Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Morphine (MOR) is widely used in the clinical management of pain. However, prolonged MOR administration may be limited by the high abuse liability of the drug, which is also true of other analgesics that produce MOR-like subjective effects (Jasinski, 1977). Abrupt termination of chronic opioid treatment or the administration of an opioid antagonist during treatment produces the physical withdrawal/abstinence signs and subjective symptoms that define physical dependence in humans.

Schedule-controlled behaviors maintained by food or brain stimulation reinforcement are sensitive indicators of both spontaneous and antagonist-precipitated opioid withdrawal in animals. Without prior MOR exposure, opioid antagonists (30-100 mg/kg), such as naloxone (NX) and naltrexone (NTX), have behavioral effects only at significantly higher (10- to 100-fold) doses than those required to antagonize the acute behavioral effects of even large doses of MOR (Adams and Holtzman, 1990). In contrast, after prolonged MOR exposure, opioid antagonists significantly suppress schedule-controlled behaviors at 100- to 1000-fold lower doses than those having effects in animals not treated with MOR (Valentino et al., 1983).

In humans given a single dose of MOR, NX precipitates a syndrome comprising both physical signs and dysphoric subjective effects that are qualitatively similar to those experienced after withdrawal from chronic MOR treatment (Bickel et al., 1988; Heishman et al., 1989). In rats, increased behavioral sensitivity to antagonists also occurs after exposure to a single dose of MOR (Meyer and Sparber, 1977; Young, 1986). Indeed, as little as 2 to 4 h after an acute MOR (3-10 mg/kg) pretreatment, low doses of NTX (0.1 mg/kg) decrease rates of responding by as much as 75% (Easterling and Holtzman, 1997), and slightly higher doses produce somatic signs that are qualitatively similar to but less intense than those seen during spontaneous withdrawal from chronic MOR administration (Schulteis et al., 1997). These data support the conclusion that antagonist-induced disruptions in operant responding, in part, operationally define withdrawal and an acute dependence syndrome in the rodent.

Drug discrimination affords an animal model for studying subjective drug effects, including those associated with MOR withdrawal (Holtzman, 1990). Rats that are maintained chronically on MOR can be trained to discriminate saline (SAL) from as little as 0.1 mg/kg of NTX (Gellert and Holtzman, 1979; Holtzman, 1985) orders of magnitude lower than the discriminable dose in animals not maintained on MOR (France and Woods, 1985). This potent antagonist drug cue generalizes dose dependently to other opioid antagonists but not to opioid agonists or to nonopioid drugs. Most important, when rats are abruptly withdrawn from the chronic MOR regimen, they respond on the NTX-appropriate lever, indicating similarities in the interoceptive states produced by NTX-precipitated and abrupt/spontaneous opioid withdrawal.

As with chronic MOR administration, a single dose of MOR (0.1-1 mg/kg) can also produce a subsequent increase in the potency of the interoceptive effects of NTX in pigeons trained on a discrimination task (France and Woods, 1985). In rats responding on an FR10 schedule of food reinforcement, a single MOR pretreatment (40 mg/kg, 8 h) increases the stimulus potency of a lower dose of NX (1.25 mg/kg), making it effective as a dose-dependent discriminative stimulus (versus SAL; Miksic et al., 1981).

The studies reviewed above suggest that when opioid antagonists are given after a single or prolonged exposure to MOR, a common and distinctive interoceptive stimulus is produced that coincides with physical withdrawal signs and, in part, defines opioid dependence. However, the antagonist-precipitated interoceptive effects that follow a single MOR exposure remain incompletely characterized. Therefore, the present experiments were designed to provide data on the acute NTX-precipitated withdrawal stimulus produced after a single dose of MOR. First, we sought to determine whether or not subjects could be trained to discriminate NTX (0.3 mg/kg, 0.25 h) alone from a MOR (10 mg/kg, 4 h) + NTX (0.3 mg/kg, 0.25 h) combination. Once it was determined that they could be, we generated antagonist stimulus-generalization curves after various MOR pretreatment doses (1-10 mg/kg) or times (0.5-6 h), and before, during, and after chronic MOR administration. Specifically, we sought to determine whether the acute NTX-induced stimulus that follows a single dose of MOR is 1) dose and time dependent, 2) mediated by opioid receptors, 3) generalizable to another opioid antagonist, or 4) to abrupt/spontaneous withdrawal following the termination of chronic MOR treatment.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects. The subjects were male Sprague-Dawley-derived rats (Charles-River; Raleigh, NC) initially weighing 250 to 350 g. The rats were housed individually in polycarbonate cages with continuous access to food and water. The colony room was maintained on a 12:12 h light/dark cycle with lights on at 7:00 AM.

