Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Heroin (3,6-diacetylmorphine) is an opioid agonist that produces a morphine-like profile of physiological and behavioral effects, including analgesia and abuse-related effects (Gutstein and Akil, 2001). Although heroin itself has low affinity for opioid receptors, it is rapidly metabolized to 6-acetylmorphine and morphine, which have high affinity for µ-opioid receptors (Way et al., 1960; Inturrisi et al., 1983, 1984; Bertalmio et al., 1992). Morphine is further metabolized to morphine-6-glucuronide (M6G), which also has relatively high affinity for µ-opioid receptors (Pasternak et al., 1987; Abbott and Palmour, 1988; Paul et al., 1989). On the basis of these and related findings, it has been suggested that heroin may function as a highly lipophilic prodrug for the active metabolites 6-acetylmorphine, morphine, and M6G (Way et al., 1960; Inturrisi et al., 1983, 1984; Paul et al., 1989).

Although all these heroin metabolites have high affinity for µ-opioid receptors, accumulating evidence suggests that different receptor populations may mediate the antinociceptive effects of heroin, 6-acetylmorphine, and M6G compared with the effects of morphine in rodents. An early behavioral study that supported this position reported a lack of antinociceptive cross-tolerance to heroin in morphine-tolerant mice (Lange et al., 1980). This finding was confirmed and extended to include a lack of cross-tolerance to 6-acetylmorphine and M6G in morphine-tolerant mice (Rossi et al., 1996). Manipulation of the expression or structure of the MOR-1 gene, using antisense probes and knockout models, also resulted in differential blockade of the antinociceptive effects of heroin, 6-acetylmorphine, and M6G compared with morphine in rodents (Rossi et al., 1995a,b, 1996, 1997; Schuller et al., 1999). In an effort to identify the molecular mechanisms underlying these differences, it was found that [³H]M6G bound with higher affinity to a novel site than to the site labeled by µ-radioligands, such as [³H]morphine, and it was suggested that this novel site corresponded to a novel M6G opioid receptor (Brown et al., 1997). A role for these receptors in heroin antinociception was suggested by studies in exon-1 MOR-1 knockout mice because these mice retained high-affinity [³H]M6G binding in brain and sensitivity to the antinociceptive effects of heroin but not of morphine (Schuller et al., 1999).

Studies with the opioid antagonist 3-methoxynaltrexone have provided additional evidence to suggest a role for M6G receptors in mediating the antinociceptive effects of heroin in rodents. 3-Methoxynaltrexone bound with higher affinity to the site labeled with [³H]M6G than to [³H]morphine-labeled sites in membranes from both mouse brain and Chinese hamster ovary cells transfected with the MOR-1 receptor, and these results suggested that 3-methoxynaltrexone may act as a relatively selective ligand for the M6G receptor (Brown et al., 1997a,b). In behavioral studies in CD-1 mice, 3-methoxynaltrexone was more potent as an antagonist of the antinociceptive effects of heroin and M6G than of morphine (Brown et al., 1997a; Rady et al., 2000). 3-Methoxynaltrexone was also more potent in antagonizing the antinociceptive effects of heroin and 6-acetylmorphine compared with those of morphine in rats (Walker et al., 1999). Overall, these results have been interpreted to suggest that 1) a novel M6G opioid receptor may mediate, at least in part, the antinociceptive effect of heroin, 6-acetylmorphine, and M6G but not of morphine in rodents; and 2) 3-methoxynaltrexone may serve as a moderately selective antagonist of heroin, 6-acetylmorphine, and M6G under some conditions. On the basis of these in vitro studies and behavioral studies of antinociception, it also has been suggested that the selectivity of 3-methoxynaltrexone may lead to new approaches to the treatment of opioid abuse (Brown et al., 1997a). However, it is of interest to note that 3-methoxynaltrexone was equipotent in altering the self-administration of heroin and morphine in rats (Walker et al., 1999).

