Xanomeline treatment attenuates cocaine self-administration in rats and nonhuman primates

Abstract

The lack of an FDA-approved pharmacotherapy to combat cocaine use disorder (CUD) is an ongoing and urgent public health challenge. Emerging evidence suggests that the muscarinic acetylcholine system modulates mesolimbic dopamine release and thus may serve as a suitable target for novel CUD medications. The M1/M4-preferring muscarinic agonist xanomeline was recently approved by the Food and Drug Administration for schizophrenia management, and a previous study in male rats suggested that xanomeline treatment attenuated cocaine self-administration in a cocaine-vs-food choice procedure. The present study was conducted to further examine xanomeline treatment effectiveness on cocaine self-administration in male and female rats and nonhuman primates. Both male and female rats and monkeys were trained to self-administer cocaine during daily behavioral sessions. Repeated xanomeline treatment significantly decreased cocaine choice in rats similar to both pharmacological (amphetamine maintenance) and non-pharmacological (increasing alternative reinforcer value) positive controls. In separate groups of monkeys, acute xanomeline pretreatment decreased cocaine-vs-food choice in three out of four monkeys and selectively decreased cocaine-, but not food-maintained responding, under a multiple schedule of cocaine and food reinforcement in three out of four monkeys. Overall, the consistent effectiveness of xanomeline to reduce IV cocaine self-administration in both rodents and nonhuman primate supports its further evaluation as a CUD medication in humans.

Graphical Abstract

1.0. Introduction

Cocaine use disorder (CUD) remains a public health crisis that affects an estimated 1.4 million individuals according to a 2021 Substance Abuse and Mental Health Services Administration (SAMHSA) study (“Key Substance Use and Mental Health Indicators in the United States: Results from the 2021 National Survey on Drug Use and Health,” 2021). Epidemiological data indicate that cocaine was the third most frequently identified drug in overdose-related deaths in 2022 (National Institute on Drug Abse, 2024). Furthermore, a report from the National Forensic Laboratory Information System (NFLIS) shows that, from 2022 to 2024, there has been a 5% increase in cocaine reports with a simultaneous 1.5% decrease in fentanyl reports, highlighting the extent of the current stimulant crisis occurring concurrently with the opioid crisis (Ciccarone and Shoptaw, 2022; Fischer et al., 2021; Drug Enforcement Administration, 2025). Unfortunately, there are no Food and Drug Administration (FDA) approved CUD treatment strategies, emphasizing the need for continued preclinical research evaluating novel candidate pharmacotherapies in translational animal models.

Muscarinic acetylcholine (ACh) receptors are one potential modulator of cocaine reinforcement. Cholinergic cell bodies project to or are located within key regions of the mesocorticolimbic dopamine pathway, including the ventral tegmental area (VTA), nucleus accumbens (NAc), and prefrontal cortex (PFC) (Eckenstein et al., 1988; Meredith et al., 1989; Walaas and Fonnum, 1979; Yeomans, 1995). Both M1 and M4 muscarinic receptors are highly expressed within the cortex and dopamine reward circuitry, suggesting their potential to modulate dopamine neurotransmission (Moran et al., 2019; Seo and Lee, 2009; Tzschentke, 2000; Yeomans, 1995). For example, activation of excitatory M1 Gq-coupled receptors on efferent glutamatergic projections from the PFC to the NAc reduces NAc dopamine release (Hirose et al., 2021; Oda et al., 2018). Additionally, M1 receptor knockout mice show elevated striatal dopamine levels (Gerber et al., 2001). Furthermore, M4 inhibitory Gi/o-coupled receptors co-express with D1 excitatory Gs-coupled receptors in NAc medium spiny neurons; suggesting that ACh acting at M4 receptors inhibits DA-stimulation within striatum (Jeon et al., 2010; Klawonn et al., 2018; Onali and Olianas, 2002; Schmidt et al., 2011). In summary, there is extant neurobiological evidence to suggest that M1 and M4 muscarinic receptor activation may attenuate dopaminergic transmission in ways that support their utility as candidate CUD pharmacotherapies.

Consistent with this neurochemical evidence, behavioral data also support a role for M1 and M4 receptor modulation of cocaine reward and reinforcement. For example, amphetamine-induced locomotion and increased dopamine levels were enhanced in M1 receptor knockout mice (Gerber et al., 2001). Additionally, acute M1 agonist VU0364572 (VU’72) administration produced long-lasting attenuation of both cocaine choice and cocaine-induced increases in dopamine levels within the mPFC and the NAc (Weikop et al., 2020). Furthermore, repeated treatment with the M1/M4 preferring muscarinic agonist xanomeline decreased cocaine-vs-food choice and cocaine discrimination in rodents (Thomsen et al., 2014, 2012, 2010). Overall, these behavioral results, in conjunction with neurochemical evidence, support the continued evaluation of the role of M1 and M4 muscarinic receptors in cocaine reinforcement.

