Effects of muscarinic receptor antagonists on cocaine discrimination in wild-type mice and in muscarinic receptor M₁, M₂, and M₄ receptor knockout mice

Abstract

Muscarinic M₁/M₄ stimulation can reduce abuse-related effects of cocaine and may represent avenues for treating cocaine addiction. Muscarinic antagonists can mimic and enhance effects of cocaine, including discriminative stimulus (S^D) effects, but the receptor subtypes mediating those effects are not known. A better understanding of the complex cocaine/muscarinic interactions is needed to evaluate and develop potential muscarinic-based medications. Here, knockout mice lacking M₁, M₂, or M₄ receptors (M₁^-/-, M₂^-/-, M₄^-/-), as well as control wild-type mice and outbred Swiss-Webster mice, were trained to discriminate 10 mg/kg cocaine from saline. Muscarinic receptor antagonists with no subtype selectivity (scopolamine), or preferential affinity at the M₁, M₂, or M₄ subtype (telenzepine, trihexyphenidyl; methoctramine, AQ-RA 741; tropicamide) were tested alone and in combination with cocaine. In intact animals, antagonists with high affinity at M₁/M₄ receptors partially substituted for cocaine and increased the S^D effect of cocaine, while M₂-preferring antagonists did not substitute, and reduced the S^D effect of cocaine. The cocaine-like effects of scopolamine were absent in M₁^-/- mice. The cocaine S^D attenuating effects of methoctramine were absent in M₂^-/- mice and almost absent in M₁^-/- mice. The findings indicate that the cocaine-like S^D effects of muscarinic antagonists are primarily mediated through M₁ receptors, with a minor contribution of M₄ receptors. The data also support our previous findings that stimulation of M₁ receptors and M₄ receptors can each attenuate the S^D effect of cocaine, and show that this can also be achieved by blocking M₂ autoreceptors, likely via increased acetylcholine release.

1. Introduction

There are currently no approved or widely effective medications to curb addiction to psychostimulant drugs like cocaine. Manipulations targeting muscarinic cholinergic systems can modulate abuse-related effects of cocaine, and muscarinic receptors are thus emerging as potential targets for medications development in cocaine addiction [1]. We have previously shown that agonists or positive allosteric modulators that selectively stimulate either the M₁ or the M₄ muscarinic receptor subtypes attenuate the discriminative stimulus (S^D) ¹effects and reinforcing effects of cocaine in rats and mice ([2-5]; [6]). Studies in knockout mice that lack M₁ receptors (M₁^-/-), M₄ receptors (M₄^-/-), or both receptors, suggested that both receptor subtypes mediate the attenuation of cocaine's S^D effects, while other receptors, likely M₃, mediate the rate-suppressing side effects that were observed with less selective ligands [2, 5]. Conversely, muscarinic receptor antagonists increased the S^D effect of cocaine, and can produce cocaine-appropriate responding when substituted for cocaine [2, 7-9].

Based on rat studies that used intracranial infusions or lesions, striatal areas appear to be central to mediating both the S^D effects of cocaine [10, 11], and the modulation of the cocaine S^D effect by muscarinic receptor ligands [12-15]. Striatal tissues express predominantly the M₁ and M₄ muscarinic receptor subtypes, and lower densities of the M₂ subtype: M₁ receptors mostly postsynaptically, M₄ receptors both pre-and postsynaptically, and M₂ receptors mostly as presynaptic inhibitory autoreceptors [16-20]. M₂ and M₄ receptors modulate the tonic acetylcholine release by striatal cholinergic interneurons, which in turn modulates striatal dopamine release via a nicotinic receptor-dependent mechanism [21-25]. Postsynaptically, M₁ and M₄ receptors modulate the excitability and activity patterns of GABAergic medium spiny neurons, the striatum's main output neurons. M₁ and M₄ receptor activation has long been known to produce functional dopamine antagonism, but more recent studies are showing the reciprocal modulation of striatal dopamine and acetylcholine release to be quite complex (for review, see [26-29]).

