Introduction

Top Abstract Introduction Methods Results Discussion References

Mg⁺⁺ is a mineral nutrient that is abundant in the brain as well as in the periphery (Aikawa, 1971). Several pharmacological studies have shown that systemically administered MgCl₂ has a profile of behavioral effects that are similar to those of indirect and direct monoamine receptor agonists. These include a heightening of aggression and an enhancement of cocaine-induced aggression, apomorphine-induced sniffing and amphetamine-induced locomotion in mice (Izenwasser et al., 1986; Kantak, 1989; Kantak and Adlerstein, 1990). In a mouse conditioned place preference procedure, MgCl₂ was shown to induce a place preference (Lawley and Kantak, 1990b). Furthermore, MgCl₂ and amphetamine, but not haloperidol or pentobarbital, increased the magnitude of place preference induced by cocaine (Lawley and Kantak, 1990a). In groups of rats self-administering cocaine under FR 1 or FR 5 schedules of drug delivery, MgCl₂ was found to dose-dependently substitute for cocaine and to maintain a constant level of drug intake over a 10-day period (Kantak et al., 1991). MgCl₂ also maintained responding above saline rates in substitution tests conducted under progressive-ratio schedules of drug delivery in cocaine-trained rats (Kantak et al., 1991). Progressive-ratio breakpoints, however, were greater after cocaine availability than they were after MgCl₂ availability.

MgCl₂ can engender stimulus effects that share some common features with cocaine and NMDA associated ion channel blockers. Drug discrimination studies demonstrated that under a 2 mg/kg cocaine training dose condition, MgCl₂, as well as dizocilpine and PCP, engendered full substitution (90% responses) for cocaine in the majority of subjects tested (Kantak et al., 1995). These same drugs failed to substitute for cocaine when subjects were trained to discriminate a higher dose (10 mg/kg) of cocaine. A competitive NMDA antagonist did not substitute for cocaine under either training dose condition. Others also have reported that the NMDA-associated channel blockers dizocilpine, PCP and ketamine partially substituted in cocaine-trained rats with a rank order corresponding to their potency for blocking the NMDA-associated ion channel (Koek et al., 1989).

These findings raise the question of whether MgCl₂ itself would have discriminative stimulus effects independent of a prior drug history. Previous work, for example, has shown that rats will not self-administer MgCl₂ if they had no prior drug experience (Kantak et al., 1990). However, conditioned place preference studies demonstrated that MgCl₂ could induce a change in preference in drug-naive mice (Lawley and Kantak, 1990b). The overlap in the behavioral effects of MgCl₂ with those of cocaine and NMDA antagonists described above prompted us to characterize the discriminative stimulus effects of MgCl₂ with respect to NMDA receptor and monoamine transporter functions. Neurochemical studies indicate that Mg⁺⁺ functions not only as a noncompetitive NMDA antagonist by blocking the NMDA-associated ion channel (Wong and Kemp, 1991), but also as an important cofactor for monamine neurotransmitter binding to transporters (Amejdki-Chab et al., 1992; Hendley et al., 1988; White, 1975) and receptors (Hamblin and Creese, 1982; Lefkowitz et al., 1976; Norman et al., 1985; Salama et al., 1982). Substitution tests were therefore conducted with the NMDA-associated ion channel blockers dizocilpine and PCP (Wong and Kemp,1991), the competitive NMDA antagonist NPC 17742 (Ferkany et al., 1993), the nonselective monoamine uptake inhibitor cocaine (Koe, 1976), the selective DA uptake inhibitor GBR 12909 (van der Zee et al., 1980), the selective NE uptake inhibitor talsupram and the selective 5-HT uptake inhibitor citalopram (Hyttel, 1982).