The present study was designed to extend these previous findings in two ways. First, this study compared the ability of 3-methoxynaltrexone to antagonize the antinociceptive effects of heroin and morphine in rhesus monkeys. Although opioid receptor populations in rodents and monkeys are similar, there are known species differences in both the relative proportions and distributions of µ-, -, and -opioid receptors (Mansour et al., 1988). Furthermore, it is not known whether primates express M6G receptors or whether these receptors play a role in mediating the antinociceptive effects of heroin. Second, this study compared the ability of 3-methoxynaltrexone to antagonize the discriminative stimulus and rate-suppressant effects of heroin and morphine in rhesus monkeys trained to discriminate heroin from saline. The discriminative stimulus effects of drugs in animals are believed to model the subjective effects of drugs in humans (Schuster and Johanson, 1987), and these effects may contribute to the abuse liability of drugs such as heroin. Thus, drug discrimination procedures afford an opportunity to assess the ability of 3-methoxynaltrexone to antagonize an abuse-related effect of heroin and morphine. To provide a context for interpreting the effects of 3-methoxynaltrexone, naltrexone antagonism of heroin- and morphine-induced antinociception and discriminative stimulus effects was also examined.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects

Six male rhesus monkeys (Macaca mulatta) weighed 8.0 to 14.0 kg and were maintained on a diet of biscuits (Lab Diet Jumbo Monkey Biscuits, PMI Feeds, Inc., St. Louis, MO), fresh fruit and vegetables, and multiple vitamins. Monkeys in the discrimination experiment could obtain up to 50 1-g banana pellets (Precision Primate Pellets Formula L/I Banana Flavor; P.J. Noyes Co., Lancaster, NH) during operant sessions. Water was available continuously. A 12-h light/dark cycle was in effect with lights on at 7:00 AM.

Animal maintenance and research were conducted according to the guidelines provided by the National Institutes of Health Committee on Laboratory Animal Resources. The research facility was licensed by the United States Department of Agriculture. Research protocols were approved by the McLean Hospital Institutional Animal Care and Use Committee. A staff veterinarian monitored the health of the monkeys on a regular basis. Monkeys had visual, auditory, and olfactory contact with other monkeys throughout the study. Environmental enrichment was provided by toys, music, and nature videotapes. For monkeys in the discrimination experiment, the operant procedure provided additional opportunities for environmental manipulation.

Thermal Nociception

General Procedure. Three monkeys were used in the thermal nociception study, and all of the monkeys had prior experience in this procedure (Brandt et al., 2001). During experimental sessions, monkeys were seated in standard primate chairs, and the lower 10 cm of the shaved tail of each monkey was immersed in water heated to 42 or 50°C. An Apple IIe microcomputer was used to measure and record the latency (in seconds) for monkeys to remove their tails from warm water. If a monkey failed to remove its tail within 20 s, the timer was stopped, the monkey's tail was removed from the warm water, and a latency of 20 s was assigned to that measurement.

Test Procedure. Test sessions were conducted on Mondays and Thursdays, and they consisted of multiple 30-min cycles. After the monkeys were seated in chairs, the experimental session began with determination of baseline tail-withdrawal latencies from water heated to 42 or 50°C. Water heated to 42°C is an innocuous stimulus in this procedure (Negus et al., 1993b), and this stimulus was included herein only during baseline determinations to ensure that tail immersion alone did not elicit the tail-withdrawal response. Monkeys in this study never withdrew their tails from 42°C water during these baseline determinations. For the remainder of each session, tail-withdrawal latencies were evaluated only from water heated to 50°C. After baseline measurements, the first cycle began with an intramuscular injection of vehicle (sterile water) or a dose of an antagonist. The remaining cycles began with i.m. administration of cumulative doses of heroin or morphine, and each dose increased the total cumulative dose by one-quarter or one-half log units. Tail-withdrawal latencies from 50°C water were determined beginning 25 min after each injection. Heroin or morphine was administered up to doses resulting in at least 50% of the maximum possible effect (see below) in each monkey. Heroin doses ranged from 0.01 to 10.0 mg/kg, and morphine doses ranged from 0.32 to 100.0 mg/kg. The drugs and doses tested in pretreatment experiments were naltrexone (0.01-0.1 mg/kg) and 3-methoxynaltrexone (0.1-3.2 mg/kg).