Here, we used subtype-preferring muscarinic receptor antagonists and knockout mice lacking M₁, M₂, or M₄ receptors to investigate the contributions of specific muscarinic receptor subtypes in the muscarinic/cocaine interaction in the drug discrimination assay. Scopolamine is a non-selective muscarinic antagonist with comparable affinities at the M₁ and M₃-M5 subtypes and a marginally lower affinity at the M₂ subtype [30-32]. Telenzepine is a moderately M₁-preferring antagonist, and trihexyphenidyl has about equal affinity at M₁ and M₄ receptors, with modest selectivity over M₂, M₃, and M5 subtypes [30, 33, 34]. Tropicamide was reported to be a modestly M₄-preferring antagonist with comparable affinities across M₁-M₃ subtypes [35]. Methoctramine and AQ-RA 741 are M₂-preferring antagonists, with methoctramine generally showing a more selective profile than AQ-RA 741 [33, 36-39]. Methoctramine binds at both the orthosteric site and an allosteric site on M₂ receptors, but the latter only at high concentrations less likely to be relevant in vivo [40, 41]. Each antagonist was either shown to be fully brain penetrant, or has been inferred to penetrate the central nervous system based on producing effects known to be centrally mediated, following systemic administration [32, 42-44].

2. Materials and Methods

2.1 Animals

Male Swiss-Webster, C57BL/6NTac, M₁^-/-, M₂^-/- and M₄^-/- mice were acquired from Taconic Farms (Germantown, NY) at 4-8 weeks of age, the knockout mice with the permission of Dr. Jürgen Wess. M₁^-/-, M₂^-/- and M₄^-/- mice were generated as described previously [45-47] and backcrossed 11 generations to C57BL/6NTac females. Age- and sex-matched wild-type C57BL/6NTac (Taconic Farms) mice served as controls. Mice were acclimated to the housing facilities for at least 7 days before training began, at no earlier than 7 weeks of age. Mice were kept in a 12-h light/dark cycle, group housed up to four per cage. Experiments were conducted during the light phase of the circadian cycle. Water was accessible ad libitum and food (rodent diet 5001; PMI Feeds, Inc., St. Louis, MO) was provided daily after training/testing sessions, 4 g/mouse/day. Rodent “treats”, nesting material, and exercise/nesting devices were provided for enrichment. Some mice had been tested previously with muscarinic receptor ligands (M₁ agonist and/or M₄ positive allosteric modulator) before the tests reported here. All procedures were approved by the McLean Hospital Institutional Animal Care and Use Committee and were carried out in accordance with the Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research and US laws.

2.2 Training and evaluation in cocaine discrimination

Operant-conditioning chambers and the experimental procedure were as previously described [2]. In brief, each chamber contained two nose-poke holes each equipped with a photocell and a cue light, and a cup into which liquid food was delivered from a syringe pump. Mice were trained to discriminate 10 mg/kg cocaine from saline (i.p.), reinforced with Vanilla-flavored Ensure nutrition drink. 30 reinforcers were available per 20-min session. Mice were trained initially under a FR 1 schedule, then, the ratio was gradually increased to FR 10, with increasing pretreatment time spent in the chamber rather than home cage. Eventually, sessions were preceded by the 10-min pretreatment period in the chamber, during which all lights were off, and responding had no scheduled consequences. Cocaine and saline were presented in pseudorandom order across daily training sessions, typically five days/week, and mice were counterbalanced with cocaine trained on the left or right nose-poke. Stable discrimination was defined as at least 7 of 8 consecutive sessions satisfying the following criteria: 1) ≥10 reinforcers earned per session, 2) ≥80% correct responses for the first reinforcer, and 3) ≥90% correct total responses.