	Methods

Top Abstract Introduction Methods Results Discussion References

Subjects. Between experimental sessions, male Wistar rats (Charles River Breeding Labs, Portage, MI) were housed in individual stainless steel cages (24 × 18 × 18 cm). Home cages were located within a temperature- (74 ± 4° F) and a light- (0800 hr on, 2000 hr off) controlled vivarium. All rats were experimentally naive at the beginning of the study. The rats had continuous access to water in their home cages. Food (Agway Prolab Rodent Chow, Syracuse, NY) was restricted to 16 g/day and was provided after daily sessions. This ration of food maintained body weights at approximately 85% of ad libitum values. For the study duration, body weights ranged from 275 ± 12 to 435 ± 9 g in experiments involving substitution with various monoamine uptake inhibitors and NMDA antagonists and from 327 ± 5 to 478 ± 10 g in experiments involving peripheral vs. central administration of MgCl₂.

Apparatus. Each of four identical experimental chambers for rats (Gerbrands, model A, Waltham, MA) was equipped with two response levers that were mounted 7.6 cm apart. A pellet dispenser, which emitted an audible click when operated, delivered 45 mg food pellets (Noyes, Traditional Formula, Lancaster, NH) to a receptacle located between the levers. A sound-attenuating cubicle enclosed each chamber. In each cubicle, an overhead light provided general illumination and a fan provided ventilation and masked extraneous sounds. Experimental events were controlled by a 286 AT-compatible computer that was programmed in Medstate Notation and connected to an interface (Med Associates, East Fairfield, VT).

Drugs. The drugs studied were: cocaine hydrochloride and PCP hydrochloride (NIDA, Rockville, MD), dizocilpine maleate [(+)-MK 801] (Merck, Sharp and Dohme, West Point, PA), NPC 17742 (NOVA Pharmaceutical Corp., Baltimore, MD), MgCl₂ ·6H₂O (Fisher Scientific, Medford, MA), GBR 12909 dihydrochloride (Research Biochemicals Inc., Natick, MA), citalopram hydrobromide and talsupram hydrochloride (Lundbeck A/S, Copenhagen, Denmark). Except as noted below, drugs were dissolved and diluted to desired concentrations in either sterile 0.9% saline or sterile distilled water. Doses are expressed as salts (the anhydrous salt for MgCl₂). GBR 12909 (30 mg/ml) was dissolved in 0.9% saline and 1 M acetic acid (8:2), heated to 55°C and diluted to desired concentrations with 0.9% saline. NPC 17742 (30 mg/ml) was dissolved in 100 mM sodium hydroxide and diluted to desired concentrations with distilled water. MgCl₂ was either administered s.c. in a volume of 3.0 ml/kg body weight 15 min before the session (Kantak et al., 1992, 1995) or infused i.c.v. in a volume of 5 µl over a 5-min period immediately before the session (Buck et al., 1979; Willetts and Balster, 1988). Rats were gently held by the experimenter during infusions. The infusion cap was left in place for one min after the infusion. Other drugs were administered i.p. 15 min before the session in a volume of 1.0 ml/kg body weight.

Drug discrimination procedure. A total of 14 rats was trained to discriminate s.c. injections of 100 mg/kg MgCl₂ from saline. After administration of MgCl₂, 10 responses (FR 10) on one lever resulted in delivery of a food pellet, whereas after administration of saline, 10 responses on the other lever produced a food pellet. Training sessions were conducted daily (Monday to Friday) and lasted for a total of 15 min. Different sequences of randomly assigned saline and drug training sessions were used for each subject to ensure that any olfactory cues associated with the two levers would not bias the discrimination (Extance and Goudie, 1981). The training criteria consisted of at least 10 consecutive sessions in which 90% of responses were made on the injection-appropriate lever and the total number of responses on both levers did not exceed 12 before the first food pellet was delivered.

