Psychopharmacology (1999) 144:45Ð53 © Springer-Verlag 1999

ORIGINAL INVESTIGATION

Ellen A. Walker á Michael J. Tiano á Steven I. Benyas Linda A. Dykstra á Mitchell J. Picker and β-funaltrexamine antagonism of the antinociceptive and response rate-decreasing effects of , , and d-propoxyphene

Received: 14 August 1998 / Final version: 4 December 1998

Abstract Rationale: Patterns of competitive and insur- d-propoxyphene, confirming other reports that dezocine mountable antagonism provide important data to guide is a lower efficacy than morphine. Additionally, the classification and characterization of different types the degree of antagonism produced by β-FNA was great- of as well as infer the mechanism of ac- er for the antinociceptive effects of all three compounds tion for agonists. Objective: Experiments with the com- than for the rate-decreasing effects. petitive antagonist, naltrexone, and the insurmountable antagonist, β-funaltrexamine (β-FNA), were conducted Key words Morphine á β-Funaltrexamine á Dezocine á to determine whether the antinociceptive and rate-de- d-Propoxyphene á Naltrexone á Efficacy creasing effects of the opioid agonists dezocine and d- propoxyphene are 1) mediated through µ opioid recep- tors in rats, and 2) differ from morphine with respect to Introduction relative efficacy. Methods: The rat tail-withdrawal assay was used to measure antinociception and a fixed ratio 20 Experiments with competitive and insurmountable an- (FR20) schedule of food delivery was used to measure tagonists provide important data to guide the classifica- rate suppression. Results: Naltrexone (0.01Ð1.0 mg/kg) tion and characterization of different types of opioid ago- was approximately equipotent as an antagonist of the an- nists and to infer mechanisms of action for these ago- tinociceptive and rate-decreasing effects of both mor- nists. For example, experiments with competitive antag- phine and dezocine and as an antagonist of the antinoci- onists show that the behavioral effects of different opioid ceptive effects of d-propoxyphene. Naltrexone failed to agonists such as morphine, U50,488, and BW373U86 block the rate-decreasing effects of d-propoxyphene. β- are mediated through different types FNA (5 and 10 mg/kg) also antagonized the antinocicep- (Dykstra et al. 1988; Negus et al. 1993). Additionally, tive and rate-decreasing effects of morphine and dezo- experiments with competitive antagonists provide evi- cine as well as the antinociceptive effects of d-propoxy- dence that some classes of opioid agonists produce dis- phene. β-FNA failed to produce a dose-dependent antag- criminative-stimulus, rate-decreasing, and antinocicep- onism of the rate-decreasing effects of d-propoxyphene. tive effects through a common, presumably µ, opioid re- Conclusions: These data suggest that the antinociceptive ceptor (Walker et al. 1994). Experiments with insur- effects of morphine, dezocine, and d-propoxyphene and mountable antagonists such as β-funaltrexamine (β- the rate-decreasing effects of morphine and dezocine are FNA) can provide data on the relative efficacy of opioid mediated through µ opioid receptors. Overall, high doses agonists (Zimmerman et al. 1987; Adams et al. 1990; of β-FNA produced a greater degree of antagonism of Pitts et al. 1996; Morgan and Picker 1998). If a larger the behavioral effects of dezocine than morphine or dose of β-FNA is required to decrease the maximum ef- fects produced by morphine than dezocine, it can be in- ferred that dezocine requires a larger fraction of the re- E.A. Walker (✉) á M.J. Tiano á S.I. Benyas á M.J. Picker ceptor population available to produce effects than mor- Department of Psychology, CB# 3270, Davie Hall, phine. University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-3270, USA The purpose of the present series of experiments was e-mail: [email protected], Fax: +1-919-962-2537 to use competitive and insurmountable antagonists to compare the receptor selectivity and relative efficacy of L.A. Dykstra Department of Psychology and Pharmacology, two opioid agonists, d-propoxyphene and dezocine. Al- University of North Carolina at Chapel Hill, though d-propoxyphene and dezocine are used clinically, Chapel Hill, NC 27599-3270, USA relatively little is known about their behavioral pharma- 46 cology. The radioligand binding profiles for d-propoxy- non-opioid activity and can be used to assess a drug’s se- phene and dezocine are similar in that both reveal an ap- lectivity over a behaviorally active range of doses (cf. proximately 10-fold selectivity of µ over δ opioid recep- Negus et al. 1993). tor subtypes and a 30-fold selectivity of µ over κ opioid receptor subtypes (Neil 1984; O’Brien and Benfield 1989; Chen et al. 1993; Yoburn et al. 1995). Both d-pro- Materials and methods poxyphene and dezocine produce morphine-like anti- nociception in a number of assays (Malis et al. 1975; Subjects Hayes et al. 1987a; Tiano et al. 1998) as well as mor- Forty-six male, Long-Evans rats weighing 300Ð360 g were used. phine-like discriminative stimulus effects (Shannon and Rats were housed individually in a humidity- and temperature- Holtzman 1977; Overton and Batta 1979; Young et al. controlled room under a 12-h light/dark cycle. Rats were fed a dai- 1992). Early characterization indicates that d-propoxy- ly ration of 20Ð25 g to maintain body weights and were given con- phene’s relative efficacy as an agonist is similar to that tinuous access to water. Testing of individual rats occurred ap- proximately once a week in the tail-withdrawal assay and once or of morphine and higher than that of (Mar- twice a week in the schedule-controlled responding assay. The rats tin et al. 1976; Hayes et al. 1987b). In contrast, observa- in the tail-withdrawal experiments were habituated to the restraint tions in animal (Malis et al. 1975; Rowlingson et al. tubes and the testing room on two occasions before drug testing 1983; Young et al. 1984) and human subjects (Rom- began. agnoli and Keats 1984; O’Brien and Benfield 1989; Zacny et al. 1992; Strain et al. 1996) indicate that dezo- Apparatus cine has lower efficacy that morphine. Although reports indicate that dezocine and d-propox- In the tail-withdrawal assay, six rodent restraint tubes were used. yphene produce their behavioral effects through the A Precision Model 182 water bath maintained water temperature at approximately 80¡C. Water temperatures of 40 and 55¡C were same opioid receptor subtype as