Discriminative stimulus effects and metabolism of ethanol in rhesus monkeys

Abstract

Background:

Animal models are an essential feature of drug and pharmacotherapy development for treating alcohol use disorders (AUDs). The rhesus macaque is a robust animal model for many aspects of AUDs particularly in exploiting individual differences in oral self-administration of ethanol, endocrine orchestration of stress response and menstrual cycle characteristics. However, the clearance rates of ethanol have not been reported in this species, and the GABA_A and NMDA receptor involvement in ethanol’s discriminative stimulus effects have not been fully characterized.

Methods:

Ethanol clearance rates following two doses of ethanol on separate days (0.5 and 1.0 g/kg, i.g.) were determined in eight young adult male rhesus macaques. The ethanol was given by nasogastric gavage and repeated blood samples were taken over 5 hours without sedation. Next, all subjects were trained on a two-choice 1.0 g/kg ethanol (i.g.) vs. water discrimination with a 60 min pre-treatment period to capture peak blood ethanol concentration (BEC). Substitution testing was conducted with GABA_A ligands pentobarbital (i.g. and i.m.) and midazolam (i.g.), as well as NMDA antagonist MK-801 (i.m.).

Results:

Peak blood ethanol concentrations (BECs) were 34 and 87 mg/dl for 0.5 and 1.0 g/kg doses respectively, and occurred at 66 and 87 minutes following gavage. All GABA_A and NMDA ligands tested resulted in responding on the ethanol-appropriate lever with the potency ranking of MK-801 (ED₅₀: 0.017 mg/kg) > midazolam (ED₅₀: 1.6 mg/kg) > pentobarbital (ED₅₀: 3.7 mg/kg) > ethanol (ED₅₀: 700 mg/kg, or 0.7 g/kg) in these subjects.

Conclusions:

These results suggest that the compound discriminative stimulus effects of ethanol are highly consistent across species, providing further support for the rhesus macaque as strong model for pharmacotherapy development for AUD.

Introduction

Interoceptive drug effects are strongly based in receptor pharmacology and can be queried behaviorally using a drug discrimination procedure (Grant, 1999). With this approach, studies have established that ethanol is a stimulus complex with concurrent activity at multiple receptor systems (Grant, 1999; Green and Grant, 1998; Stolerman et al., 1999; Stolerman and Olufsen, 2001). Specifically, in drug discrimination, interoceptive drug cues are trained as discriminative stimuli, enabling an animal to reliably report the presence of a specific drug cue through a discrete behavioral response (Stolerman, 2014). Ethanol’s discriminative stimulus properties are primarily mediated by activity at the GABA_A and NMDA receptor systems. Specifically, stimulus effects of ethanol are most strongly associated with positive modulatory action at the GABA_A receptor and non-competitive antagonism at the NMDA receptor (see Grant, 1994). The relative contribution of each of these receptor systems to ethanol’s discriminative stimulus effects is dependent on ethanol training dose in rodents, with lower ethanol doses being most similar to GABA_A receptor positive modulators, and higher ethanol doses being most similar to NMDA receptor antagonists (Stolerman et al., 2011). However, this dose relationship does not directly translate to macaque monkeys, with GABA_A receptor positive modulation and NMDA receptor antagonism remaining prominent at both low and high ethanol doses in cynomolgus macaques (Grant et al., 2000; Grant et al., 2008a; Vivian et al., 2002; Allen et al., 2017).

Importantly, there is evidence that the alcohol’s subjective effects (i.e., interoceptive effects) can perpetuate continued drinking, indicating that the way alcohol makes you feel may contribute to its abuse potential (Holdstock et al., 2000; Kelly et al., 2003; Childs et al., 2011). Thus, investigation of the mechanisms that mediate alcohol’s interoceptive effects can inform our understanding of alcohol’s actions in the brain to improve targeting strategies for potential pharmacotherapies.