Drug testing procedure. Drug test sessions were conducted once or twice per week, with training sessions scheduled on intervening days. Test sessions were conducted only if performance during the preceding training session met the criteria previously described. Test sessions were identical to training sessions except that 10 responses on either lever resulted in delivery of a food pellet. In experiments involving substitution tests with monoamine uptake inhibitors and NMDA antagonists, a range of doses of s.c. MgCl₂ initially was examined in 10 subjects. Next, a range of doses of each of the four monoamine uptake inhibitors and each of the three NMDA antagonists was examined in subsets of six subjects to determine the degree to which each compound substituted for the 100 mg/kg training dose of MgCl₂. Most subjects received three or four of the seven test compounds; two subjects received six of the test compounds and one subject received all seven test compounds. Among the test compounds examined in each subject, at least one was an uptake inhibitor and one was a NMDA antagonist. Drugs were studied in different orders with different subjects, and saline test sessions were conducted on two or three occasions in each subject at various times during the experiment. In experiments involving peripheral vs. central administration of MgCl₂, the remaining subjects were implanted with a chronic indwelling 23-ga. stainless steel guide cannula ending 1 mm above the right lateral ventricle of the brain. Before surgery, rats were anesthetized with a combination of ketamine and xylazine (90 and 10 mg/kg, i.p., respectively). Coordinates for cannula placement were according to the atlas of Pellegrino et al. (1979): 0.6 mm posterior to bregma, 2.0 mm lateral from midline and 2.7 mm ventral from the surface of the skull. A 30-ga. stainless steel stylus was inserted into the guide cannula between infusions. The rats recovered for 4 days and then training with s.c. injections resumed. A range of doses of s.c. MgCl₂ and saline was subsequently examined in all four subjects. Next, a range of doses of i.c.v. MgCl₂ and saline was examined on two or three occasions in each subject to determine the degree to which i.c.v. MgCl₂ substituted for the 100 mg/kg training dose of MgCl₂.

Analysis of drug effects. The percentage of MgCl₂-appropriate responses was determined for each subject by dividing the number of MgCl₂-associated lever responses by the total number of responses emitted on both levers. Data were not included in the analysis if less than 10 responses were made during a test session. When multiple determinations were obtained for a particular dose in the same animal (see above), values were averaged together for individual subjects before group averages were calculated. Where appropriate, the dose of a drug estimated to engender 50% MgCl₂-appropriate responses (ED₅₀ ± 95% CI) was calculated by linear regression analysis over the ascending linear portion of the log dose-response curve. Using Pharmacologic Calculation System software, the log dose-response curves were further analyzed with the parallel line assay of Finney (1964) to ascertain differences in potency relative to MgCl₂. In the first study, each mg/kg dose of the test compounds was first converted to µmol/kg to account for differences in their molecular weights. In the second study, the relative potency between s.c. and i.c.v. MgCl₂ was computed after converting mg/kg s.c. MgCl₂ to µg/rat. For purposes of this study, full substitution for MgCl₂ was defined as 90% MgCl₂-appropriate responses.

	Results

Top Abstract Introduction Methods Results Discussion References

Discrimination acquisition and effects of s.c. administered MgCl₂. The 10 rats used in this study met the training criteria after an average of 35 ± 2 training sessions (range 19-40 sessions). At criteria, rats made an average of 99% ± 0.3 responses on the MgCl₂-appropriate lever after injections of 100 mg/kg MgCl₂ and 98% ± 0.6 responses on the saline-appropriate lever after injections of saline. Throughout training, the rate of responding was slightly lower after MgCl₂ injections (.84 ± .01 responses per second) than after saline injections (1.1 ± .01 responses per second).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Effects of MgCl2 in rats trained to discriminate 100 mg/kg MgCl2 from saline. Abscissae: dose, log scale; ordinates: percentage of responses on the MgCl2-appropriate lever (top panel) and rate of responding (bottom panel). Points are the means (± S.E.M.) based on 10 rats, except as noted. The numbers above each point refer to the proportion of rats that made 90% MgCl2-appropriate responses after administration of the dose indicated on the abscissa. Points above S show the effects of saline.

