Buprenorphine-induced alterations of cocaine's reinforcing effects in rhesus monkey: a dose-response analysis

DEPENDENCE ELSEVIER

Drug and Alcohol Dependence 40 (1995) 87-98

Buprenorphine-induced alterations of cocaine’s reinforcing effects in rhesus monkey: a dose-response analysis Scott E. Lukas*,

Nancy K. Mello,

John M. Drieze, Jack H. Mendelson

Alcohol and Drug Abuse Research Center, Harvard Medical School-McLean

Hospiral, 11.5 &fill Street, Belmont, MA 02178, uSA

Received 13 November I994 accepted 19 September 1995

AbSlTPCt Buprenorphine reduces cocaine self-administration by rhesus monkeys, opiate- and cocaine-dependent men and polydrug abusers,but the mechanismsunderlying thesecocaine-opiate interactions are not well understood. In the present study, the effects of daily placebo or buprenorphine (0.1,0.3 and 1.0 mg/kg) treatment on cocaine self-administration (0.001-0.3 mg/kg/inject) were examined in five cocaine-experiencedrhesus monkeys. Saline and each of six cocaine doseswere available in an irregular order. Responding for cocaine (or saline) and food was maintained on a second order FR4 (VR 165) schedule of reinforcement. During placebo treatment, the daily number of cocaine injections increased as the unit dose was increased and then decreasedat higher doses.Cocaine dosesthat maintained the highest rates of responding during placebo treatment were more resistant to buprenorphine’s effects. The typical increase in response rate during the first five cocaine injections of a session also was attenuated by buprenorphine. The ascending limb of the cocaine dose-responsecurve was shifted downward and approximately one log unit to the right during low-dose buprenorphine treatment (0.1 mg/kg/day). In contrast, individual responserates for food pellets were unaffected. We conclude that buprenorphine selectively decreasesself-administration of some unit dosesof cocaine at dosesthat have minimal effects on food-maintained responding. Keywords: Buprenorphine; Cocaine; Drug abuse; Drug self-administration; Rhesus monkey; Schedule-controlled behavior; Drug abuse treatment

1. Introduction

Buprenorphine, an opioid analgesic with partial Jo agonist and K antagonist properties (Lewis, 1974; Lewis et al., 1983; Dykstra et al., 1987; Negus and Dykstra, 1988; Negus et al., 1989) reduces cocaine selfadministration by rhesus monkeys and human polydrug abusers (for review see Mello and Mendelson, 1995; Mello et al., 1993c).In rhesus monkeys, daily buprenorphine treatment for 15 days at doses between 0.237 to 0.70 mg/kg significantly reduced cocaine selfadministration (0.05 mg/kg/inject) by 72 to 93% below placebo treatment baseline levels (Mello et al., 1989, Abbreviations:

CPP, conditioned place preference; FI, fixed

interval, FR, fixed ratio. l

Corresponding author, Tel.: (617) 855 2767; Fax: (617) 855 3711.

1990). When the duration of treatment was extended from 30 to 120days, buprenorphine continued to reduce cocaine self-administratioti by 70 to 94% (Mello et al., 1992).Buprenorphine’s ef&cts were selectivefor cocaine self-administration since food-maintained responding was initially reduced, but returned to or remained slightly below control levels suggesting that tolerance develops to buprenorphine’s suppressive effects on food-maintained responding (Lukas et al., 1988; Mello et al., 1989, 1990, 1992; for review see Mello and Mendelson, 1992). These reports that buprenorphine reduces cocaine self-administration have recently been replicated in rhesus monkeys (Carroll et al., 1992; Winger et al., 1992) and rats (Carroll and Lac, 1992). Clinical trials of buprenorphine in men who fulfill DSM-III-R criteria for both cocaine and opiate dependence confirm findings in animal drug self-adminis-

0376-8716/95/$09.50 0 1995 Elscvier Science Ireland Ltd. All rights reserved SSDI 0376-8716(95)01195-5

