Sensitivity to the Discriminative Stimulus and Antinociceptive Effects of μ Opioids: Role of Strain of Rat, Stimulus Intensity, and Intrinsic Efficacy at the μ Opioid Receptor1

Materials and Methods

Subjects.

Experimentally naive rats were obtained from Charles River Suppliers (Raleigh, NC). Rats of the LE, SD, Lewis, and F344 strains were maintained at approximately their 80% free feeding body weight by restricting daily intake of Purina Rat Chow, had unlimited access to water, and were housed individually in a climate-controlled colony maintained on a 12-h light/dark cycle.

Apparatus.

Antinociceptive tests were conducted with 40, 50, 52, and 56°C water, maintained at the particular temperature using hot water baths (Fisher Scientific, Inc., Fairlawn, NJ). Tail withdrawal latencies were measured using a hand-operated digital stopwatch with a time resolution of 0.01 s. Drug discrimination sessions were conducted in seven plastic and aluminum operant conditioning chambers: four chambers were 23 cm long, 19 cm high, and 20 cm wide and four chambers were 25 cm long, 31 cm high, and 25 cm wide. Each chamber was equipped with two 5-cm-long response levers located 9 cm from the chamber floor and either 1 or 3 cm from either wall, with either two or three stimulus lights located above them. When operated, a pellet dispenser delivered a 45-mg Noyes food pellet (P.J. Noyes, Co., Lancaster, NH) into a pellet trough that was centrally mounted between the two levers and approximately 1 cm above the chamber floor. House lights were centrally mounted on the ceiling 2.5 cm from the rear wall in four chambers and 1 cm below the ceiling on the front wall in four chambers. All chambers were equipped with an exhaust fan that supplied ventilation and white noise to mask extraneous sounds. Scheduling of experimental events and data collection were accomplished through the use of a microcomputer, using software and interfacing supplied by MED Associates, Inc. (St. Albans, VT).

Antinociception Testing.

In this procedure, rats were gently restrained and approximately half of the distal portion of the tail placed in either 40, 50, 52, or 56°C water, and the latency to remove the tail was recorded using a hand-held stopwatch. An increase in the latency to remove the tail from the warm water was taken as a measure of antinociception with an upper cutoff time of 15 s. In one series of tests, trials with the 40°C water were conducted once in each rat, followed by two baseline latency scores for both the 50 and 52°C water. The order of warm water tests was counterbalanced across rats, with at least 3 min separating each trial. In a second series of tests, baseline trials were conducted with 40 and 56°C water.

After baseline tail withdrawal latencies were determined under these conditions, the testing procedure was initiated. For tests conducted with the 50 and 52°C water, 30 min after administration of the opioid, one test was conducted using one water temperature, and then 3 min later the other temperature was used, with the order of testing counterbalanced across rats. Three min later, the next injection was administered, and the cycle began again. Each successive dose increased the total opioid concentration by 0.25 or 0.5 log units. A similar procedure was used when tests were conducted with the 56°C water, with the exception that only one water temperature was used. For any given rat, no more than five tests were conducted with a minimum of 1 week separating tests. Each drug was tested in 7 to 10 rats.

Drug Discrimination Training and Testing.

Rats were initially trained to press either of two response levers for food delivery, with responding maintained by a fixed ratio (FR) 1 schedule. During preliminary sessions, the lever designated as correct varied from session to session, and over several sessions the ratio value on each lever was increased gradually until a FR20 schedule was in effect. Discrimination training began when the FR value on each lever was approximately 10 (FR10). During these training sessions, injections of the training drug (3.0 or 5.6 mg/kg morphine) or 1.0 ml/kg water were administered 30 min before the session. A random sequence was used to determine which injection was administered, with the restriction that the same injection was not given on more than two consecutive sessions and that the number of drug and water injections was approximately equal over a 30-session period. After injection of the training drug, responses on one lever were reinforced, whereas responses on the other lever were recorded but had no programmed consequences. After injection of water, the contingencies were reversed. For approximately half of the rats, the left lever was designated morphine-appropriate, and the right lever was designated water-appropriate. These conditions were reversed for the other rats. Sessions lasted 15 min and were conducted 5 days per week. Training continued until the mean percentage of injection-appropriate responding before delivery of the first reinforcer was equal to or greater than 80% over 10 consecutive sessions.

