocaine are thought to be mediated by the inhibition of dopamine transport (Heikkila et al., 1979); those actions are thought to be critical components in the abuse of cocaine (Kuhar et al., 1991). This latter hypothesis is supported by several findings. First, dopamine antagonists appear to specifically increase rates of cocaine self-administration in laboratory animals, whereas noradrenergic antagonists do not (deWit and Wise, 1977). Furthermore, lesions of dopamine rich sites in the brain lead to changes in cocaine self-administration that are not produced by lesions in other areas (Roberts et al., 1977; Koob et al., 1987). Moreover, the potencies of monoamine uptake inhibitors in maintaining self-administration behavior are directly related to their affinities for the dopamine transporter and more closely related to those affinities than their affinities for the norepinephrine or serotonin transporters (Ritz et al., 1987; Bergman et al., 1989). Taken together, these results suggest that the inhibition of dopamine reuptake is the critical action of cocaine that leads to its abuse.

Despite considerable evidence supporting the dopamine hypothesis of the abuse liability of cocaine, some conflicting evidence remains. One apparent inconsistency is that not all dopamine uptake inhibitors are subject to abuse as is cocaine (Rothman, 1990). Both bupropion and benztropine (BZT) are dopamine uptake inhibitors that are used therapeutically but are seemingly devoid of any significant abuse liability. The reason for the differences between these drugs and cocaine is currently not clear, although there are several potential explanations. BZT has clinical use in the treatment of Parkinson's disease and has several known actions in addition to its dopamine uptake inhibiting effects (Coyle and Snyder, 1969). Notable among other actions of BZT are its antimuscarinic effects (Richelson, 1979).

BZT has some stuctural similarities to cocaine, particularly the N-methyl-tropane ring. Rather than the 3-ester-linked phenyl ring of cocaine, BZT has a diphenyl ether system (see Fig. 1). This functional group is also a component of another dopamine uptake inhibitor, GBR 12909. Initial studies of analogs of BZT have indicated that their structure-activity relations differ from those of cocaine analogs (Newman et al., 1994, 1995; Meltzer et al., 1994, 1996). For example, the tropane ring of the BZT analogs lacks a substituent at the 2-position that is necessary for high-affinity binding of cocaine to the dopamine transporter (Carroll et al., 1997; Xu et al., 1997). In addition, when the diphenyl ether system of the BZT analogs is in the axial () configuration, higher affinity binding results than when that moiety is in the equatorial () configuration. Conversely, the benzoyl ester of cocaine is preferred in the equatorial conformation. Furthermore, substitutions on the phenyl rings of BZT alter the binding affinity more dramatically than do similar substitutions on the cocaine analog, WIN 35,065-2 (Newman et al., 1994). Moreover, the displacement of [³H]WIN 35,428 binding by 3-(diphenylmethoxy)tropane analogs models better for a single site than it does for two sites, whereas the binding of cocaine and several of its analogs is better fit by a two-site model than by a single-site model (Madras et al., 1989; Izenwasser et al., 1994). Finally, initial reports indicated differences between BZT analogs and cocaine with regard to their behavioral effects. For example, the 4'-Cl substituted analog of BZT had minimal efficacy as a locomotor stimulant and did not substitute for cocaine in rats trained to discriminate cocaine from saline (Newman et al., 1994). Similar results have been obtained with the parent compound, BZT (Colpaert et al., 1979; Acri et al., 1996), and BZT has been reported to be less efficacious than d-amphetamine in producing stereotyped behavior (Scheel-Krüger, 1972). These structural points of comparison between BZT and cocaine, along with the divergence in pharmacology, led to the present study in which we examined the pharmacology of a series of substituted phenyl ring analogs of BZT.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Basic structures of cocaine and 3-diphenylmethoxytropane analogs with substitutions in the 3', 4' or 4" positions.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

[³H]WIN 35,428 Binding Assay. Details of the procedures used have been published previously (Izenwasser et al., 1993). Briefly, male Sprague-Dawley rats (200-250 g; Taconic Farms, Germantown, NY) were decapitated and their brains removed to an ice-cooled dish for dissection of the caudate putamen. The tissue was homogenized in 30 volumes of ice-cold modified Krebs-HEPES buffer (15 mM HEPES, 127 mM NaCl, 5 mM KCl, 1.2 mM MgSO₄, 2.5 mM CaCl₂, 1.3 mM NaH₂PO₄, and 10 mM D-glucose, pH adjusted to 7.4) using a Brinkman polytron and centrifuged at 20,000g for 10 min at 4°C. The resulting pellet was then washed two more times by resuspension in ice-cold buffer and centrifuged at 20,000g for 10 min at 4°C. Fresh homogenates were used in all experiments.