Heroin Discrimination

Three monkeys were trained to discriminate 0.1 mg/kg heroin (i.m.) from saline (i.m.) in a two-key, food-reinforced drug discrimination procedure. All of the monkeys were trained previously to discriminate a cocaine/heroin mixture (i.e., "speedball"; 0.4 mg/kg cocaine + 0.04 mg/kg heroin i.m.; Negus et al., 1998a), and this stimulus maintained responding for approximately 2 to 3 years. Speedball discrimination training was stopped 2 weeks to 2 months before training of the heroin discrimination began.

Apparatus. The drug discrimination procedure used in the present study was similar to that used in a previous experiment (Negus et al., 1998a). Each monkey was housed in a ventilated, stainless steel cage (56 × 71 × 69 cm). The front wall of each cage was adapted to fit an operant panel (28 × 28 cm) that included three square translucent response keys (6.4 × 6.4 cm) arranged 2.54 cm apart horizontally and 3.2 cm from the top of the panel. Each response key could be transilluminated by red or green stimulus lights (Superbright LEDs; Fairchild Semiconductor, San Jose, CA). A food pellet dispenser (model G5210; Ralph Gerbrands Co., Arlington, MA) was mounted on the top of the operant panel, and delivered 1-g banana-flavored food pellets to a receptacle beneath the panel. Operant panels were controlled, and data were collected, with an IBM-compatible computer interface and power supply obtained from MED Associates (Georgia, VT).

Discrimination Training. Training sessions consisted of one to five cycles, and each cycle consisted of a 15-min time-out period followed by a 5-min response period. Monkeys were given an intramuscular injection of either vehicle (saline) or the heroin-training dose (0.1 mg/kg) at the beginning of the 15-min time-out period. All stimulus lights were turned off and responding had no scheduled consequences during the time-out period. During the response period, the right and left response keys were transilluminated red or green, and the positions of the red and green keys were counterbalanced across monkeys. Depending upon the training condition, monkeys could respond on the stimulus-appropriate key under a fixed ratio 30 (monkeys 90B164 and 163F) or 40 (monkey 90B147) schedule to obtain up to 10 food pellets per cycle. After vehicle administration, responding on only the green key resulted in the delivery of a food pellet. After heroin administration, responding on only the red key resulted in the delivery of food. Inappropriate responses reset the fixed ratio requirement on the stimulus-appropriate key. The center key was not illuminated during operant sessions, and responses on the center key had no scheduled consequences. If all of the available food pellets were delivered in less than 5 min then the stimulus lights were extinguished, and responses had no scheduled consequences for the remainder of the 5-min response period. Training sessions consisted of zero to five saline cycles followed by zero to one drug cycles. If the training drug was administered, it was given only during the last cycle. This design ensured a constant interval between drug administration and the onset of response periods during which responding on the drug-appropriate key produced food. Monkeys were considered to have acquired the discrimination when the following criteria were met for seven of eight consecutive training sessions: 1) the percentage of injection-appropriate responding before the delivery of the first reinforcer was greater than or equal to 80% for all cycles; 2) the percentage injection-appropriate responding over the entire response period was greater than or equal to 90% for all cycles; and 3) response rates during vehicle training cycles were greater than 0.5 responses/s, which required between 30 and 90 training sessions, depending upon the animal. Experimental sessions were conducted 5 days/week.

Discrimination Testing. Once monkeys met criterion levels of heroin discrimination, testing began. Test sessions were conducted only if the three criteria listed above were met during the training session immediately preceding the test session. If responding did not meet criterion levels of discrimination performance then training was continued until criterion levels of performance were obtained for at least two consecutive sessions. In general, test sessions were conducted on Tuesdays and Fridays, and training sessions were conducted on Mondays, Wednesdays, and Thursdays.