morphine, there is also obtained by mixing water from the water bath with tap water in a evidence to suggest differences among these agonists. wide-mouth thermos (diameter=11 cm). Water temperature was For example, several studies indicate that the doses of an measured with a VWR Digital Thermometer Model 100 A and required to block the antinociceptive, tail-withdrawal latency was measured with a hand-operated digital stopwatch with a time resolution of 1/100 s. rate-decreasing, or discriminative-stimulus effects of de- In the schedule-controlled responding assay, experiments took zocine and d-propoxyphene are similar to the doses re- place in eight standard rat operant-conditioning chambers (Ger- quired to block the effects of morphine (Nickander et al. brands model G7322) enclosed in ventilated sound-attenuating cu- 1977; Lehman and Peterson 1978). Nevertheless, other bicles. Each chamber was equipped with two response levers mounted 13 cm apart and 9 cm above the floor. Two stimulus studies (Gal and DiFazio 1984) indicate that opioid an- lamps were located 7 cm above each lever. Only the left lever and tagonists are less potent in blocking these effects of de- stimulus lights were used in the present experiments. A food re- zocine and d-propoxyphene than morphine (Neil and ceptacle was centered 6 cm below the two levers. A Sonalert tone Terenius 1981; Schaefer and Holtzman 1981; Picker et generator was mounted behind the front panel and two houselights al. 1993). Other reports indicate that opioid antagonists were located on the ceiling above the rear wall. White noise was present within each chamber and throughout the room. Experi- are ineffective in blocking the rate-decreasing effects of mental events were programmed and data were collected using a dezocine and d-propoxyphene (Leander 1979; Picker microcomputer controlled by software and interfacing supplied by 1997). Moreover, d-propoxyphene reveals very little Med Associates, Inc. (Georgia, Vt., USA). cross-tolerance to morphine (Neil 1982) or supersensitiv- ity to naltrexone (Paronis and Holtzman 1992; Yoburn et Tail-withdrawal procedure al. 1995) in antinociceptive assays, suggesting that other opioid receptors or non-opioid receptors may be in- Thirty-six rats were used in the tail-withdrawal procedure. Antino- volved in some of d-propoxyphene’s effects. ciceptive testing occurred within a 4-h time period, between 10 a.m. and 2 p.m. Rats were placed into the restraining tubes with In order to examine these differences further, the rate- their tails hanging freely. The last 5Ð10 cm of the tail was im- decreasing and antinociceptive effects of dezocine and d- mersed into either 40¡ or 55¡C water, and the latency for tail with- propoxyphene were examined alone and in combination drawal was measured. A 15-s cut-off time was imposed to prevent with the competitive antagonist naltrexone and the insur- tissue damage. The first three test trials were conducted with 40¡C mountable antagonist, β-FNA. β-FNA is a highly-selec- water. If a rat removed its tail from the 40¡C water on more than one test trial, that rat did not continue in the experimental session tive, long-lasting µ-receptor alkylating agent that has (one rat was removed from one control morphine experiment). Af- been used in both antinociceptive and scheduled-con- ter the first test periods with 40¡C water, single tail-withdrawal la- trolled responding assays to make inferences about the tencies from 55¡C water were obtained for each rat. A 2-min inter- relative efficacy of opioid agonists (Zimmerman et al. val occurred between each test trial. After control values were obtained, the first drug injection was 1987; Adams et al. 1990; Pitts et al. 1996b). Operant and administered. Fifteen minutes later, latencies from both 40¡C and reflexive behavioral endpoints were chosen to increase 55¡C water were redetermined during a 10-min testing period. the generality of the conclusions. Antinociception, as the Each temperature was presented once, in random order, with a 2- reflexive measure, is a clinically relevant, behavioral min interval between stimulus presentations. At the conclusion of the 10-min testing period, a second drug injection occurred which endpoint that is extremely sensitive to opioid as well as increased total dose by 0.25Ð0.50 log unit. Fifteen minutes later, non-opioid activity. Schedule-controlled responding, as tail-withdrawal latencies were redetermined from 40¡C and 55¡C the operant measure, is also sensitive to both opioid and water. This cumulative-dosing procedure was terminated when 47 tail-withdrawal latencies reached maximum effect (15 s) in 55¡C In the naltrexone and β-FNA antagonism studies, a total of ten water, the solubility limits of a compound were reached, or other rats were used throughout the experiments. Two or three morphine effects interfered with the measurements (i.e. convulsions, death). or dezocine control dose-response curves were determined in sub- In the antagonism experiments, three morphine or dezocine sets of six to eight rats (naltrexone) or five rats (β-FNA). Two d- control dose-response curves were obtained in two groups of sev- propoxyphene control dose-response curves were determined in en rats. Thereafter, on a weekly basis, a morphine or dezocine cu- subsets of three to five rats. Naltrexone was administered and the mulative dose-response curve was redetermined, in the presence rate-decreasing effects of naltrexone alone were tested in the first of a dose of naltrexone, as described above. The effects of naltrex- component of the experiment. Thereafter, the agonist cumulative one alone were tested in the first component of the experiment. dose-response curve was redetermined, in the presence of naltrex- Two weeks after the last naltrexone experiment, a control dose-re- one, as described above. The effects of three doses of naltrexone sponse curve was determined for morphine and dezocine in the re- were examined over the course of 2 or 3 weeks. In the β-FNA ex- spective groups and was similar to previous control dose-response periments, a dose of 5.0 or 10 mg/kg β-FNA was administered and curves. Next, 5.0 mg/kg β-FNA was administered 24 h prior to the the rate-decreasing effects of β-FNA alone were tested for five redetermination of the morphine or dezocine dose-response curve. components. The next day, 24 h after the β-FNA injection, the ag- Morphine and dezocine were tested once a week in the respective