Once criteria were met, mice were tested with saline, 0.32, 1.0, 3.2, 10, and 18 mg/kg cocaine to generate dose-effect functions. The non subtype-selective antagonist scopolamine (0.032–10 mg/kg i.p.), the M₁/M₄-preferring antagonists telenzepine (3.2-56 mg/kg s.c.) and trihexyphenidyl (0.032– 32 mg/kg i.p.), the M₄-preferring antagonist tropicamide (0.1–10 mg/kg s.c.), and the M₂-preferring antagonists AQ-RA 741 and methoctramine (0.1–3.2 mg/kg s.c.) were each tested alone (i.e., as “substitution” for cocaine stimulus). In addition, pretreatment/combination tests were conducted, in which cocaine doses were tested in combination with scopolamine (0.32 mg/kg i.p., administered with cocaine as a single injection), tropicamide (0.1 mg/kg s.c.), AQ-RA 741 (0.1–1 mg/kg s.c.), and methoctramine (0.0032–1.0 mg/kg s.c.). Pretreatment times before cocaine injection (when tested in combination) or before session start (when tested alone as substitution) were: telenzepine, 10 min; methoctramine, 15 min; all other drugs, immediately before. Doses were tested within-subjects in a counterbalanced sequence. At least one training session was interspersed between each test session, and tests were only performed when mice satisfied discrimination criteria. If responding was suppressed to less than 10 responses in a session, the quantity of behavior was considered insufficient to evaluate response selection and the percentage of drug-appropriate responding (DAR) was not included in the data set.

2.3 Drugs

Cocaine hydrochloride was supplied by the National Institute on Drug Abuse (National Institutes of Health, Bethesda, MD), Scopolamine hydrobromide, telenzepine dihydrochloride hydrate, trihexyphenidyl hydrochloride, tropicamide, and methoctramine hydrate were purchased from Sigma-Aldrich (St. Louis, MO). AQ-RA 741 was purchased from R&D Systems, Inc (formerly Tocris, Ellisville, MO). Cocaine, scopolamine, telenzepine, and AQ-RA 741 were dissolved in 0.9% saline, methoctramine, in sterile water. Trihexyphenidyl was dissolved by gentle heating in sterile water. Tropicamide was dissolved in ethanol and diluted to ≤1% ethanol in sterile water. All drug doses refer to the weights of the respective salts. Vehicles, route of administration, pretreatment times, and initial dose ranges were selected based on published reports, and adjusted empirically in initial studies [2, 44, 48-51].

2.4 Data analysis

The %DAR for the whole session and total rates of responding (i.e., in both holes) are presented. Comparable effects were observed in %DAR for the first reinforcer, unless stated otherwise. Repeated measures ANOVA were performed with dose of pretreatment drug and/or cocaine or dose of substitution drug, as variables, on %DAR and response rate. For knockout strain studies, ANOVA were performed with genotype as between-subjects variable and drug doses/pretreatments as repeated-measures variables. Occasionally, responding was eliminated or suppressed to the point that no reinforcers were earned during the session; in those cases, no %DAR was calculated for that mouse (i.e., missing value). Data are reported as group means with standard error of the mean. Significance level was set at P<0.05; statistical software was Stata/SE for Mac.

3. Results

3.1 “Cocaine-like” muscarinic receptor antagonists in Swiss-Webster mice

In outbred Swiss-Webster mice, the M₁/M₄-preferring antagonists telenzepine and trihexyphenidyl, and the M₄-preferring antagonist tropicamide, each produced some cocaine-appropriate responding, with peak averages between 38% and 61% (Fig. 1). DAR was related to antagonist dose for telenzepine [F(3,21)=4.07, P<0.05], trihexyphenidyl [F(3,39)=3.68, P<0.05], and tropicamide [F(5,39)=3.82, P<0.01]. Trihexyphenidyl decreased rates of responding at the highest dose tested (main effect [F(3,39)=14.3, P<0.0001]; 32mg/kg vs. vehicle P<0.001). Tropicamide and telenzepine did not affect rates of responding significantly.

Figure 1.

Dose-effect functions for cocaine and muscarinic M₁/M₄ receptor -preferring antagonists in Swiss-Webster mice, tested as substitutions for the training dose of 10 mg/kg cocaine: telenzepine (N=8), tropicamide (N=9), trihexyphenidyl (N=14), and cocaine in the same cohort (N=19). Data are %DAR (top) and responses per second (bottom), as a function of substitution drug dose.