Substitution tests with monoamine uptake inhibitors. Dose-related increases in the percentage of MgCl₂-appropriate responses (fig. 2, top panel) were engendered by 0.3 to 10.0 mg/kg cocaine, 3.0 to 17.8 mg/kg GBR 12909, 0.1 to 5.6 mg/kg talsupram and 1.0 to 10 mg/kg citalopram. Full substitution for MgCl₂ was observed in all six subjects after one or more doses of cocaine in the range of 1.0 to 10.0 mg/kg. For the group of six rats, the average percentage of MgCl₂-appropriate responses reached a maximum of 82% after 3.0 mg/kg cocaine and was 50% after 10.0 mg/kg cocaine. One or more doses of GBR 12909 engendered full substitution for MgCl₂ in five of six subjects in the dose range of 3.0 to 17.8 mg/kg. Averaged for the group of rats, the maximum percentage of MgCl₂-appropriate responses reached 74% after 17.8 mg/kg GBR 12909. One or more doses of talsupram also engendered full substitution for MgCl₂ in five of the six subjects in the dose range of 0.1 to 5.6 mg/kg. For the group of six rats, the maximum average percentage of MgCl₂-appropriate responses reached 72% after 5.6 mg/kg talsupram. Finally, full substitution for MgCl₂ was observed in five of the six subjects after citalopram in the dose range of 3.0 to 10.0 mg/kg. The maximum average percentage of MgCl₂-appropriate responses reached 82% after 10.0 mg/kg citalopram for the group of six rats. Based on ED₅₀ and relative potency analysis (table 1), the rank order of potency of these drugs for producing MgCl₂-appropriate responses was talsupram = cocaine > citalopram = GBR 12909 > MgCl₂. Each monoamine uptake inhibitor, except cocaine, produced dose-related decreases in the response rate, with significant (P < .05) reductions in the rate after the highest doses tested (fig. 2, bottom panel).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Effects of cocaine, GBR 12909, talsupram and citalopram in rats trained to discriminate 100 mg/kg MgCl2 from saline. Points are the means (± S.E.M.) based on six rats, except as noted. Saline values and other details are as in figure 1.

Substitution tests with NMDA antagonists. Administration of 0.03 to 0.18 mg/kg dizocilpine and 0.3 to 3.0 mg/kg PCP resulted in dose-related changes in the percentage of MgCl₂-appropriate responses (fig. 3, top). Full substitution for MgCl₂ was observed in three of six subjects after one or more doses of dizocilpine in the range of 0.056 to 0.18 mg/kg. The maximum average percentage of MgCl₂-appropriate responses reached 50% after 0.1 mg/kg dizocilpine for the group of six rats. One or more doses of PCP in the range of 1.0 to 3.0 mg/kg engendered full substitution for MgCl₂ in four of six subjects and the maximum average percentage of MgCl₂-appropriate responses reached 65% after 1.0 mg/kg PCP for the group of six rats. The competitive NMDA antagonist NPC 17742 (0.3-30 mg/kg) did not produce any dose-related change in the percentage of MgCl₂-appropriate responses (fig. 3, top). Responses were mainly made on the saline-associated lever in most animals after each dose of NPC 17742. The average percentage of MgCl₂-appropriate responses was 49% or less across the range of doses for the group of rats. Although NPC 17742 produced full substitution for MgCl₂ in four of six subjects, its occurrence was irregularly related to dose (between 0.3-17.8 mg/kg for individual animals). As shown in the bottom of figure 3, each NMDA antagonist produced dose-related decreases in the response rate, with significant (P < .05) reductions in the rate after the two highest doses tested.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Effects of dizocilpine, PCP and NPC 17742 in rats trained to discriminate 100 mg/kg MgCl2 from saline. Points are the means (± S.E.M.) based on six rats, except as noted. Saline values and other details are as in figure 1.