Once the discrimination criterion was met, substitution tests were initiated. These tests were typically conducted on Tuesdays and Fridays, whereas training sessions continued on Mondays, Wednesdays, and Thursdays. If discrimination performance was <80% injection-appropriate before the delivery of the first reinforcer on Monday or Thursday, the next scheduled test session was replaced with a training session. During all test sessions, the conditions were the same as during the training sessions, except that completion of the FR20 requirement on either lever was reinforced.

Rats from all four strains were trained with the 3.0 mg/kg training dose. Based on the antinociception experiments, F344 and Lewis appeared the most and least sensitive to the effects of these opioids. Therefore, additional groups of rats (F344 and Lewis) were trained to discriminate 5.6 mg/kg morphine from saline. Because no differences were observed at this higher training dose, rats of the LE and SD strains were not trained with this dose. In addition, it was determined (data not shown) that rats could not be trained with 10.0 mg/kg morphine, as this dose completely suppressed responding and endangered the health of the animals. There were eight rats in each 3.0 mg/kg training dose group and 12 rats in each 5.6 mg/kg training dose group. Each drug was tested in five to eight rats, except morphine, which was tested in all rats.

Data Analysis.

Tail withdrawal latencies (test latency) were converted to percent antinociceptive effect using the following equation:In the drug discrimination procedure, the number of responses emitted on both response levers before delivery of the first reinforcer was recorded; these data were then converted to percent drug-appropriate responding. Rate of responding was expressed as average responses per second over a session. Each of these measures are expressed as a function of the drug dose for each strain of rat tested.

In general, each figure displays doses ranging from those that had no effect on a given dependent measure up to the lowest dose to produce >80% antinociception or morphine-appropriate responding. ED₅₀ values were estimated by log-linear interpolation from the regression line using at least three points on the ascending limb of the group dose-effect curve. ED₅₀ values were considered significantly different if the 95% confidence limits (95% CL) did not overlap. With nalbuphine in the tail withdrawal procedure, ED₅₀values could not be reliably determined; therefore, area under the dose-effect curve was estimated by the Trapezoidal Rule; a repeated measures ANOVA was then used to determine differences across strains of rat. Control values in both the tail withdrawal and drug discrimination procedures were also subjected to an ANOVA. For all analyses, the alpha level was set at 0.05, and post hoc comparisons were made using a Bonferroni adjustment for repeated comparisons. Finally, based on the absolute ED₅₀ values, each strain of rats was assigned a ranking of 1 (lowest ED₅₀ value) through 4 (highest ED₅₀ value). These rankings were averaged across temperatures, resulting in a mean rank order of potency across strains for each drug.

Drugs.

The following drugs were used: morphine sulfate, buprenorphine hydrochloride (both provided by the National Institute on Drug Abuse), butorphanol tartrate (generously supplied by Bristol-Meyers, Wallingford, CT), levorphanol tartrate, and nalbuphine hydrochloride (both purchased from Research Biochemicals Inc., Natick, MA). All drug doses are expressed in terms of the salts. All drugs were dissolved in distilled water and administered i.p. in an injection volume of 0.5 to 1.0 ml/kg.

Results

Tail Withdrawal Procedure.

Baseline performance. When tested with the 40°C control water, all latencies were 15 s. Table 1 shows that the control withdrawal latencies varied as a function of stimulus intensity (F _2,44 = 97.0, p < .05) but not as a function of rat strain. That is, average latencies for 50, 52, and 56°C water were 10.7, 7.4, and 3.9 s, respectively. However, there were no differences in baseline latency across rat strains for each of the temperatures.

Effects of opioids.

Fig. 1 shows the effects of the high-efficacy μ opioids morphine and levorphanol in the tail withdrawal procedure using 50, 52, and 56°C water. In F344, SD, LE, and Lewis rats, morphine produced dose-dependent increases in percent antinociceptive effects with maximal effects (i.e., >80% antinociceptive effect) obtained at each of the water temperatures. Based on ED₅₀ values (Table2), the potency of morphine decreased with increases in the water temperature in F344, SD, and Lewis rats. At the 52 and 56°C water temperatures, morphine was approximately 2.0- and 3.0-fold more potent in the F344 than the Lewis, respectively. Although the absolute differences across strains were small, the mean rank order of morphine’s potency was F344 > SD > LE = Lewis.