onist dose-response curves were redetermined. Rats in these ex- groups until the dose-response curves were not significantly dif- periments received multiple injections of naltrexone and/or β- ferent than the initial control dose-response curves (approximately FNA over the course of 1 year. A rat was not tested after a naltrex- 4 weeks). Next, 10 mg/kg β-FNA was administered 24 h prior to one or β-FNA experiment with the next dose of naltrexone or β- the redetermination of the morphine or dezocine dose-response FNA unless the agonist dose-response curves were not significant- curve. ly different than the initial control dose-response curves. The naltrexone and β-FNA experiments with d-propoxyphene were conducted in a modified-groups design because preliminary tests in the scheduled-controlled responding assay (see below) re- Data analysis vealed that lesions at the site of injection, convulsions, and/or death were observed in some rats approximately 12 h after testing Tail-withdrawal latencies following drug administration were con- with d-propoxyphene. Thus, although two d-propoxyphene control verted into percent maximum effect by the formula: dose-response curves were tested in all rats, some rats participated β test latency− control latency in only one naltrexone or -FNA experiment. In the antagonism maximum effect = ×100 experiments in which high doses of d-propoxyphene were admin- (15 sec− control latency) istered, the rats were typically euthanized immediately after the tests. When relatively low doses of d-propoxyphene were used, as Each rat served as its own control. A value of 0% maximum effect in the 0.01 mg/kg naltrexone experiment, lethality was not ob- was assigned if tail-withdrawal latencies were less than the control served. Therefore, these rats also received a dose of β-FNA 2 latency. In the scheduled-controlled responding assay, the rate of weeks later. Other details of the antagonism experiments were as responding (responses/s) was determined for an individual rat by described above. dividing the total number of responses in a component by the time spent in the component. The effects of agonists alone or in combi- nation with the antagonists on rates of responding are expressed as Schedule-controlled responding procedure a percentage of saline control rates collected 1Ð3 days prior to the test. Control data were pooled for individual rats and then aver- Ten experimentally naive rats were trained to press the left lever aged into a group mean for both the tail-withdrawal and schedule- under an FR1 schedule of food presentation. When the stimulus controlled responding assays. lamps above the lever were lit, each lever press resulted in the de- The dose of each drug that produced 50% maximal effect or a livery of a 45 mg Noyes pellet. Whenever a pellet was delivered, 50% reduction in response rate (ED50 values) was determined for the houselight dimmed briefly and a tone was sounded. Once the each individual rat by linear regression using at least three points rats were pressing the lever reliably, the response requirement was of the ascending (tail-withdrawal assay) or descending (scheduled- increased gradually over several sessions to an FR20. When re- controlled responding assay) portions of the dose-response curves. sponse rates stabilized under the FR20 schedule, the rats were Individual ED50 values were then averaged to obtain group ED50 trained in a multiple FR20 timeout (TO) 10-min schedule. In this values and 95% CL. In the antagonism experiments, naltrexone schedule, the chamber was dark and lever presses had no pro- apparent pA2 values were determined using Schild regressions as grammed consequences during the TO component. Each FR20 previously described (Arunlakshana and Schild 1959) and with component ended after 5 min or the delivery of the tenth food pel- drug doses substituted for drug concentrations (Takemori 1974) let, whichever occurred first. If ten food pellets were earned be- using the program from Tallarida and Murray (1987). Specifically, fore 5 min, the chamber was darkened until the initiation of the in this case, the apparent pA2 value is defined as the negative log- next TO component; lever presses during this period had no pro- arithm of the molar dose of antagonist that produces a two-fold grammed consequences. Each session began with the FR20 com- shift to the right in the agonist dose-response curve. Slopes from ponent and the TO and FR20 components alternated thereafter. the Schild regression were considered to be significantly different Sessions ended after the completion of five to seven FR20 compo- from unity if the 95% CL of the slope did not include Ð1. To deter- nents and were conducted 5 days per week. mine significance in the β-FNA experiments, control individual β When response rates stabilized under the multiple schedule, ED50 values were compared to -FNA individual ED50 values us- dose-response curves for morphine, dezocine, and d-propoxy- ing a paired Student’s t-test (GraphPad Instat, V2.05a). Analysis phene were determined using a cumulative-dosing procedure. Un- of covariance of multiple regression lines was used to determine der this procedure, the lowest dose of a drug was administered. whether the slopes of the linear portions of the dose-response Ten minutes later, the first FR20 component began. At the conclu- curves were different (GraphPad Prism V1.03). The level of sig- sion of the 5-min FR20 component, the rat was removed from the nificance was set at P<0.05 for all analyses. chamber and injected with a second drug dose which increased the total dose by 0.25Ð0.50 log units. Ten minutes later, the second FR20 component began. This cumulative-dosing procedure con- Drugs tinued until a cumulative dose was reached that decreased re- sponse rates to fewer than 0.2 responses/s, the solubility limits of The following drugs were examined: morphine sulfate and β-fun- a drug were reached, or other effects of a drug interfered with the altrexamine hydrochloride (supplied by the National Institute on measurements (i.e. convulsions, death). Cumulative dose tests Drug Abuse, Rockville, Md., USA), dezocine hydrochloride (As- generally occurred on Tuesdays and Fridays and vehicle injections tra Pharmaceutical Products, Inc., Westborough, Mass., USA), d- were given before each component on Thursdays. propoxyphene and naltrexone hydrochloride (Research Biochemi- 48 cals Inc., Natick, Mass., USA). All injections were