[³H]Dopamine Uptake Assay. [³H]Dopamine uptake was measured in a chopped tissue preparation as described previously (Izenwasser et al., 1990). Briefly, rats were sacrificed by decapitation and their brains were removed to an ice-cooled dish for dissection of the caudate putamen. The tissue was chopped into 225-µm slices on a McIllwain tissue slicer with two successive cuts at an angle of 90°. The strips of tissue were suspended in oxygenated modified Krebs-HEPES buffer (see above), which was pregassed with 95% O₂/5% CO₂ and warmed to 37°C. After rinsing, aliquots of tissue slice suspensions were incubated in buffer in glass test tubes at 37°C to which either the drug being tested or no drug was added, as appropriate. After a 5-min incubation period in the presence of the test drug, [³H]dopamine (final concentration 15 nM) was added to each tube. After 5 min the incubation was terminated by the addition of 2 ml of ice-cold buffer to each tube and filtration under reduced pressure over glass fiber filters (presoaked in 0.1% polyethylenimine in water). The filters were rinsed and placed in scintillation vials to which 1 ml of methanol and 2 ml of 0.2 M HCl were added to extract the accumulated [³H]dopamine. Radioactivity was determined by liquid scintillation spectrometry at an efficiency of approximately 30%. The reported values represent specific uptake from which nonspecific binding to filters was subtracted.

[³H]Pirenzepine Binding Assay. Muscarinic M₁ binding was carried out as modified from methods described by Luthin and Wolfe (1984) and Potter et al. (1988). Frozen whole rat brains were prepared as described previously (Bowen et al., 1989). Briefly, whole brain excluding cerebellum (Taconic Farms) was thawed in ice-cold buffer (10 mM Tris-HCl, 320 mM sucrose, pH 7.4) and homogenized with a Brinkman polytron in a volume of 10 ml/g tissue. The homogenate was centrifuged at 1000g for 10 min at 4°C. The resulting supernatant was then centrifuged at 10,000g for 20 min at 4°C. The resulting pellet was resuspended in a volume of 1.53 ml/g in 10 mM Tris buffer (pH 7.4).

Locomotor Activity. Ambulatory activity of Male Swiss Webster mice (Taconic Farms) were studied in 40-cm³ clear acrylic chambers. The acrylic chambers were placed inside monitors (Omnitech Electronics, Columbus, OH) that were equipped with light-sensitive detectors spaced 2.5 cm apart along two perpendicular walls. Mounted on the opposing walls were infrared light sources that were directed at the detectors. One count of horizontal activity was registered each time the subject interrupted a single beam. Mice were injected and immediately placed in the apparatus for 60 min, with horizontal activity counts collected every 10 min. Intraperitoneal injections were administered in volumes of 1 ml/100 g. Each dose was studied in eight mice, and mice were used only once. In some experiments, the interaction of atropine or scopolamine with cocaine was assessed by administering the drugs with cocaine (i.p.) immediately before placing the subjects inside the monitors. Horizontal activity was counted as above for a 60-min time period. In other studies, the time course of the effects on locomotor activity of selected compounds was assessed. Mice were injected and immediately placed in the apparatus for 8 h. All other aspects of these experiments were identical with those which assessed activity for 60 min.

Cocaine Discrimination. Male Sprague-Dawley rats (Charles River Breeding Laboratories, Inc., Wilmington, MA) weighing 310 to 385 g were individually housed with free access to water under a 12-h light/dark cycle (lights on at 7:00 AM). The rats were experimentally naive at the beginning of the study. All testing was performed between 9:00 AM and 1:00 PM. Rats were fed 15 g of Purina chow 30 min after testing daily.