As observed with morphine, levorphanol produced dose-dependent increases in antinociception in each strain with maximal effects obtained at each of the water temperatures. Based on ED₅₀ values obtained at the 50 and 52°C water (Table 2), levorphanol was approximately 2.0- and 7.0-fold more potent in the F344 than in LE and Lewis, respectively. Similarly, when tested using 56°C, levorphanol was between 2.0- and 8.0-fold more potent in the F344 than the other strains, and approximately 4.0-fold more potent in the SD than the Lewis. When collapsed across temperatures, the mean rank order of potency for levorphanol was F344 > SD > LE = Lewis.

Figure 2 shows the effects of the intermediate-efficacy μ opioid buprenorphine in four strains of rats tested in the tail withdrawal procedure. At the 50°C water, buprenorphine produced dose-dependent increases in antinociceptive effects and was more than 10-fold more potent in F344 than in LE and Lewis (see Table 2). At 56°C water, buprenorphine produced maximal effects in F344 and SD at a 0.3 mg/kg dose and in LE at a 1.0 mg/kg dose. In contrast, even when tested up to the 1.0 mg/kg dose, buprenorphine had little antinociceptive effect in Lewis. The mean rank order of potency across temperatures for buprenorphine was F344 > SD > Lewis > LE.

The effects of the low-efficacy μ opioids butorphanol and nalbuphine in the tail withdrawal procedure using 50 and 52°C water are shown in Fig. 3. At both water temperatures, butorphanol produced maximal effects in the F344 rats at the 0.1 mg/kg dose. Somewhat different effects were obtained in the SD, where maximal effects were obtained at 1.0 mg/kg butorphanol at the 50°C. Although butorphanol produced dose-dependent increases in the percentage of antinociceptive effects in SD when tested at 52°C, the dose-effect curve was relatively shallow with near maximal effects obtained at doses ranging from 0.3 to 10 mg/kg. A shallow dose-effect curve was also obtained for butorphanol in the LE, where even the 56 mg/kg dose produced only 50% and 40% effect at the 50 and 52°C water, respectively. In the Lewis, no dose of butorphanol produced greater than 20% maximal effects. When collapsed across temperatures, the mean rank order of potency for butorphanol was F344 > SD > LE > Lewis.

Also shown in Fig. 3 is that nalbuphine failed to produce maximal effects in any of the strains when tested at the 50 and 52°C water temperatures. There were, however, differences across strains in terms of the effects produced by nalbuphine and area under the dose-effect curve. For example, at 50°C water near maximal effects were obtained in the F344 and SD at the 30 and 56 mg/kg doses, respectively. In these strains, only half-maximal effects were obtained at the highest dose of nalbuphine using the 52°C water. The maximal level of antinociception obtained with nalbuphine in the LE at both temperatures was 25%. Across the dose range examined, antinociceptive effects were not obtained in the Lewis rat. Based on the area under the dose-effect curve, the mean rank order for sensitivity to nalbuphine’s effects was F344 > SD > Lewis > LE.

Drug Discrimination.

Baseline performance. As shown in Table 1, the number of sessions required until acquisition of the discrimination (sessions to criteria) varied across strains and training doses, with F344 requiring the fewest sessions and SD the most sessions (p < .05). There was no relationship, however, between the amount of training needed to acquire the discrimination and sensitivity to the discriminative stimulus effects of morphine. Table 1 also shows that there were no consistent differences across strains in the percentage of injection-appropriate responding engendered by the training dose of morphine or water. There were differences in response rate among the strains; in particular, LE rats responded at a higher rate than both F344 and Lewis groups under morphine and water training conditions (see Table 1).

Effects of opioids.

Fig. 4 shows the effects of morphine, levorphanol, and buprenorphine in rats trained to discriminate 3.0 (left) and 5.6 (right) mg/kg morphine from water. Each of these opioids produced dose-dependent increases in the percentage of drug-appropriate responding; this effect was observed in all four strains of rats and in both training dose conditions. As shown in Table 3, there were no consistent differences in sensitivity to the discriminative stimulus effects of these opioids across these strains. Figure5 shows the effects of butorphanol and nalbuphine in rats trained to discriminate 3.0 (left) and 5.6 (right) mg/kg morphine from water. Butorphanol and nalbuphine also produced dose-dependent increases in the percentage of drug-appropriate responding with no consistent differences in the potency of these drugs or maximal effects across strains or training dose conditions (Table3).