When telenzepine (3.2 mg/kg) and tropicamide (0.1 mg/kg) were tested in combination with cocaine, each antagonist produced a small shift of the cocaine dose-effect function to the left (Fig. 2), but only the effect of tropicamide reached statistical significance by ANOVA [F(1,63)=10.6, P<0.01]. The effect of cocaine dose on response allocation was always highly significant (P<0.0001). In two mice, telenzepine profoundly suppressed responding at the highest doses (despite no to minimal effects on rates of responding in other mice), which resulted in missing values and consequently reduced statistical power. As an alternative analysis method less affected by missing values, potencies of cocaine were calculated by interpolation in each mouse with and without telenzepine, which confirmed a leftward shift (1.34 mg/kg, 95% confidence interval 0.68 – 2.65 vs. 0.39 [0.23 – 0.65] mg/kg, P<0.01 by paired-sample t-test). Telenzepine produced a small decrease in rates of responding regardless of the cocaine dose [F(1,62)=6.90, P<0.05], while tropicamide did not affect rates.

Figure 2.

Effects of adding 3.2 mg/kg telenzepine (A, N=9) or 0.1 mg/kg tropicamide (B, N=8) to cocaine in Swiss-Webster mice. Data are %DAR (top) and responses per second (bottom), as a function of cocaine dose. In the top panel, exceptions to the group sizes are indicated on the figure when some mice failed to respond at a given dose.

3.2 “Cocaine-like” muscarinic receptor antagonists in muscarinic receptor knockout mice

Wild-type mice and knockout mice lacking M₁, M₂, or M₄ receptors acquired cocaine discrimination after on average 89.3±5.4, 85.5±5.2, 96.3±5.8, and 129.9±10.5 sessions, respectively. The M₄^-/- mice required significantly longer training to meet criteria relative to wild-type mice (χ²=13.1, P<0.001). The genotypes did not differ in their cocaine dose-effect functions (Supplemental Fig. 1).

To test the hypothesis that M₁ and/or M₄ receptors mediated the cocaine-like discriminative stimulus effect of the non subtype-selective muscarinic antagonist scopolamine, scopolamine was tested as substitution in each knockout strain. Tests were aborted in the M₂^-/- strain due to unexpected toxicity, as the first few M₂^-/- mice tested with doses of scopolamine above 0.32 mg/kg died within a day. Thus, dose-effect functions for scopolamine substitution were obtained in wild-type mice, M₁^-/- mice, and M₄^-/- mice (Fig. 3). DAR was related to both genotype [F(2,71)=3.54, P<0.05] and scopolamine dose [F(4,71)=8.47, P<0.0001]. Scopolamine produced partial substitution, and significant effects on DAR, in the wild-type mice [F(4,25)=7.28, P<0.001] and in the M₄^-/- mice [F(4,24)=3.41, P<0.05], but not in the M₁^-/- mice. Rates of responding were also affected differentially (genotype effect [F(2,74)=6.41, P<0.01], scopolamine effect [F(4,74)=6.08, P<0.001].

Figure 3.

Dose-effect functions for the non subtype-selective antagonist scopolamine in wild-type mice (WT, N=8), M₁^-/- mice (N=7), and M₄^-/- mice (N=7), tested as substitutions for the training dose of 10 mg/kg cocaine. No data in the M₂^-/- mice due to toxicity. Data are %DAR (top) and responses per second (bottom), as a function of scopolamine dose. In the top panel, exceptions to the group sizes are indicated on the figure when some mice failed to respond at a given dose.

A low dose of scopolamine can shift the cocaine discrimination dose-effect function to the left in intact mice [2]. To test the hypothesis that M₁ receptors also mediated the potentiation of the cocaine discriminative stimulus effect by scopolamine, we determined cocaine dose-effect functions with and without 0.32 mg/kg scopolamine in wild-type mice, M₁^-/- mice, M₂^-/- mice, and M₄^-/- mice. We observed a qualitative difference between strains, in that scopolamine produced the expected shifts to the left in the wild-type mice, M₂^-/- mice, and M₄^-/- mice, but produced a shift to the right in the M₁^-/- mice (Fig. 4). A three-way ANOVA with the factors genotype, cocaine, and scopolamine confirmed a significant genotype by scopolamine interaction [F(3,237)=6.11, P<0.0001], and follow-up simple effects confirmed a significant effect of scopolamine in each knockout line (P<0.05 to P<0.001). No evidence of toxicity was observed in the M₂^-/- mice with this dosing. The effect of cocaine dose on DAR was significant in all pretreatment/combinations tests (P<0.0001).

