The Discriminative Stimulus Effects of Epibatidine in C57BL/6J Mice

Abstract

The α4β2* nicotinic acetylcholine receptor (nAChR) subtypes are targeted for the development of smoking cessation aids, and the use of drug discrimination in mice provides a robust screening tool for the identification of drugs acting through nAChRs. Here, we established that the α4β2* nAChR agonist epibatidine can function as a discriminative stimulus in mice. Male C57BL/6J mice discriminated epibatidine (0.0032 mg/kg, s.c.) and were tested with agonists varying in selectivity and efficacy for α4β2* nAChRs. The discriminative stimulus effects of epibatidine were characterized with the nonselective, noncompetitive nicotinic antagonist mecamylamine, with the selective β2-substype-containing nAChR antagonist DHβE, and the α7 antagonist MLA. Nicotine (0.32–1.0 mg/kg, s.c.), the partial nAChR agonist cytisine (1.0–5.6 mg/kg, s.c.), and the α7 nAChR agonist N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]-4-chlorobenzamide (PNU-282987, 10–56 mg/kg, i.p.) produced no more than 33% epibatidine-appropriate responding. The partial α4β2* nAChR agonists varenicline RTI-102 produced 61% and 69% epibatidine-appropriate responding, respectively. DHβE and mecamylamine, but not MLA, significantly antagonized the discriminative stimulus effects of epibatidine. These results show that epibatidine may be trained as a discriminative stimulus in mice and has utility in elucidating the in vivo pharmacology of α4β2* nAChR ligands.

Introduction

Disease and mortality related to tobacco use remains a significant health problem in the United States (Wang et al., 2018). Although pharmacotherapies introduced to the market in the last 15 years, such as varenicline, have improved tobacco-cessation rates, relapse to tobacco use remains significantly high (Hays et al., 2008). As a consequence, the development of more effective smoking cessations aids is an important goal. Drug discrimination is a behavioral assay that can be used preclinically to characterize the in vivo pharmacology of CNS drugs (Colpaert, 1999), and as a result, has utility for identifying potential candidate medications for the management of substance abuse. In this regard, drug discrimination has had a significant role in development of smoking cessation aids such as varenicline (Rollema et al., 2007a,b).

Several compelling lines of evidence have identified the α4β2* nicotinic acetylcholine receptor (nAChR; * denotes the potential for the presence of other nAChR subtypes included in the pentomeric receptor stoichiometry) as the predominant receptor that mediates the abuse-related effects of nicotine (Picciotto et al., 1998; Exley et al., 2011). Consequently, drugs with actions at α4β2* nAChRs may be viable candidate medications, and due to its capacity to elucidate the in vivo receptor pharmacology of drugs, drug discrimination may have significant utility in the discovery of α4β2* nAChR ligands. However, by examining patterns of asymettrical generalization and antagonism with the various nAChR antagonists, previous drug discrimination studies conducted with nicotine and other nicotinic ligands (e.g, varenicline and cytisine) as discriminative stimuli demonstrate that the discriminative stimulus effects of nicotine are mediated by more than just the α4β2* nAChR subtype (Chandler and Stolerman, 1997; de Moura and McMahon, 2017). As the α4β2* nAChR has been identified as a key target for novel pharmacotherapeutic interventions for tobacco use, developing a drug discrimination procedure that is more functionally selective for α4β2* nAChRs may increase the effectiveness of identifying effective smoking cessation aids preclinically.

Epibatidine is an alkaloid isolated from the skin of the Epipedobates anthonyi dart frog (Fitch et al., 2010). Compared to nicotine, epibatidine possesses as much as 100-fold higher binding affinity for α4β2* nAChRs, and is more selective for β2-containing nAChRs than other nAChR subtypes in vitro (Sullivan et al., 1994; Carroll et al., 2001). Because of its high potency at and selectivity for α4β2* nAChRs, epibatidine has been used to identify high-affinity nicotine bindings sites in the brains of mammals (Dukat and Glennon, 2003), and has been used as the structural backbone in the construction of investigational smoking cessation drugs (Lloyd and Williams, 2000). Recently, epibatidine has been trained as a discriminative stimulus in squirrel monkeys (Desai and Bergman, 2015; Desai et al., 2016; Withey et al., 2018). Those studies provide strong evidence that α4β2* nAChRs mediate the discriminative stimulus effects of epibatidine.