administered uled-controlled responding assay. All agonist dose-re- SC in the dorsal flank in a volume of 0.5Ð2.0 ml/kg. Doses are ex- sponse curves obtained for morphine, dezocine, and d- pressed in terms of the salt. propoxyphene in combination with naltrexone were par- allel to the initial dose-response curves for morphine, de- Experimental subjects zocine, and d-propoxyphene alone in both procedures. Linear Schild regressions (Fig. 2) and apparent pA The rats used in this study were maintained in accordance with the 2 guidelines of the Institutional Animal Care and Use Committee of analysis (Table 1) revealed naltrexone’s potency as an the University of North Carolina and the “Guide for the Care and antagonist of the antinociceptive and rate-decreasing ef- Use of Laboratory Animals” (Institute of Laboratory Animal Re- fects of morphine, dezocine, and d-propoxyphene. Nal- sources, National Academy Press 1996; NIH publication No. 85- trexone was approximately equipotent (i.e., the 95% CL 23, revised 1985). of the apparent pA2 values overlapped) as an antagonist of the antinociceptive and rate-decreasing effects of mor- phine and dezocine and as an antagonist of the antinoci- Results ceptive effects of d-propoxyphene. A Schild regression for naltrexone in combination with d-propoxyphene Naltrexone antagonism studies could not be determined because the effects of naltrex- one were not dose-dependent (Fig. 1). The slopes of the Prior to naltrexone pretreatment, cumulative doses of Schild regressions that were not different from unity morphine, dezocine, and d-propoxyphene each produced were constrained to Ð1 and the constrained apparent pA2 dose-dependent increases in tail-withdrawal latency (Fig. values for naltrexone were determined. The constrained 1, top panels). Morphine, dezocine, and d-propoxyphene apparent pA2 values were not significantly different than also produced dose-dependent decreases in response rate the unconstrained values. (Fig. 1, bottom panels). Naltrexone produced a dose-de- pendent, surmountable antagonism of the antinociceptive and rate-decreasing effects of morphine and dezocine. Naltrexone also antagonized the antinociceptive effects Fig. 1 Naltrexone antagonism of the antinociceptive and rate-de- creasing effects of morphine , dezocine and d-propoxyphene. Or- of d-propoxyphene although d-propoxyphene did not dinate, upper panels: % maximum antinociceptive response (15 s). produce a maximal antinociceptive effect when com- Baseline tail-withdrawal latencies ranged from 1.15 to 11.3 s. bined with a dose of 1.0 mg/kg naltrexone. Although low Control dose-response curves (●) are the average of two to three doses of naltrexone (0.01 and 0.1 mg/kg) shifted the d- experiments in seven rats. Other points represent the mean of one observation in seven rats. Ordinate, lower panels: response rate propoxyphene dose-response curve rightward in the expressed as a percentage of the saline control session 1Ð3 days scheduled-controlled responding procedure, 1.0 mg/kg prior to the test. Saline control values ranged from 0.47 to 1.54 re- naltrexone produced a smaller shift in the d-propoxy- sponses/s. Control dose-response curves are the average of two to phene dose-response curve than 0.1 mg/kg naltrexone. three experiments in six to eight rats. Other points represent the mean of one observation in six to eight (morphine and dezocine) Low doses of naltrexone blocked the convulsant and le- or three to five (d-propoxyphene) rats. Abscissa: dose of drug in thal effects of intermediate doses of d-propoxyphene ob- mg/kg. Points above N represent tests of naltrexone alone. Vertical served in 30% of the rats 12 h after testing in the sched- bars represent ±SEM. ● 0.01, ▲ 0.1 and ■ 1.0 mg/kg naltrexone 49

Fig. 2 Schild plots for naltrexone as an antagonist of antinocicep- fect was obtained in four of five rats in the d-propoxy- tive (left panel) and response-decreasing effects of morphine, de- phene experiment and after 10 mg/kg β-FNA, a maxi- zocine, and d-propoxyphene. Ordinate: logarithm of the quantity mum effect was obtained in three of four rats in the de- [ED50 of the agonist in the presence of naltrexone divided by the ED50 of the agonist alone]Ð1. The data for the ED50 values are zocine and d-propoxyphene experiment. from Fig. 1. Abscissae: negative logarithm of molar doses of nal- In the schedule-controlled responding assay, mor- trexone. ● Morphine, ▲ dezocine, ■ d-propoxyphene phine, dezocine, and d-propoxyphene also produced dose-dependent decreases in response rate (Fig. 3, bot- tom panels), similar to those decreases observed in earli- β-FNA antagonism studies er experiments (Fig. 1, bottom panels). β-FNA (5.0 and 10 mg/kg) alone failed to alter response rate when tested Prior to β-FNA pretreatment, cumulative doses of mor- after 15 min, 120 min, or 24 h (data not shown). Pre- phine, dezocine, and d-propoxyphene each produced treatment with 5.0 mg/kg β-FNA produced a 5-fold shift dose-dependent increases in tail-withdrawal latency (Fig. in the dezocine dose-response curve. This dose of β- 3, top panels) similar to those increases observed in the FNA failed to alter significantly the d-propoxyphene naltrexone experiments (Fig. 1, top panels). Pretreatment dose-response curve and was not tested in combination with 5.0 mg/kg β-FNA produced approximately 5- and with morphine. Increasing the pretreatment dose to 10 6-fold shifts in the morphine and dezocine dose-response mg/kg β-FNA produced approximately 6- and 4-fold curves, respectively (Fig. 3, top panels; Table 2). In- shifts in the morphine and dezocine dose-response creasing the pretreatment dose to 10 mg/kg β-FNA pro- curves, respectively, although the shift in the dezocine duced a larger degree of antagonism: 20- and 90-fold dose-response curve was not significant. A dose of 10 shifts in the morphine and dezocine dose-response mg/kg β-FNA did not significantly alter the d-propoxy- curves, respectively. Pretreatment doses of 5.0 and 10 phene dose-response curve. All agonist dose-response mg/kg β-FNA produced approximately 2- and 6-fold curves obtained for morphine, dezocine, and d-propoxy- shifts in the d-propoxyphene dose-response curves, re- phene in combination with β-FNA were parallel to the spectively, although the shift after 10 mg/kg β-FNA was initial dose-response curves for morphine, dezocine, and not significant. After 5.0 mg/kg β-FNA, a maximum ef- d-propoxyphene alone in both procedures.