Figure 4.

Effects of adding 0.32 mg/kg scopolamine to cocaine in wild-type mice (A, N=8), M₁^-/- mice (B, N=6), M₂^-/- mice (C, N=6), and M₄^-/- mice (D, N=8). Note the qualitatively different effect in the M₁^-/- mice. Data are %DAR (top) and responses per second (bottom), as a function of cocaine dose. In the top panel, exceptions to the group sizes are indicated on the figure when some mice failed to respond at a given dose.

3.3 “Cocaine-attenuating” muscarinic receptor antagonists in Swiss-Webster mice

The M₂-preferring antagonist methoctramine produced little cocaine-appropriate responding per se, with a maximal DAR of 35% at the highest dose (main effect [F(4,26)=3.43, P<0.05]; 3.2mg/kg vs. vehicle P<0.05), (Fig. 5A). Methoctramine was tested over a range of doses, up to doses that produced rate-decreasing effects [F(4,28)=7.21, P<0.001].

Figure 5.

A: Dose-effect functions for cocaine (N=12) and the muscarinic M₂ receptor-preferring antagonist methoctramine (N=8) in Swiss-Webster mice, tested as substitutions for the training dose of 10 mg/kg cocaine. B: Effect of adding a 0.01 mg/kg methoctramine to a range of cocaine doses (N=7). C: Effect of adding 1 mg/kg methoctramine to a range of cocaine doses (N=8). D: Dose-response relationship of adding methoctramine to 3.2 mg/kg cocaine in Swiss-Webster mice (N=6). Data are %DAR (top) and responses per second (bottom), as a function of cocaine dose or methoctramine dose. In the top panel, exceptions to the group sizes are indicated on the figure when some mice failed to respond at a given dose.

We then tested methoctramine in combination with cocaine, testing for a shift in the cocaine dose-effect function at two doses of methoctramine, and testing a range of methoctramine doses in combination with 3.2 mg/kg cocaine. As opposed to scopolamine and the M₁/M₄ receptor-preferring antagonists telenzepine, trihexyphenidyl, and tropicamide, methoctramine moderately attenuated the cocaine discriminative stimulus effect in Swiss-Webster mice: administration of 0.01 mg/kg methoctramine produced a small shift of the dose-effect function to the right (Fig. 5B; methoctramine effect [F(1,53)=4.34, P<0.05], cocaine by methoctramine interaction [F(4,53)=2.67, P<0.05]. At 1.0 mg/kg, methoctramine had no signignificant effect on DAR (in fact showed a trend to shift the cocaine curve to the left), and moderately decreased rates of responding ([F(1,63)=17.6, P=0.0001; Fig. 5C). This “biphasic” effect was also apparent when testing a range of methoctramine doses with 3.2 mg/kg cocaine, in which doses from 0.01 to 0.32 mg/kg methoctramine decreased DAR (main effect of methoctramine dose [F(6,30)=3.28, P<0.05]), but 1 mg/kg had no effect (Fig. 5D).

In an effort to replicate those findings, the putative M₂-preferring antagonist AQ-RA 741 was also tested. AQ-RA 741 alone produced no cocaine-appropriate responding in a range of doses previously shown to be active in vivo (no effect of dose on DAR or rate; see Supplemental Fig. 2A). However, the selectivity of AQ-RA 741 is modest, and the usefulness of those data is limited by the fact that higher, rate-suppressing doses were not tested, out of concerns for potential toxicity. Administration of 1.0 mg/kg AQ-RA 741 produced a small shift of the cocaine dose-effect function to the right (Supplemental Fig. 2B), although the effect only reached statistical significance for the first reinforcer (AQ-RA 741 dose [F(1,62)=4.45, P<0.05], AQ-RA 741 by cocaine interaction [F(4,62)=2.98, P<0.05]), but not for total-session responding. A range of AQ-RA 741 doses were tested with 3.2 mg/kg cocaine (Supplemental Fig. 2C). Again, AQ-RA 741 produced moderate decreases in DAR that was significant for first-reinforcer responding [F(3,15)=4.43, P<0.05] (total session analysis: P=0.06). Rates of responding were not significantly affected.