The aim of the present study was to extend the previous evaluation of xanomeline on cocaine self-administration across three primary independent variables. First, previous behavioral studies cited above only used male subjects to evaluate the effects of genetic and pharmacological M1 and M4 manipulations (Gerber et al., 2001; Klawonn et al., 2018; Miyakawa et al., 2001; Onali and Olianas, 2002; Thomsen et al., 2010; Weikop et al., 2020). However, sex is one biological variable that can influence both the expression and modulation of cocaine’s effects. Thus, to address this gap, the present study included both males and females subjects across all conditions with the a priori hypothesis that there would not be sex differences in experimental manipulations (Bagley et al., 2019; Kerstetter et al., 2012; Yoest et al., 2014). Second, there are species differences in the neurocircuitry for both dopamine and acetylcholine, yet the effects of xanomeline on cocaine self-administration have only been examined in rodents (Coppola and Disney, 2018; Smith and Porrino, 2008; Tillet and Kitahama, 1998; Weerts et al., 2007; Yelnik et al., 1993). Accordingly, the present study compared the effects of xanomeline treatment on cocaine self-administration and cocaine-vs-food choice in rats with its effects in rhesus monkeys as a nonhuman-primate species. Lastly, because there are no clinically available CUD pharmacotherapies to serve as positive control comparators to xanomeline, the present study also determined the effects of both an environmental positive control (e.g., altering magnitude of an alternative food reinforcer) and a pharmacological positive control (i.e., amphetamine maintenance) to facilitate interpretation and enhance experimental rigor when interpreting xanomeline treatment effects. These experimental comparators were selected because the extant literature consistently reports that both increasing the magnitude of a nondrug alternative reinforcer and amphetamine maintenance decrease cocaine choice in both preclinical and human laboratory studies (Johnson et al., 2016; Lile et al., 2016, 2020; Negus, 2003; Negus and Mello, 2003; Rush et al., 2009, 2010; Thomsen et al., 2013; Townsend et al., 2021).

2.0. Methods

2.1. Subjects

A total of 36 Sprague-Dawley rats (18 Male/ 18 Female) were purchased from a commercial supplier (Envigo, Fredrick, MD, USA). At the start of the study, all rats were 6-8 weeks of age and weighed 225 – 275g for females and 250–300g for males. All subjects were single-housed in a temperature-controlled and AAALAC-accredited vivarium under a 12 h light:dark cycle with the dark cycle running from 6 pm to 6 am. Subjects were given ad libitum access to both food (Tekland Rat Diet, Envigo) and water in their home cage and were weighed weekly. Animal research and maintenance were conducted following the 2011 NIH Guide for the Care and Use of Laboratory Animals, eighth edition. All enrichment and experimental protocols were approved by the Virginia Commonwealth University Institute for Animal Care and Use Committee.

A total of eight adult rhesus monkeys (Macaca mulatta; 7 Male/1 Female) served as nonhuman primate subjects. Monkeys were fed enough food to maintain stable and healthy body weights, and the diet included food pellets obtained during the experimental sessions, monkey biscuits (Lab Diet Monkey Chow #5038, Tekland Diets, Envigo), and one piece of fresh fruit daily. Individual monkeys had visual, auditory, and olfactory contact with other monkeys, and enrichment devices were provided in the housing chamber. All nonhuman primate enrichment and experimental protocols were approved by the McLean Hospital Institute for Animal Care and Use Committee.

2.2. Catheter Implantation and Maintenance

An indwelling intravenous jugular catheter and vascular access port were aseptically implanted in rats as previously described (Townsend et al., 2021). Catheters were flushed daily with 0.1 ml of cefazolin (50 mg/ml) and heparin (250 U/ml). Catheter patency was confirmed at the end of each experiment in rats by administering intravenous (IV) methohexital (1.6 mg) and observing instant loss of muscle tone and ambulation. Also under aseptic conditions, monkeys were surgically implanted with either a single-lumen intravenous catheter attached to a subcutaneous port or a double-lumen intravenous catheter that exited through an incision in the midscapular region of the back and was connected to two external lines through which either saline or IV cocaine could be administered.

2.3. Apparatus

Rat modular operant chambers (Med Associates, St. Albans, VT) housed in sound-attenuating chambers (Med Associates) were utilized for cocaine-choice studies as described previously (Townsend et al., 2021). In brief, each retractable lever had tricolor LED lights (red, yellow, green) located directly above the lever. The syringe pump (PHM-100, Med-Associates) was connected to a fluid swivel (375/22PS, Instech Laboratories, Plymouth Meeting, PA), and the IV line was protected by a stainless steel, magnetic tether (Instech Laboratories). The operant chamber also contained a retractable dipper (Med Associates) with a 0.1 ml cup for liquid food delivery (vanilla flavored Ensure^® in tap water, Abbot Laboratories, Chicago, IL). All behavioral experiments were conducted using custom programs written in Med-State notation (Med Associates).

Monkeys were chaired for daily cocaine-food choice sessions conducted within ventilated, sound-attenuating chambers or were housed in ventilated stainless-steel chambers that also served as experimental chambers for the multiple schedule of cocaine and food reinforcement as described previously (Caine et al., 2000). An operant panel (28 x 28 cm) and a food receptacle were mounted on the front wall of each experimental chamber. Three square translucent response keys (5.1 cm x 5.1 cm) were arranged 3.5 cm apart in a horizontal row 9 cm from the top of the operant panel. Each key could be trans-illuminated by red, green, or yellow stimulus lights (LEDs). A syringe pump (Model B5P-lE, Braintree Scientific, Braintree, MA) was mounted above each chamber for IV delivery of saline or drug solutions to the monkey through the external catheter. A pellet dispenser (Gerbrands, Model G5210) also was externally positioned to deliver 1 gm food pellets (Formula 5TUR banana flavor grain-based pellet, Purina Mills Test Diet, Richmond, IN) into the food receptacle within the chamber. All experimental operations were arranged via hardware interface (Med Associates, Inc.) using customized software programmed in MedState Notation.