Drugs. The following drugs were dissolved in normal SAL (0.9%): MOR sulfate (Penick Corp., Newark NJ), NTX hydrochloride and NX hydrochloride (Research Biochemicals Inc., Natick, MA). All doses are expressed as the free base. MOR was administered in a volume of 1 ml/kg s.c. or by osmotic minipump before training or testing. When given s.c., NX and NTX were always administered in a volume of 1 ml/kg.

Osmotic Pump Implantation. During chronic MOR administration experiments, two osmotic pumps (Models 2 ML1 or 2 ML2; Alza Corp., Palo Alto, CA) were implanted in each 500 to 700 g rat while it was under light methoxyflurane anesthesia. A small incision was made in the mid-scapular region, and pumps were inserted in a rostral-caudal direction, with their flow-moderator entering first. Wounds were closed with 9-mm stainless steel wound clips. These wound clips were removed when the pumps were removed while the rats were under light methoxyflurane anesthesia. At this time, a second set of wound clips was installed and then removed 14 days later.

Discrimination Testing Apparatus. The apparatus has been described in detail previously (Shannon and Holtzman, 1976). Briefly, a standard two-lever operant test chamber was modified by adding one lever (the "observing" lever) to the wall opposite the two original levers (the "choice" levers). The choice levers were separated by a 5-cm-wide clear polycarbonate partition that extended from floor to ceiling of the chamber. A constant current generator delivered a scrambled electric current to the grid floor of the chamber, which was housed in a ventilated, light- and sound-attenuating outer enclosure. All contingencies of the behavioral schedule were controlled via a desktop computer.

Drug Discrimination Training. Rats were trained to discriminate between s.c. injections of SAL [(1 ml/kg, 4 h) and NTX (0.3 mg/kg, 0.25 h)] and MOR [(10 mg/kg, 4 h) and NTX (0.3 mg/kg, 0.25 h)] in a discrete-trial avoidance/escape procedure in which a two-lever-press response chain terminated a trial. The doses of MOR and NTX and the pretreatment duration (4 h) were chosen on the basis of published data characterizing antagonist-induced operant response-rate disruptions following an acute MOR injection (Young, 1986; Adams and Holtzman, 1990). In this training procedure, the observing lever and two choice levers were always available. The beginning of each of the 20 trials composing a session was signaled by the simultaneous illumination of the house light and the onset of white noise. Unless the appropriate two-response chain (observing-choice) occurred beforehand, 5 s after a trial began, a 1-s, 1- to 3-mA current was delivered to the grid floor of the chamber every 3 s for 300 s or until terminated by the appropriate response chain. The first observing response made in a trial terminated the white noise, and a response on the correct choice lever extinguished the house light and ended the trial. If a response on the incorrect choice lever followed an observing response, the white noise resumed and another observing response and correct choice response was required to end the trial. Because NTX was always given 0.25 h before testing, the correct choice response was determined by what the animal was injected with 4 h before the session, i.e., SAL or 10 mg/kg MOR. An incorrect choice response was the alternative for that session. The intertrial interval was 50 s, and during this time the chamber was dark. Each session ended after 20 trials or 30 min.

Drug Discrimination Testing. Following training, individual rats were tested twice weekly with doses of drug administered according to the sequence shown in Table 1. In test sessions, a response on either choice lever after a response on the observing lever terminated the trial. As with training sessions, all test sessions consisted of 20 trials. As each subject met the training criterion for stable discrimination performance (90% correct responding), it was randomly assigned to a test drug series until each drug series included six to eight subjects.