3.4 “Cocaine-attenuating” muscarinic receptor antagonists in muscarinic receptor knockout mice

Because methoctramine has relatively low selectivity for the M₂ subtype, we wanted to test the hypothesis that effects on cocaine discrimination were mediated through M₂ receptors. To this end, we tested 0.032 mg/kg methoctramine as pretreatment to cocaine in wild-type mice, M₁^-/- mice, M₂^-/- mice, and M₄^-/- mice (Fig. 6). A three-way ANOVA with the factors genotype, cocaine dose and methoctramine dose confirmed a significant genotype by methoctramine interaction [F(3,205)=2.97, P<0.05]. As hypothesized, methoctramine had no effect in the M₂^-/- mice (Fig. 6C), while the moderate rightward shift observed in the Swiss-Webster mice was confirmed in the wild-type mice (P=0.001; Fig. 6A). The effect appeared intact in the M₄^-/- mice (P<0.01; Fig. 6D), but was diminished in the M₁^-/- mice, in which the shift was not significant (Fig. 6B).

Figure 6.

Effects of adding 0.032 mg/kg methoctramine to cocaine in wild-type mice (A, N=8), M₁^-/- mice (B, N=6), M₂^-/- mice (C, N=6), and M₄^-/- mice (D, N=6). Data are %DAR (top) and responses per second (bottom), as a function of cocaine dose. In the top panel, exceptions to the group sizes are indicated on the figure when some mice failed to respond at a given dose.

4. Discussion

Muscarinic antagonists can partially mimic and increase the S^D effects of cocaine, but the muscarinic receptor subtypes involved in these effects are not known. We tested muscarinic receptor antagonists with no subtype selectivity, or moderately preferential affinity at the M₁, M₂ or M₄ subtype, in wild-type mice, M₁^-/- mice, M₂^-/- mice and M₄^-/- mice trained to discriminate 10 mg/kg cocaine from saline. We found that the non-selective antagonist scopolamine or antagonists with relatively higher affinity at M₁/M₄ receptor subtypes produced some cocaine-appropriate responding per se, and produced leftward shifts in the cocaine S^D effect curve in intact animals. Experiments in the knockout mice further indicated that M₁ receptors, rather than M₄ receptors, mediate the cocaine-like S^D effects of scopolamine. In contrast, antagonists with relatively higher affinity at the M₂ receptor subtype attenuated the cocaine S^D effect. Experiments in knockout mice also supported the notion that M₂ receptors mediate the cocaine S^D attenuating effects of methoctramine.

4.1 Cocaine-like effects of non-selective and M₁/M₄-preferring antagonists in intact mice

We previously showed that scopolamine produced cocaine-appropriate responding in Swiss-Webster mice (partial substitution with an inverted U-shaped curved), and that a sub-threshold dose of scopolamine shifted the cocaine dose-effect curve to the left [2]. Methylscopolamine, which has poor brain penetration, was less potent than scopolamine in producing these effects, indicating a centrally mediated effect [2]. This is in agreement with earlier studies in rats, in which the non subtype-selective muscarinic antagonists atropine and scopolamine produced leftward shifts in cocaine's S^D effects [7, 8]. Potentiation of psychostimulant effects by muscarinic receptor blockade has been observed consistently in mice, rats, and non-human primates across a range of endpoints, including locomotor activity, stereotypies, and intravenous self-administration [8, 52-55]. Isobolographic analyses confirmed that the effects of cocaine-scopolamine combinations were more than additive both in drug discrimination in mice, and in locomotor activity in rats, implying that muscarinic receptor antagonists and cocaine produce their effects through different brain pathways [2, 56]. The interaction likely involves dopamine release, as systemic administration of muscarinic antagonists induces striatal dopamine release in humans and rats, and potentiates cocaine-induced dopamine increases [34, 57, 58].