88

S.E. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

tration models that buprenorphine reduces cocaine intake in human subjects(Mendelson et al., 1991;Mello and Mendelson, 1992, 1995; Gastfriend et al., 1993; Schottenfeld et al., 1993). Similarly encouraging results were not observed during methadone treatment suggesting that buprenorphine may be more effective than methadone in reducing cocaine abuse by polydrug abusers(Kosten et al., 1989a,b).Preliminary data from a double-blind comparison of buprenorphine (8 mg/day) and methadone (60 mg/day) treatment of opiatedependent patients found no differences in occasional cocaine use, but these patients did not meet DSM-III-R criteria for cocaine dependence (Fudala et al., 1991). Recently, Foltin and Fischman (1994) reported that buprenorphine pretreatment reduced choice of higher cocaine dosesin healthy volunteers in a controlled clinical study. Buprenorphine also reduces heroin selfadministration by opiate-dependent men (Mello and Mendelson, 1980; Mello et al., 1982) and is effective in the outpatient treatment of opiate-dependent patients (Kosten et al., 1989a,b; Rosen and Kosten, 1991; Johnson et al., 1992). Thus, the apparent successof buprenorphine in reducing cocaine abuse by opiatedependent persons represents a significant advance in drug abuse treatment. Currently, the mechanisms by which buprenorphine reduces cocaine-administration are unknown. Several possibilities have been suggestedincluding blockade of cocaine-reinforcing effects; enhancement of cocaine’s reinforcing-effects; substitution for cocaine; alteration of the metabolic disposition of cocaine; aversive effects when combined with cocaine or a non-specific suppression of behavior (Carroll and Lac, 1992; seeMello and Mendelson, 1995 for review). It is also possible that somecombination of the above accounts for buprenorphine’s effects on cocaine self-administration. The present study was designed to examine the effects of buprenorphine on the cocaine dose-effect curve and to more completely characterize buprenorphine/cocaine interactions. The effects of chronic treatment with a range of dosesof buprenorphine (0. l- 1.Omg/kg/day) on responding maintained by several doses of cocaine (0.001-0.3 mg/kg/inject) and 1 g food pellets were examined in rhesus monkeys. 2. Methods

2.1. Subjects Three male and two female rhesus monkeys (Macacu muiatta) weighing between 7.7 to 13.5 kg served as subjects. Subjects had a history of i.v. cocaine selfadministration that averaged304.2 f 197.2days (range 24 to 1101days) at the beginning of this study. All monkeys except 12918and 497T had exposure to buprenorphine before the present study began. Monkeys were maintained at ad libitum weight and given multiple vita-