Drugs. The drugs tested were ()-cocaine HCl (Sigma Chemical Co.); GBR 12909 diHCl (Research Biochemicals, Inc., Natick, MA), atropine sulfate (Sigma), methylatropine (Sigma), scopolamine (Sigma), methylscopolamine (Sigma), and the diphenylmethoxytropane analogs synthesized in our laboratories (Newman et al., 1994, 1995). The basic skeleton of the diphenylmethoxytropane analogs is shown in Fig. 1. Substitutions examined in the present study were exclusively on the 4'-, 4"-, or 3'-positions. All drugs were dissolved in water or 0.9% NaCl. The drugs were administered i.p. at 1 ml/kg b.wt. All drugs were injected immediately before testing, with the exceptions as noted below.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

[³H]WIN 35,428 Binding Assay. All of the drugs fully displaced [³H]WIN 35,428 from caudate putamen membranes. The affinities (K_i values) for this activity ranged from 11.8 to 2000 nM (Table 1). The highest affinity was obtained with the 4',4"-difluoro-substituted compound which had an affinity for the dopamine transporter that was 10-fold greater than that for the parent compound, BZT. In addition, this compound had an affinity that was 3-fold greater than that of cocaine (Izenwasser et al., 1994). The comparison to cocaine is based on the high-affinity component of WIN 35,428 binding, as WIN 35,428 binding to the dopamine transporter can be modeled significantly better for two sites than one (Madras et al., 1989; Izenwasser et al., 1993). None of the present 3-diphenylmethoxytropane analogs produced a displacement profile that was better fit to a two-site model than a one-site model.

View this table: [in this window] [in a new window]   TABLE 1 Potencies of 3-diphenylmethoxytropane analogs in binding to the dopamine transportera and M1 muscarinic receptors, inhibiting dopamine uptakea, and in effecting locomotor activityb

[³H]Dopamine Uptake Assay. Each of the compounds also inhibited the uptake of dopamine in chopped caudate-putamen tissue (Table 1). The potencies for this effect ranged from 24 nM (IC₅₀ value) for the 3',4'-di-chloro,4"-fluoro-substituted compound to 3520 nM for the 4-chloro-substituted analog with the diphenyl ether system in the  conformation. Previous studies have indicated that the inhibition of dopamine uptake produced by cocaine and several cocaine analogs under these procedures is not linear with respect to concentration (Izenwasser et al., 1990). In contrast to the effects obtained with cocaine and its congeners, the results with the present diphenylmethoxytropane analogs did not significantly deviate from linearity. In general, the potencies of these compounds for the inhibition of dopamine uptake were well correlated with their affinities for the dopamine transporter labeled with [³H]WIN 35,428 (r² = 0.770; p < .001).

[³H]Pirenzepine Binding Assay. All of the drugs displaced [³H]pirenzepine from whole-brain membranes (Table 1). The affinities (K_i values) for this activity ranged from 2.1 nM for the parent compound, BZT, to 120 nM for the 4',4"-dimethoxy-substituted analog. All of the compounds displaced [³H]pirenzepine with a profile that was best fit to a one-site model. The relation between structure and affinity for the M₁ muscarinic site was distinctly different from that between structure and affinity for the dopamine transporter as indicated by a lack of a significant correlation between the two groups of affinities (r² = 0.126, p = .137).

Locomotor Activity. Cocaine, as has been demonstrated previously, increased horizontal ambulatory activity, and under the present conditions 29.4 µmol/kg produced a maximum of approximately 15,200 counts over 30 min, with higher doses producing less stimulation (Fig. 2, filled circles; see Table 1 for absolute count numbers). The diphenylmethoxytropane analogs showed different degrees of stimulant activity (Fig. 2). The compounds with a para-fluoro substitution as a group (Fig. 2A) had the highest efficacy in stimulating activity among the 3-diphenylmethoxytropane analogs, with the 4',4"-difluoro compound showing the greatest effect. Compounds with a para-chloro substituent were generally less efficacious than the corresponding fluoro-substituted analogs (Fig. 2B). For example, 4'-fluoro-3-(diphenylmethoxy)tropane produced a maximum stimulation to approximately 10,200 counts over 30 min, whereas 4'-chloro-3-(diphenylmethoxy)tropane (4'-Cl-BZT) produced a maximum stimulation to approximately 7,800 counts. Similarly, the maximum stimulation produced by 4',4"-difluoro-3-(diphenylmethoxy)tropane (4',4"-diF-BZT) was approximately 12,400 counts, but the maximum stimulation produced by 4',4"-dichloro-3-(diphenylmethoxy)tropane (4',4"-diCl-BZT) was 5,500. As noted previously (Newman et al., 1994), 4'-Cl-BZT, which has the diphenyl-ether system attached to the tropane ring with an axial () stereochemistry, had higher affinity than its stereoisomer with the diphenyl-ether system in the equatorial () configuration (Table 1). In the present study the 4'-chloro-3-(diphenylmethoxy)tropane [4'-Cl-BZT ()] produced less stimulation than its  analog, 4'-Cl-BZT (Table 1); this is consistent with its lower affinity for the dopamine transporter. Various other substituents, including methyl and methoxy, yielded compounds with minimal, if any, stimulant activity. ("Minimal" stimulant activity did not achieve statistical significance across a range of doses, from those having no effect to those virtually eliminating locomotor activity.) (Fig. 2C; italicized ED₅₀ values in Table 1).