Here, M₁ and/or M₄-preferring antagonists produced some cocaine-appropriate responding, comparable to results obtained previously with scopolamine [2]. Previous investigations similarly found that the moderately M₁-preferring antagonists telenzepine and trihexyphenidyl potentiated the locomotor stimulant and/or S^D effects of cocaine in rats [9, 34]. Tropicamide was described as somewhat M₄-preferring, although recent data suggest that is a relatively nonspecific muscarinic antagonist [35, 59]. Effects of M₄ receptor blockade on cocaine's S^D effects have not been reported previously, but our findings are consistent with a report that tropicamide increased or prolonged cocaine-induced stereotypies in rats [52]. The notion that blocking either M₁ receptors or M₄ receptors increases effects of cocaine is also consistent with our findings that, conversely, pharmacological stimulation of M₁ and/or M₄ receptors attenuated the S^D effects and reinforcing effects of cocaine in rats and mice ([2-5]; [6]). Taken together, this first data set suggests that M₁ and/or M₄ receptor blockade at least partly mediate the effects of scopolamine in the cocaine discrimination assay.

4.2 Cocaine-like effects of muscarinic antagonists in muscarinic receptor knockout mice

Brain-penetrant muscarinic receptor antagonists with a high degree of selectivity for each of the five receptor subtypes are still being developed, and we therefore used a combination of subtype-preferring ligands and receptor knockout mice to determine the contributions of individual receptor subtypes to the above effects. Cocaine produced comparable S^D dose-effect functions in all four strains – wild-type, M₁^-/-, M₂^-/-, and M₄^-/-. In contrast, substitution of scopolamine for cocaine revealed striking differences: scopolamine produced over 80% cocaine-appropriate responding in the wild-type mice, but no appreciable cocaine-appropriate responding in the M₁^-/- mice in the same dose range. Partial substitution was observed in the M₄^-/- mice. Although we cannot exclude that higher doses of scopolamine could have produced some cocaine-appropriate responding in the knockout mice, these findings indicate that M₁ receptors are necessary for scopolamine to produce cocaine-S^D like effects, while M₄ receptors may contribute partially to the effect.

When we tested scopolamine in combination with cocaine, the M₁^-/- mice again showed a qualitatively different effect: whereas scopolamine produced leftward shifts in the cocaine dose-effect function in wild-type mice, M₂^-/- mice, and M₄^-/- mice, M₁^-/- mice showed a rightward shift. Thus, the cocaine-potentiating effect of scopolamine appears to be dependent upon blockade of M₁ receptors.

These findings illustrate the opposing modulatory effects of different muscarinic receptor subtypes on striatal dopaminergic transmission (for review, see [28]; Thomsen et al. 2017 under review). In the absence of M₁ receptors, scopolamine would block M₂-M5 receptors, and the fact that scopolamine produced a rightward shift, not leftward shift, in M₁^-/- mice, suggests that blocking M₄ receptors alone is not sufficient to potentiate the effects of cocaine. Rather, antagonism of M₂ receptors and the resulting increase in acetylcholine tone appears to produce sufficient stimulation of M₄ receptors to attenuate cocaine's effects, consistent with the effect of an M₄ positive allosteric modulator in the same assay [6]. It is possible that effects at M5 receptors play a role as well [66, 67]. M₁^-/- mice (but not M₂^-/- mice or M₄^-/- mice) showed plasma levels of scopolamine roughly twice as high as the wild-type controls 30 min after an intraperitoneal injection of 1 mg/kg, as well as a trend for higher brain levels [32]. However, given that scopolamine produced cocaine-like effects up to at least 10 mg/kg in wild-type mice, it seems unlikely that this slightly higher blood level would account for the complete reversal of effect observed here between the wild-type mice and the M₁^-/- mice.

The scopolamine-induced leftward shift in cocaine S^D was preserved or if anything larger in the M₂^-/- mice and M₄^-/- mice relative to wild-type mice. Loss of inhibitory autoreceptors may be expected to lead to increased extracellular levels of acetylcholine, and/or reduced effect of scopolamine on acetylcholine release. Indeed, these effects were observed in M₂^-/- mice, M₄^-/- mice, and M₂^-/-M₄^-/- double knockout mice [25, 68]. Thus, loss of scopolamine-increased acetylcholine tone (which would tend to attenuate the cocaine S^D effect) would unmask the M₁ receptor-mediated potentiation of the cocaine S^D effect. These data suggest that M₁ receptors, not M₄ receptors, play the major role in scopolamine-potentiated cocaine S^D effects. Studies using M₁^-/- mice, M₄^-/- mice, and double M₁^-/-M₄^-/-mice similarly indicated that M₁ receptors, not M₄ receptors, mediated the disruption of prepulse inhibition of the startle response by scopolamine, another effect that is shared between muscarinic receptor antagonists and psychomotor stimulant drugs [69].