mins, fresh fruit, vegetablesand Purina@monkey chow to supplement a banana pellet diet. Water was continuously available. A 12-h light/dark cycle was in effect with lights on from 0700- 1900h. The lights were turned off during food and drug self-administration sessions. Monkeys were surgically implanted with double lumen Silicone@rubber catheters (I.D. 0.7 mm, O.D. 2.0 mm) to facilitate buprenorphine or placebo pretreatment and concurrent cocaine self-administration. Catheters were implanted in the jugular or femoral vein and exited in the mid scapular region. All surgical procedures were performed under strict aseptic conditions using pentobarbital (30 mgkg, i.v.) anesthesia. After surgery, monkeys were given 200 000 Units of Cornbiotic@ (Dilhydrostreptomycin and Penicillin G) i.m. every other day for a total of 5 injections. The intravenous catheter was protected by a tether system consisting of a custom-fitted nylon vest connected to a flexible stainless steel cable and fluid swivel (Spaulding Medical Products, Arroyo Grande, CA). This flexible tether systempermits monkeys to move freely. Catheter patency was verified periodically by administering a short-acting barbiturate, methohexital sodium (3 mg/kg), through the catheter lumen. A sleep onset time of lessthan 15 s after injection was considered a positive indication of catheter patency. Animal maintenance and research was conducted in accordancewith the guidelines provided by the Committee on Laboratory Animal Resources. The facility is licensedby the U.S. Department of Agriculture and protocols were approved by the Institutional Animal Care and Use Committee. 2.2. Procedures Monkeys worked at an operant task for food and intravenous cocaine injections in a well-ventilated chamber equipped with an operant panel, a pellet feeder and a water dispenser.Drug injections were delivered by a syringe pump in a single pulse which dispensed0.1 ml of fluid over 0.9 s. The operation of the syringe pump (Model 981210,Harvard Apparatus, Inc., South Natick, MA) was audible to the monkey. Schedulesof reinforcement were programmed by custom-designed Software and run on Apple IIe microcomputers. Cocaine self-administration training was initiated at a unit doseof 0.05 mg/kg/inject. If high responserates and high number of injections per sessionwere not maintained then the unit dose was increased to 0.1 mg/kg/inject. These cocaine unit doses were identical to those previously used in our studies of daily buprenorphine treatment (Mello et al., 1989, 1990, 1992, 1993a).In the present study, three monkeys self-administered 0.05 mg/kg/inject of cocaine and two monkeys selfadministered 0.1 m&/inject during training. Once training had been completed, food (1 g banana pellet) and cocaine (0.05 or 0.1 mg/kg/inject) self-

S.E. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

89

sessionsand all monkeys returned to pre-treatment levels of responding for cocaine. One monkey (606.5) repeated the entire cocaine dose-responsecurve; it was identical to that in the present study.

administration were maintained on a second-order scheduleof reinforcement (FR4 [VR 1651). An average of 16responseson a variable ratio schedule(VR 16) produced a brief red or green stimulus light (S+) located at eye level on the operant panel. The red light was associated with cocaine and green with food. However, a drug injection or a food pellet was delivered only after a fixed ratio of four (FR 4) of the VR 16 responserequirements had been completed, i.e. each food pellet or drug injection required an average of 64 responses. Each experimental day consisted of four food and four drug availability sessions. Food sessions were signaled by transilluminating the response key with a colored light. Food sessionsbegan at 1100 h, 1500 h, 1900h and 0700 h each day. Drug sessionswere signaled by transilluminating the response key with a different color. Drug sessionsbegan 1 h later at 1200 h, 1600 h, 2000 h and 0800 h. Each food or drug sessionlasted for 1 h or until 25 food pellets or 20 drug injections were delivered. Cocaine injections were limited to 20 per session to minimize the possibility of adverse drug effects.

In order to simulate the absorption profile of an i.m. or S.C.injection, buprenorphine or an equal volume of placebo control solution was administered each day as a slow injection at a rate of 1 ml of solution every 10min over a 50 min interval and flushed through with sterile saline in a volume which exceededthe estimated catheter dead space.This injection began at 0930 h each day. The buprenorphine administration procedure was identical to that usedin each of our previous studies (Mello et al., 1989, 1990, 1992, 1993a,c). Buprenorphine was administered through the second lumen of a double lumen catheter unlessone lumen becameblocked. In that case, buprenorphine was administered through the same lumen as cocaine and flushed through with saline in a volume that exceededthe catheter dead space.