Table 1 Apparent pA2 values a for naltrexone as an antagonist Agonist Slope unconstrained Slope constrained of antinociceptive and rate-de- creasing effects of ago- pA2 (95% CL) ÐSlope (95% CL) pA2 (95% CL) nists. Values are mol/kg and slope of the Schild plot Morphine Antinociceptive effects 8.4 (6.7Ð10) 0.68 (1.3Ð0.091) 7.8 (7.0Ð8.6) Rate effects 7.6 (5.2Ð9.9) 0.84 (2.4 to +0.72) 7.4 (6.9Ð7.9) Dezocine a Slopes of the Schild regres- Antinociceptive effects 8.5 (7.5Ð9.6) 0.51 (0.76Ð0.27) Ð sions were constrained to Ð1 if Rate effects 8.0 (4.9Ð11) 0.69 (2.0 to +0.60) 7.6 (6.8Ð8.4) the 95% CL of the slope were not different than unity d-Propoxyphene bThe effects of naltrexone were Antinociceptive effects 7.3 (4.7Ð9.9) 1.1 (3.9 to +1.6) 7.4 (6.7Ð8.0) not dose-dependent, so values Rate effectsb ÐÐ Ð were not calculated 50

β Fig. 3 -FNA antagonism of the antinociceptive and rate-decreas- through µ opioid receptors. The apparent pA2 values of ing effects of morphine, dezocine, and d-propoxyphene. β-FNA 7.3Ð8.5 for naltrexone are similar to values obtained for was administered 24 h prior to the redetermination of the agonist dose-response curves. Ordinate, upper panels: Control dose-re- naltrexone in other studies using behavioral preparations sponse curves (●) are the average of one to two experiments in such as tail-withdrawal (7.2Ð8.4), drug discrimination seven rats. Other points represent the mean of one observation in (7.3Ð7.8), and scheduled-controlled responding (7.1Ð7.9) five to seven (5.0 mg/kg β-FNA) or three or four (10 mg/kg β- (Young et al. 1992; Walker et al. 1994; Pitts et al. 1996). FNA) rats. Ordinate, lower panels: Control dose-response curves are the average of two experiments in five rats. Other points re- It is important to note, however, that the Schild re- present the mean of one observation in five (morphine and dezo- gression slope for naltrexone as an antagonist of the anti- cine) or three to four (d-propoxyphene) rats. Other details as in nociceptive effects of dezocine was significantly differ- Fig. 1. ▼ 5 mg/kg, ■ 10 mg/kg β-FNA ent than unity. When examining the Schild plot for nal- trexone in combination with dezocine, the shallow slope of this regression is predominantly due to a smaller than Discussion predicted shift in the dezocine dose-response curve after the high dose of naltrexone. This observation suggests In the present series of experiments, the receptor selec- that additional receptors may be involved in dezocine’s tivity of dezocine and d-propoxyphene was characterized antinociceptive effects or that a non-equilibrium steady by comparing in vivo pA2 values obtained for dezocine state exists between naltrexone, dezocine, and presum- and d-propoxyphene to those obtained for morphine in ably the µ opioid receptors (cf. Kenakin 1997). Previous two behavioral preparations. Naltrexone was approxi- studies using the same assay but conducted in Sprague- mately equipotent as an antagonist of the antinociceptive Dawley rats revealed a similar apparent pA2 value for and rate-decreasing effects of morphine and dezocine naltrexone in combination with dezocine to that value and of the antinociceptive effects of d-propoxyphene, obtained in the present study with Long Evans rats. In suggesting that these produce these effects the study with Sprague-Dawley rats, however, the slope

Table 2 ED50 values (95% CL) β β for opioid agonists alone and Agonist Control +5.0 mg/kg -FNA +10 mg/kg -FNA after 24 h pretreatment with β- FNA. Values are in mg/kg Morphine Antinociceptive effects 1.4 (1.1Ð1.7) 7.6 (5.2Ð11)* 28 (13Ð62)* Rate effects 4.2 (0.89Ð20) − 27 (8.9Ð84)* Dezocine Antinociceptive effects 0.31 (0.16Ð0.58) 1.9 (0.29Ð13)* 28 (12Ð64)* Rate effects 0.48 (0.13Ð1.8) 2.3 (0.25Ð20)* 1.9 (0.29Ð12) d-Propoxyphene *ED50 values significantly dif- Antinociceptive effects 10 (8.2Ð13) 23 (10Ð54)* 58 (11Ð303) ferent than control values Rate effects 20 (2.5Ð163) 25 (7.0Ð92) 33 (1.0Ð1064) (P<0.05) 51 of the Schild regression included unity (E. A. Walker, nor naltrexone blocked the rate-decreasing ef- unpublished observations). It is unlikely that the devia- fects of dezocine or d-propoxyphene in pigeons (Leander tion from unity for the Schild regression for naltrexone 1979; Picker 1997); however, in the present study, nal- and dezocine is due to species differences. A more plau- trexone clearly antagonized the rate-decreasing effects of sible explanation probably lies in the inherent variability dezocine in Long-Evans rats. Similarly, the rate-decreas- associated with quantifying the antagonism of the behav- ing effects of usually are not antagonized by ioral effects of low efficacy agonists (e.g., Koek and opioid antagonists in Sprague-Dawley rats or in pigeons Woods 1989). (Walker and Young 1993; Gerak and France 1996; Pick- The observation that naltrexone failed to antagonize er et al. 1996); however, naltrexone has been shown to the rate-decreasing effects of d-propoxyphene in a dose- antagonize the rate-decreasing effects of nalbuphine in dependent manner suggests that either other receptors Long-Evans rats (Pitts et al. 1996). Perhaps the relevant are involved in this behavioral effect or that the toxicity difference between these experiments is not the species of d-propoxyphene may have limited the degree of an- or strains per se but rather the potency of the agonist in a tagonism observed. Clearly, d-propoxyphene has affinity given species or a given type of behavioral assay. For ex- for other opioid receptors (Neil 1984; Chen et al. 1993; ample, the doses of dezocine that decreased response Yoburn et al. 1995), suggesting the rate-decreasing ef- rates in the present study (1 mg/kg) are slightly lower fects may be mediated through a combination of opioid than those that have been shown to decrease response and perhaps non-opioid mechanisms. The toxicity of d- rates in a pigeon drug discrimination procedure (3.2 and propoxyphene also may have limited the degree of an- 10 mg/kg) (Picker 1997). Other studies indicated that tagonism produced by naltrexone. This compound occa- nalbuphine is