Despite the longstanding use of rhesus macaques in ethanol self-administration experiments (Winger and Woods, 1973; Kornet et al., 1990; Grant et al., 2008b), the discriminative stimulus effects of an ethanol training dose have only been examined in one published study to date (Berro et al., 2019). Long-term ethanol self-administration experiments have indicated that rhesus monkeys can model many aspects of AUD in humans, particularly the individual risks to become a heavy drinker (Grant et al., 2008b; Baker et al., 2014; Baker et al., 2017), including the age of onset of drinking (Helms et al., 2014a), the adrenal response to long term drinking (Helms et al., 2014b), and the response to repeated abstinence periods (Allen et al., 2018). Identifying the pharmacological basis of ethanol’s discriminative stimulus effects in rhesus monkeys may help identify candidate receptor systems for the development of targeted pharmacotherapies (Grant, 1999). In addition to limited understanding of ethanol’s discriminative stimulus effects in rhesus macaques, there have also been minimal studies characterizing the absorption and elimination rate of ethanol in this species. These data are essential to understanding the dose-dependent receptor basis of alcohol through drug discrimination, but also inform our understanding of outcomes associated with binge-level alcohol consumption. Thus, in order to improve our understanding of alcohol receptor pharmacology and improve translatability of alcohol-related research in rhesus monkeys, the current study examined alcohol pharmacokinetics in low-to-moderate doses (0.5 g/kg and 1.0 g/kg) as well as discriminative stimulus effects of a 1.0 g/kg training dose of ethanol. This dose was selected for the discrimination experiments because it was expected to result in blood alcohol levels of ~80 mg/dl (Green et al., 1999a), which translates to legal intoxication in humans. Additionally, other non-human primate species (cynomolgus macaque and squirrel monkeys) have demonstrated GABA_A and NMDA receptor substitution at this training dose, allowing us to compare across species within both receptor systems (Platt et al. 2005; Grant et al., 2000; Vivian et al., 2002).

Material and Methods

Animals

Eight experimentally-naïve late adolescent male rhesus monkeys (Macaca mullata) were used in the current study (3.9–4.2 years old, 5.5–7.6 kg at assignment), and the experiments took place over 28 months (6.3–6.6 years old, 7.8–10.2 kg at end). This age range was selected for this study based on a population analysis from our laboratory that found exposure to alcohol during late adolescence and early adulthood represents the highest risk for developing a heavy drinking phenotype in rhesus (Helms et al., 2014a). All monkeys were born and raised at the Oregon National Primate Research Center (ONPRC; Beaverton, OR) and confirmed to not have common parents or grandparents. They were housed in stainless steel one-over-one cages (32 × 28 × 32 in) that were attached along the vertical axis into quad (2 × 2) cages to allow for side-by-side pair housing. All monkeys lived in a single housing room and were pair-housed at all times, except for during behavioral testing and feeding (3–4 hours/day). Daily ration of food for each monkey was adjusted throughout the experiment to maintain weight gain at a level of ~100–200g per month. The housing room was temperature (20–22°C) and humidity (65%) controlled, with a 12-hr on/off light cycle (lights on at 7AM). All monkeys had visual, auditory, and olfactory contact with other members of the study. Monkeys were weighed weekly without sedation and were monitored throughout the experiment by veterinary staff. All procedures were conducted in accordance with NIH and the Guide on the Care and Use of Laboratory Animals and approved by the IACUC at ONPRC.

Behavioral testing apparatus

Discrimination training and testing sessions were conducted 4–6 days/week in four ventilated, sound attenuating operant chambers (1.50×0.74×0.76m; Med Associates, Inc., St. Albans, VT) in a behavioral suite down the hall from the housing room. The apparatus has been described in detail previously (Helms and Grant, 2011). Each chamber had an operant panel (0.48×0.69 m) equipped with two retractable levers, three lights (red, amber, green) above each lever, and a centrally located white light above a food magazine. The red and green lights were not active during discrimination training and testing, and only the center amber light was illuminated when the associated lever was available. Two house lights and a fan were located in the top rear of the chamber. The panel was accessible from a primate chair (1.17×0.61×0.61 m; Plas Labs, Lansing, MI) that had a food magazine tray. One-gram banana flavored pellets (Bio-Serv) were delivered through vinyl tubing attached to a feeder (Med Associates, ENV-203–1000) located outside the chamber, and the central white stimulus light was illuminated at the time of delivery. All events were programmed and recorded by LabView (version 4.0.1., National Instruments, Austin, TX) connected to a computer interface (Med Associates, Fairfax, VT) attached Mac computer.