Rodents often are the species chosen for preclinical whole animal pharmacological research due to their relatively low cost and small size, as well as opportunities to genetically modify drug targets (Hoffman, 2016). Whereas the discriminative stimulus effects of epibatidine have been systematically characterized in squirrel monkeys (Desai and Bergman, 2015; Desai et al., 2016; Withey et al., 2018), the present study demonstrates inter-species generalizability of the epibatidine discriminative stimulus by exhibiting that epibatidine can, in fact, function as discriminative stimulus in mice. Furthermore, this study begins to characterize the discriminative stimulus effects of epibatidine, and provides a potential foundation for epibatidine to be used as a tool in mice to screen for α4β2* nAChR ligands. Here, mice were trained to discriminate 0.0032 mg/kg epibatidine from saline. This dose of epibatidine was chosen based on previous studies published in mice characterizing the discriminative stimulus effects of various nicotinic ligands (de Moura and McMahon, 2017). Test drugs that varied in vitro for their selectivity for and efficacy at α4β2* nAChRs were tested including nicotine, the low efficacy α4β2 and high efficacy α7 nAChR agonist varenicline, the selective low efficacy α4β2* nAChR agonist 2-fluoro-3-(4-nitro-phenyl)deschloroepibatidine (RTI-102), the low efficacy nAChR agonist cytisine, and the α7 nAChR agonist N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]-4-chlorobenzamide (PNU-282987) (Hajos et al., 2005; Abdrakhmanova et al., 2006; Grady et al., 2010).

Methods

Subjects

Eight male C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME) were received into our laboratory at eight weeks of age. The mice were individually housed under a 14/10-h light/dark cycle, and maintained in home cages at 85% of free-feeding body weight according to the growth curve provided by the commercial breeder. Mice were fed 2.5 g of food (Dustless Precision Pellets 500 mg, Rodent Grain-Based Diet, Bio-Serv, Frenchtown, NJ) immediately after experimental sessions. Water was available ad libitum in each home cage. The experiments, which were conducted 7 days per week, were approved by The University of Texas Health Science Center at San Antonio’s Institutional Animal Care and Use Committee. The National Institutes of Health’s Guide for the Care and Use of Laboratory Animals (Institute for Laboratory Animal Research, 2011) was followed.

Drugs

The drugs used in the present study and their salt and enantiomeric forms were as follows: (±)-epibatidine dihydrochloride hydrate (Sigma-Aldrich, St. Louis, MO), (−)-nicotine hydrogen tartrate salt (Sigma-Aldrich), with a pH adjusted to 7, (±)-varenicline dihydrochloride (National Institute on Drug Abuse, Rockville, MD), (−)-cytisine (Atomole Scientific, Hubei, China), 2-fluoro-3-(4-nitro-phenyl)deschloroepibatidine (RTI-102; Dr. F. Ivy Carroll, Research Triangle Institute, NC), N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]-4-chlorobenzamide hydrochloride (PNU-282987; National Institute on Drug Abuse), (±)-mecamylamine hydrochloride (Waterstone Technology, LLC, Carmel, IN), dihydro-β-erythroidine hydrobromide (DHβE; Tocris Biosciences, Bristol, UK), methyllycaconitine citrate salt (MLA; National Institute on Drug Abuse), midazolam hydrochloride (Ben Venue Labs, Bedford, OH), and (−)-cocaine hydrochloride (Sigma-Aldrich). Epibatidine, nicotine, varenicline,, cytisine, RTI-102, DHβE, and midazolam were administered subcutaneously (s.c.). PNU-282987, mecamylamine, MLA, and cocaine were administered intraperitoneally (i.p.). All drugs were administered in a volume of saline equivalent to 10 mL/kg immediately prior to placing the mouse in the operant chamber for the 10-min timeout (i.e., 10 min prior to the period of food availability) with the exception of DHβE (administered 5 min prior to placement of each subject in the operant conditioning chamber) and mecamylamine and MLA (administered 10 min prior to placement of each subject in the operant conditioning chamber). Drug doses were expressed as the weight of the forms listed above except for nicotine, which was expressed as the weight of the base.