Global Rating of Physical Signs. Signs of physical dependence were assessed by visual observation (Gellert and Holtzman, 1978) of rats in the home cage over the 0.25 h pretreatment period that preceded discrimination training or testing. Before this visual scoring, subjects were weighed and given an injection of SAL or NTX (0.03-30 mg/kg), and then weighed again following discrimination testing an hour after the initial weighing. In the acute dependence experiments, SAL or MOR (1-10 mg/kg) was given 3.75 h before scoring. While subjects were receiving MOR (20-40 mg/kg day) through osmotic pumps, they were given a SAL or a NTX (0.001-1 mg/kg) injection and then scored (0.25 h). To assess spontaneous withdrawal signs, visual scoring and discrimination testing continued at intervals for 3 to 72 h after pump removal. In the Pump MOR 20 group, physical signs were scored 24 h after pumps were removed, whereas in the group that followed, the Pump MOR 40 group, subjects were scored at 3, 6, 9, 12, and 24 h.

Data Analysis. Discrimination data are presented as an average number of trials to the MOR  NTX-appropriate choice lever. The remaining trials of the session were completed on the choice lever appropriate for SAL  NTX. Discrimination data were excluded from analysis if subjects did not complete the entire range of doses within a drug series. The discrimination data were used to calculate ED₅₀ or antagonist-dose 50 (AD₅₀, NTX) values by linear regression of the ascending or descending portion of each individual dose-response curve. The time until discrimination responding reached or returned to one-half its maximal value (T_1/2) was calculated by linear regression of the ascending or descending part of the time course curve. Subsequently, group means for these data were calculated and t tests (protected for planned multiple comparisons) were used for comparison among drug conditions. The  level chosen for all analyses was .05. Subjects received osmotic pumps delivering 40 mg/kg/day on two different occasions. During each replication, one SAL and two NTX challenges were given (days 5 and 6). These discrimination data were combined for the purposes of generating ED₅₀s and figures based on four drug doses.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Stimulus Generalization Curves

Subjects (n = 20) reached the discrimination testing criterion of 90% correct responding in an average (±S.E.M.) of 43 ± 6 training sessions. Initially, stimulus generalization curves for NTX were determined after both of the pretreatments used in training (Fig. 1). After SAL pretreatment, NTX (100 mg/kg) did not produce substantial MOR  NTX-like responding. Consequently, no ED₅₀ was calculated for this condition. Likewise, MOR (10 mg/kg)-pretreated rats responded on the SAL  NTX-appropriate lever when they were injected with SAL 0.25 h before a session. When graded doses of NTX were given 3.75 h after 10 mg/kg MOR, the number of trials completed on the MOR  NTX-appropriate lever increased in a dose-dependent manner. The ED₅₀ of NTX was 0.05 ± 0.02 mg/kg. The group completed an average of 90% of the trails on the MOR  NTX-appropriate choice lever when MOR pretreatment was followed by either 0.3 (training dose) or 3 mg/kg NTX (Fig. 1).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Comparison of discriminative stimulus effects of NTX (s.c., 0.25 h) following SAL (4 h) or MOR (10 mg/kg, s.c., 4 h) administration. Each point represents mean of six to eight rats. Broken line at top represents criteria for full MOR  NTX-appropriate responding, and broken line at bottom represents full SAL  NTX-appropriate responding. Data indicate that NTX is a potent training stimulus in acutely MOR-treated rats (ED50 = 0.05 mg/kg), and that this stimulus is distinct from that produced by MOR alone or up to 2000-fold higher doses of NTX alone.