2.4. Cocaine-vs-Food Choice in Rats

All rat behavioral sessions were conducted between 7 AM and 11 AM Monday through Friday. Subjects were initially trained to respond for 0.32 mg/kg/infusion cocaine under a fixed-ratio 1 (FR1) schedule of reinforcement during daily 2-h sessions. Before each session, subjects received a noncontingent 0.32 mg/kg cocaine infusion followed by a 60-s timeout. Once subjects earned ≥ 20 cocaine infusions, the FR value was initially increased to FR3 and then FR5. Following three consecutive days of ≥ 20 cocaine infusions at FR5, training for food-maintained responding was initiated. Initially, the response requirement was FR1, and response requirement completion resulted in presentation of 0.1 ml of 32% vanilla flavored Ensure diluted in tap water. Training at FR1 continued until >100 reinforcers were earned during the 2-h session. The response requirement was then increased first to FR3 and then FR5. One subject had the Ensure concentration increased to 100% due to failure to meet acquisition criteria. Subsequently, subjects were trained on the terminal cocaine-vs-food choice procedure. The terminal choice procedure had five separate 20-min components with a 5-min timeout between each component. During the timeout, subjects received a noncontingent food presentation and a noncontingent infusion of the cocaine dose available during the subsequent response component as previously described (Townsend et al., 2021). Increasing cocaine doses (0, 0.32, 0.1, 0.32, 1.0 mg/kg/infusion) were available as the alternative to 32% Ensure during successive components under a concurrent FR5:FR5 schedule of reinforcement. Subjects could complete up to 10 ratio requirements across both levers, and responding on one lever reset the FR for the other lever. Cocaine choice was considered stable when the smallest cocaine dose that maintained ≥80 percent cocaine choice did not vary by more than 0.5 log units over three consecutive days.

2.5. Cocaine Self-Administration in Monkeys

Two monkey self-administration procedures were conducted through NIDA’s Addiction Treatment Discovery Program (ATDP) under contract No. N01DA-24-8946. One procedure consisted of a 2h daily behavioral session that began at 11 AM. At the session start, the room lights were extinguished, and an illuminated response panel was activated. Each session consisted of three response components separated by five-minute timeout periods. During the first and last components (the food components; Food 1 and Food 2), 1g banana-flavored food pellets were available under an FR 30; Timeout (TO) 10 sec schedule for five minutes. During the second component (the drug component), intravenous cocaine or saline infusions were available under an FR30; TO 60 sec schedule for 100 minutes. The response key was illuminated with red lights during the food components and with green lights during the drug component. During the timeout following each reinforcer delivery, the response key was illuminated with yellow lights. During inter-component timeout periods, the lights illuminating the response key were turned off. During TO periods, responding had no scheduled consequences. The drug self-administration component was preceded by illumination of the yellow light for 10 sec together with the non-contingent delivery of a single “priming” infusion of the cocaine or saline solution that was available during the ensuing test component. This priming infusion served to signal the unit dose available during the subsequent drug self-administration component and facilitate stable performance during sessions of saline or IV cocaine availability. Saline and different cocaine doses (0.001 – 0.1 mg/kg/infusion) were available for IV self-administration under a double-alternation schedule (e.g., two sessions of IV cocaine self-administration followed by two sessions of IV saline availability, i.e., Coc-Coc-Sal-Sal).

In the second monkey cocaine self-administration procedure, monkeys were trained to respond under a concurrent FR30:FR30 schedule of IV cocaine and food availability. Saline substitution was introduced on a double-alternation basis as above, and saline was made available less frequently when necessary (see below for criterion). Also, the FR requirement for each reinforcer was adjusted, as necessary, for individual subjects so that unit cocaine doses greater than 0.032 mg/kg/infusion reliably maintained responding on the injection-associated manipulandum. Subjects continued in this phase of training until they met the following criteria for stable drug self-administration during three consecutive cycles of double alternation between cocaine and saline: (a) difference of no more than 20% between the number of infusions during any of the second days of IV cocaine availability and the mean number for those sessions and (b) redistribution of responding (>80%) to the food-associated key when saline was available for IV self-administration. After reliable cocaine-vs-food choice behavior was maintained, the training procedure was modified to permit self-administration of different unit cocaine doses (0.003 – 0.32 mg/kg/infusion). Saline and different unit doses of cocaine were each available for 2 consecutive sessions, with saline sessions interspersed periodically and at least once weekly.

2.6. Experimental Manipulations

First, we determined effects of repeated 5-day xanomeline (1.0 – 10 mg/kg, SC) treatments on cocaine-vs-food choice in rats. Xanomeline was administered as a 5-min pretreatment to the daily choice session. Different xanomeline doses were tested during different weeks and were separated by at least one no-treatment week.