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Dose-dependent effects of 3-diphenylmethoxytropane analogs on locomotor activity in mice. Ordinates, horizontal activity counts after drug administration. Abscissae, dose of drug in µmol/kg, log scale. Each point represents the average effect determined in eight mice. The data are from the 30-min period during the first 60 min after drug administration, in which the greatest stimulant effects were obtained. Note that the fluoro-substituted compounds (A) were generally more efficacious than the other compounds, and that the bromo-substituted compounds (C) were generally the least efficacious.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effects of atropine and scopolamine on locomotor activity in mice. Ordinates, horizontal activity counts after drug administration. Abscissae, dose of drug in µmol/kg, log scale. Each point represents the average effect determined in eight mice. The data are from 30 to 60 min after drug administration, which was the time at which the greatest stimulant effects were obtained. Note that each of the antimuscarinics stimulated locomotor activity, although not with the efficacy of cocaine (see Fig. 2). Further note that scopolamine was approximately 10-fold more potent than atropine.

Cocaine Discrimination. As has been shown in the past, subjects trained to discriminate 10.0 mg/kg (29.4 µmol/kg) cocaine showed a dose-related increase in the percentage of responses emitted on the cocaine-appropriate response key as dose was increased up to the training dose (Fig. 4, filled circles). In contrast, none of the 3-diphenylmethoxytropane analogs produced a maximum level of drug appropriate responses that exceeded 70% (Fig. 4, A and B; Table 2; column A). The most efficacious of the 3-diphenylmethoxytropane analogs were generally those with a para-F substitution on at least one ring of the diphenyl ether system (Fig. 4A; Table 2, column A). The greatest efficacy was achieved with 3',4'-dichloro, 4"-F-3-(diphenylmethoxy)tropane, which had a maximal effect of 68% cocaine-appropriate responding (Fig. 4A, triangles) with evidence of a plateau occurring at doses of 7 and 13 µmol/kg. Other compounds (e.g., 4',4"-diF-BZT and 3',4'-difluoro-3-(diphenylmethoxy)tropane) had comparable efficacy at the highest dose at which reliable responding was obtained (Table 2). Compounds with para Br, Cl, or methoxy substitutions were generally less efficacious than those with at least one F substituent in producing cocaine-appropriate responses (Fig. 4B; Table 2, column A). The exceptions to this generalization are 4'-Br,4"-F-BZT, which was among the least efficacious of compounds with a fluoro substituent (Fig. 4A, hexagons; Table 2), and 4'-Cl-BZT, which showed a maximal effect of 55% cocaine-appropriate responses (Fig. 4B, squares). Among the remaining compounds, the maximum effect was the 35% cocaine-appropriate responses produced by 4',4"-dibromo-3-(diphenylmethoxy)tropane (4',4"-diBr-BZT) (Fig. 4B, open circles; Table 2, column A).

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Effects of 3-diphenylmethoxytropane analogs in rats trained to discriminate injections of cocaine from saline. Ordinates for A and B, percentage of responses on the cocaine-appropriate key. Ordinates for C and D, rates at which responses were emitted (as a percentage of response rate after saline administration). Abscissae, drug dose in µmol/kg (log scale). Each point represents the effect in four or six rats. The percentage of responses emitted on the cocaine-appropriate key was considered unreliable and not plotted if fewer than half of the subjects responded at that dose. Note that the fluoro-substituted compounds (A) were generally more efficacious than the other compounds and that the bromo-substituted compounds (B) were among the least efficacious.