Apparatus

Operant conditioning chambers (ENV-307A-CT, Med Associates Inc., St. Albans, VT) were stored in cubicles that attenuated external noise; each cubicle was equipped with an exhaust fan. External noise was masked further with a white-noise generator. Each chamber contained a stand-alone light on one wall, a 2.2 cm-diameter hole recessed on one wall for presentation of liquid food, and three identical nose-poking holes horizontally arranged and spaced 5.5 cm apart on the opposite wall. A stimulus light-emitting diode (LED) was installed in each nose-poking hole. The center of each hole was 1.6 cm above from the floor. When one or both the left and right holes were illuminated, disruption of a photobeam in one of the holes resulted in access to 0.01 cc of 50% v/v unsweetened condensed milk/water through the dispenser hole on the opposite wall. The operant conditioning chambers were connected through an interface (MED-SYST-8, Med Associates Inc.) to a PC. Med-PC software (Med Associates Inc.) was used to control experimental events and provide a record of responses.

Discrimination Training

The current procedure was modelled on a previous study that established varenicline as a discriminative stimulus in mice (de Moura and McMahon, 2017). After seven days of habituating the mice to the housing room, experimental sessions were conducted approximately at the same time during the light period each day. Initially, mice were placed into the operant conditioning chamber for 60 min with both the left and right holes illuminated. A photobeam disruption in either hole resulted in 10-s access to the condensed milk through presentation of a dipper. For 10 s, the milk dipper was presented to the mouse, the light on the wall was illuminated, and the lights inside the nose-poking holes were extinguished. Any disruption of a photobeam during milk dipper presentation was not associated with a programmed consequence. After 100 reinforcers per session were earned for four consecutive sessions, session duration was shortened to 25 min and the response requirement was systematically increased to a fixed ratio 10-response (FR10) schedule of reinforcement. Experimental sessions were then divided into a 10-min timeout; during the timeout, any disruption of a photobeam had no programmed consequence. Following the timeout, milk was available under the FR10 schedule for a total of 15 min.

Drug discrimination training was initiated by administering 0.0032 mg/kg of epibatidine at the beginning of the timeout during some training sessions (i.e., epibatidine training sessions) and saline at the beginning of the timeout during other sessions (i.e., saline training sessions). The epibatidine training dose was selected as that which produced greater than 80% varenicline-appropriate responding in mice (de Moura and McMahon, 2017). During the 10-min timeout, all stimulus and house lights in the operant conditioning chamber were off. Following the 10-min timeout, illumination of the lights in the left and right nose-poke holes provided a stimulus indicating that milk was available for delivery pending completion of the FR10. During epibatidine-training sessions, ten responses in the correct hole (left following epibatidine for half of the mice; right following epibatidine for the other half) resulted in presentation of milk. During saline-training sessions, ten responses in the opposite hole were required to obtain milk delivery. The correct holes were permanently assigned for that mouse. The sequence of training included two consecutive days of epibatidine training followed by two consecutive days of saline training. To pass a training session, two criteria had to be satisfied: 1) a minimum of 80% of the total responses during the 15-min response period needed to be correct; and 2) any incorrect responses made prior to delivery of the first reinforcer needed to be less than 10 (i.e., the first reinforcer had to be obtained with fewer than a total of 20 responses in both holes). Responses in the incorrect hole had no scheduled consequences, i.e., did not reset the ratio requirement. Testing commenced once the criteria were satisfied in five consecutive or six out of seven training sessions, as determined per mouse. After the first test, inter-test training alternated between one epibatidine- and one saline-training session; however, the first training condition (i.e., epibatidine or saline) following each drug test varied non-systematically. If a mouse failed to meet the criteria, a training condition was repeated. However, the next training session was switched to the alternative training condition to avoid more than two consecutive days of a particular training condition. Subsequent test sessions were conducted after the mouse passed at least three consecutive training sessions.