Second, we determined acute xanomeline treatment effects on cocaine- and food-maintained responding in rhesus monkeys under the multiple schedule of cocaine and food reinforcement and under the cocaine-vs-food choice procedure. Two types of experiments were conducted involving xanomeline pretreatments: 1) xanomeline dose-ranging studies, and 2) effects of single xanomeline dose on cocaine self-administration dose-effect functions. In the xanomeline dose-ranging studies, xanomeline (0.1 – 3.2 mg/kg) was administered intramuscularly as a 15-min pretreatment before the session during 0.03 mg/kg/infusion cocaine availability under the multiple schedule. Once xanomeline dose-ranging studies were completed, a single intramuscular xanomeline dose was individually determined for each monkey and tested as a pretreatment during availability of saline or a range of unit cocaine doses (0.001 – 0.1 mg/kg/infusion).

Third, as a pharmacological comparator, we determined the effects of continuous 7-day amphetamine (vehicle, 0.32, and 0.56 mg/kg/h) treatment via osmotic pump (2ML1, Alzet, Campbell, CA) on cocaine-vs-food choice in rats. Osmotic pumps were implanted directly under the skin via an incision behind the left or right scapula, moving caudally with the spine, or via an incision near the posterior ilium joint, with the pump moving rostrally with the spine. Starting the day after the implantation, subjects were given access to the cocaine-vs-food choice procedure and weighed daily for 7 consecutive days. After the last choice session, the minipump was surgically extracted. During both experiments, the completion of one dose of amphetamine was followed by at least 4 days of no treatment, during which subjects were baselined; an additional week was given if behavior had not returned to previous baseline conditions.

Fourth, as a nonpharmacological comparator, we determined the effects of different Ensure concentrations on cocaine-vs-food choice in rats. Each Ensure concentration (Water, 10, 32, and 100 % Ensure) was presented for five consecutive days, and the order of Ensure concentrations were randomized using a Latin square design. Different Ensure concentrations were presented during successive weeks.

2.7. Data Analysis

For each dependent measure, data were averaged across rats within a given treatment and analyzes as we have described previously for preclinical studies that include both sexes but are not intended to examine sex as the primary measure of interest (Diester et al., 2019; Santos et al., 2022). First, the primary dependent measures for the rat studies were (1) percent cocaine choice, defined as [(the total number of cocaine reinforcers earned/ total number of reinforcers earned)*100], and (2) reinforcers per component defined as the total number of cocaine and food reinforcers earned in each component. These measures were plotted as a function of the unit cocaine dose. Additional dependent measures were the number of reinforcers (total, cocaine, and food) earned per session. Data from the last two days of each treatment period in rats were analyzed using either a one-way or two-way repeated-measures ANOVA, mixed-effect analysis, or t-test, as appropriate, with cocaine, xanomeline, and amphetamine dose or Ensure concentration as the main factors. Post-hoc comparisons following a significant treatment effect and/or interaction were made to vehicle for drug treatments using a Dunnet’s post-hoc test or pairwise comparisons were made for Ensure concentration using a Tukey’s post-hoc test. The Geisser-Greenhouse correction was used to account for any sphericity violations. Second, data were segregated by sex and analyzed by two-way ANOVA, with sex as one of the variables. A significant main effect of sex or sex × treatment interaction was followed by a Holm-Sidak post hoc test. Lastly, two-way ANOVA results were submitted to post hoc power analyses to calculate the Cohen’s f effect size, achieved power (1-b). Statistical significance was established a priori at the 95% confidence level (p<0.05).

The primary dependent measures for the monkey studies were: (1) the number of reinforcers (food pellets or infusions) delivered during each component of the test session and (2) the distribution of responding, i.e., number of cocaine infusions as a percentage of total reinforcers. Data are presented and interpreted for each individual subject, and data were not grouped for statistical analysis.

2.8. Drugs

Cocaine HCl was provided by the National Institute on Drug Abuse Drug Supply Program (Bethesda, MD). Xanomeline oxalate was purchased from a commercial vendor (Tocris, Minneapolis, MN) for the rat studies. Xanomeline base was supplied by the NIDA Drug Supply Program for the monkey studies. d-Amphetamine hemisulfate was purchased from a commercial vendor (Millipore Sigma, St. Louis, MO). Cocaine and d-amphetamine solutions were dissolved in a bacteriostatic saline and IV solutions passed through a 0.22 um sterile filter before use. Xanomeline oxalate was dissolved in bacteriostatic water for daily administration in the rat studies and xanomeline base was dissolved in 2:1:1:2 solution of dimethyl sulfoxide (DMSO):95% ethanol:Tween 80: sterile water. All drug doses are expressed as the salt or base forms listed above.

3.0. Results

3.1. Repeated 5-day xanomeline treatment on cocaine choice in male and female rats

Figure 1 shows the effects of repeated 5-day vehicle and xanomeline (1.0 – 10 mg/kg/day) treatment on cocaine-vs-food choice in male and female rats. Xanomeline treatment attenuated cocaine choice and post-hoc analysis showed both 5.6 (p = 0.01) and 10 mg/kg/day (p = 0.04) xanomeline significantly decreased cocaine choice (cocaine dose: F(2, 18) = 257, p<0.0001; xanomeline dose: F(2.9, 26.4) = 6.5, p = 0.002; interaction: F(3.9, 32.6) = 4.8, p = 0.004; Fig. 1A). Xanomeline treatment did not significantly alter session total, cocaine, or food reinforcers (Fig. 1B).There was a trend for xanomeline treatment to attenuate cocaine choice when males and females were analyzed separately (Fig. 1C and 1D); however, power analyses showed the experiments were underpowered in both males (ηp²: 0.08, Cohens F: 0.29, Current Power: 0.06) and females (ηp²:0.11, Cohens F: 0.36, Current Power: 0.07; Table S1) to detect a xanomeline treatment effect when data were separated by sex. There was no significant effect of sex on percent cocaine choice across all xanomeline doses (Fig. S1).