Figure 6 shows the effects of morphine, levorphanol, buprenorphine, butorphanol, and nalbuphine on rates of responding in rats trained to discriminate 3.0 and 5.6 mg/kg morphine from saline. In each strain and in both training dose conditions, morphine, buprenorphine, and butorphanol decreased rates of responding in a dose-dependent manner, whereas levorphanol and nalbuphine failed to markedly alter rates of responding. There was no evidence of differential sensitivity across the strains in terms of the rate-decreasing effects of these drugs.

Discussion

The present findings are consistent with a growing body of evidence indicating that the level of antinociception produced by a particular opioid is dependent on the intrinsic efficacy of the opioid and the nociceptive stimulus intensity. Furthermore, these findings demonstrate that F344, SD, LE, and Lewis rats differ in terms of the potency and effectiveness of various opioids. The magnitude of these differences is increased when tests are conducted using a relatively high-intensity nociceptive stimuli or lower-efficacy μ opioids. In contrast to the antinociceptive effects of μ opioids, a differential sensitivity (in either potency or maximal effects produced) across rat strains was not apparent in the drug discrimination procedure, regardless of the intrinsic efficacy of the opioid or the training dose of morphine.

One dimension along which μ opioids differ is their relative degree of intrinsic efficacy at the μ opioid receptor. In antinociceptive assays, the relative efficacy of an opioid can often be inferred on the basis of its effectiveness across different intensities of the nociceptive stimulus (O’Callaghan and Holtzman, 1975; Walker et al., 1993; Butelman et al., 1995; Morgan and Picker, 1996; Morgan et al., 1999). For example, whereas high-efficacy opioids typically produce maximal antinociceptive effects regardless of the intensity of the nociceptive stimulus, the potency and effectiveness of lower-efficacy opioids typically decrease with increases in the intensity of the nociceptive stimulus. Findings obtained in various assays indicate the relative ranking of intrinsic efficacy for the μ opioids used in the present study was: morphine ≥ levorphanol > buprenorphine > butorphanol > nalbuphine (Adams et al., 1990; France and Woods, 1990; Paronis and Holtzman, 1992; Young et al., 1992; Picker et al., 1993). That the same ranking was obtained in the present study in four strains of rats attests to the robustness of this phenomenon. It should be noted that each of these drugs has significant affinity for other types of opioid receptors (e.g., κ receptors); however, studies from our laboratory and others using irreversible and competitive antagonists and cross-tolerance regimens have demonstrated that the antinociceptive effects of these drugs in this procedure are primarily mediated by actions at the μ opioid receptor (Zimmerman et al., 1987; Walker et al., 1994; Tiano et al., 1998). Furthermore, it has been demonstrated that when these opioids are administered in combination they produce either additive effects or act as competitive antagonists. Both of these outcomes are consistent with activity at the same receptor site (probably the μ opioid receptor;Morgan et al., 1999).

A major goal of the present study was to evaluate the antinociceptive effects of opioids in different rat strains and to determine if the magnitude of the differences in sensitivity (in potency or maximal effects) across rat strains is influenced by the sensitivity of the task, which can be altered by changing the intensity of the nociceptive stimulus. In general, the potency of the high-efficacy μ opioids morphine and levorphanol to produce antinociception decreased with increases in the intensity of the nociceptive stimulus. Differences in the potency of these opioids across strains became apparent as the intensity of the nociceptive stimulus increased. For example, morphine was equipotent across strains at the low-intensity stimulus. At the highest intensity however, morphine was approximately 2- and 3-fold more potent in the F344 than in the SD and Lewis rats, respectively. Most studies show little or no differences across rat strains in response to morphine’s antinociceptive effects (Woolfolk and Holtzman, 1995), although consistent differences have been noted (Vaccarino and Couret, 1995). The present findings demonstrate that these differences are most pronounced when high-intensity nociceptive stimuli are used.

Differences between the various strains of rats were also apparent when tested with opioids that have low efficacy at the μ opioid receptor. Both the potency and effectiveness of buprenorphine, butorphanol, and nalbuphine varied considerably across strains, with the F344 generally being the most sensitive and Lewis the least sensitive. The magnitude of this differential sensitivity was influenced by stimulus intensity. For example, although buprenorphine was approximately 18-fold more potent in the F344 when tested at 50°C water, maximal effects were obtained in both strains. In contrast, at the 56°C water, buprenorphine produced maximal effects in the F344 and was ineffective across the dose range examined in the Lewis. The most dramatic differences across strains were obtained with the lower-efficacy μ opioid butorphanol, where F344 rats showed maximal effects at a dose of 0.1 mg/kg and the Lewis failed to show any antinociceptive effects even when tested up to a dose as high as 56 mg/kg.