Procedural training

Upon arrival to the laboratory, all monkeys were trained with positive reinforcement (fruit and seeds) to sit in a primate chair with the guided pole-and-collar technique. The monkeys were then transported in chairs to the testing chambers and trained to press an extended lever which resulted in the delivery of a food pellet. Training sessions were initially conducted with only one of the two levers extended into the chamber, but once responding was stable and consistent on both levers under a fixed ratio 1 (FR1) schedule, the FR was escalated on an individual basis. Terminal FRs were selected that resulted in delivery of 25 food pellets in approximately 10 minutes, which provided similar blood ethanol concentrations (BECs) across monkeys when performing the ethanol discrimination.

Following this initial training, the monkeys were trained to accept a nasogastric infant feeding tube (5 French, 36” length), for accurate placement into the stomach and verified by endoscopy as 14–16” depth placement from the nostril in adult male rhesus monkeys. During this time, monkeys were also trained to comply with awake venipuncture for blood collection from the medial saphenous vein to determine BECs following testing.

Ethanol pharmacokinetics

Following training to accept the nasogastric tube, but before training the discrimination (less than 7 alcohol administrations per monkey), ethanol (20% w/v ethanol in water) was gavaged and repeated blood samples (20μl from medial saphenous vein) were taken for BEC analysis over a 5-hour period to capture the absorption and elimination phases. Ethanol administration and blood sampling were performed without sedation following an overnight fast, and at least 72 hours after the last ethanol administration. Two doses of ethanol were tested at least 2 weeks apart: 0.5 g/kg (n=4) and 1.0 g/kg (n=5, same 4 subjects, plus an additional monkey). Blood samples were collected at the following time points following ethanol gavage: 15, 30, 45, 60, 75, 90, 120, 180, 240, and 300 min.

Discrimination training

Next, all monkeys were trained to discriminate 1.0 g/kg ethanol (20% w/v in water) from water (equivalent volume to 1.0 g/kg ethanol) with a 60 minute pre-treatment interval. During both ethanol and water gavages, a flavored 1-gram pellet (Bio-Serv) was given halfway through the gavage and immediately after to mask any taste cues. The animal was then immediately placed in a darkened operant chamber for a programmed 60-minute pretreatment time, after which the house lights turned on, two levers were extended into the chamber and associate stimulus lights turned on, signaling the start of the session. Sessions ended when 25 pellets were earned under the terminal FR, or at 30 minutes, whichever came first. Terminal FRs ranged from FR20 to FR110 over the course of the experiment. For the first 5 training sessions, water was administered and only the water-appropriate lever was extended (forced choice procedure). The same conditions were repeated for the next five sessions, except ethanol was administered before the session, and only the ethanol-appropriate lever was extended. Lever assignments associated with ethanol (left or right) were counterbalanced across animals. For the remaining training sessions, both levers were extended into the chamber and ethanol or water was administered on a double-alternating schedule (e.g., 2 water days followed by 2 ethanol days, and so on). Successful completion of the terminal FR on the condition-appropriate lever resulted in the delivery of a 1 g banana-flavored food pellet. Responding on the inappropriate lever reset the FR requirement and was not reinforced. Discrimination training was complete once the monkeys met the following criteria for 5 consecutive sessions: 1) ≥90% of total session responding on the condition-appropriate lever, and 2) ≥70% of the first FR responses on the condition-appropriate lever.

Substitution Testing

Test sessions were identical to training sessions, with two key differences: 1) completion of consecutive FR responses on either lever resulted in the delivery of a food pellet, and 2) the route of administration varied based on the drug administered. The pretreatment time was kept constant at 60 minutes following drug administration by adjusting the dose of the test drugs. In general, test sessions occurred 1–2 days/week, with training sessions on the intervening days. If performance on a training day did not meet criteria, then training sessions were continued until criteria was met for 3 consecutive sessions. Each drug and dose combination was tested on a single day (single dosing procedure) and each test dose was double determined, counterbalancing for the training session on the day prior to each test. Negative control tests (morphine and muscimol) were not double determined.