Effects of peripherally vs. centrally administered MgCl₂. In experiments involving s.c. vs. i.c.v. injections of MgCl₂, the rats met the training criteria after an average of 46 ± 4 training sessions (range 24-57 sessions). After training, s.c. injections of MgCl₂ (3-178 mg/kg) engendered dose-related increases in the percentage of MgCl₂-appropriate responses, with an ED₅₀ = 31.6 mg/kg (95% CI = 15.1-66.1). Full substitution was observed in all subjects after doses that were equal to or greater than the training dose (fig. 4, top left). These results are quite similar to those reported in figure 1.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Effects of s.c. MgCl2 (left) and i.c.v. MgCl2 (right) in rats trained to discriminate 100 mg/kg MgCl2 from saline. Points are the means (± S.E.M.) based on four rats. Other details are as in figure 1.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Our results indicate that MgCl₂ has discriminative stimulus effects that were reproducibly acquired and involved central mechanisms. Furthermore, the discriminative stimulus effects of MgCl₂ were less similar to other drugs with NMDA antagonist effects and more similar to monoamine uptake inhibitors. After a moderate number of training sessions, 100 to 300 mg/kg MgCl₂ fully substituted for the 100 mg/kg training dose. Lower doses engendered fewer responses on the MgCl₂-appropriate lever. The S-shaped dose-response function that resulted from training rats to discriminate 100 mg/kg MgCl₂ is a general feature of most drugs having discriminative stimulus effects (see Jarbe, 1989 for review).

An important question that can be raised relates to the pharmacological specificity of the discriminative stimulus effects of MgCl₂. Because the concentration of the 100 mg/kg training dose was 350 mM, it is possible that the solution's hypertonicity rather than its pharmacological properties caused the cuing effects of MgCl₂. Controls for hypertonicity were not used as part of our experiments. Arguments against this possibility include the fact that the 30 µg solution was hypotonic (62.5 mM), yet its i.c.v. infusion fully substituted for the 100 mg/kg training dose. Other studies have shown that the i.v. delivery of hypertonic (335 mM) NaCl did not alter self-administration responding any differently than physiological (154 mM) NaCl (Kantak et al., 1990; 1991). Furthermore, the differential effects of the NMDA-associated ion channel blockers vs. monoamine uptake inhibitors and the relatively weak effects of NPC 17742 in substitution tests suggest some degree of pharmacological specificity to the discriminative stimulus effects of MgCl₂ (Colpaert et al., 1979; Terry et al., 1994).

The biphasic, dose-related changes in drug-appropriate responses after i.c.v. administration of MgCl₂ have generally not been observed after the i.c.v. administration of other drugs (e.g., Sannerud et al., 1991; Willetts and Balster, 1988; Wood et al., 1987). In these other studies, the dose-related changes in drug-appropriate responses tended to be linear. In our study, drug-appropriate responses were linearly related to MgCl₂ dose, up to 100 µg. The i.c.v. infusion of our highest dose of MgCl₂ (300 µg) resulted in much less MgCl₂-appropriate responses than lower doses (30 or 100 µg). The effects of 300 µg MgCl₂ were most likely nonspecific, which could be related to a severe depression of cortical neurons that is produced by high concentrations of centrally applied Mg⁺⁺ ions (Kelly et al., 1969).

In other drug discrimination studies, i.c.v. infusions of 80 µg cocaine (Wood et al., 1987), 150 µg PCP (Willetts and Balster, 1988) and 44 µg midazolam (Sannerud et al., 1991) fully substituted for their systemic injections, suggesting central mediation of their discriminative stimulus effects. Furthermore, the potency of these drugs was up to 40-fold greater after direct infusion into the lateral ventricle of the brain. From comparing s.c. and i.c.v. administration of MgCl₂, it is apparent that all doses of i.c.v. administered MgCl₂ studied would have engendered saline-appropriate responses if systemically administered. Therefore, diffusion out of the brain could not have accounted for the full substitution by 30 µg i.c.v. MgCl₂ for the 100 mg/kg s.c. training dose. The i.c.v. administration of 17 and 50 µg MgCl₂ has previously been shown to increase CSF concentrations of Mg⁺⁺ by 2- and 6-fold, respectively, and to dose-dependently decrease audiogenic seizure severity in rats without altering serum concentrations of Mg⁺⁺ (Buck et al., 1979).