Discrimination Testing

Test sessions were the same as training sessions, except that 10 responses in either hole resulted in presentation of milk and different drugs could be administered. Dose-response tests were conducted with epibatidine, followed by varenicline, nicotine, RTI-102, cytisine, and PNU-282987 in a non-systematic order. However, within each drug test series the doses of each were studied in ascending order. Cocaine and midazolam were studied next. Drugs were administered from doses that produced less than 20% epibatidine-appropriate responding up to doses that produced greater than or equal to 80% epibatidine-appropriate responding, antagonized epibatidine-appropriate responding produced by a test drug, decreased response rate to less than 20% of the saline control, or were deemed potentially toxic. Dose-response functions for all the nAChR agonists were conducted prior to testing with antagonists. Cocaine and midazolam were studied after all the nAChR agonists. Doses were selected based on their capacity to substitute for, or antagonize, the discriminative stimulus of nicotine (1 mg/kg; de Moura and McMahon, 2017). DHβE (3.2 mg/kg, s.c.), mecamylamine (3.2 mg/kg, i.p.), and MLA (10 mg/kg, i.p.) were combined with epibatidine. Dose-response functions for epibatidine were determined twice: once before and again after completion of all tests with other drugs.

Data analyses

Data on discriminative stimulus effects and response rate were expressed as the mean ± standard error of the mean (S.E.M.). The discrimination data were expressed as the percentage of epibatidine-appropriate responses out of the total number of responses per subject. The response rate data were expressed as a percentage of control for each animal, defined as the mean response rate from the five saline training sessions immediately preceding the test. When the response rate was less than 20% of control for an individual, the corresponding discrimination data (assuming at least one response was made) were not plotted or analyzed. Only when all mice in the group responded greater than 20% of saline control were the corresponding discrimination data graphed and analyzed. Every response rate data point was plotted and included for analysis.

Statistical analyses were conducted using GraphPad Prism version 5.0 for Windows (San Diego, CA). Dose-response data were fitted with straight lines using linear regression. The linear portion of a dose-response function, as determined per individual mouse, included not more than one dose producing less than 20% effect. Larger doses of epibatidine, nicotine, varenicline, cytisine, RTI-102, PNU-282987, midazolam, and cocaine were included up to the dose that increased epibatidine-appropriate responding to greater than 80%, decreased response rate to 20% of control or less, or were deemed potentially toxic. The slopes and intercepts of different functions were determined by grouping individual data in a single analysis and were compared with an F-ratio test. If the F-ratio value was significant, then the dose-response functions were considered significantly different from each other. When the mean of epibatidine-appropriate responding was greater than 50%, the ED₅₀ values, potency ratio, and corresponding 95% confidence limits were calculated according to Tallarida (2000). If the 95% confidence limits of the ED₅₀ values did not overlap, or the 95% confidence limits of the potency ratio of the drug in combination with an antagonist did not include 1, the drugs were considered to have significantly different potencies. The effects of antagonists alone were examined with an ANOVA or t-test. Antagonism of epibatidine with mecamylamine and DHβE was examined with F-ratio tests, while antagonism of epibatidine with MLA was examined using a t-test. Drugs were deemed to have fully substituted for the epibatidine discriminative stimulus with ≥80% epibatidine-appropriate responding, partially substituted with epibatidine-appropriate responding between 21–79%, and not substituting with ≤20% epibatidine-appropriate responding.