Following training, ethanol substitution testing was conducted across a range of doses, in all monkeys (0.0–2.0 g/kg), beginning with the training dose (1.0 g/kg). Following ethanol dose-response determination tests, selected doses of morphine (μ-opioid receptor agonist, 0.01–1.7 mg/kg; i.m.) were tested. Ethanol does not have any direct activity at the μ-opioid receptor, so including morphine (μ-opioid agonist) substitution tests will confirm that the ethanol discrimination is specific for ethanol-like discriminative stimulus effects (Grant, 1994; 1999). Next, substitution testing was conducted with pentobarbital (barbiturate, GABA_A receptor positive allosteric modulator, 0.56–10.0 mg/kg; i.g.), midazolam (benzodiazepine, GABA_A receptor positive allosteric modulator 0.30–5.6 mg/kg; i.g.), and MK-801 (NMDA receptor antagonist, 0.003–0.10 mg/kg; i.m.). Muscimol (GABA_A receptor agonist, 0.3–0.56 mg/kg; i.m.) was also tested in a subset of monkeys to confirm ethanol’s stimulus effects were selective for positive modulation at the GABA_A receptor, rather than direct agonism (Grant et al., 2000). In general, a new drug was not introduced until the all doses of the previous drug were tested. For each drug, testing began at an intermediate dose and then escalated incrementally until a dose was found that either substituted fully (≥80% ethanol-appropriate responding) or decreased response rates (commonly <65% baseline). Lower doses were tested until a dose was found that did not produce substitution (≤20% ethanol-appropriate responding). In some cases, higher doses of morphine did not produce a significant effect on response rate, but did produce an increase in scratching (operationally defined as 3–4 fold increase in scratching compared to water session) to verify a behaviorally active dose was given. Additionally, not all subjects showed a decrease in response rate following muscimol, but did vomit during or after the session. If side effects were observed, dose levels were not increased further. For all i.m. test sessions, monkeys first received a water gavage to match the training procedures.

In addition to blood samples to determine ethanol pharmacokinetics, 20μl blood samples were also collected immediately following ethanol test sessions (approx. 75 minutes following ethanol administration) for BEC analysis. All BEC samples were collected in a capillary tube and diluted with 500μl sterile water, placed in airtight containers and stored at − 4°C until assayed using headspace gas chromatography (Agilent Technologies, Santa Clara, CA). Samples were analyzed using linear regression against a standard curve that included 25, 50, 100, 200, and 400 mg/dl.

Drugs

In general, all drugs were prepared fresh on the morning of the test session or the night before. Ethanol (95%) was diluted to 20% w/v in water for doses ≤ 1.0 g/kg, and to 25% and 30% for 1.5g/kg and 2.0g/kg tests respectively, leading to gavage volumes between 30–50ml for all doses tested. Pentobarbital was purchased in prepared form (Nembutal, 50 mg/ml), and midazolam hydrochloride (Sigma Aldrich) was diluted in saline to 3 mg/ml. Both drugs were administered through the nasogastric gavage, followed by a gavage of water up to the training dose volume (30–45ml). (+)-MK-801 hydrogen maleate (Sigma Aldrich) was diluted in saline to 0.2–0.5 mg/ml, morphine maleate salt (Sigma Aldrich) was diluted in saline to 5 mg/ml, and muscimol (Tocris) was diluted in saline to 7 mg/ml. Drug concentrations were based on the salt form of the drug. All i.m. drug preparations were less than 3ml, and if the injection volume exceeded 1ml, it was given across two injection sites. Vehicle injections matched the maximum drug volume given. All drugs administered i.m. (pentobarbital, MK-801, morphine, muscimol) were filtered through a 20μm millipore filter into a sterile vial prior to administration.

Data analysis

Data from the BEC time course were used for three analyses: 1) determination of peak BEC, 2) time to peak, and 3) calculation of elimination rates (β). Elimination rates were calculated using the linear portion of each elimination curve (0.5 g/kg ethanol: 90–180 min; 1.0 g/kg ethanol: 120–300 min). These time points were put into a linear regression and the slope of the line for each monkey was used to calculate an individual elimination rate per hour. Elimination rates were then averaged across the group for between subjects and within-subjects comparisons.