There are several possible explanations for the differential potency and effectiveness across strains, including different pharmacokinetic properties of these opioids (but see Guitart et al., 1992; Gosnell and Krahn, 1993), differences in absolute numbers and/or densities of μ opioid receptors (Baran et al., 1975; Sudakov et al., 1993; Elmer et al., 1995), or different neurochemical attributes (e.g., endogenous opioid levels; Beitner-Johnson et al., 1991; Guitart et al., 1992,1993; Strecker et al., 1995). At this point, it is impossible to conclude with certainty that any of these possibilities account for the differential sensitivity observed in this study.

The findings that the antinociceptive effects of opioids were dependent upon the stimulus intensity and the intrinsic efficacy of the opioid suggest that the strain differences may reflect some fundamental difference in the physiological mechanisms underlying efficacy. For example, it is possible that the intracellular mechanisms determining μ opioid efficacy (e.g., the degree and/or the amount of G protein activation; Selley et al., 1997) differ across strains. That is, the efficiency of the coupling mechanism between receptor and intracellular signaling proteins in the F344 may be considerably greater than in the Lewis. When tested with high-efficacy μ opioids, this difference may not be apparent because these opioids require so few receptors to produce maximal effects. Under conditions where the physiological system is taxed (e.g., by increasing the intensity of the nociceptive stimulus) or when using low-efficacy opioids that require occupation of close to all available receptors to produce a given effect, these intrinsic differences may result in measurable differences in potency and/or effectiveness, as observed in the present study.

The findings obtained in the antinociception procedure contrast with those obtained in the drug discrimination procedure where no consistent differences across strains were observed. Several possibilities may account for the lack of this differential sensitivity, including the possibility that antinociceptive effects of these opioids are mediated by different anatomical loci or pathways (e.g., the spinal cord) than the discriminative stimulus and rate-decreasing effects (e.g., the brain). A number of studies indicate that various spinal tracts and neural structures, such as the periaqueductal gray, are important in mediating opioid antinociception (for a review, see Yaksh, 1997;Jensen, 1997). Although the neural correlates of the discriminative stimulus or rate-decreasing effects of these opioids have not been clearly identified, many of the substrates that mediate opioid-induced antinociception have been eliminated (e.g., Jaeger and van der Kooy, 1993; Shoaib and Spanagel, 1994).

An alternative explanation is that the drug discrimination procedure is more sensitive to opioid agonist effects compared with the tail withdrawal procedure (Holtzman, 1997). In the drug discrimination procedure, the lower-efficacy opioids nalbuphine and butorphanol produced maximal or near maximal effects in both the low and high training dose conditions. In contrast, at the lowest water temperature used in the tail withdrawal procedure, nalbuphine and butorphanol produced no antinociception in the Lewis and only low levels of effects in the F344. Furthermore, comparisons of ED₅₀ values suggest that these opioids were more potent in the drug discrimination procedure compared with the antinociception procedure. Taken together, these data suggest that the drug discrimination procedure is more sensitive to agonist effects. It is possible that differences between the strains would be more apparent had higher training doses of morphine been used. Unfortunately, higher doses of morphine (e.g., 10 mg/kg) eliminated responding in both F344 and Lewis. A previous study (Young et al., 1992) found that nalbuphine failed to substitute for the same training dose of morphine used in the present study. It has been previously suggested (Picker et al., 1993; Morgan and Picker, 1996) that the level of substitution observed with these opioids (i.e., lower-efficacy opioids) across studies may vary substantially given the differences in sensitivity across individuals. Whether the failure to directly replicate Young et al.’s findings with nalbuphine are due to individual differences or to procedural differences (e.g., route of administration) is unclear at this point.

Taken together, the present findings suggest that there are particular situations where very profound differences in sensitivity to a drug effect exist between several strains of rats, and in some instances that these differences can be as large or larger than those observed in divergent lines of rodents selected for low- and high-sensitivity to a particular opioid effect. In particular, these differences are largest under extreme conditions (i.e., low-efficacy opioids or insensitive tasks are used).