Data Analysis

ED₅₀ values were used to evaluate the effects of opioid agonists in both antinociception and drug discrimination procedures. In the antinociception procedure, ED₅₀ values were defined as the doses of heroin or morphine that produced 50% MPE. In the drug discrimination procedure, ED₅₀ values were determined for both heroin-appropriate responding and response rate-suppression, and were defined as the doses of heroin or morphine that produced 50% heroin-appropriate responding and a 50% reduction in response rates compared with saline control values, respectively. ED₅₀ values were calculated by interpolation when only two data points were available (one below and one above 50%) or by linear regression when at least three data points were available on the linear portion of the dose-effect curve. Individual ED₅₀ values were calculated and averaged to yield a mean ED₅₀ value (± S.E.M.). Individual substitution ED₅₀ values were calculated only if a monkey responded during at least the first three cycles of a test session. Because drug doses were incremented on a logarithmic scale, ED₅₀ values were converted to their log values for calculation of mean and S.E.M. and for statistical analysis. Mean ED₅₀ values (± S.E.M.) were converted back to their linear values for presentation in Table 1.

Antinociception, substitution, and rate-suppression ED₅₀ values for morphine and heroin administered alone and after opioid antagonist pretreatments were compared using one-way repeated measures analysis of variance (SuperAnova; Abacus Concepts, Berkeley, CA) with antagonist dose as the within-subjects factor. The statistical analyses included only antagonist doses for which there were ED₅₀ values for all three of the monkeys. A significant analysis of variance was followed by linear contrasts comparing individual means. For all statistical analyses, the criterion for significance was set a priori at p < 0.05.

For each monkey, dose ratios were calculated as the ED₅₀ of an opioid agonist in the presence of some dose of antagonist divided by the ED₅₀ of the agonist alone. Dose ratios were then used to calculate in vivo apparent pA₂ and pK_B values for naltrexone and 3-methoxynaltrexone antagonism of the antinociceptive, discriminative stimulus and rate-suppressant effects of heroin and morphine. pA₂ and pK_B values are defined as the negative logarithm of the molar dose of antagonist required to produce a 2-fold rightward shift in an agonist dose-effect curve, and these values provide an in vivo estimate of the affinity of the antagonist for the receptor that mediates the effects of the agonist (Negus et al., 1993a). pA₂ values and Schild plot slopes were determined using Schild regression analysis (Tallarida and Murray, 1987) when at least three doses of the antagonist produced dose-dependent increases in agonist ED₅₀ values. If the 95% confidence limits of the Schild plot slope included the theoretical value of 1 and did not include positive numbers then the pA₂ value (±95% confidence limits) was redetermined with the slope constrained to 1. pA₂ values could not be calculated for 3-methoxynaltrexone antagonism of heroin and morphine in the heroin discrimination study due to individual differences in the effects of the highest dose of 3.2 mg/kg 3-methoxynaltrexone. However, pK_B values were determined from data collected with 1.0 mg/kg 3-methoxynaltrexone, as described previously (Negus et al., 1993a). Calculation of apparent pK_B values assumes that the slopes of the Schild plots are equal to 1. pA₂ and pK_B values were considered significantly different if 95% confidence limits did not overlap. Our hypothesis predicted that pA₂/pK_B values for naltrexone antagonism of heroin and morphine would be similar (i.e., naltrexone would be equipotent as an antagonist of heroin and morphine), whereas pA₂/pK_B values for 3-methoxynaltrexone would be greater for heroin than for morphine (i.e., 3-methoxynaltrexone would be more potent as an antagonist of heroin than of morphine).

Drugs

Heroin hydrochloride, morphine sulfate, naltrexone hydrochloride, and 3-methoxynaltrexone were supplied by the National Institute on Drug Abuse (Bethesda, MD). U50,488 was purchased from Sigma/RBI (Natick, MA). Ketamine (Ketaset; Fort Dodge Animal Health, Fort Dodge, IA) was purchased as a 100-mg/ml stock solution. Unless otherwise noted, drugs were dissolved in sterile water. To get 3-methoxynaltrexone into solution, 1% lactic acid was added. Drug stock solutions were diluted to the appropriate concentrations with sterile saline. Doses were based on the salt forms of the drugs, and i.m. injections were administered in volumes of 0.1 to 3.8 ml.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Thermal Nociception. Under baseline conditions, the mean baseline tail-withdrawal latency (± S.E.M.) from 50°C water was 2.20 ± 0.21 s. Both morphine and heroin administered alone produced dose-dependent antinociception (Fig. 1), and the mean antinociception ED₅₀ values of morphine and heroin are shown in Table 1. Heroin was approximately 10- to 20-fold more potent than morphine in this warm water tail-withdrawal assay. Neither naltrexone nor 3-methoxynaltrexone alone altered baseline tail-withdrawal latencies (Fig. 1, points above PT).