Following each session, the percentage of ethanol-appropriate responding and response rate (responses/second) were calculated for each subject. In cases where substitution was double-determined, the ethanol-appropriate responses and response rates were averaged for the two sessions before further analyses and served as the primary dependent variables. Full substitution was defined as ≥80% responding on the ethanol-appropriate lever and no substitution was defined as ≤20% ethanol-appropriate responding. Partial substitution was between 21–79% ethanol-appropriate responding. For all dose response curves that reached full substitution, the ED₅₀ (50% effective dose) was calculated using linear interpolation with the two doses that encompassed the 50% effect. ED₅₀ was then used in paired t-tests or RM ANOVAs comparing drug potency.

Baseline response rates were calculated as a rolling average of three water sessions prior to, or at the beginning of, a new dose response determination for the given drug. These baseline response rates were compared to water or saline (i.g. or i.m. routes, respectively) and were tested for equivalency. This method accounted for variance in a single subject’s response rate over the duration of the experiment (28 months).

Results

Ethanol time course

Under fasted conditions, peak BEC following 1.0 g/kg ethanol (i.g.) was 86 ± 6 mg/dl (range: 80–95 mg/dl) and occurred between 75–90 minutes following ethanol gavage (87 ± 6.7 min) (Figure 1). Peak BEC following 0.5 g/kg ethanol (i.g.) was 34 ± 5 mg/dl (range: 28–40 mg/dl) and occurred at variable time points between 45–90 min following ethanol administration across monkeys (67.5 ± 19.4 min). The effect of dose on elimination rate was compared using a paired t-test of the four subjects that were tested at both dose levels. There was no significant effect of dose on elimination rate (t(3)=1.6, p=0.2) and the group averages were 14.8 ± 1.7 mg/dl/hr following 0.5 g/kg ethanol (n=4), and 14.7 ± 3.6 mg/dl/hr following 1.0 g/kg ethanol (n=5). The time to peak and peak BEC recorded was used to define our discrimination training parameters. We selected a 60 minute pretreatment interval following a 1.0 g/kg ethanol gavage to capture the final rising phase of BEC between 70–83 mg/dl during the 30 minute testing period (Figure 1).

Figure 1.

Blood ethanol concentration (BEC) time course following 0.5 g/kg (n=4) and 1.0 g/kg (n=5) ethanol gavage (i.g.). Four subjects were tested at both doses, and one additional subject was included in the 1.0 g/kg ethanol group. All data are plotted as mean ± SD.

Ethanol discrimination and substitution

All monkeys successfully acquired the discrimination in 81 ± 21 sessions (mean ± SD, range: 52–114 sessions, n=8). A representative acquisition curve is shown in Figure 2. Following training, responding on the ethanol-lever increased as a function of ethanol dose (F(4,28)=270.3, p<0.0001), with only two monkeys showing partial substitution of 0.5 g/kg ethanol (Figure 3a; ED₅₀=0.7 ± 0.1 g/kg). Ethanol ED₅₀ was positively correlated with BEC immediately following the 0.5 g/kg ethanol test session, indicating that higher BECs were associated with increased ethanol-appropriate responding (r=0.74, p=0.03). All monkeys showed generalization of higher test doses (1.5–2.0 g/kg) to the 1.0 g/kg ethanol training dose. BEC following the testing session also increased as a function of ethanol dose (Figure 3b; F(3,21)=52.9, p<0.0001). Since BECs during testing were taken at approximately 75 minutes following gavage, and samples were collected at 75 minutes in the metabolism study, BEC following 1.0 g/kg could be compared across long term exposure (approximately 4–8 months). BEC was not significantly different between early ethanol administrations (mean BEC: 78 mg/dl) and 4–8 months of ethanol gavage 2–3 times per week (mean BEC: 73 mg/dl) (n=5, t(4) = 0.85, p=0.43). There was no effect of ethanol dose on response rate (Table 1; F(4,28)=1.4, p=0.26). In order to confirm the specificity of the discrimination for ethanol-like stimulus effects, morphine substitution tests were conducted. Morphine did not substitute for ethanol in any of the eight monkeys tested (mean percent ethanol-appropriate responding below 2%), even at doses that were behaviorally active (Table 2).