Fig. 1:

Effects of 5-day repeated xanomeline treatment on cocaine-vs-food choice in male and female rats (N=9-10, 4M/5-6F. A) Percent cocaine choice during repeated 5-day xanomeline treatment. B) Session food, cocaine, and total reinforcers. C-D) Percent cocaine choice during repeated 5-day xanomeline treatment separated by males and females, respectively. All points and bars represent group mean±SEM of the last two treatment days. Filled data points represent a statistically significant (p<0.05) difference in percent cocaine choice compared to saline treatment.

3.2. Acute xanomeline pretreatment on cocaine self-administration in monkeys

Because nonhuman primates show significant individual differences in response to both pharmacological and environmental manipulations (Nader et al., 2002; Vivian et al., 2001), initial xanomeline dose-ranging studies were conducted to identify xanomeline treatment doses that altered responses for the peak cocaine dose, followed by a redetermination of the cocaine dose-effect function in the presence of that selected xanomeline dose. Xanomeline pretreatment doses ranged from 0.1 – 3.2 mg/kg in the multiple schedule of cocaine and food reinforcement and 0.1 – 1.0 mg/kg in the cocaine choice procedure (Fig. S2-S5). Larger xanomeline doses in monkeys disrupted food-maintained responding and produced stupor, hypersalivation, and pupillary constriction. Figure 2 shows acute xanomeline pretreatment effects under the multiple schedule of cocaine and food reinforcement in individual monkeys. Xanomeline pretreatment produced a downward shift in the cocaine self-administration dose-effect function without altering food-maintained responding in two subjects L958M and RIB9M. Xanomeline produced a rightward shift in the cocaine self-administration dose-effect function in subjects RO3041M and 97D113M. Figures 3 and 4 show xanomeline pretreatment effects on cocaine-vs-food choice in individual monkeys. Xanomeline pretreatment right shifted the cocaine choice dose-effect function in two out of three male monkeys (1344M and 20-781M), decreased percent cocaine choice in the female monkey 012F, and did not alter the cocaine choice dose-effect function in monkey 091M. In general, decreases in percent cocaine choice reflected decreases in the number of cocaine infusions, accompanied by increases in food deliveries.

Fig. 2:

Effects of acute xanomeline pretreatment on cocaine- and food-maintained responding under the multiple schedule of cocaine and food reinforcement in individual monkeys. Abscissae: unit cocaine dose available for self-administration in milligrams per kilogram per infusion or the availability of food reinforcement during the first and last components of the session (Food 1 and Food 2). Ordinates: Number of cocaine infusions earned during 100-min drug component or number of food pellets earned during the 5-min food components (Food 1 and Food 2). Open symbols show monthly data averaged over the most recent three months of cocaine dose-effect curve determinations. Filled symbols represent single determinations of xanomeline pretreatments (1.0 mg/kg in monkeys 97D113M and L958M, or 1.8 mg/kg in monkeys RIB9M and RO3041M) on the cocaine dose effect curve and on food-maintained behavior.

Fig. 3:

Effects of acute xanomeline pretreatment (1 mg/kg in 1344M; 0.32 mg/kg in 012F) on cocaine-vs-food choice in individual monkeys. Abscissae: Unit cocaine dose available in milligrams per kilogram per infusion as the alternative to food pellets. Ordinates: (top) percentage of all reinforcers that were injections, (middle) number of cocaine infusions, or (bottom) number of food pellets earned during the 100-min drug-choice component. Open symbols show mean self-administration data averaged over the most recent three months of cocaine dose-effect curve determinations. Filled symbols represent single xanomeline pretreatment effects.

Fig. 4:

Effects of acute xanomeline (0.1 mg/kg in 20-781M and 0.3 mg/kg in 091M) on cocaine-vs-food choice in individual monkeys. Abscissae: Unit cocaine dose available in milligrams per kilogram per infusion as the alternative to food pellets. Ordinates: (top) percentage of all reinforcers that were injections, (middle) number of cocaine infusions, or (bottom) number of food pellets earned during the 100-min drug-choice component. Open symbols show mean self-administration data averaged over the most recent three months of cocaine dose-effect curve determinations. Filled symbols represent single xanomeline pretreatment effects.