Although cannulae placements were not verified in this study, central mediation of the discriminative stimulus effects of MgCl₂ also is suggested by the finding that MgCl₂ was more than 4000 times more potent after its central vs. peripheral administration. The greater increase in potency after direct infusion of MgCl₂ into the lateral ventricle of the brain compared to other drugs (Sannerud et al., 1991; Willetts and Balster, 1988; Wood et al., 1987) may be related to the fact that only small amounts of MgCl₂ are absorbed and distributed in the brain after its peripheral administration (Hilmy and Somjen, 1968), yet central neurons are quite sensitive to a relatively small concentration (1 mM) of iontophoretically applied Mg⁺⁺ (Somjen and Kato, 1968). This may account for the relatively large dose of MgCl₂ (100 mg/kg or an average of 44,000 µg/rat) necessary for maintaining discriminative stimulus control after a s.c. injection and the relatively small dose of MgCl₂ (30 µg) necessary for full substitution after infusion into the lateral ventricle of the brain.

Among its biochemical effects in the brain, Mg⁺⁺ functions as a noncompetitive NMDA antagonist by binding to a specific Mg⁺⁺ recognition site and blocking the NMDA-associated ion channel (Wong and Kemp, 1991). However, the results of our study suggest that the discriminative stimulus effects of MgCl₂ were distinct from other NMDA antagonists. The NMDA-associated ion channel blockers PCP and dizocilpine only partially substituted for MgCl₂. The MgCl₂-like effects of NPC 17742 were weak and not systematically related to dose. Responses were mostly made on the saline-associated lever over the entire dose range. A previous study in PCP-trained rats demonstrated that 80 mg/kg MgCl₂ engendered only 21% PCP-appropriate responses, yet the rate of responding was significantly decreased, suggesting that a behaviorally effective dose of MgCl₂ was used (Jortani et al., 1992). In other studies, a competitive NMDA antagonist did not substitute for PCP (Willetts and Balster, 1988) and PCP did not substitute for a competitive NMDA antagonist (Willetts et al., 1989). Taken together, these findings show that the discriminative stimulus effects of MgCl₂ do not completely overlap with those of other NMDA-associated ion channel blockers or competitive NMDA antagonists, which also do not completely overlap with each other.

The discriminative stimulus effects of MgCl₂ do appear to overlap with the nonselective monoamine uptake inhibitor cocaine and the selective monoamine uptake inhibitors GBR 12909, talsupram and citalopram. The nonselective monoamine uptake inhibitor cocaine engendered 82% MgCl₂-appropriate responses after a dose of 3 mg/kg. These results are consistent with the study showing that 30 to 300 mg/kg MgCl₂ engendered 73 to 96% cocaine-appropriate responses in rats trained to discriminate 2 mg/kg cocaine (Kantak et al., 1995). The selective DA, NE and 5-HT uptake inhibitors fully substituted for MgCl₂ in most animals and engendered a higher degree of MgCl₂-appropriate responses than the NMDA antagonists. These findings suggest that the discriminative stimulus effects of MgCl₂ may involve more than its neurochemical effect at the NMDA receptor. Indeed, Mg⁺⁺ has multiple biochemical sites of action in the brain (Ebel and Gunther, 1980). The discriminative stimulus effects of MgCl₂ could involve interactions of MgCl₂ at either monoamine uptake sites, by inhibiting monoamine transporter functions, or monoamine receptor sites, by increasing monoamine binding and receptor activation. Although low mM concentrations of Mg⁺⁺ are necessary for maintaining normal rates of monoamine uptake (e.g., Amejdki-Chab et al., 1992; Hendley et al., 1988; White, 1975), Mg⁺⁺ concentrations of more than 10 mM can inhibit the uptake of DA and reduce specific binding of [³H]GBR 12783 or [³H]GBR 12935 with an IC₅₀ value of 16 mM (Amejdki-Chab et al., 1992; Janowsky et al., 1986). From the i.c.v. infusion of 30 µg MgCl₂, which fully substituted for the 100 mg/kg training dose, the concentration of MgCl₂ delivered to the lateral ventricle was 62.5 mM. This concentration would be diluted at least 100-fold after the distribution of MgCl₂ into brain tissue. If DA, NE and 5-HT transporters are similarly inhibited by concentrations of Mg⁺⁺ of more than 10 mM in vivo, then the concentration of MgCl₂, either from s.c. injections or i.c.v. infusions, may not have reached high enough levels to inhibit monoamine transporter functions in brain tissue.