Figure 2.

Representative discrimination training acquisition curve for a single subject (monkey 8 in Figures 3 and 4). Total session ethanol-appropriate responding for ethanol (black circles) and water (grey triangles) discrimination training sessions for a single subject. Dotted lines at 10% and 90% represent discrimination criteria for water and ethanol sessions respectively. 10% ethanol-appropriate responding during water sessions corresponds to 90% water-appropriate responding and vice versa.

Figure 3.

Ethanol dose response function and post-session BECs. a) Ethanol dose response curves plotted for each individual subject. Each data point represents an average for each subject (double determination, n=8). Dotted lines represent the threshold for full substitution (≥80% on ethanol-appropriate lever) and no substitution (≤20% on ethanol-appropriate lever). The area between the dotted lines indicates partial substitution. b) BEC samples taken immediately following the test session, between 70–90 min post-ethanol administration for each monkey (single determination).

Table 1.

Response rates during test sessions (mean ± SD)

Test Drug	Dose (mg/kg)	Resp. rate (% baseline)	# of subjects
Ethanol	0	100 ± 6	8
	500	102 ± 22	8
	1000	119 ± 38	8
	1500	128 ± 70	8
	2000	122 ± 58	8
Pentobarbital	0.56	130	1
	1.0	135 ± 28	4
	1.7	136 ± 63	4
	3.0	139 ± 80	8
	5.6	119 ± 32	6
	10.0	78	1
Midazolam	0.3	79 ± 15	2
	0.56	115 ± 26	5
	1.0	141 ± 69	5
	1.7	129 ± 59	6
	3.0	73 ± 16	4
	5.6	74	1
MK-801	Saline	93 ± 24	5
	0.003	114 ± 16	2
	0.0056	147	1
	0.01	105 ± 17	5
	0.017	107	1
	0.03	74 ± 21	5
	0.056	53	1

Table 2.

Morphine and muscimol substitution and response rates. (mean ± SD)

Test Drug	Dose (mg/kg)	% Ethanol responses	Full substitution	Resp. rate (% saline)	Beh. Active dose1	# of subjects
Morphine (n=8)	0.1	1.7 ± 2.4	0/2	93 ± 31	0/2	2
	0.3	0.8 ± 1.8	0/8	77 ± 28	2/8	8
	1.0	0.3 ± 0.3	0/6	67 ± 32	5/6	6
	1.7	0.0	0/1	84	1/1	1
Muscimol (n=4)	0.3	26 ± 46	1/4	79 ± 15	3/4	4
	0.56	0.1 ± 0.2	0/3	59 ± 27	3/3	3

¹Number of monkeys in which a given dose demonstrated to be behaviorally active.

GABA_A and NMDA substitution

Pentobarbital produced dose-dependent increases in ethanol-appropriate responding in 7 out of 8 subjects following i.g. administration (Figure 4a, left panel). The ED₅₀ for pentobarbital (i.g.) substitution was 3.7±1.6 mg/kg and there was no significant effect of dose on response rate relative to baseline (Table 1; F(5,11)=2.6, p=0.09). A subset of subjects (n=3) were tested with pentobarbital (i.m.) and the ED₅₀ was 3.8±0.7 mg/kg (Figure 4a, right panel) and was not significantly different from the i.g. route of administration (p=0.9). Midazolam fully substituted for ethanol in six out of seven subjects tested (ED₅₀ = 1.6 ± 0.4 mg/kg; Figure 4b) and did not have a significant group effect on response rate (F(5,11)=1.9, p=0.2; Table 1). Muscimol substituted for ethanol in 1 out of 4 subjects tested (% ethanol-appropriate responding; Table 2). MK-801 produced full substitution in 4 out of 5 subjects, with one subject showing partial substitution (44–54%, Figure 4c) at doses that significantly lowered response rate (MK-801 ED₅₀ = 0.017 ± 0.009 mg/kg; Table 1).