Temporal Dependency

To assess the temporal dependencies of the discrimination, the duration of the MOR (10 mg/kg) pretreatment was systematically varied around the training duration of 4 h (0.25-8 h), while holding the NTX dose (0.3 mg/kg) and pretreatment time (0.25 h) constant (Fig. 2A). When the duration of the MOR pretreatment was less than 1 h, responding was fully SAL  NTX appropriate. Three hours after MOR, responding was fully MOR  NTX-like, but by 8 h, responding had returned to less than 50% of its peak value. Thus, the onset of MOR  NTX-appropriate responding occurred with a T_1/2 = 1.75 ± 0.2 h.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Temporal dependence of MOR  NTX discrimination. A, MOR  NTX-appropriate responding as a function of time between MOR and NTX injections. Discriminative effects were produced by NTX (0.3 mg/kg, 0.25 h) when a single MOR (10 mg/kg) pretreatment was given 0.25 to 8 h before testing. Although 0.3 mg/kg NTX produced full stimulus effects 3 to 4 h after MOR, partial (50%) effects developed in as little as 1.75 h (T1/2 = 1.75 ± 0.2 h). B, MOR  NTX-appropriate responding in MOR-pretreated (10 mg/kg, 3.5 h) rats as a function of time after injection of NTX. When a SAL injection preceded (0.25 h) the first, third, fourth, and fifth of the consecutive test trials and NTX (0.3 mg/kg) was given before the second, stimulus effects of NTX were short-lived (T1/2 = 1 ± 0.25 h). Other details are as in Fig. 1.

Next, the duration of the stimulus effects of the training dose of NTX was examined by giving an acute MOR pretreatment (10 mg/kg), followed (3.75 h) by the training dose of NTX (0.3 mg/kg, 0.25 h) and repeatedly testing the animals in consecutive sessions until responding was SAL  NTX appropriate (Fig. 2B). As it did in training sessions, NTX had full stimulus effects 0.25 h after injection, but these effects were almost completely absent by 2 h. Thus stimulus control of behavior by MOR  NTX declined with a T_1/2 of 1 ± 0.25 h after the injection of NTX.

Antagonism of Stimulus Effects

If the interoceptive effects of the MOR pretreatment are indeed mediated by opioid receptors, they should be attenuated by prior opioid antagonist administration. We tested this hypothesis by giving SAL or an additional NTX injection (0.001-1 mg/kg) 4.25 h before the test session, which was 0.25 h before the MOR injection (10 mg/kg, 4 h); NTX (0.3 mg/kg) was administered a second time 0.25 h before the test session, as usual.

SAL given 4.25 h before testing and before the training drug combination, did not significantly affect the full MOR  NTX-like stimulus effects of the MOR  NTX combination (Fig. 3). However, the increasing doses of NTX that were given (4.25 h) dose dependently attenuated MOR  NTX-appropriate responding, with a dose of 1 mg/kg completely blocking the stimulus effects of the combination. The AD₅₀ of NTX for blocking the discriminative effects of MOR  NTX was 0.11 ± 0.04 mg/kg.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Administration of NTX (0.001-1 mg/kg) 0.25 h before 4-h pretreatment with 10 mg/kg MOR dose dependently blocks discriminative effects of MOR  NTX. Discrimination dose-response curve for NTX given 4.25 h before testing, and before training drug combination MOR (10 mg/kg, 4 h)  NTX (0.3 mg/kg, 0.25 h). NTX potently antagonized full stimulus effects of MOR  NTX combination (AD50 = 0.11 mg/kg). Other details are as in Fig. 1.