3.3. Effects of continuous amphetamine maintenance on cocaine choice in rats

Figure 5 shows the effects of continuous amphetamine maintenance via osmotic pump on cocaine-vs-food choice in male and female rats. 0.32 mg/kg/h amphetamine maintenance significantly decreased 0.32 (p = 0.01) and 1 mg/kg/infusion (p = 0.03) cocaine choice and 0.56 mg/kg/h amphetamine maintenance significantly decreased 1 mg/kg/infusion (p = 0.01) cocaine choice (cocaine dose: F(1.9, 20.6) = 26.6, p<0.0001; amphetamine dose: F(1.7, 18.2) = 6, p = 0.01; interaction: F(2,16.9) = 4.8, p = 0.02; Fig. 5A) without significantly altering total cocaine or food reinforcers (Fig. 5B). Additionally, when data were separated by sex there was a trend for amphetamine maintenance to attenuate cocaine choice in males; however, a retrospective power analysis (ηp²:0.331, Cohens F: 0.70, Current Power: 0.23; Table S1) reported the experiment was underpowered to detect an amphetamine maintenance effect when the data were separated by sex (cocaine dose: F(2.1, 10.4) = 14.9, p = 0.0008; Fig. 5C). Amphetamine maintenance significantly attenuated cocaine choice in females (cocaine dose: F(1.6, 7.9) = 18.1, p = 0.002; amphetamine dose: F(1, 5.1) = 1.8, p = 0.23; interaction: F(1.6, 6.8) = 8.8, p = 0.02; Fig. 5D) with post-hoc analysis showing a significant decrease of 1 mg/kg/infusion cocaine choice during both 0.32 (p = 0.046) and 0.56 (p = 0.004) mg/kg/h amphetamine maintenance. Although there was a significant cocaine dose × sex interaction for 0.56 mg/kg/h amphetamine maintenance (cocaine dose: F(1.8, 7) = 5.6, p = 0.04; sex: F(1, 4) = 0.2, p = 0.66; interaction: F(1.7, 4.5) = 6.7, p = 0.05; Fig. S6C), post-hoc comparisons failed to detect any significant differences of amphetamine maintenance between males and females.

Fig. 5:

Continuous 7-day amphetamine (Amph.) treatment on cocaine-vs-food choice in male and female rats (N=8-12, M3-6/F5-6). A) Effects of 7-day amphetamine maintenance on cocaine choice. B) Effects of 7-day amphetamine maintenance treatment on session food, cocaine, and total reinforcers. C-D) Effects of 7-day amphetamine maintenance on cocaine choice separated by males and females, respectively. All points and bars represent the group mean±SEM of the last two test days. Filled data point denote statistically significance (p<0.05) compared to saline treatment.

3.4. Ensure concentration manipulations on cocaine choice in rats

Figure 6 shows that manipulating the Ensure concentration (0% – 100%) as the alternative to cocaine significantly altered cocaine choice in both males and females. Increasing Ensure concentrations produced a concentration-dependent decrease in cocaine choice compared to water (cocaine dose: F(2, 17.8) = 56.9, p<0.0001; Ensure concentration: F (2.1, 18.7) = 22.2, p<0.0001; interaction: F(3.6, 25.7) = 7.1, p = 0.0008; Fig. 6A); post-hoc analysis showed that changing the alternative reinforcer to water increased 0.032 (p = 0.005) and 0.1 (p = 0.04) mg/kg/infusion cocaine choice, whereas changing the alternative reinforcer to 100% Ensure attenuated 0.32 mg/kg/infusion cocaine choice (p = 0.008) compared to 32% Ensure. Increasing Ensure concentrations also significantly increased session food reinforcers and decreased session cocaine reinforcers in a concentration-dependent manner without altering session total reinforcers. (Ensure concentration: F(1.9, 18.5) = 4.2, p = 0.04; interaction: F(3, 20.7) = 28, p < 0.0001; Fig. 6B); post-hoc analysis showed that changing the alternative reinforcer to water significantly reduced food reinforcers (p = 0.02) with a corresponding increase in cocaine reinforcers (p = 0.04) compared to 32% Ensure. Changing the alternative reinforcer to 100% Ensure significantly increased food reinforcers (p = 0.002) compared to 32% Ensure. After separating the data by sex, there was a main effect of Ensure concentration on cocaine choice in males (cocaine dose: F(1.9,9.5) = 17.8, p = 0.0007; Ensure concentration: F(1.6, 4.7) = 9.7, p = 0.02; Fig. 6C); however, post-hoc comparisons failed to detect any significant effects likely due to the experiment being underpowered when the data were separated by sex (ηp²:0.277, Cohens F: 0.62, Current Power: 0.14; Table S1). In females, there was a significant effect of Ensure concentration (cocaine dose: F(1.2, 3.6) = 150, p = 0.0004; Ensure concentration: F(2.1, 6.2) = 34.3; p = 0.0004; interaction: F(2, 6) = 10.3, p = 0.01; Fig. 6D), with post-hoc analysis showing increased 0.032 mg/kg/infusion cocaine choice (p = 0.03) when water was available and decreased 0.32 mg/kg/infusion cocaine choice (p = 0.03) when 100% Ensure was available relative to 32% Ensure. There were no sex differences in sensitivity of cocaine choice to Ensure concentration manipulations (Fig. S7).

Fig. 6:

Effects of Ensure concentration on cocaine-vs-food choice in male and female rats (N= 7-10, 3-6M/ 4F). A) Percent cocaine choice as a function of Ensure concentration. B) Session, food, cocaine, and total reinforcers. C-D) shows percent cocaine choice across Ensure concentration separated by males and females, respectively. All points and bars represent means±SEM of the last two days of each 5-day experiment. Filled data points represents a significant (p <0.05) difference in cocaine choice compared to 32% Ensure. # represents a significant difference in total reinforcers earned compared to 32% Ensure.