Figure 4.

GABA_A and NMDA ligand substitution profiles. a-c) Dose response curves are plotted for individual subjects for pentobarbital (i.g., left panel, i.m., right panel), midazolam and MK-801. Each data point is an average value for each subject (double determination). Dotted lines represent the threshold for full substitution (>80% on ethanol-appropriate lever) and no substitution (<20% on ethanol-appropriate lever). The area between the lines represents partial substitution. d) Average ED₅₀ values for each test drug to compare drug potency. Each data point is an individual monkey, and the bar graphs are the group mean ± SD. The numbers at the bottom of the bar graphs indicate the proportion of subjects tested that had full substitution with the specific drug.

The potency ranking of substitution in these subjects was MK-801 (0.017 mg/kg) > midazolam (1.6 mg/kg) >pentobarbital (3.7 mg/kg) > ethanol (700 mg/kg, or 0.7 g/kg) (Figure 4d). Interestingly neither pentobarbital nor midazolam substituted for ethanol in one monkey, whereas in this same monkey, there was full generalization of MK-801 to ethanol. These data suggest that the NMDA component of the ethanol cue was more prominent in this subject. Pentobarbital did not have a significant effect on response rate in the dose range tested. Midazolam decreased response rates below 60% of baseline in 1 out of 7 subjects, and MK-801 decreased response rates in 3 out of 5 monkeys tested (Table 2).

Discussion

The current study is the first study to our knowledge that has characterized the absorption and clearance rates of low to moderate ethanol doses as well as determined the GABA_A and NMDA receptor involvement in moderate ethanol doses. These data fill an important gap in understanding the psychopharmacology of ethanol in the rhesus monkey, which is used in translational studies of alcohol drinking relevant to alcohol use disorders (Grant et al., 2008b; Baker et al., 2014; Jimenez and Grant, 2017; Allen et al., 2017). The pharmacokinetic data indicate that the time to peak BEC following 1.0 g/kg (i.g.) is approximately 90 minutes in male rhesus monkeys, which is 30 minutes longer than was previously reported in the cynomolgus macaque (Green et al., 1999a). However, the previous experiment did not measure BEC between 60 and 120 minutes following ethanol administration, and thus, differences in time to peak BEC between rhesus and cynomolgus macaques may be related to sampling resolution. The magnitude of peak BEC was not significantly different between species, reported at at 86 mg% for male cynomolgus macaques (Green et al., 1999a) and 86 mg% in male rhesus macaques here. Importantly, BEC after a small number of ethanol exposures was not significantly different after months of training following 1.0 g/kg ethanol. A recent study reported 120 minutes to peak BEC following 2.0 g/kg ethanol, i.g., in rhesus monkeys, but blood sampling was done under sedation (Berro et al., 2019). For the 0.5 g/kg ethanol dose, the time to peak BEC was highly variable, ranging from 45–90 minutes even under fasted conditions. For both doses tested in the current experiment, the time to peak BEC was slower for rhesus macaques relative to reports in human subjects, which found that time to peak BEC for 0.5 g/kg and 1.0 g/kg was 29 and 52 minutes, respectively (approximately 30 min faster than the male rhesus macaques examined here) (Dubowski, 1985; Zorzano and Herrera, 1990).

The elimination phase of the pharmacokinetic time course supports a zero order kinetics, such that the elimination rate was nearly identical at both doses tested. The elimination rates in male rhesus monkeys in this study (~15mg/dl/hr) are somewhat slower than rates reported previously in cynomolgus monkeys (20–30 mg/dl/hr) (Green et al., 1999a). Relative to the clinical literature, however, young adult rhesus monkeys appear to eliminate alcohol at slightly faster rates compared to a sample of healthy adults, which have been reported between 8–17 mg/dl/hr for both men and women (Taylor et al., 1996; Baraona et al., 2001). Overall, our data indicate that the pharmacokinetic time course of ethanol in rhesus monkeys is similar to reports in human literature, providing additional support for the translational strengths of the rhesus monkey in alcohol research.