Reductions in MOR Pretreatment Time or Dose

If the interoceptive effects of acute dependence are a function of the dose and duration of MOR exposure, a lower dose of MOR or a shorter pretreatment interval might produce less dependence and smaller interoceptive effects. Consequently, a larger antagonist dose would be required to produce discriminative stimulus effects equivalent to those occurring during training. Pretreatment with 1 mg/kg of MOR compared with the training dose (10 mg/kg) produced a significantly (p < .05) different NTX stimulus-generalization curve (Fig. 4A). The ED₅₀ of NTX after pretreatment with 1 mg/kg of MOR (2.1 ± 0.62 mg/kg) was 42 times greater than the ED₅₀ of NTX after pretreatment with 10 mg/kg MOR, and trials completed on the MOR  NTX-appropriate lever did not reach 90%, even after 10 mg/kg of NTX. Only one dose of NTX (0.3 mg/kg) was given following a 3-mg/kg MOR pretreatment (4 h); its effects were intermediate between those of the same dose of NTX given after the lower (1 mg/kg) and higher (10 mg/kg) doses of MOR.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Reducing acute MOR pretreatment dose or duration decreases potency of NTX for occasioning MOR  NTX-appropriate responding. A, Discrimination dose-response curves for NTX (s.c., 0.25 h) given after a 4-h pretreatment with SAL or various doses of MOR (1, 3, or 10 mg/kg). Only one dose of NTX was tested after 3 mg/kg of MOR. B, Discrimination dose-response curves for NTX (s.c., 0.25 h) given following MOR (10 mg/kg) pretreatment of varied durations (0.5-4 h). Compared with training conditions, a decrease in MOR pretreatment dose or duration of pretreatment (10 mg/kg) produced a significantly smaller increase in potency of NTX. SAL and 10 mg/kg MOR curves (gray symbols and dashed lines) are reproduced from Fig. 1. Other details are as in Fig. 1.

When MOR (10 mg/kg) pretreatments were shorter (0.5 and 1 h) than the training pretreatment (4 h), the stimulus-generalization curves for NTX (s.c., 0.25 h) changed (Fig. 4B). After the 0.5-h MOR pretreatment, the NTX dose-effect curve was not different from the one determined after SAL pretreatment. Following the 1-h MOR pretreatment and compared with the SAL pretreatment NTX produced substantial MOR  NTX-like responding; comparing ED₅₀s, it was only 0.014 times as potent as it was after the 4-h MOR pretreatment (ED₅₀ = 3.66 ± 1 ). Peak (80%) MOR  NTX-appropriate responding was produced by 10 mg/kg NTX, and a higher dose of NTX (30 mg/kg) resulted in less responding on the MOR  NTX-appropriate lever than occurred after the lower dose. Therefore, reducing the pretreatment dose of MOR to 1 mg/kg or reducing the pretreatment time to 1 h resulted in approximately equivalent (p > .05) decreases in the potency of NTX relative to its potency under the conditions used for training (Table 1).

Generalization to NX

The dose-response curves for the discriminative stimulus effects of NTX and NX were compared following a 4-h pretreatment with SAL or MOR (Fig. 5). As with NTX, after SAL pretreatment, NX (30 mg/kg) did not produce MOR  NTX-like responding. When MOR (10 mg/kg) was given 4 h before the antagonists, NX (ED₅₀ = 0.35 ± 0.09 mg/kg) fully and dose dependently generalized to the MOR  NTX cue, but was significantly (7-fold) less potent than NTX.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   A comparison of dose-response curves for discriminative stimulus effects of NTX and NX following a 4-h pretreatment with MOR (10 mg/kg). MOR NTX curve (gray symbols and dashed line) is reproduced from Fig. 1. Under these conditions, NX (ED50 = 0.35 mg/kg) was significantly (7-fold) less potent than NTX. Other details are as in Fig. 1.