View this table: [in this window] [in a new window]   TABLE 2 Efficacy of 3-diphenylmethoxytropane analogs in substituting for cocaine in rats

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effects of various doses of atropine on the discriminative stimulus effects of selected doses of cocaine. Ordinates for top panels, percentage of responses on the cocaine-appropriate key. Ordinates for bottom panels, rates at which responses were emitted (as a percentage of response rate after saline administration). Abscissae, cocaine dose in mg/kg (log scale). Each point represents the effect in four or six rats. The percentage of responses emitted on the cocaine-appropriate key was considered unreliable and not plotted if fewer than half of the subjects responded at that dose. Filled triangles represent the effects of atropine alone. Note that atropine did not substitute for cocaine. Note also that the discriminative stimulus effects of 5.6 and 10.0 mg/kg cocaine were not altered by coadministration of atropine (top), however, the response-rate disruptive effects of cocaine were potentiated by atropine (bottom).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Changes in the cocaine dose-effect curve for discriminative stimulus effects produced by pretreatments with the muscarinic antagonists, atropine or scopolamine. Ordinates, percentage of responses on the cocaine-appropriate key. Abscissae, cocaine dose in mg/kg (log scale). Each point represents the effect in four or six rats. The percentage of responses emitted on the cocaine-appropriate key was considered unreliable and not plotted if fewer than half of the subjects responded at that dose. Note that either muscarinic antagonist shifted the cocaine dose-effect curve to the left.

Time Course of Behavioral Effects. The study of the time course of the effects on locomotor activity of selected 3-diphenylmethoxytropane analogs revealed lasting locomotor stimulant effects of 4',4"-diF-BZT that were of a greater duration than those for cocaine (Fig. 7; compare A and B). Maximal stimulant effects of cocaine across the 8-h study were obtained in the first 30 min (Fig. 7A, shaded portion), as they were in the 1-h study, and as would be expected, the parameters of the effect (maximal effects and ED₅₀ values) were similar in the two studies (Table 5). Two-way ANOVA of results of the 8-h study revealed significant effects of time, cocaine dose, and the interaction of the two [F(time)_47,1680 = 51.0, p < .001; F(dose)_4,1680 = 9.1, p < .001; F(txd)_188,1680 = 2.7, p < .001]. As can be seen (Fig. 7A), at the doses of cocaine (29 and 59 µmol/kg) that produced maximal stimulant effects during the 1-h study, the stimulant effects were brief, clearly diminishing over the first 30-min period (shaded portion of Fig. 7A). In addition, the highest dose (118 µmol/kg; Fig. 7A, diamonds) produced comparable stimulation during the subsequent 30-min period, with the degree of stimulation decreasing thereafter, but lasting up to 150 min after injection.

View larger version (73K): [in this window] [in a new window]   Fig. 7.   Time course of effects of 3-diphenylmethoxytropane analogs on locomotor activity in mice. Ordinates, horizontal activity counts after drug administration. Abscissae, time since injection and placement of subject in experimental chamber. Effects of several doses are shown (see symbol keys on each panel for individual pretreatment times). Each point represents the average effect determined in eight mice. Columns represent ± 1 S.E.M. Points represent average total counts for successive 10-min time periods up to 480 min (8 h) after injection. The shaded portions correspond to portions of the sessions in which the greatest effects were observed, and for which detailed data are provided in Table 5. Note that the duration of action of cocaine was relatively short, noticeably diminished after 30 min (A). In contrast, the duration of action of the difluoro-substituted compound was considerably greater, with maximal stimulation after the 25 µmol/kg dose occurring in the second 30 min after injection and not diminishing until approximately 440 min after injection (B). None of the other compounds showed appreciable locomotor stimulant like activity.