2.3. Dose-response curve determinations

2.5. Drug solution preparation

Once the monkeys’ responding stabilized on the second order schedule, cocaine dose-responsecurves were determined. Monkeys were given accessto each cocaine dose or to saline for at least 5 days (20 sessions),but not more than 10 days. Monkeys were studied at each cocaine unit until stable responding was achieved. Stable responding was defined as 3 consecutive days during which the number of drug injections each day did not differ by more than 20% from the average of those 3 days and there was no upward or downward trend. If fewer than 15 cocaine injections per day were selfadministered for 5 consecutive days, then the data from the last 3 days were used in the analysis and the next cocaine dose was studied. The cocaine dose-responsecurves were first established during daily placebo (physiologic saline) treatment. Subsequently, the effects of daily buprenorphine treatment (0.1, 0.3 and 1.0 mg/kg) on the cocaine doseresponsecurve were studied under identical conditions. Each buprenorphine treatment condition was separated by at least 10 days. Although buprenorphine’s half-life is quite long, responding for cocaine returned to control levels by the third day after termination of chronic treatment (Mello et al., 1993a).Cocaine doseswere presented sequentially without intervening intervals of training dose availability. The order of presentation of the unit doses of cocaine (0.001, 0.003, 0.01, 0.03, 0.1, and 0.3 mg/kg/inject) and saline was irregular and varied across monkeys. However, the order of cocaine doses and saline were consistent within a monkey across placebo and buprenorphine treatment conditions. Several cocaine unit doses were studied again in four of the live monkeys 3-6 months after buprenorphine treatment

All drugs were administered i.v. on a mg/kg basis and were prepared as stock solutions which were filtersterilized using an 0.22 pm Millipore filter and stored in sterile, pyrogen-free vials. Stock solutions were diluted, as necessary,to provide the neededconcentration for a particular monkey. 2.5.1. Buprenorphine. Buprenorphine hydrochloride was obtained from the National Institute on Drug Abuse (NIDA) and was dissolved in sterile water at a concentration of 8.0 or 16.7 mg/ml. Solutions were checked daily to ensure that no precipitate had formed. Fresh solutions were prepared every 14 to 21 days. Doses were calculated on the basis of the monkeys’ weights so that a final dilution of stock solution (with Sterile Saline for Injection, USP) resulted in a daily injection of each buprenorphine dose in a volume of 5.0 ml/injection. Placebo treatment consisted of an equal volume of sterile saline. 2.5.2. Cocaine. Cocaine hydrochloride was obtained in crystalline form from NIDA. The purity was certified by Research Triangle Institute to be greater than 98%. Cocaine was dissolved in Sterile Saline for Injection, USP, to make a stock solution of 50 mg/ml. Doses were calculated on the basis of monkeys’ weights so that a final dilution of the stock solution resulted in unit doses of 0.001 to 0.3 mg/kg/inject. Injection volume and injection duration were constant at 0.1 ml and 10 s, respectively, and the unit doses(mg/kg/inject) were altered by changing the concentration.

2.4. Daily buprenorphine or placebo treatment

2.6. Data analysis

The major dependent variables used to assessbuprenorphine’s effects on cocaine’s reinforcing effects

SE. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

90

were daily number of cocaine injections, daily rates of responding and daily number of food pellets. Response rates were calculated for each of the four daily sessions and then averaged to yield an overall daily rate. Threeday averageswere used in all measures.In an attempt to obtain a measure of buprenorphine’s initial effects on cocaine- and food-maintained behavior, we devised a new analytic procedure that quantified the response rates for the first live injections (or first live food pellets) of each session during the first 3 days of exposure to each dose of cocaine (i.e. the first 12 sessions).The response rate for the very first injection was calculated from the onset of the first response to avoid any confounds due to initial pausing. Becauseresponserates for the initial injection were the sameacross all conditions, data from this injection were included in the analysis. Restricting the analysesto the first five injections of the sessionwas designed to eliminate the possible confounding effects of cumulative cocaine doses on response rates which invariably occur during the middle and end of each cocaine self-administration session. The potential sensitivity of responsesearly in the session to the reinforcing effects of drugs was shown by Mansbach and Balster (1990) who reported that latency to the first

few injections of cocaine were much shorter than those for saline. Similar local response rate data were generated for food-maintained responding. Cocaine dose-response curves determined during placebo treatment were compared with the cocaine dose-responsecurves that were determined during treatment with various doses of buprenorphine. Changes in responserates for cocaine injections were recorded for individual monkeys and shifts in the ascending limb of the dose-responsecurves were quantified using a parallel line bioassay technique that employed linear regression models for assessingdrug effects as a function of log dose (Tallarida and Jacob, 1979). All individual data points were used for this analysis. Each curve was constructed using a minimum of three cocaine doses. Repeated measures ANOVAs were performed on the number of injections and response rates as the dependent variables in separateanalyses.To correct for the inherent correlations of repeated measurements,degrees of freedom for all F-tests were corrected using the Greenhouse-Geissermethod. When overall significant effects were detected (significance level of P < 0.05) then Dunnett’s post-hoc follow-up were performed to determine the source of the variance. All calculations