Chronic MOR Administration

Drug Discrimination. In the next set of experiments, the discriminative stimulus effects (Fig. 6) and physical withdrawal signs (Table 2) produced by NTX were compared following acute pretreatment with SAL or MOR (1-10 mg/kg) or during the chronic infusion of either of two doses (Pump MOR 20 or Pump MOR 40) of MOR by osmotic pump. Between the implantation of the first set of osmotic pumps (n = 10, Pump MOR 20) and the second set of pumps (n = 7, Pump MOR 40), two rats died of causes unrelated to the experiment, and one rat failed to meet the training criteria. Before the implantation of the third set of pumps (n = 5), two more rats died. The data from these rats were excluded from the analysis.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   A comparison of discriminative stimulus effects produced by SAL (S) or NTX (0.001-100 mg/kg) following either acute pretreatment with SAL or MOR or during chronic infusion of MOR by osmotic minipump (open symbols). Subjects received NTX 0.25 h before discrimination testing and 4 h after an acute MOR injection or after a minimum of 5 days of continuous MOR infusion. SAL and 10 MOR curves (gray symbols and dashed lines) are reproduced from Fig. 1. Highest infused dose of MOR (Pump MOR 40; 40 mg/kg/day) produced a significantly greater increase in stimulus potency of NTX than either lower infused dose (20 mg/kg/day) or acutely administered MOR (10 mg/kg, 4 h). Other details are as in Fig. 1.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   A time course of discriminative stimulus effects (A) and physical withdrawal signs (B) seen while MOR-filled (Pump MOR 20 or Pump MOR 40) osmotic pumps were in operation and in first 3 to 72 h after their removal. Subjects were not tested in each procedure at all time points. MOR-filled pumps were removed after 7 to 14 days of operation, and a SAL injection preceded each discrimination session or 0.25-h scoring period. A, MOR  NTX-appropriate responding was not seen while pumps were in operation (Pre). After pumps were removed, responding was partially MOR  NTX appropriate and dose dependent and in both experimental groups it peaked 6 to 12 h after pumps were removed. *Significantly greater than Pre (p  .05). +Significantly greater than Pump MOR 20 (p  .05). Other details are as in Figs. 1 and 6.

Physical Withdrawal Signs. As determined by our scoring procedure, no significant physical signs of withdrawal were produced by NTX (10 mg/kg) 4 h after an acute injection of SAL or 1 mg/kg of MOR (Table 2). A significant number of physical withdrawal signs were produced by NTX (0.03-30 mg/kg) after an acute 4-h pretreatment with 3 or 10 mg/kg of MOR. These effects were dependent on both the MOR and NTX doses, and regardless of MOR or NTX dose, appeared to plateau with a global score of 10 to 12. The specific signs seen included profuse salivation, ptosis, facial fasciculations/teeth chattering, and genital grooming; weight loss (above a 1.6% control loss) and other physical signs, were almost completely absent.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the current two-choice drug discrimination training procedure, rats were able to discriminate between daily doses of NTX (0.3 mg/kg) that were either preceded by a single dose of MOR or the alternative (SAL). Up to 100 mg/kg of NTX alone produced relatively little MOR  NTX-appropriate responding. However, when MOR preceded NTX, stimulus control of behavior was an orderly function of both the MOR and NTX doses and of the time between those doses. During chronic MOR treatment, NTX also produced full MOR  NTX-like stimulus effects, with its stimulus potency being dependent on the concentration of the MOR infusion. Thus, the discriminative stimulus effects of acute MOR pretreatment followed by NTX appear to be qualitatively similar to the discriminative effects of NTX in rats made physically dependent by chronic treatment with MOR.

In rats responding on an FR10 schedule of food reinforcement, stimulus control of behavior can be established (versus SAL) by a single MOR pretreatment (40 mg/kg, 8 h) that is followed by a lower dose of NX (1.25 mg/kg; Miksic et al., 1981) than that used in the present study. In this case, the discriminative effects of NX are not evident 1 h after MOR pretreatment but peak at the training interval (8 h), and then disappear progressively over the next 24 to 96 h. The stimulus effects that followed acute MOR treatment in the present study appeared earlier, also at the training interval (3-4 h), and were only half-maximal by 8 h. In this respect, they corresponded with the previously reported time course of MOR-induced sensitization to antagonist-induced disruption of food-reinforced operant responding (Young, 1986). These results suggest that training parameters are important determinants of the time course of stimulus generalization. Consistent with the studies reviewed above, NX shared full stimulus properties with NTX in MOR-pretreated rats. The 7-fold potency difference between NTX and NX in this study is virtually identical with their potency difference in producing NTX-appropriate responding in MOR-treated rhesus monkeys (France et al., 1990).