View larger version (34K): [in this window] [in a new window]   Fig. 8.   Time course of effects of selected 3-diphenylmethoxytropane analogs in rats trained to discriminate injections of cocaine from saline. Each panel shows effects for a different compound at various pretreatment times. Ordinates for top panels, percentage of responses on the cocaine-appropriate key. Ordinates for the bottom panels, rates at which responses were emitted (as a percentage of response rate after saline administration). Abscissae, drug dose in µmol/kg (log scale). Each point represents the effect in four or six rats. The percentage of responses emitted on the cocaine-appropriate key was considered unreliable and not plotted if fewer than half of the subjects responded at that dose. Note that the effects of cocaine diminish with increased time between injection and testing. In contrast, the effects of the difluoro-substituted compound increases with increased time between injection and testing. When tested 5 min after injection, 4',4"-diF-BZT did not fully substitute for cocaine; however, with greater time periods the compound fully substituted. None of the other compounds showed this increased efficacy with greater time between injection and testing.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The present study examined the pharmacology of some novel 3-diphenylmethoxytropane analogs. Like cocaine, these compounds had varying affinities for the dopamine transporter and inhibited dopamine uptake in vitro. However, as shown with 4'-Cl-BZT in our initial report (Newman et al., 1994), the behavioral effects of these drugs differed from those of cocaine. Generally, the 3-diphenylmethoxytropane analogs had varying efficacies for stimulating locomotor activity, but none had an efficacy equivalent to that of cocaine. Furthermore, with the exception of the 4',4"-difluoro substituted compound, the drugs commonly had lower efficacy than cocaine in producing discriminative-stimulus effects in rats trained to discriminate cocaine from saline.

The present data correspond to previous studies that have indicated differences between the behavioral effects of BZT and cocaine. For example, McKearney (1982) showed that both BZT and cocaine increased operant behavior; however, cocaine generally had greater efficacy (see also Acri et al., 1996). Furthermore, several studies have indicated that BZT did not fully substitute for cocaine in rats trained to discriminate cocaine from saline (Colpaert et al., 1979; Acri et al., 1996). Differences from cocaine in the behavioral effects of BZT (Newman et al., 1994; Acri et al., 1996) or other dopamine uptake inhibitors (Rothman, 1990) have prompted suggestions that some of these compounds may serve as leads for the development of pharmacological treatments for cocaine abuse. Moreover, dopamine uptake inhibitors that have in vivo effects that differ from those of cocaine may serve to better our understanding of the pharmacological mechanisms that underlie the behavioral effects of cocaine and lead to its abuse.

Compounds with fluoro substitutions in the para- and meta-positions generally were among the compounds with the highest affinity for the dopamine transporter and the highest potency for inhibition of dopamine uptake. These compounds generally had behavioral effects most closely resembling those of cocaine, producing the greatest stimulation in locomotor activity and the highest efficacy in substituting for the discriminative stimulus effects of cocaine. In contrast, several of the chloro-substituted analogs exhibited relatively lower behavioral efficacy but had affinities at the dopamine transporter comparable to those of the corresponding fluoro-substituted compounds. This comparison suggests that cocaine-like in vivo effects were not related to affinity at the dopamine transporter in a simple manner. Consistent with that conclusion is the finding of an inverse relation between the affinity of these drugs at the dopamine transporter and potency for producing locomotor stimulation. Among close structural analogs of cocaine and WIN 35,428 there is typically a direct relation between affinity at the dopamine transporter and potency for locomotor stimulation (Cline et al., 1992; Izenwasser et al., 1994). Although this relationship is drawn across species in the study by Izenwasser et al. (1994), it was drawn within species by Cline et al. (1992). In agreement with the present findings, Vaugeois et al. (1993) and Izenwasser et al. (1994) showed that among dopamine uptake inhibitors that are structurally dissimilar to cocaine, the relationship between dopamine transporter affinity and in vivo behavioral effects is distinctly different from that for analogs of cocaine (see also Rothman et al., 1992). There currently is no mechanistic explanation for the differences between cocaine-like dopamine uptake inhibitors and other dopamine uptake inhibitors with respect to these relations.

The 3-diphenylmethoxytropane analogs that have effects most like those of cocaine all contain at least one fluoro group in the para-position of one of the phenyl rings of the diphenyl ether moiety. These substituents are sterically small halogens that appear to provide the optimum combination of volume and electronic character to achieve maximum behavioral response. In addition, these compounds are among those in this series with the highest affinity at the dopamine transporter. Cocaine-like activity is somewhat reduced in the compounds that have at least one chloro group in the para-position of one of the phenyl rings, despite comparable binding affinities (see above). Additional chloro groups (e.g., 4',4"-diCl-BZT), or in particular substitution with a sterically bulkier bromo group (e.g., 4'-Br,4"-F-BZT), results in diminished cocaine-like activity as compared to the analogous fluoro-substituted compounds. The larger halogenated compounds demonstrate relatively lower binding potencies at the dopamine transporter. The least efficacious compounds in this series do not have small halogens on the phenyl rings but rather have electronically neutral (CH₃) or electron withdrawing groups (NO₂, CF₃, CN), some of which are relatively bulky in size as well. Clearly, these compounds did not produce a cocaine-like stimulant profile. In addition, these compounds were relatively less potent in binding to the dopamine transporter. Therefore, the compounds in this series most like cocaine possess at least one fluoro group in the para-position of one of the phenyl rings. Interestingly, the meta-Cl-substituted compound has significantly greater cocaine-like efficacy than its para-substituted analog, both in stimulation of locomotor activity and in producing cocaine-like discriminative-stimulus effects (Kline et al., 1997). However, this observation may be attributed to the fact that the 3'-Cl analog, in addition to its relatively high potency at the dopamine transporter, is also a potent muscarinic M₁ ligand.