371A

12918

1 Buprenorphine Treatment : o Placebo;

go80706050403020lo01

371A

606.5

90: EO706050403020. lo-

l

0.1 mglkglday;

679C

9.

A 0.3 mglkglday;

901

0-S

n

497T

1.O mgikglday

90 80 70 60 50 40 30 20 IO

y -

o-: S

1

497T

T

--h.t I so I.001 0.01

0.1

1.0

Cocaine Unit Dose (mg/kg) Fig. I. Mean number of cocaine injections ( f SD.) per day as a function of cocaine unit dose in individual monkeys during placebo and 0. I mg/ltg/day, 0.3 mg/kg/day and I.0 m&/day buprenorphine treatment. Data are the mean values of 3 consecutivedays (12 sessions).Overall effects of buprenorphine were significant at P < 0.9091.Some of the SD. are hidden by the corresponding symbols. *Denote individual data points that are significantly different from placebo treatment at P c 0.05.

S.E. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

were conducted using Super Anova (Abacus Concepts, Berkeley, CA).

3.Results 3.1. Effects of buprenorphine on cocaine injections and food pellets Mean daily cocaine self-administration during the last 3 days of placebo and buprenorphine treatment for the five monkeys is shown in Fig. 1. During placebo treatment all monkeys took the maximal number of cocaine injections available at at least 1 unit dose, and an inverted U-shaped dose-responsecurve was evident. Monkeys could self-administer a maximum of 80 injections per day. Cocaine dosesof 0.01-0.032 mg/kg/inject maintained peak performance (75430 injections per day) in all five monkeys while the highest unit dose (0.32 mg/kg/inject) maintained moderate levels of responding (27-40 injections per day). In general, buprenorphine treatment reduced the number of daily cocaine injections, but self-administration of cocaine doses that maintained maximal responding was reduced in only one monkey (606.5). Two monkeys (371A and 12918)

Placebo

took more cocaine injections at low unit doses (0.001 and 0.003 mg/kg/inject) during buprenorphine treatment. A more detailed analysis of buprenorphine’s effects was obtained by analyzing the data acrossall animals by sessionrather than by day. The effects of various doses of buprenorphine on the number of cocaine injections or food pellets earned per sessionare shown in Fig. 2. During placebo treatment, the maximal number of cocaine injections at a unit dose of 0.032 mg/kg earned per session was 19 out of a possible 20 injections (Fig. 2, upper left panel). All three doses of buprenorphine caused a significant (P < 0.005) decreasein the number of cocaine injections earned at lower and higher cocaine unit doses (e.g. 0.001, 0.0032, 0.1 and 0.32 mg/kg/inject). These doses typically maintained only moderate responding during placebo treatment. However, the doses of cocaine that maintained the highest intake (0.01 and 0.032 mg/kg) were less affected by buprenorphine. As the buprenorphine treatment dose was increased the resultant cocaine dose-effectcurves had fewer unit doses that were self-administered above 15 injections per day. Buprenorphine did not affect saline self-administration (indicated by ‘S’ in all four top panels of Fig. 2).

Buprenorphine 0.1 mglkglday 203

91

Buprenorphine 0.3 mglkgfday

Buprenorphine 1.O mg/kg/day

201

, I I I I , 6 0.001 0.01 0.1

454035. 30. 2520- Q 15. lo5O-

454035. 30. 25. 20’ 0 15 IO5. O.5 obol

Oil

a 0:1

454035. 30. 25.