3.5. Comparative effectiveness of xanomeline and positive-control treatments to attenuate cocaine choice

Figure 7 compares the decreases in total session cocaine-vs.-32% Ensure choice that could be achieved with xanomeline, amphetamine, and Ensure manipulations. Analysis compared baseline cocaine choice for each manipulation (repeated saline treatment, osmotic minipump saline treatment, 32% Ensure availability) with the lowest treatment magnitude to produce a significant downward/rightward shift in the cocaine-vs.-32% Ensure dose-effect curve (i.e., 5.6 mg/kg/day xanomeline, 0.32 mg/kg/h amphetamine, and 100% Ensure). There was no significant difference between repeated saline injections, saline pumps, and 32% Ensure on cocaine choice (cocaine dose: F(2, 53.5) = 120.3, p<0.0001; Fig. 7A). There was also no difference between 5.6 mg/kg/day xanomeline, 0.32 mg/kg/h amphetamine, and 100% Ensure on cocaine choice (cocaine dose: F(2.1, 59.7) = 60.7, p<0.0001; Fig. 7B). When maximal xanomeline, amphetamine, and Ensure treatment effectiveness were compared, all three experimental manipulations attenuated cocaine choice to a similar magnitude (Treatment: F(1,22) = 5.5, p = 0.03; Fig. 7C).

Fig. 7:

Comparison of xanomeline effectiveness to attenuate cocaine self-administration compared to both pharmacological and nonpharmacological comparators in male and female rats (N=10-12). A-B) shows percent cocaine choice for both control conditions and the smallest experimental manipulation that significantly attenuated cocaine choice. C) shows total session percent cocaine choice across control and treatment conditions (5.6 mg/kg/day xanomeline; 0.32 mg/kg/h amphetamine; 100% Ensure). * represents a significant difference (p<0.05) between control and treatment condition. All points and bars represent the group means±SEM of the last two days of each 5-day (xanomeline or Ensure) or 7-day (amphetamine) experiment.

4.0. Discussion

This study determined the effectiveness of the M1/M4-preferring agonist xanomeline to attenuate cocaine self-administration across two species and sexes. There were three main findings: First, xanomeline treatment significantly attenuated cocaine-vs.-food choice in male rats, consistent with previously published effects in male rats, and extended these findings to show that xanomeline also attenuated cocaine choice in female rats and cocaine self-administration under multiple and concurrent schedules of cocaine- and food-maintained responding in rhesus monkeys. Second, both amphetamine maintenance and increased Ensure concentrations significantly attenuated cocaine choice in rats, consistent with the extant literature in both rats and nonhuman primates. Lastly, the magnitude of xanomeline effectiveness to attenuate cocaine choice was comparable to effects that could be achieved with either amphetamine maintenance or 100% Ensure. Overall, these results indicate a role for muscarinic receptors in modulating cocaine reinforcement and support further consideration of xanomeline as a candidate cocaine use disorder (CUD) pharmacotherapy.

4.1. Xanomeline effects on cocaine self-administration

Repeated 5-day xanomeline treatment significantly decreased cocaine-vs.-Ensure choice in both male and female rats. Specifically, 5.6 and 10 mg/kg/day xanomeline attenuated cocaine choice consistent with previous findings in male rats (Thomsen et al., 2014), and the present data expanded upon the literature by including females. Xanomeline treatment effectiveness does not appear to be sex specific (Fig. S1). Furthermore, the present results are consistent with a prior study demonstrating xanomeline attenuation of cocaine-induced reinstatement and xanomeline-induced facilitation of extinction (Stoll et al., 2018). Overall, the present results and the extant literature support further experimentation to determine xanomeline treatment effects on cocaine self-administration in higher order species.

In addition to the rat cocaine-choice studies, acute xanomeline treatment decreased cocaine self-administration and cocaine choice in both male and female rhesus monkeys. Xanomeline demonstrated some degree of behavioral selectivity with relatively smaller xanomeline doses selectively attenuating cocaine self-administration under the multiple schedule and reducing cocaine choice, whereas relatively larger xanomeline doses decreased both cocaine- and food-maintained responding. Unlike the rat choice studies, larger xanomeline doses decreased food-maintained responding and produced undesirable effects such as stupor, hypersalivation, and pupillary constriction. These physiological effects of xanomeline in monkeys are consistent with clinical reports of parasympathomimetic undesirable effects by xanomeline in humans (Shekhar et al., 2008). One future direction would be to determine repeated xanomeline treatment effects alone or in combination with the peripheral muscarinic antagonist trospium. A xanomeline+trospium combination has been approved for treatment of schizophrenia with fewer parasympathomimetic effects than xanomeline alone (Syed, 2025), and this combination might also enhance behavioral selectivity to attenuate cocaine vs. nondrug-maintained responding. In summary, these preclinical results in both rats and nonhuman primates support the continued evaluation of xanomeline as a candidate CUD pharmacotherapy.

4.2. Amphetamine maintenance and alternative reinforcer magnitude effects on cocaine choice

Cocaine dose-dependently increased cocaine choice in both male and female rats and rhesus monkeys consistent with the extant literature in rats, non-human primates, and humans (Donny et al., 2003; Hart et al., 2000; Hutsell et al., 2016; Nader and Woolverton, 1992; Negus, 2005; Thomsen et al., 2017, 2008; Townsend et al., 2021). Furthermore, amphetamine maintenance decreased cocaine choice in the present study, consistent with prior results in rats, nonhuman primates, and humans (Chiodo et al., 2008; Hutsell et al., 2015; Lile et al., 2020; Negus and Mello, 2003; Thomsen et al., 2013). Thus, amphetamine served as a pharmacological positive control to facilitate interpretation of xanomeline clinical effectiveness as a CUD pharmacotherapy.