Although both MOR and NTX can easily be discriminated from SAL and each other (France and Woods, 1985), animals cannot readily be trained to discriminate (from SAL) a dose of NTX that is coadministered with MOR (Jarbe et al., 1979). Therefore, under the proper temporal contingencies, NTX can competitively block the development of MOR-like stimulus effects by an opioid-dependent mechanism. Consistent with this finding, in the present study NTX given shortly (0.25 h) before MOR completely blocked the development of MOR  NTX-like stimulus effects. Because later MOR  NTX-appropriate responding was completely attenuated by this prior NTX treatment, it is clear that MOR must occupy opioid receptors for a period of time before NTX is given to produce the MOR  NTX stimulus.

In spite of the fact that MOR pretreatment is a prerequisite for NTX-induced stimulus effects, the mere presence or absence of MOR during testing is not the primary determinant of MOR  NTX-appropriate responding. When SAL was given 4 h after the training dose of MOR alone or during MOR infusion via osmotic pump, no MOR  NTX-appropriate responding occurred. Therefore, the discrimination was not simply based on the presence versus absence of MOR (agonist)-like effects at the time of testing. NTX pretreatment blocked the MOR  NTX discrimination, also ruling out the possibility that MOR  NTX-appropriate responding was based on a nonopioid effect of MOR that was simply "unmasked" by the administration of an opioid antagonist. Rather, it indicates the discrimination was dependent upon MOR occupying opioid receptors, presumably µ, for a finite period of time.

It is possible that acute MOR pretreatment simply enhances some opioid-specific stimulus effect of a lower (0.3 mg/kg) NTX dose, effectively producing a high-dose stimulus. However, because doses of NTX (100 mg/kg) that were 2000-fold higher than the NTX ED₅₀ after MOR pretreatment produced less than 50% MOR  NTX-appropriate responding, this explanation appears unlikely. Similarly, a lower dose of NTX (0.3 mg/kg) preceded by SAL was the alternative choice during training, and it clearly did not share significant MOR  NTX-like stimulus effects.

Repeated opioid antagonist administration sometimes results in progressive sensitization to the behavioral effects of the antagonist (Goldberg et al., 1981; Dykstra, 1983; Adams and Holtzman, 1990; Schindler et al., 1993). Therefore, NTX might become a more potent stimulus, either alone or after MOR, over the course of a lengthy study such as the present one. However, there was little change in the effects of NTX in MOR-pretreated rats over the course of this year-long study (Table 1, first and second determinations). Moreover, the continued efficacy of SAL  NTX as an alternative training stimulus argues against any significant increase in the stimulus potency of NTX alone over the course of the experiments.

Previous drug discrimination studies have yielded generalization profiles for opioid antagonists that were dependent on previous MOR exposure, and it has been suggested that compared with antagonists alone, a novel and presumably opioid-mediated stimulus state results when opioid antagonists are given following MOR (Frey and Winter, 1978; France and Woods, 1985). For example, naive animals can easily discriminate among MOR (0.1-10 mg/kg), SAL, and NTX (10-100 mg/kg; France and Woods, 1985). However, at least in naive pigeons, opioid antagonists such as nalorphine, diprenorphine, or cyclazocine do not generalize to the interoceptive cue produced by a high training dose of NX or NTX, suggesting that the interoceptive effects of high doses of NTX or NX are not opioid specific (Valentino et al., 1983). In contrast, rats that are maintained chronically on MOR can be trained to discriminate SAL from as little as 0.1 mg/kg of NTX, a cue that generalizes dose dependently to other opioid antagonists (Gellert and Holtzman, 1979; Holtzman, 1985).

The present data add further support to the theory that acute MOR pretreatment does not simply enhance the existing stimulus effects of NTX but instead induces a state of physical dependence that is unmasked by an antagonist that also produces attendant interoceptive stimuli (Meyer and Sparber, 1977). In humans, physical dependence is assessed by the emergence of a withdrawal syndrome upon removal of chronic MOR treatment or by administration of an opioid antagonist after chronic or acute (single-dose) MOR treatment (Jaffe, 1990). There are qualitative similarities in the negative interoceptive states associated with abrupt/spontaneous withdrawal of chronic MOR treatment or with antagonist administration in acutely MOR-treated subjects (Bickel et al., 1988; Kanof et al., 1992). In this study, NTX produced signs of physical opioid withdrawal, in conjunction with full MOR  NTX-like stimulus effects after acute or during chronic MOR administration, providing further support for the existence of an acute opioid dependence syndrome with interoceptive and somatic components.