Results

Substitution Tests

Saline (s.c.) produced 6% epibatidine-appropriate responding (Figure 1, open circle above Saline). The dose-response functions determine for epibatidine, once at the beginning and at the end of the study, were not significantly different from each other; these were averaged for further analyses. Maximum epibatidine-appropriate responding, and ED₅₀ values for epibatidine-like discriminative stimulus and rate-decreasing effects of the test drugs are summarized in Table 1. Epibatidine produced a maximum of 96% epibatidine-appropriate responding at the training dose of 0.0032 mg/kg epibatidine; a dose of 0.01 mg/kg epibatidine decreased response rate to 11% of control (Figure 1, filled circle above Saline). The ED₅₀ values (95% confidence limits) of epibatidine to produce discriminative stimulus and rate-decreasing effects were 0.0026 (0.0012–0.0058) and 0.0042 (0.0032–0.0056) mg/kg (s.c.), respectively (Table 1). Nicotine produced a maximum effect of 18% epibatidine-appropriate responding at 0.56 mg/kg; a dose of 1 mg/kg nicotine decreased response rate to 20% of control (Figure 1, upward triangles). The ED₅₀ value of nicotine to decrease response rate was 0.55 (0.40–0.76) mg/kg (s.c.) (Table 1). Varenicline produced a maximum of 61% epibatidine-appropriate responding at 3.2 mg/kg, and suppressed responding at 5.6 mg/kg (Figure 1, upside-down triangles). The ED₅₀ values of varenicline to produce epibatidine-like discriminative stimulus and rate-decreasing effects were 1.0 (0.62–1.6) and 3.1 (1.3–7.3) mg/kg (s.c.), respectively (Table 1). Cytisine produced a maximum of 33% epibatidine-appropriate responding at a dose of 3.2 mg/kg, and decreased response rate to 14% of control at 5.6 mg/kg (Figure 1, squares). The ED₅₀ value of cytisine to decrease response rate was 3.7 (3.1–4.5) mg/kg (s.c.) (Table 1). RTI-102 produced a maximum effect of 69% epibatidine-appropriate responding at a dose of 1.78 mg/kg, and at a dose of 3.2 mg/kg decreased response rate to 2% control (Figure 1, open diamonds). The ED₅₀ values of RTI-102 to produce epibatidine-like discriminative stimulus and rate-decreasing effects were 1.4 (0.85–2.2) and 1.3 (1.0–1.7) mg/kg (s.c.), respectively (Table 1). PNU-282987 produced a maximum of 29% epibatidine-appropriate responding at a dose of 32 mg/kg, and reduced the response to 11% of control at a dose of 56 mg/kg (Figure 1, circles with cross hatch). The ED₅₀ value of PNU-2828987 to produce rate-decreasing effects was 25 (19–33) mg/kg (i.p.) (Table 1).

Figure 1:

Discriminative stimulus effects (top) and rate-decreasing effects (bottom) of epibatidine (filled circles; 0.001–0.0032 mg/kg (n=8), 0.0056 mg/kg (n=5) and 0.01 mg/kg (n=4)), saline (open circles; n=8), nicotine (upward triangles; n=7 for % epibatidine-appropriate responding, n=8 for response rate), varenicline (downward triangles; all doses (n=8) except 5.6 mg/kg (n=3)), cytisine (squares; all doses (n=8) except 5.6 mg/kg (n=7)), RTI-102 (squares; 0.56 and 1.0 mg/kg (n=8), 1.78 mg/kg (n=7), 3.2 mg/kg (n=3)), and PNU-282987 (circles with cross hatch; 10–32 mg/kg (n=8) and 56 mg/kg (n=7)) in mice trained to discriminate 0.0032 mg/kg epibatidine from saline. Top panel shows epibatidine-appropriate responding on the ordinate and drug dose in mg/kg (log scale) on the abscissae. Bottom panel shows response rate normalized to saline control on the ordinate as function of dose in mg/kg (log scale) on the abscissae. Epibatidine, saline, nicotine, varenicline, RTI-102, and cytisine were administered s.c. immediately prior to placing each subject in the box for the 10-min pretreatment interval; PNU-282987 was administered i.p. immediately prior to placing each subject in the box for the 10-min pretreatment interval.

Table 1.

Maximum percentage of epibatidine-appropriate responding (SEM), and ED₅₀ values (95% confidence intervals) for discriminative stimulus and rate-decreasing effects for individual test drugs in mice trained to discriminate 0.0032 mg/kg epibatidine from saline as shown in the Figure 1.

Test drug	Maximum % Drug-appropriate Responding (95% CI)	Discriminative stimulus ED50 (95% CI)	Response rate ED50 (95% CI)
Saline	6 (3)	ND	ND
Epibatidine	96 (±1) @ 0.0032 mg/kg	0.0026 (0.0012–0.0058)	0.0042 (0.0032–0.0056)
Nicotine	18 (±10) @ 0.56 mg/kg	ND	0.55 (0.40–0.76)
Varenicline	61 (±16) @ 3.2 mg/kg	1.0 (0.62–1.6)	3.1 (1.3–7.3)
Cytisine	33 (±13) @ 3.2 mg/kg	ND	3.7 (3.1–4.5)
RTI-102	69 (±11) @ 1.78 mg/kg	1.4 (0.85–2.2)	1.3 (1.0–1.7)
PNU-282987	29 (±13) @ 32 mg/kg	ND	25 (19–33)
Midazolam	45 (±13) @ 5.6 mg/kg	ND	11 (6.4–18)
Cocaine	14 (±10) @ 5.6 mg/kg	ND	4.6 (2.6–8.3)

ND; not determined because effect did not reach 50%.