Altering Ensure concentrations also significantly altered cocaine-vs.-Ensure choice in both male and female rats. Specifically, increasing the Ensure concentration to 100% attenuated cocaine choice, whereas decreasing the Ensure concentration increased cocaine choice. These results are also congruent with previous cocaine-vs.-food choice experiments in rats and nonhuman primates (Carroll et al., 1989; Michael A. Nader and Woolverton, 1992; Nader and Woolverton, 1991; Negus, 2003; Thomsen et al., 2013). Moreover, Ensure manipulations in this study were consistent with published opioid-vs-food choice studies in rats suggesting that effects of alterations in alternative reinforcer value are consistent across addictive drugs (Townsend et al., 2021). Overall, this nonpharmacological treatment modeling aspects of contingency management provides another positive control to facilitate interpretation of the present results.

4.3. Xanomeline comparative effectiveness to attenuate cocaine choice

Xanomeline treatment effectiveness to attenuate cocaine choice was similar to the effectiveness of both amphetamine maintenance and 100% Ensure in rats. Although amphetamine maintenance has not been approved as a CUD pharmacotherapy, it has reduced metrics of cocaine self-administration in rodents, nonhuman primates, human laboratory studies, and clinical trials, and it is arguably the most effective candidate pharmacotherapy yet tested (Grabowski et al., 2004; Negus and Henningfield, 2015). Contingency management strategies that include strategic use of alternative non-drug reinforcers have also been consistent in their effectiveness to reduce cocaine use, and in the absence of approved pharmacotherapies, contingency management is currently among the most effective tools for CUD treatment (Davidson et al., 2025). The finding that xanomeline was as effective in the present study as amphetamine maintenance and availability of Ensure as an alternative non-drug reinforcer suggests that xanomeline might also be effective clinically for CUD treatment. Moreover, xanomeline might offer advantages relative to amphetamine maintenance or contingency management. For example, xanomeline does not fall under the controlled substance act (CSA) regulation because of a lack of abuse potential, whereas amphetamine is a CSA Schedule II substance. Although there are few published preclinical abuse potential assessment studies for xanomeline, xanomeline did not substitute for cocaine in a drug discrimination procedure in mice, and abuse potential assessment studies in both nonhumans and humans did not lead to scheduling of xanomeline/trospium (Cobenfy^®) (Thomsen et al., 2012). The lack of CSA scheduling could facilitate clinical deployment of xanomeline as a CUD pharmacotherapy by avoiding abuse potential concerns for amphetamine as a CUD pharmacotherapy (Negus and Henningfield, 2015). Ultimately, clinical studies will be important for establishing both patient compliance and tolerability of xanomeline in CUD patients.

4.4. Potential xanomeline treatment effectiveness mechanism of action

The pharmacological mechanism of xanomeline effectiveness to attenuate cocaine choice may require both M1 and M4 muscarinic receptor activation. Inhibitory M4 receptors co-express with D1 receptors on NAc medium spiny neurons, and M4 activation should oppose D1 receptor activation in the direct dopamine pathway (Glick et al., 1994; Klawonn et al., 2018; Onali and Olianas, 2002; Roberts et al., 1977). In addition, M1 receptors are expressed on glutamatergic projections from the PFC to NAc and modulate NAc dopamine release (Hirose et al., 2021; Oda et al., 2018). Genetic knockout of both M1 and M4 eliminated xanomeline effectiveness to attenuate cocaine discrimination and self-administration in mice, whereas genetic deletion of either M1 or M4 alone only attenuated xanomeline effects on cocaine discrimination (Thomsen et al., 2010, 2012). In addition, co-administration of both the M1-selective agonist VU0364572 and the M4-selective PAM VU0152100 was more effective in decreasing acquisition of cocaine self-administration in mice compared to either muscarinic ligand treatment alone (Balakrishnan et al., 2024). These results suggest potential additive or synergistic interactions between M1 and M4 muscarinic receptor activation on cocaine self-administration

Although genetic deletion of M1 or M4 receptors alone and in combination have been examined for xanomeline attenuation of cocaine discrimination and self-administration, these studies have not been conducted for xanomeline effects on cocaine choice. Thus, one future direction would be to determine whether an M1- or M4-selective negative allosteric modulator (NAM) or orthosteric antagonist alone or in combination attenuated xanomeline effects on cocaine choice in rats or monkeys. The working hypothesis is that antagonism of both M1 and M4 receptors would be necessary to block xanomeline-induced reductions in cocaine choice. If confirmed, subsequent studies could determine the neuroanatomical location of M1 and M4 receptor mediation of xanomeline-induced attenuation of cocaine choice towards the potential development of next generation muscarinic ligands as CUD pharmacotherapies with reduced undesirable effects.

5.0. Declaration of generative AI and AI-assisted technologies in the writing process.

The first author used Gemini, a large language model from Google, to correct grammatical and verb tense agreement errors as well as to improve the readability of the initial manuscript draft. Subsequently, all authors reviewed and edited the manuscript and take full responsibility for article content.

• Xanomeline treatment attenuated cocaine choice in male and female rats

• Acute xanomeline treatment decreased cocaine SA in 7 out of 8 monkeys

• Xanomeline effects were similar to amphetamine maintenance in rats