Dissociations between the physical and interoceptive signs of opioid dependence have been reported, and this fact must be considered in interpreting the dependence-predictive relevance of the physical signs reported herein (Higgins and Sellers, 1994). Compared with the NTX-induced signs seen after acute MOR pretreatment (Wei et al., 1973; Schulteis et al., 1994), a larger number of signs (i.e., wet-dog shakes and abdominal contractions) were seen 6 to 12 h following the removal of the MOR-filled osmotic pumps. At these same time points, only partial MOR  NTX-appropriate responding occurred. The dose- and time-dependence of the physical signs accompanying abrupt/spontaneous withdrawal are consistent with previous literature on withdrawal from chronic MOR treatment (Wei et al., 1973). In fact, only when comparing among MOR pretreatment conditions is it evident that full MOR  NTX-appropriate responding, although dependent on the presence of NTX, is seen (Fig. 1, training curve) when physical withdrawal signs are comparatively minimal (Table 2).

The potency of NTX to occasion MOR  NTX-appropriate responding was comparable in rats pretreated with a single dose of MOR (10 mg/kg) and in rats treated with MOR chronically (20 mg/kg/day). The present surprising results are consistent with previous findings that the potency of NTX in producing effects on behaviors maintained by food or i.c. self-stimulation is comparable after acute MOR injection or during chronic MOR infusion (Adams and Holtzman, 1990; Easterling and Holtzman, 1997). The generality of these previous observations is now extended to interoceptive effects. In all of the experiments reported herein, antagonist injection produced full MOR  NTX-like interoceptive effects in MOR-treated rats. Following either acute or chronic MOR treatment, the similarity in the stimulus potency of NTX (Fig. 6) suggests that acute MOR pretreatment probably results in a relatively high-intensity interoceptive stimulus. The fact that antagonist-precipitated abstinence syndromes are generally more intense and of shorter duration than the spontaneous withdrawal syndrome (Wei et al., 1973) might explain the failure of spontaneous withdrawal to generalize completely with MOR  NTX. Regardless, our results suggest that acute MOR treatment produces changes in the cellular substrates for these antagonist-induced effects that are comparable with those changes produced by chronic MOR treatment.

One possible mechanism that might account for the effects of NTX that follow acute or prolonged MOR treatment is MOR-induced conversion of opioid receptors to a constitutively active state (Cruz et al., 1996). The existence of constitutively active opioid receptors, which couple to G proteins in the absence of an agonist, has been proposed (Bilsky et al., 1996). In a system with a large number of constitutively active opioid receptors (a dependent rat) an antagonist might have negative intrinsic efficacy (Wang et al., 1994). In such a system, an antagonist, by rapidly binding constitutively active receptors, might produce quantitatively more intense or qualitatively different effects than those that would be expected to follow the abrupt withdrawal of chronic MOR treatment. Although the model requires experimental validation, it would provide an explanation of how low doses of opioid antagonists can have such profound effects on behavior hours after a single dose of MOR.

In humans who are physically dependent upon a MOR-like drug, the desire to prevent the emergence of aversive withdrawal symptomatology is an important factor in the perpetuation of drug self-administration (Jaffe, 1990). Animal models are often criticized for failing to provide a subjective measure of the motivational changes occurring during opioid withdrawal that have been documented in humans. In the procedure described herein, changes in MOR  NTX-appropriate responding were noted when subjects were undergoing NTX-precipitated withdrawal from acute or chronic MOR and when MOR-filled pumps were abruptly removed. Therefore, this drug discrimination procedure should afford a valuable animal model for studying behavioral, and ultimately, cellular events that reflect the early drug-receptor interactions underlying the development of physical dependence upon MOR-like drugs.