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Effects of naltrexone (NTX; top) and 3-methoxynaltrexone (3-MeONTX; bottom) on the antinociceptive effects of morphine (left) and heroin (right). Abscissae, dose of drug in milligrams per kilogram (log scale). Ordinates, antinociception quantified as %MPE. All of the points show mean data (±1 S.E.M.) from three monkeys.

Heroin Discrimination. Figs. 2 and 3 show the discriminative stimulus effects of heroin and morphine alone in individual monkeys. The cumulative administration of heroin (open circles, top) or morphine (open circles, bottom) alone produced dose-dependent and complete substitution for the training dose of heroin in each of the three subjects. Both morphine and heroin also produced dose-dependent decreases in response rates (Fig. 4). Responding was eliminated in all of the monkeys by 0.32 mg/kg heroin, and a dose of 3.2 mg/kg morphine eliminated responding in two monkeys and decreased response rates in the third monkey. The mean substitution and rate-suppression ED₅₀ values of morphine and heroin are shown in Table 1. With respect to both discriminative stimulus and rate-suppressant effects, heroin was approximately 10- to 20-fold more potent than morphine. The pharmacological selectivity of this discrimination was evaluated by administering the -opioid receptor agonist U50,488 and the noncompetitive N-methyl-D-aspartate glutamate receptor antagonist ketamine. Both U50,488 and ketamine produced less than 10% heroin-appropriate responding up to doses that eliminated responding in at least two of the three monkeys (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Effects of naltrexone (NTX) on the heroin-like discriminative stimulus effects of heroin (top) and morphine (bottom). Abscissae, dose of drug in milligrams per kilogram (log scale). Ordinates, percentage of heroin-appropriate responding. All of the points show individual data from three monkeys. , 0 mg/kg NTX; , 0.0032 mg/kg NTX; , 0.01 mg/kg NTX; , 0.032 mg/kg NTX.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Effects of 3-methoxynaltrexone on the heroin-like discriminative stimulus effects of heroin (top) and morphine (bottom). Abscissae, dose of drug in milligrams per kilogram (log scale). Ordinates, percentage of heroin-appropriate responding. All of the points show individual data from three monkeys. , 0 mg/kg NTX; , 0.32 mg/kg NTX; , 1.0 mg/kg NTX; , 3.2 mg/kg NTX.

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Effects of naltrexone (NTX; top) and 3-methoxynaltrexone (3-MeONTX; bottom) on the rate-suppressant effects of morphine (left) and heroin (right). Abscissae, dose of drug in milligrams per kilogram (log scale). Ordinates, response rates in responses per second. All of the points show mean data (±1 S.E.M.) from three monkeys.

In Vivo Apparent pA₂ and pK_B Analysis of Antagonist Effects. Fig. 5 shows the Schild plots and Table 2 shows Schild plot slopes and in vivo affinity estimates for naltrexone and 3-methoxynaltrexone antagonism of the antinociceptive, discriminative stimulus, and rate-suppressant effects of morphine and heroin. For naltrexone, all Schild plot slopes included the value of 1 and did not include positive values, so pA₂ values also were determined with slopes constrained to 1. The constrained in vivo apparent pA₂ values were similar for naltrexone antagonism of the antinociceptive, discriminative stimulus and rate-suppressant effects of both morphine and heroin, as determined by overlapping 95% confidence limits.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Schild plots for naltrexone (NTX) and 3-methoxynaltrexone (3-MeONTX) antagonism of the antinociceptive, discriminative stimulus, and rate-suppressant effects of morphine and heroin. Abscissae, negative log of the antagonist dose in moles per kilogram. Ordinates, log (dose ratio  1). All of the points show mean data (±1 S.E.M.) from two to three monkeys. , NTX + morphine; , NTX + heroin; , 3MeONTX + morphine; , 3MeONTX + heroin.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The main finding of this study was that 3-methoxynaltrexone was equipotent in antagonizing the effects of heroin and morphine in thermal nociception and heroin discrimination procedures in rhesus monkeys. The opioid antagonist naltrexone also was equipotent in antagonizing the effects of heroin and morphine, although naltrexone was approximately 100-fold more potent than 3-methoxynaltrexone. These findings are consistent with the conclusion that the effects of heroin and morphine are mediated by pharmacologically similar populations of µ-opioid receptors in rhesus monkeys. These results also suggest that some of the differences in the pharmacology of heroin and morphine that have been observed in rodents may not extend to studies in nonhuman primates.