CI; confidence intervals

Midazolam produced a maximum effect of 45% epibatidine-appropriate responding at a dose of 3.2 mg/kg, and abolished responding at a dose of 32 mg/kg (Figure 2, squares with cross hatch). The ED₅₀ value of midazolam to produce rate-decreasing effects was 11 (6.4–18) mg/kg (s.c.) (Table 1). Cocaine produced a maximum effect of 14% epibatidine-appropriate responding at a dose of 5.6 mg/kg, and reduced the rate of responding to 1% of saline control at a dose of 17.8 mg/kg (Figure 2, diamonds with cross hatch). The ED₅₀ value of cocaine to decrease response rate was 4.6 (2.6–8.3) mg/kg (i.p.) (Table 1).

Figure 2:

Discriminative stimulus effects (top) and rate-decreasing effects (bottom) of midazolam (circles with cross hatch; 1.0 and 3.2 mg/kg (n=8), 10 mg/kg (n=5 for % epibatidine-appropriate responding, n=7 for response rate) and 17.8 mg/kg (n=1)) and cocaine (diamonds with cross hatch; 3.2–10 mg/kg (n=8) and 17.8 mg/kg (n=7)). Top panel shows epibatidine-appropriate responding on the ordinate and drug dose in mg/kg (log scale) on the abscissae. Bottom panel shows response rate normalized to saline control on the ordinate as function of dose in mg/kg (log scale) on the abscissae. Midazolam and cocaine were administered s.c. and i.p., respectively, immediately prior to placing each subject in the box for the 10-minute pretreatment interval; cocaine was administered i.p.

Effects of epibatidine in combination with DHβE, mecamylamine, and MLA

DHβE (3.2 mg/kg, s.c.) alone produced 11% epibatidine-appropriate responding and a response rate of 89% control (Figure 3, upward triangles above Saline). DHβE produced a 2.1 (1.9–2.5) fold rightward shift in the epibatidine discriminative stimulus dose-response function (Figure 3, upward triangles). However, DHβE did not significantly alter the rate-decreasing effects of epibatidine (Figure 3, upward triangles). Mecamylamine (3.2 mg/kg, i.p.) alone produced 4% epibatidine-appropriate responding, and decreased response rate to 51% of control (Figure 3, downward triangles above Saline). When mecamylamine was administered in combination with epibatidine, mecamylamine produced a significant antagonism of the discriminative stimulus effects of epibatidine (F_2,29=28, p<0.0001; Figure 3, downward triangles). For epibatidine alone and in combination with mecamylamine, the slopes of the two functions were different (F_1,29=16, p=0.0004). In contrast, mecamylamine did not antagonize the rate-decreasing effects of epibatidine (F_2,33=1.9, p=0.16). The largest dose of MLA (10 mg/kg, i.p.) administered in combination with epibatidine that did not decrease response rate to less than 20% of control produced 1% epibatidine-appropriate responding when substituted for epibatidine; response rate was 79% of control (Figure 3, squares above Saline). MLA (10 mg/kg, i.p.) did not significantly antagonize the discriminative stimulus (t₆=1.19, p=0.29) or the rate-decreasing (t₆=0.55, p=0.61) effects of epibatidine (Figure 3, squares).

Figure 3:

Discriminative stimulus effects (top) and rate-decreasing effects (bottom) of epibatidine alone (filled circles, replotted from Figure 1; 0.001–0.0032 mg/kg (n=8), 0.0056 mg/kg (n=5) and 0.01 mg/kg (n=4)) and in combination with 3.2 mg/kg DHβE (downward triangles; saline and 0.0032 mg/kg (n=8) and 0.0056 mg/kg (n=4 for % epibatidine-appropriate responding, n=8 for response rate)), 3.2 mg/kg mecamylamine (upward triangle; n=8 for all data points), or 10 mg/kg MLA (squares; saline (n=5) and 0.0032 mg/kg (n=6)). Top panel shows epibatidine-appropriate responding on the ordinate and drug dose in mg/kg (log scale) on the abscissae. Bottom panel shows response rate normalized to saline control on the ordinate as function of dose in mg/kg (log scale) on the abscissae. Epibatidine and DHβE were administered s.c. at 10- and 15-min pretreatment intervals, respectively. Mecamylamine and MLA were administered i.p. at 20-min pretreatment intervals.