The relatively high affinity of fluoro-substituted analogs concurrent with their relatively greater cocaine-like efficacy compared to analogs substituted with other moieties suggests that the general absence of cocaine-like effects in the latter compounds might be due to a lack of sufficient affinity for the dopamine transporter. Mitigating against this interpretation are the findings that several of the compounds showing little or no cocaine-like behavioral activity had affinities for the dopamine transporter that were similar to those of several of the relatively efficacious fluoro-substituted 3-diphenylmethoxytropane analogs. For example, 4'-Cl-BZT, 4',4"-diCl-BZT, and 3',4'-dichloro-3-(diphenylmethoxy)tropane) had affinities for the dopamine transporter ranging from 20 to 30 nM. In comparison, the affinities of the fluoro-substituted analogs ranged from 11.8 to 32 nM, encompassing those of the chloro-substituted compounds. Thus, affinity for the dopamine transporter alone cannot account for the reduced cocaine-like efficacy of many of the 3-diphenylmethoxytropane analogs.

One objective of the present study was to examine potential mechanisms contributing to the in vivo differences between the 3-diphenylmethoxytropane analogs and cocaine. The comparative actions of these drugs at the dopamine transporter were examined using displacement of [³H]WIN 35,428 binding and the inhibition of dopamine uptake. Because BZT is known for its antimuscarinic effects, these compounds were also examined for their affinity at muscarinic M₁ receptors; that activity was examined as a potential reason for the differences between these compounds and cocaine. In addition, the durations of action of several of the drugs were assessed in order to determine if cocaine-like behavioral effects could be obtained with a longer time for absorption and distribution to relevant sites of action.

It is tempting to speculate that the reduced cocaine-like activity of the compounds, particularly those with lower affinity for the dopamine transporter, was due to other prepotent effects predominating and interfering with the expression of cocaine-like behavioral effects. For example, cocaine-like behavioral effects of meperidine can be obscured by prepotent opioid activity that decreases response rates. When the opioid activity is blocked by coadministration of an opioid antagonist, higher meperidine doses, which produce cocaine-like effects, can be administered (Izenwasser et al., 1996). For the present compounds one possible action that might interfere with the expression of cocaine-like effects is antagonist effects at muscarinic receptors. Antimuscarinic effects of these drugs might decrease response rates at doses lower than those that produce cocaine-like effects, obscuring their cocaine-like pharmacology. Support for this interpretation, however, was not provided by the linear regression analysis of M₁ affinity and ED₅₀ values. Furthermore, more direct evidence indicates that antimuscarinic effects did not preclude the expression of cocaine-like behavioral activity. The muscarinic antagonists, atropine and scopolamine, stimulated locomotor activity, albeit with less efficacy than cocaine. This result is inconsistent with an interference with cocaine-like locomotor stimulant effects and suggests that, if anything, antimuscarinic activity of the 3-diphenylmethoxytropane analogs would potentiate rather than diminish cocaine like activity. Evidence for a potentiation of cocaine-induced stimulation of locomotor activity by the antimuscarinics was obtained in the studies of the interactions among these drugs. Further in the cocaine-discrimination procedure, atropine was ineffective in attenuating the discriminative stimulus effects of cocaine. Moreover, in further studies of these interactions, both atropine and scopolamine potentiated rather than antagonized the discriminative stimulus effects of cocaine. This potentiation occurred at doses of the two muscarinic antagonists that are consistent with their relative in vivo potencies for producing other antimuscarinic effects (McKeon, 1967; Malick and Barnett, 1975), suggesting that the potentiation was indeed due to antimuscarinic effects. These results are consistent with several previous studies that clearly demonstrate a similar potentiation of stimulants by antimuscarinics. For example, atropine and scopolamine augment the stimulation of avoidance behavior produced by cocaine (Scheckel and Boff, 1964). Together with the present findings, these results indicate that the antimuscarinic actions of 3-diphenylmethoxytropane analogs are incapable of attenuating any cocaine-like actions that they might have.