Mg⁺⁺ has been shown to be necessary for the binding of monoamine neurotransmitters at D₁, D₂, ₂-NE, -NE and 5-HT_1a receptors (e.g., Hamblin and Creese, 1982; Lefkowitz et al., 1976; Norman et al., 1985; Salama et al., 1982). Dopamine binding studies, for example, have demonstrated that Mg⁺⁺ increases the B_max associated with ³H-spiroperidol binding in rat striata by 200% above control (Usdin et al., 1980). This increase is seen with concentrations of Mg⁺⁺ as low as 0.1 to 1.0 mM. Mg⁺⁺ does not actually increase the number of receptors, but prevents the time-dependent degradation of ligand binding to the receptor. Other studies (De Vries and Beart, 1985; Hamblin and Creese, 1982) demonstrated that Mg⁺⁺ enhances DA affinity at D₂ receptors by a factor of 1000 (a µM to nM potency change). Given that the monoamine uptake inhibitors used in our study could be regarded as indirect monoamine receptor agonists, it is more likely that monoamine receptor activation, rather than monoamine uptake inhibition, could be the common mechanism by which MgCl₂, cocaine, GBR 12909, talsupram and citalopram engendered a high degree of MgCl₂-appropriate responses. Further characterization of the discriminative stimulus effects of MgCl₂ should include antagonism studies with NMDA and selective monoamine receptor blockers to evaluate the role of the Mg⁺⁺ site on the NMDA receptor complex vs. monoamine receptor activation in mediating the effects of MgCl₂. Furthermore, NE mechanisms may be particularly important for the discriminative stimulus effects of MgCl₂. The potency of talsupram was equal to that of cocaine and more than that of citalopram and GBR 12909 in engendering MgCl₂-appropriate responses. Interestingly, as with MgCl₂ (Kantak et al., 1995), selective NE uptake inhibitors engendered cocaine-appropriate responses under a low-dose, but not a high-dose, training condition (Spealman, 1995; Terry et al., 1994). To eliminate any pharmacokinetic differences among talsupram, cocaine, citalopram and GBR 12909, additional substitution studies with the i.c.v. administration of these uptake inhibitors should also be conducted to determine if talsupram is still the most potent in engendering MgCl₂-appropriate responses. If the rank order of potency is the same in this context as it was when these drugs were given systemically, then these potency differences would indicate a predominance of NE mechanisms in discriminative stimulus effects of MgCl₂.

In summary, the discriminative stimulus effects of MgCl₂ are novel findings and add to the growing evidence that MgCl₂ can have pharmacological effects (e.g., Buck et al., 1979; Kantak et al., 1992; 1995; McIntosh et al., 1989; Smith et al., 1993). The discriminative stimulus effects of MgCl₂ appear to be centrally mediated and may involve interactions of MgCl₂ at monoamine receptor sites.