Discussion

The present study established epibatidine as a discriminative stimulus in mice. The β2-containing nAChR antagonist DHβE (3.2 mg/kg) and the nonselective nAChR non-competitive antagonist mecamylamine (3.2 mg/kg) both significantly antagonized the discriminative stimulus effects of epibatidine, but not its rate-decreasing effects. In contrast, the α7 nAChR antagonist MLA (10 mg/kg) did not antagonize any of the effects of epibatidine. In addition, the α7 nAChR agonist PNU282987 did not fully generalize to the epibatidine discriminative stimulus, supporting the hypothesis that α7 nAChRs are minimally involved in the discriminative stimulus effects of epibatidine. This profile of discriminative stimulus effects is consistent with in vitro studies demonstrating that epibatidine is selective for β2-containing nAChRs (Grady et al., 2010). On the other hand, up to doses that significantly decreased the rate of responding, the low efficacy α4β2* nAChR agonists varenicline and RTI-102 partially substituted for the epibatidine discriminative stimulus (maximum 61 and 69% epibatidine-appropriate responding, respectively). Nicotine and cytisine, both relatively non-selective nAChR agonists, produced no greater than 33% epibatidine-appropriate responding.

Nicotine did not substitute for the epibatidine discriminative stimulus, which contrasts with data from squirrel monkeys trained to discriminate epibatidine (Desai and Bergman, 2015; Desai et al., 2016). However, besides the difference in species, the reinforcer type in the current study (milk presentation) differed from that in studies by Desai and colleagues (stimulus shock termination). It is possible that squirrel monkeys responding under a stimulus-shock termination schedule of reinforcement are more resistant to the rate-decreasing effects of nicotine. It is important to consider the possibility that various behavioral effects of nicotinic ligands could be mediated by differing subsets of nAChRs. In support of this view, studies in which tolerance and cross-tolerance were studied in mice treated daily with nicotine or varenicline suggest that the behaviorally disruptive effects of nicotine and epibatidine may be mediated by different nAChR populations (de Moura and McMahon, 2017; 2019). Inasmuch as epibatidine has previously been demonstrated to substitute for a nicotine discriminative stimulus in mice (Rodriguez et al., 2014; de Moura and McMahon, 2017), failure of nicotine to substitute for the epibatidine discriminative stimulus may reflect different receptor mechanisms that mediate the rate-decreasing, but not necessarily the discriminative stimulus, effects of epibatidine and nicotine. In that regard, similar to varenicline-trained mice (de Moura and McMahon, 2017), the rate-decreasing effects of nicotine may have prevented evidence of its substitution for epibatidine in the present study.

Previous studies have demonstrated that the training dose of the drug used as a discriminative stimulus can impact the results obtained with test drugs (Young et al., 1992; Jutkiewicz et al., 2011; Cunningham and McMahon, 2013). For instance, in rats trained to discriminate either 3.2 or 5.6 mg/kg morphine, both morphine and etorphine were more potent in the 3.2 mg/kg discrimination compared to the 5.6 mg/kg discrimination (Young et al., 1992). Furthermore, in the 3.2 mg/kg morphine discrimination, nalbuphine substituted for morphine, and decreased the dose of morphine needed to substitute (i.e. potentiation) (Young et al., 1992). However, in rats trained to discriminate 5.6 mg/kg morphine, nalbuphine failed to substitute for morphine and antagonized the discriminative stimulus effects of morphine. In Jutkiewicz et al. (2011), cytisine partially substituted for nicotine in rats trained to discriminate 0.32 mg/kg nicotine, but substituted fully in rats trained to discriminate 1.78 mg/kg nicotine. Therefore, varying the epibatidine training dose might yield a different substitution and antagonism profile. If the training dose were to be increased, then nicotine might be more likely to substitute for epibatidine. This may occur because increasing the epibatidine training dose would increase the probability that the receptors that mediate the rate-decreasing effects of nicotine would now be activated and rendered less sensitive due to agonist-induced desensitization (Giniatullin et al., 2005), i.e., cross-tolerance from epibatidine to nicotine might occur in measures of rate-decreasing effects.