The current study demonstrated that rhesus monkeys reliably learned to discriminate a moderate dose of ethanol from water in a two-choice discrimination task. There is only one recent study that trained a high dose of ethanol (2.0 g/kg) vs. water in rhesus monkeys, but none have reported lower doses. Additionally, there is a previous report of ethanol being used as a test drug for substitution in a midazolam versus water discrimination in rhesus monkeys (McMahon and France, 2005). Mean sessions to criteria for 1.0 g/kg ethanol vs. water discrimination was lower for this cohort relative to the cynomolgus macaque (average of 81 sessions to criteria for rhesus versus 137 sessions for cynomolgus macaques) (Grant et al., 2000), but similar to squirrel monkeys trained to discriminate 1.0 g/kg ethanol intravenously (82 sessions, Platt et al., 2005). Importantly, tests with morphine established that the ethanol discrimination was specific for the ethanol cue, as the μ-opioid receptor system is not directly involved in producing the discriminative stimulus effects of ethanol in primates (Platt and Bano, 2010).

In general, our findings that GABA_A receptor positive modulators and NMDA receptor antagonists are both sufficient to produce ethanol discriminative stimulus effects are consistent with previous reports in Old and New World monkeys (Grant et al., 2000; Grant et. al., 1999; Helms and Grant, 2011, 2011; Vivian et al., 2002; Platt et al., 2005; Berro et al., 2019). However, there were several cases that appear to be a departure from previous reports. There was one monkey in the current group in which muscimol produced full substitution for ethanol, which has not been reported in monkeys or rodents following peripheral muscimol administration (Shelton and Balster, 1994; Grant et al., 2000). The only case in which muscimol has substituted for ethanol is when it was directly administered into specific brain nuclei associated with ethanol discrimination in rats (Hodge and Aiken, 1996; Hodge and Cox, 1998). Additionally, there was one subject that did not demonstrate any ethanol substitution for either pentobarbital or midazolam, indicating that the GABAergic component of the ethanol cue was not prominent in this monkey. Interestingly, however, MK-801 fully substituted for ethanol in this monkey, suggesting that the glutamatergic cue was guiding the discrimination. This finding is contrary to several other published reports on ethanol discrimination in cynomolgus macaques that have concluded that the GABAergic component is more prominent in macaques relative to rodents across several training doses (Grant et al., 2000; Grant et al., 2008a; Stolerman et al., 2011; Allen et al., 2017). One explanation is the relatively large sample size (8 subjects relative to groups of 4 monkeys in previous studies) and choosing the monkeys to be relatively genetically heterogeneous (no common grandparents) allowed this individual difference to be captured. Further studies with additional subjects are necessary to confirm if the proportion of monkeys showing a strong glutamatergic bias (12.5%) or direct GABAA agonist bias (25%) are representative of the rhesus macaque population.

In conclusion, data presented here demonstrate that ethanol’s discriminative stimulus effects in rhesus monkeys are largely consistent with reports in many other species including pigeons (Grant and Barrett, 1991), rats (Shelton and Balster, 1994), mice (Shelton and Grant, 2002), squirrel monkeys (Platt et al., 2005), cynomolgus macaques (Grant et al., 1999; Grant et al., 2000; Vivian et al., 2002; Helms and Grant, 2011), and humans (Duka et al., 1998) (for review: Grant 2003, Stolerman et al., 2011; Allen et al., 2017). Additionally, we have conducted a thorough blood ethanol time course with 15-min sampling intervals during the rising phase of BEC to capture the time and value of peak BEC following low and moderate ethanol doses. The rate of ethanol elimination is within the range of reports in human subjects, providing face validity for future research on alcohol self-administration in rhesus macaques. One limitation of the current experiment is the inclusion of only male subjects, as there are known differences between males and females in ethanol pharmacokinetics (Zorzano and Herrera, 1990; Green et al., 1999a; Baraona et al., 2001) and sex-specific effects in females related to different phases of the menstrual cycle (Grant et al., 1996; Green et al., 1999b; Dozier et al., 2019). Future studies examining these variables in female rhesus macaques would allow for complete cross-species and cross-sex comparisons in translational alcohol research.