'

Cocaine Unit Dose (mg/kg/inj) Fig. 2. Group mean (f S.D.) number of cocaine injections (top panels) and food pellets (bottom panels) earned during a l-h sessionas a function of cocaine unit dose during treatment with placebo and three doses of buprenorphine. Data are from the last I2 consecutive l-h sessions.Each data point representsan average of five monkeys, except for the 0. I and I .Omg/kg dosesof buprenorphine where n = 4 and 2, respectively. *Denote individual ‘data points that are significantly different from placebo treatment at P < 0.05.

S.E. Lukas et al. /Drug and All *ohol Dependence 40 (1995) 87-98

92

During placebo treatment, the number of food pellets taken per sessioninitially increased and then decreased as the cocaine unit dose was increased(Fig. 2, lower left panel). Cocaine alone reduced the number of food pellets taken (F(6,18) = 13.68, P = O.OOOl),however, buprenorphine did not have an overall effect on food pellets (F(2,18) = 1.189,P = 0.3672). Except for a single

Buprenorphine

Dose: 0 Placebo

data point (an increase in the number of food pellets), buprenorphine alone did not significantly affect responding for food pellets under any condition. This effect is most clearly evident by comparing the responding for food during saline self-administation. The number of food pellets earned per session during saline self-administration were similar during placebo treat-

Ei 0.1 mg/kg/day

n 0.3 mg/kg/day

0.001 mg/kg/inj cocaine

0.0032 mg/kg/inj cocaine

20

20-

15

1

t l”ilii 15-

T

T

5 0

g

.= $! ai & n : -Z=

--8c -

0.01 mg/kg/inj cocaine

0.032 mg/kg/inj cocaine

2015-

lo5-

o-

Iii4 t I Iiilil!!i Iiilb1

0.1 mg/kg/inj cocaine

0.32 mg/kg/inj cocaine

20

20-

1

r

15

15-

10

lo-

5-

5

,.b

L

0 A

12-1

p.m.

Buprenorphine (9:30-i 0:20 a.m.)

4-5

p.m.

8-9 p.m.

8-9 a.m.

12-1 p.m. Buprenorphine (9:30-l 0:20 a.m.) A

4-5

p.m.

8-9 p.m.

8-9 a.m.

Individual Daily Sessions Fig. 3. Group mean ( l SD.) number of cocaine injections during each of four individual daily sessionsas a function of cocaine unit dose and buprenorphine treatment. Data are from the last 12consecutive l-h sessions.Each data point representsthe mean (+ S.D.) of four monkeys who receivedall treatments. tDenotes individual data points that are significantly different from placebo treatment within the samesessionat P < 0.05. *Denote data points that differ from the 1200-1300 h sessiondata at P < 0.05.

S.E. Lakas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

ment (22.0 f 6.63) and the 0.1, 0.3 and 1.Omg/kg/day doses of buprenorphine (22.3 f 5.4, 17.9 f 6.5 and 33.4 f 18.0, respectively). These values were not significantly different. In order to determine the time-course of buprenorphine’s effects on the number of cocaine injections, the data were further analyzed using each of the four individual daily sessions(Fig. 3). The dose by sessioninteraction during placebo treatment was statistically significant (fl6, 15) = 4.072, P = 0.011). Post-hoc analysis revealedthat only the two lowest cocaine unit doses (0.001and 0.0032 mg/kg) actually displayed the diurnal pattern of cocaine self-administration (P = O.OOl0.033). The highest number of injections was taken during the 1200-1300 h session,while the lowest number of injections was taken during the 2000-2100 h session. However, this diurnal variation was not maintained when higher cocaine unit doses were self-administered;