4.3 Cocaine-attenuating effects of M₂-preferring antagonists

In contrast to the M₁/M₄ receptors, we did not anticipate that blockade of the (primarily presynaptic) M₂ receptors would produce cocaine-like effects in the drug discrimination assay. Indeed, neither methoctramine nor AQ-RA 741 substituted for cocaine in intact mice, with the caveat that neither ligand is highly selective for the M₂ receptor, and that neither ligand was tested at doses that produced strong behavioral suppression, due to expected risks of toxicity. To evaluate the effects of methoctramine in combination with cocaine, we therefore tested a wide range of doses (two and a half log units), including doses that may bind M₂ receptors preferentially, and doses that may have off-target effects. This produced a biphasic dose-effect function, from an ineffective, sub-threshold dose, over a range of doses that decreased cocaine-appropriate responding, to a complete reversal of effect this at the highest dose. A biphasic dose-effect function of methoctramine has been observed previously, similarly attributed to recruitment of non-M₂ receptors [60]. M₂ receptor antagonists reduced the S^D effects of 3.2 mg/kg cocaine by up to 50% in Swiss-Webster mice, an effect similar to that produced by M₁/M₄ agonists [2, 5]. When tested against a range of cocaine doses, a small rightward shift was obtained.

Brain M₂ receptors are mainly inhibitory autoreceptors [61, 62]. M₂ receptors serve as autoreceptors throughout the brain with the probable exception of the nucleus accumbens, while M₄ receptors appear to serve this function only in the striatum [18, 21, 23, 25]. Consistent with this function, in vitro and in vivo studies have shown that M₂-preferring antagonists increase acetylcholine in dorsal striatum, hippocampus, and cortex [39, 60, 63-65]. We therefore hypothesize that the M₂ antagonists attenuated the cocaine S^D effect indirectly, through stimulation of brain M₁ and/or M₄ receptors by endogenous acetylcholine. Nicotinic receptors, which also modulate dopamine release in the dorsal and ventral striatum, could also be involved [23]. A complete lack of effect of methoctramine in the M₂^-/- mice supports the interpretation that methoctramine attenuated the cocaine S^D effect in intact mice via antagonism of M₂ receptors. Although rat studies indicated that the nucleus accumbens is important in mediating the S^D effects of cocaine [10, 11], the above results suggest that the effect of methoctramine is likely mediated at least partially in dorsal striatum (and/or other brain regions) rather than the nucleus accumbens, because M₄ receptors, not M₂ receptors, serve as autoreceptors in the accumbens [23]. While methoctramine produced a rightward shift of the cocaine curve in wild-type mice and in the M₄^-/- mice, the effect was strongly attenuated in the M₁^-/- mice. This is perhaps surprising in the light of the effects of scopolamine in the knockout mice, but may suggest that stimulation of postsynaptic M₁ receptors plays a more important role relative to M₄ receptors in attenuating the cocaine S^D.

4.4 Conclusions

Taken together, the present findings suggest that cocaine-like S^D effects of muscarinic antagonists are primarily mediated through muscarinic M₁ receptors, with a minor contribution of M₄ receptors. The data are in agreement with the notion that stimulation of M₁ receptors and M₄ receptors can each attenuate the S^D effect of cocaine, not only through direct pharmacological stimulation as shown previously ([2, 5]; [6]), but also through increased cholinergic tone due to blockade of M₂ autoreceptors. Finally, the findings also suggest that the cocaine-modulating effects of muscarinic receptor manipulations are likely not mediated entirely in the nucleus accumbens, but involve dorsal striatum or other brain regions as well.

Highlights.

• Muscarinic antagonists produce cocaine-like effects mainly via M₁ receptors

• M₂-preferring antagonists attenuate cocaine effects

• M₂ antagonist effects are likely mediated by M₁/M₄ receptors via increased acetylcholine