As with the stimulation of locomotor activity, the potencies of the 3-diphenylmethoxytropane analogs for decreasing response rates in the cocaine-discrimination studies were not related to their K_i values for displacement of [³H]WIN 35,428. These results suggest that neither of these behavioral effects are simply related to activity at the dopamine transporter, as might have been expected. However, the relation between transporter binding and ED₅₀ value for decreasing response rates approached significance (p = .099), suggesting that activity at the dopamine transporter is a factor influencing the effects of these drugs on rates of responding.

Initial observations that 4'-Cl-BZT lacked cocaine-like discriminative stimulus effects (Newman et al., 1994) led to studies of time course of action to determine whether the relatively lower efficacy of these compounds was due to an insufficient time for distribution of drug to relevant central nervous system sites of action. 4',4"-diF-BZT produced a long-lasting stimulation of locomotor activity, although it was not significantly greater than the stimulation obtained in the first 60 min after injection, and it was less than that produced by cocaine. For those compounds that have been studied to date, there is no evidence that the present estimates of locomotor stimulant potency or efficacy would be increased with longer pretreatment times. With the exception of 4',4"-diF-BZT, a similar conclusion could be made for the discriminative stimulus effects of these drugs. In contrast to the other compounds studied, 4',4"-diF-BZT showed increased efficacy in the cocaine-discrimination procedure when the time between injection and testing was between 30 and 90 min compared to when subjects were tested 5 min after injection. Structurally 4',4"-diF-BZT, more than the other 3-diphenylmethoxytropane analogs, resembles the selective dopamine uptake inhibitor, GBR 12909, which also has a 4',4"-difluoro-substituted diphenyl ether system. GBR 12909 also has been reported to share discriminative stimulus effects with cocaine (e.g., Witkin et al., 1991). However, it is currently unclear what mechanism may account for the enhanced cocaine-like pharmacology of 4',4"-diF-BZT compared to others of the 3-diphenylmethoxytropane analogs.

The absence of substitution for cocaine by the other 3-diphenylmethoxytropane analogs in the discrimination procedure and the low efficacy of these drugs in stimulating locomotor activity contrasts with their affinity for the dopamine transporter and their in vitro inhibition of dopamine uptake. Previous studies with BZT (e.g., Church et al., 1987) and preliminary studies with 4'-Cl-BZT (Tolliver et al., 1998) also indicate that these drugs can inhibit dopamine uptake in vivo. These results suggest behavioral effects similar to those of cocaine. It is possible that other actions or pharmacological properties of the 3-diphenylmethoxytropane analogs interfered with their cocaine-like activity; however, it is also possible that the regulation of behavioral activity through their actions at the dopamine transporter is different for the 3-diphenylmethoxytropane analogs and cocaine-like drugs. Consistent with that interpretation is the finding that, in contrast to analogs of cocaine and WIN 35,428, the binding of 3-diphenylmethoxytropane analogs to the dopamine transporter modeled better for single-site binding than it did for two-site binding. As mentioned above, for the cocaine and WIN 35,428 analogs there is a good correlation between affinity for the dopamine transporter and in vivo potency (e.g., Cline et al., 1992; Izenwasser et al., 1994); however that relationship is not found with structurally diverse dopamine uptake inhibitors (the present study; Vaugeois et al., 1993; Izenwasser et al., 1994). Finally, the structure-activity relations among these drugs are distinctly different from those obtained with cocaine and WIN 35,428 analogs (Newman et al., 1994, 1995), suggesting that these compounds are accessing a different binding domain than that accessed by cocaine. Identification of the precise domain on the dopamine transporter accessed by different classes of small molecular structures and the functional consequences of these binding interactions will greatly advance our understanding of how different dopamine uptake inhibitors act, why some are subject to abuse and others are not, and may provide leads for the development of pharmacotherapies for the treatment of cocaine dependence.