Antinociceptive and Discriminative Stimulus Effects of Heroin and Morphine Alone. The warm-water tail-withdrawal procedure has been used extensively to examine the antinociceptive effects of opioids in rhesus monkeys (Dykstra et al., 1987; Negus et al., 1993b, 2002; Gatch et al., 1996; Negus and Mello, 1999; Brandt et al., 2001). In agreement with previous studies, heroin and morphine produced dose-dependent antinociception in this procedure, and heroin was 10- to 20-fold more potent than morphine (Dykstra et al., 1987; Negus et al., 1998a,b). We also demonstrated previously that heroin has a more rapid rate of onset and a shorter duration of action than morphine in this procedure (Negus et al., 1998b).

Antagonism of Heroin and Morphine Antinociception by 3-Methoxynaltrexone. 3-Methoxynaltrexone was equipotent as an antagonist of the antinociceptive effects of heroin and morphine as indicated by similar pA₂ values. Moreover, the slopes of the Schild plots for 3-methoxynaltrexone antagonism of heroin and morphine were similar to 1, which is the theoretical value predicted by receptor theory for a competitive agonist and a competitive antagonist interacting at a homogenous population of receptors (Kenakin, 1993). These results are consistent with the conclusion that the antinociceptive effects of heroin and morphine were mediated by a single population of pharmacologically similar opioid receptors.

Antagonism of Discriminative Stimulus and Rate-Suppressant Effects of Heroin and Morphine by 3-Methoxynaltrexone. 3-Methoxynaltrexone was equipotent as an antagonist of the discriminative stimulus and rate-suppressant effects of heroin and morphine. Although pA₂ values could not be determined due to individual differences in the effects of 3.2 mg/kg 3-methoxynaltrexone (see below), pK_B values for 1.0 mg/kg 3-methoxynaltrexone antagonism of the discriminative stimulus and rate-suppressant effects of heroin and morphine were similar. Moreover, these pK_B values were similar to the pA₂ values for 3-methoxynaltrexone antagonism of the antinociceptive effects of heroin and morphine.

Antagonist Effects of Naltrexone. The Schild plot slopes for naltrexone antagonism of heroin and morphine were similar to 1, and naltrexone was equipotent as an antagonist of the antinociceptive, discriminative stimulus, and rate-suppressant effects of heroin and morphine as indicated by similar pA₂ values. The in vivo apparent pA₂ values for naltrexone antagonism of heroin and morphine are consistent with those reported in other studies of naltrexone antagonism of the behavioral effects of heroin and morphine in rhesus monkeys (France et al., 1990; Platt et al., 2001; Rowlett et al., 1998). Furthermore, these in vivo apparent pA₂ values are similar to pA₂ values obtained for naltrexone antagonism of other selective µ-agonists in behavioral studies in rhesus monkeys (France et al., 1990; Gerak et al., 1994; Ko et al., 1998). Together, the results with naltrexone and 3-methoxynaltrexone suggest that the effects of heroin and morphine assessed in this study were mediated by a single population of pharmacologically similar µ-opioid receptors. The finding that naltrexone pA₂ values were consistently and significantly higher than 3-methoxynaltrexone pA₂ and pK_B values suggests that naltrexone has higher affinity than 3-methoxynaltrexone for the µ-opioid receptors that mediate these effects of heroin and morphine in rhesus monkeys. This in vivo finding agrees with binding studies that suggest that 3-methoxynaltrexone has relatively low affinity for µ-opioid receptors (Brown et al., 1997b).