far more potent when responding was sionally produced lesions at the site of injection, which maintained under an FR30 schedule of food delivery may be a result of the local anesthetic properties of d- (Pitts et al. 1996) than when responding is maintained propoxyphene (Nickander et al. 1977). Also, the LD50 under an FR20 schedule in the same strain of rats (M. A. for d-propoxyphene in rats was within the range of a Smith, personal communication). Large differences in control d-propoxyphene dose-response curve (Emmerson potency such as these may determine the extent to which et al. 1971), resulting in death for 30% of the scheduled- an opioid antagonist can alter the rate-decreasing effects controlled responding rats approximately 12 h after test- of opioids, especially those that may have multiple opio- ing. Therefore, a modified-groups design was adopted id and nonopioid mechanisms of action at very high and the number of exposures to d-propoxyphene was re- doses. duced for an individual subject. Interestingly, while nal- The results from the β-FNA antagonism experiments trexone and β-FNA failed to dose-dependently block the paralleled the results obtained in the naltrexone antago- rate-decreasing effects of d-propoxyphene, these antago- nism experiments in a number of ways. First, β-FNA an- nists blocked the lethal effects of intermediate and high tagonized both the antinociceptive and rate-decreasing doses of d-propoxyphene. Other studies have demon- effects of morphine and dezocine as well as the antinoci- strated that naloxone pretreatment or morphine chronic ceptive effects of d-propoxyphene, further supporting the treatment reversed the acute lethality produced by d-pro- hypothesis that these effects are mediated through µ opi- poxyphene (Nickander et al. 1977). Pentobarbital but not oid receptors. Second, the pattern of antagonism pro- naloxone prevented convulsions induced by d-propoxy- duced by β-FNA in blocking d-propoxyphene’s rate-de- phene in rodents without altering the threshold dose for creasing effects was similar to the pattern observed in lethality (Fiut et al. 1966). These data, as well as the ob- the naltrexone experiments. That is, whereas a low dose servations from the present experiments, suggest that the of β-FNA (5.0 mg/kg) blocked the rate-decreasing ef- acute lethal effects but not the convulsant effects of d- fects of d-propoxyphene, a higher dose of β-FNA (10 propoxyphene may be mediated by opioid receptors in mg/kg) was ineffective. An examination of data from in- rodents. Other opioid effects of d-propoxyphene in hu- dividual rats reveals that the dose of 5.0 mg/kg β-FNA mans or animals that are reversed by opioid antagonists produced a modest antagonism of d-propoxyphene in all include sedation, nausea, vomiting, constipation, hypo- rats. In contrast, the 10 mg/kg dose of β-FNA produced tension, and respiratory depression. The other effects of complete antagonism in two rats and no antagonism in d-propoxyphene in humans or animals that are not re- another two rats. This variability in the susceptibility of versed by opioid antagonists, and therefore most likely individual rats to β-FNA antagonism was not observed non-opioid effects, include tremors, central nervous with dezocine or morphine. system stimulation, and cardiac depression (Nickander et Since higher doses of β-FNA (40 mg/kg) have been al. 1984). shown to block other effects of d-propoxyphene (Hayes A number of studies have shown that the rate-de- et al. 1987b), it is possible that larger doses of β-FNA creasing effects of d-propoxyphene or lower efficacy would have completely antagonized the rate-decreasing opioid agonists are not blocked by opioid antagonists effects of d-propoxyphene in all of the rats. If higher (e.g., Leander and McMillan 1977; Izenwasser et al. doses of β-FNA were tested in the present experiments 1996; Picker 1997). Whether an opioid antagonist will and these doses blocked the effects of d-propoxyphene, block the rate-decreasing effects of an agonist may de- d-propoxyphene would be considered a higher efficacy pend on species or strain tested. For example, neither agonist than morphine or dezocine. However, definitive 52 conclusions about d-propoxyphene’s relative efficacy given effect than this scheduled-controlled responding as- must be postponed because the results regarding the re- say. Therefore, agonists, especially lower efficacy ago- ceptor selectivity of d-propoxyphene were not straight- nists, might be more susceptible to β-FNA antagonism in forward. Nevertheless, d-propoxyphene is probably not the tail-withdrawal assay than the scheduled-controlled re- higher efficacy than morphine because d-propoxyphene sponding assay under the present schedule conditions. is a poor substitute for morphine in addicts and pro- In conclusion, the opioid antagonists naltrexone and β- nounced abstinence is not observed after chronic admin- FNA dose-dependently antagonized the antinociceptive istration of d-propoxyphene (Fraser and Isbell 1960). effects of morphine, dezocine and d-propoxyphene and Investigators have postulated that an insurmountable the rate-decreasing effects of morphine and dezocine. antagonist will produce a downward shift in the dose-re- These data suggest that these effects of morphine, dezo- sponse curve for a low efficacy agonist at doses lower cine and d-propoxyphene are mediated through µ opioid than those that produce downward shifts in the dose-re- receptors whereas the rate-decreasing effects of d-pro- sponse curves of high efficacy agonists (Adams et al. poxyphene are probably not mediated through µ opioid 1990; Burke et al. 1994). This hypothesis is based on the receptors. Furthermore, the observation that β-FNA pro- notion that low efficacy agonists require a larger avail- duced a greater antagonism of the behavioral effects of able fraction of the receptor population to produce a dezocine than morphine or d-propoxyphene suggests that maximum effect (Sosnowski and Yaksh 1990). There- dezocine is a lower efficacy agonist than morphine. Firm fore, lower efficacy