Varenicline, cytisine, and RTI-102 have all been characterized in vitro as having lower efficacy than epibatidine at α4β2* nAChRs. Consistent with those reports, all three of those drugs only partially substituted for epibatidine. Notwithstanding the influence of the training dose on patterns of generalization in drug discrimination, differences in efficacy and/or selectivity are the most likely mechanisms for partial substitution (Colpaert et al., 1980; Colpaert and Janssen, 1984). Due to constraints involving the lifespan of mice, all the tests required to discern whether partial substitution is a function of differences in efficacy or selectivity could not be conducted in the present experiment and should be conducted in future studies (e.g., combining epibatidine with varenicline, cytisine, and RTI-102); however, evidence from previous reports may provide some clarity as to why these drugs partially substituted. For example, epibatidine has been shown to fully substitute for the varenicline discriminative stimulus (de Moura and McMahon, 2017), suggesting overlapping receptor mechanisms. In that regard, assuming that epibatidine substituting for varenicline is indicative of overlapping receptor pharmacology, partial substitution of varenicline in epibatidine-trained mice could reflect differences in efficacy at those same receptors. According to in vitro studies, RTI-102 and cytisine have less efficacy than epibatidine at α4β2* nAChRs (Abdrakhmanova et al., 2006), and their partial substitution for the epibatidine discriminative stimulus is consistent with differences in efficacy. However, without a more systematic approach to address these questions, these interpretations of partial substitution for the epibatidine discriminative stimulus are, at present, speculative.

In mice trained to discriminate nicotine, mecamylamine and DHβE antagonized epibatidine (Rodriguez et al., 2014; de Moura and McMahon, 2017), suggesting that the discriminative stimulus effects of epibatidine were due to its effects at nAChRs. Antagonism of epibatidine by DHβE in both nicotine- and varenicline-trained animals, even when the training drug was not antagonized by DHβE (e.g., varenicline), strongly suggested that the discriminative stimulus effects of epibatidine, at least in part, were mediated by α4β2* nAChRs (de Moura and McMahon, 2017). The doses of mecamylamine and DHβE chosen in the present study were the same doses shown in previous studies to significantly antagonize the effects of nicotine (de Moura and McMahon, 2017). Although only a single dose of each antagonist was studied, the present results support the hypothesis that the discriminative stimulus effects of epibatidine are mediated by nAChRs, and more specifically, at least partly mediated by α4β2* nAChRs.

Failure of cocaine to generalize to the epibatidine discriminative stimulus illustrates some degree of pharmacological selectivity of the current assay. However, midazolam partially substituted for the epibatidine discriminative stimulus, a result previously observed in mice discriminating low to medium doses of nicotine (Cunningham and McMahon, 2013). It is unclear why midazolam partially substitutes for epibatidine, but future studies should vary the epibatidine training dose in order to attempt to increase the pharmacological selectivity of this discrimination.

The key findings from this study are that, like in squirrel monkeys (Desai and Bergman, 2015; Desai et al., 2016; Withey et al., 2018), epibatidine can function as a reliable discriminative stimulus in mice, and that this discrimiantive stimulus is sensitive to antagonism by both mecamylamine and DHβE. Furthermore, drugs described as low efficacy agonists in vitro (e.g., varenicline, cytisine, RTI-102) only partially substituted for the epibatidine discriminative stimulus. Whereas this study established a new foundation for preclinical research in mice using epibatidine, it is important that future research is conducted to elucidate the relative selectivity of this discrimination procedure relative to nicotine as a training drug and to optimize the conditions in which epibatidine can be discriminated. Establishing epibatidine as a discriminative stimulus in mice provides the framework for utilizing the advantages that exist in rodent research in the development of candidate medications for smoking cessation.

Non-standard abbreviations

DHβE

dihydro-β-erythroidine

FR

fixed ratio

MLA

methyllycaconitine

nAChR

nicotinic acetylcholine receptor

PNU-282987

N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]-4-chlorobenzamide

RTI-102

2-fluoro-3-(4-nitro-phenyl)deschloroepibatidine