agonists should be more susceptible conclusions regarding the relative efficacy of d-propoxy- to decreases in the receptor population produced by β- phene compared to morphine and dezocine in these two FNA than higher efficacy agonists. For example, in a behavioral procedures cannot be made at the present time similar scheduled-controlled responding experiment, 5.0 due to the variability across the procedures in the effec- mg/kg β-FNA altered the slope of the dose-response tiveness of naltrexone and β-FNA as antagonists. curve for but not or ,sug- gesting nalorphine is a lower efficacy agonist than butor- Acknowledgements The authors would like to thank Christopher phanol or fentanyl (Pitts et al. 1996). In the present ex- Matthewson for his technical assistance. This work was supported β by United States Public Health Service Grants DA02749 (L. A. periment, however, -FNA failed to produce any signifi- D.), DA07947 (E. A. W.), and DA10277 (M. J. P.). L. A. D. was cant downward shifts or any significant changes in slope supported by Research Scientist Award DA00033. in the agonist dose-response curves, suggesting that the efficacy differences among these agonists cannot be de- tected under these experimental conditions. In the same References tail-withdrawal procedure as that used in the present ex- β Adams JU, Paronis CA, Holtzman SG (1990) Assessment of rela- periment, pretreatment with 5.0 mg/kg -FNA in tive intrinsic activity of mu-opioid in vivo by using Sprague-Dawley rats produced both a rightward and beta-funaltrexamine. J Pharmacol Exp Ther 255:1027Ð1032 downward shift in the dezocine dose-response curve. Arunlakshana O, Schild H (1959) Some quantitative uses of drug This same dose of β-FNA was ineffective in antagoniz- antagonists. Br J Pharmacol 14:48Ð58 ing the effects of a number of other opioid agonists (et- Burke TF, Woods JH, Lewis JW, Medzihradsky F (1994) Irrevers- ible opioid antagonist effects of on opioid anal- orphine, l-, morphine, and buprenorphine), gesia and mu receptor binding in mice. J Pharmacol Exp Ther further supporting the notion that dezocine is a lower ef- 271:715Ð721 ficacy agonist than these agonists (Tiano et al. 1998). Butelman ER, Negus SS, Lewis JW, Woods JH (1996) Clo- Other experiments have demonstrated efficacy differ- cinnamox antagonism of opioid suppression of schedule-con- trolled responding in rhesus monkeys. Psychopharmacology ences between dezocine and morphine in both humans 123:320Ð324 and animals (Young et al. 1984; O’Brien and Benfield Chen JC, Smith ER, Cahill M, Cohen R, Fishman JB (1993) The 1989; Strain et al. 1996). These data illustrate the impor- opioid receptor binding of dezocine, morphine, fentanyl, bu- tance of examining a series of agonists in combination torphanol and nalbuphine. Life Sci 52:389Ð396 with an insurmountable antagonist in a variety of behav- Dykstra LA, Bertalmio AJ, Woods JH (1988) Discriminative and effects of mu and kappa opioids: in vivo pA2 analy- ioral preparations in order to describe the rank order of sis. Psychopharmacol Ser 4:107Ð121 agonist relative efficacy. Emmerson JL, Gibson WR, Harris PN, Todd GC, Pierce EC, An- The degree of antagonism produced by equal doses of derson RC (1971) Short-term toxicity of propoxyphene salts in β-FNA in the tail-withdrawal assay was greater than in the rats and dogs. Toxicol Appl Pharmacol 19:452Ð470 Fiut RE, Picchioni AC, Chin L (1966) Antagonism of the convul- schedule-controlled responding assay. For example, a dose sive and lethal effects induced by propoxyphene. J Pharmacol of 10 mg/kg β-FNA produced a 3- and 20-fold greater an- Sci 55:1085Ð1087 tagonism of the antinociceptive effects of morphine and Fraser HF, Isbell H (1960) Pharmacology and addiction liability of dezocine, respectively, than of the rate-decreasing effects. dl- and d-propoxyphene. Bull Narcot 12:9Ð14 Gal TJ, DiFazio CA (1984) Ventilatory and analgesic effects of Similarly, other studies have shown that the antinocicep- dezocine in humans. Anesthesiology 61:716Ð722 tive effects of opioid agonists are blocked by lower doses Gerak LR, France CP (1996) Changes in sensitivity to the rate-de- of an insurmountable agonist such as β-FNA or clo- creasing effects of opioids in pigeons treated acutely or chron- cinnamox than are the rate-decreasing effects (Butelman ically with nalbuphine. Behav Pharmacol 7:437Ð447 Hayes AG, Sheehan MJ, Tyers MB (1987a) Differential sensitivity et al. 1996; Zernig et al. 1994; Walker et al. 1996, 1998). of models of antinociception in the rat, mouse and guinea-pig These observations suggest that the tail-withdrawal assay to mu- and kappa-opioid receptor agonists. Br J Pharmacol may require more activated functional µ receptors for a 91:823Ð832 53 Hayes AG, Skingle M, Tyers MB (1987b) Evaluation of the recep- Picker MJ, Benyas S, Horwitz JA, Thompson K, Mathewson, tor selectivities of opioid drugs by investigating the block of Smith MA (1996) Discriminative stimulus effects of butorpha- their effect on urine output by beta-funaltrexamine. J Pharma- nol: influence of training dose on the substitution patterns pro- col Exp Ther 240:984Ð988 duced by mu, kappa and delta opioid agonists. J Pharmacol Izenwasser S, Newman AH, Cox BM, Katz JL (1996) The co- Exp Ther 279:1130Ð1141 caine-like behavioral effects of meperidine are mediated by Pitts RC, West JP, Morgan D, Dykstra LA, Picker MJ (1996a) activity at the dopamine transporter. Eur J Pharmacol 297: Opioids and rate of positively reinforced behavior: differential 9Ð17 antagonism by naltrexone. Behav Pharmacol 7:205Ð215 Kenakin TP (1997) Pharmacologic analysis of drug-receptor inter- Pitts RC, West JP, Hapke DM, Morgan D, Dykstra LA, Picker MJ actions, 3rd edn. Lippincott-Raven, Philadelphia (1996b) Opioids and rate of positively reinforced behavior: II. Koek W, Woods JH (1989) Partial generalization in pigeons Antagonism by β-funaltrexamine. Exp Clin Psychopharmacol trained to discriminate morphine from saline: applications of 4:389Ð395 receptor theory. Drug Dev Res 16:169Ð181 Romagnoli A, Keats AS (1984) Ceiling respiratory depression by Leander JD (1979) Effects of propoxyphene, , and dezocine. Clin Pharmacol Ther 35:367Ð373 azabicyclane on schedule-controlled responding: attenuation Rowlingson JC, Moscicki JC, DiFazio CA (1983) Anesthetic po- by pentobarbital but not naloxone. Psychopharmacology 66: tency of dezocine and its interaction with morphine in rats. 19Ð22 Anesth Analg 62:899Ð902 Leander JD, McMillan DE (1977) Meperidine effects on schedule- Schaefer GJ, Holtzman SG (1981) Morphine-like stimulus effects controlled responding. J Pharmacol Exp Ther 201:434Ð443 in the monkey: opioids with antagonist properties. Pharmacol Lehman T, Peterson GR (1978) Naloxone-reversible analgesic ac- Biochem Behav 14:241Ð245 tion of SKF 525-A in mice. Psychopharmacology 59:305Ð308 Shannon HE, Holtzman SG (1977) Further evaluation of the dis- Malis JL, Rosenthale ME, Gluckman MI (1975) Animal pharma- criminative effects of morphine in the rat. J Pharmacol Exp cology of Wy-16,225, a new analgesic agent. J Pharmacol Exp Ther 201:55Ð66 Ther 194:488Ð498 Sosnowski M, Yaksh TL (1990) Differential cross-tolerance be- Martin WR, Eades CG, Thompson JA, Huppler RE, Gilbert PE tween intrathecal morphine and in the rat. Anesthe- (1976) The effects of morphine and nalorphine-like drugs in siology 73:1141Ð1147 the nondependent and morphine-dependent chronic spinal dog. Strain EC, Preston KL, Liebson IA, Bigelow GE (1996) Opioid J Pharmacol Exp Ther 197:517Ð532 antagonist effects of dezocine in opioid-dependent humans. Morgan D, Picker MJ (1998) The µ opioid irreversible antagonist Clin Pharmacol Ther 60:206Ð217 beta-funaltrexamine differentiates the discriminative stimulus Takemori AE (1974) Determination of pharmacological constants: effects of opioids with high and low efficacy at the µ opioid use of antagonists to characterize analgesic receptors. receptor. Psychopharmacology 140:20Ð28 In: Braude M, Harris L, May E, Smith J, Villarreal J (eds) Negus SS, Burke TF, Medzihradsky F, Woods JH (1993) Effects Narcotic antagonists. Raven Press, New York, pp 335Ð344 of opioid agonists selective for mu, kappa and delta opioid re- Tallarida RJ, Murray RB (1987) Manual of pharmacological cal- ceptors on schedule-controlled responding in rhesus monkeys: culations with computer programs. Springer, New York antagonism by quadazocine. J Pharmacol Exp Ther 267:896Ð Tiano MJ, Walker EA, Dykstra LA (1998) Cross-tolerance to et- 903 orphine differentiates µ-opioid agonists in a rat tail withdrawal Neil A (1982) Morphine- and methadone-tolerant mice differ in assay. Analgesia 3:251Ð257 cross-tolerance to other . Heterogeneity in opioid Walker EA, Young AM (1993) Discriminative-stimulus effects of mechanisms indicated. Naunyn-Schmiedeberg’s Arch Pharma- the low efficacy mu agonist nalbuphine. J Pharmacol Exp Ther col 320:50Ð53 267:322Ð330 Neil A (1984) Affinities of some common opioid analgesics to- Walker EA, Makhay MM, House JD, Young AM (1994) In vivo ward four binding sites in mouse brain. Naunyn-Schmiede- apparent pA2 analysis for naltrexone antagonism of discrimi- berg’s Arch Pharmacol 328:24Ð29 native stimulus and analgesic effects of opiate agonists in rats. Neil A, Terenius L (1981) d-Propoxyphene acts differently from J Pharmacol Exp Ther 271:959Ð968 morphine on opioid receptor-effector mechanisms. Eur J Phar- Walker EA, Richardson TM, Young AM (1996) In vivo apparent macol 69:33Ð39 pA2 analysis in rats treated with either clocinnamox or mor- Nickander RC, Smits SE, Steinberg MI (1977) Propoxyphene and phine. Psychopharmacology 125:113Ð119 : pharmacologic and toxic effects in animals. Walker EA, Zernig G, Young AM (1998) In vivo apparent affinity J Pharmacol Exp Ther 200:245Ð254 and efficacy estimates for mu opiates in a rat tail-withdrawal Nickander RC, Emmerson JL, Hynes MD, Steinberg MI, Sullivan assay. Psychopharmacology 136:15Ð23 HR (1984) Pharmacologic and toxic effects in animals of dex- Yoburn BC, Shah S, Chan K, Duttaroy A, Davis T (1995) Super- tropropoxyphene and its major metabolite norpropoxyphene: a sensitivity to opioid analgesics following chronic opioid an- review. Hum Toxicol 3:13SÐ36S tagonist treatment: relationship to receptor selectivity. Pharma- O’Brien JJ, Benfield P (1989) Dezocine. A preliminary review of col Biochem Behav 51:535Ð539 its pharmacodynamic and pharmacokinetic properties, and Young AM, Stephens KR, Hein DW, Woods JH (1984) Reinforc- therapeutic efficacy. Drugs 38:226Ð248 ing and discriminative stimulus properties of mixed agonist- Overton DA, Batta SK (1979) Investigation of and anti- antagonist opioids. J Pharmacol Exp Ther 229:118Ð126 tussives using drug discrimination techniques. J Pharmacol Young AM, Masaki MA, Geula C (1992) Discriminative stimulus Exp Ther 211:401Ð408 effects of morphine: effects of training dose on agonist and an- Paronis CA, Holtzman SG (1992) Development of tolerance to the tagonist effects of mu opioids. J Pharmacol Exp Ther 261: analgesic activity of mu agonists after continuous infusion of 246Ð257 morphine, meperidine or fentanyl in rats. J Pharmacol Exp Zacny JP, Lichtor JL, de Wit H (1992) Subjective, behavioral, and Ther 262:1Ð9 physiologic responses to intravenous dezocine in healthy vol- Picker MJ (1997) Discriminative stimulus effects of the mixed- unteers. Anesth Analg 74:523Ð530 opioid agonist/antagonist dezocine: cross-substitution by mu Zernig G, Butelman ER, Lewis JW, Walker EA, Woods JH (1994) and delta opioid agonists. J Pharmacol Exp Ther 283:1009Ð In vivo determination of mu opioid receptor turnover in rhesus 1017 monkeys after irreversible blockade with clocinnamox. J Phar- Picker MJ, Yarbrough J, Hughes CE, Smith MA, Morgan D, macol Exp Ther 269:57Ð65 Dykstra LA (1993) Agonist and antagonist effects of mixed Zimmerman DM, Leander JD, Reel JK, Hynes MD (1987) Use of action opioids in the pigeon drug discrimination procedure: in- β-funaltrexamine to determine mu opioid receptor involve- fluence of training dose, intrinsic efficacy and interanimal dif- ment in the analgesic activity of various opioid ligands. J ferences. J Pharmacol Exp Ther 266:756Ð767 Pharmacol Exp Ther 241:374Ð378