93

the number of cocaine injections was equally spaced among all four sessions.There was a strong trend for an interaction between buprenorphine treatment and cocaine dose (J’(2,lO) = 2.138, P = 0.052). The evening sessionwas most sensitive to buprenorphine’s effects in that a significant decreasein injections was observed during this session for four of the six cocaine doses tested and a slight recovery was evident by the next morning’s session,approximately 22 h after treatment. Buprenorphine had no effect on the highest unit dose of cocaine (0.32 mg/kg), but intake of this unit dose was already very low during placebo treatment. 3.2. Effects of buprenorphine on rates of responding

Overall rates of responding for cocaine and food were calculated by dividing the total number of responsesper 1 h session by the total time (including mid-schedule pausing) spent responding for each reinforcer. These

Buprenorphine Treatment (mg/kg/day) 0 Placebo

a 1.0

A 0.3

l 0.1

Injections Cocaine, 0.001 mg/kg/inj

Saline

Cocaine, 0.01 mg/kg/inj

3.0 ?

12

cn 3

3

4

3.0 -

2.5 y

2.5 :

2.0 1

2.0 y

1.5:

1.5:

1.0:

1.0.

0.5

0.5.

0.0 1

j 0.0 J 12

5

3

4

5

Cocaine, 0.3 mg/kg/inj

12

3

4

5

Food Pellets Cocaine, 0.001 mg/kg/inj

Saline

g

Cocaine, 0.3 mglkglinj

Cocaine, 0.01 mg/kg/inj

8

2.0:

2.01

u

1.5:

1.5

1.0.

1.0

0.5;

0.5'

2.0

*

!w

0.0:

0.0 : 12

3

4

5

12

3

4

5

12

3

4

5

Consecutive Reinforcers Fig. 4. Individual rates (mean A S.E.M.) of responding for each of the first five saline or cocaine injections (0.001,0.01 and 0.3 m&kg/inject) and food pellets during placebo and buprenorphine treatment (0.1, 0.3 and 1.0 mg/kg/day). Data are from 12 consecutive sessionsin each treatment condition. Data were obtained by averaging response rates from all of the first reinforcers obtained during the first 12 sessionsthat a particular cocaine dose was available. This process was then repeated for all of the second reinforcers, then the third, etc. Each data point is based on an averageof five monkeys, except for the 0.01 mg/kg unit dose of cocaine during the 1.0mgikg dose of buprenotphine, where n = I. *Denote individugl data points that are significantly different (P < 0.05) from the first reinforcer.

94

SE. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

data for eachmonkey were averagedover the last 3 days of exposure to the various buprenorphine or placebo treatment conditions. Overall response rates failed to reveal any significant differences in cocaine- or foodmaintained responding (data not shown). There were no differences in overall response rates maintained by cocaineor food pellets acrossthe various buprenorphine doses. The changesin responserates for each of the first five cocaine injections (0.001, 0.01 and 0.3 mg/kg/inject) and food pellets during the first 3 days of placebo and buprenorphine treatment are shown in Fig. 4. Each data point represents the average of all monkey’s response rates for the first five cocaine and food reinforcers during the first 12 sessions(3 days) of each treatment condition. Thesedata points were analyzed to determine if the very first exposure to a cocaine unit dose was altered by buprenorphine treatment. During placebo treatment (open symbols), the rate of cocaine- and foodmaintained responding increased by the second reinforcer and then either remained constant for the next three reinforcers (food) or began to decline (saline). None of the rates increased significantly during selfadministration of the 0.001 mg/kg dose of cocaine. The 0.01 mg/kg dose of cocaine maintained the highest rates of responding (Fig. 4, upper right center panel), and the rates for the second through fifth injection were significantly faster than for the first injection (P = 0.0013). The pattern of responding for the highest cocaine unit dose (0.3 mg/kg/inject) during placebo treatment was similar to saline-maintained responding in that an initial acceleration was followed by a decreasein responserate by the fourth injection. Buprenorphine treatment did not significantly alter response rates for the very first cocaine injection suggesting that it does not simply reduce responserates in a non-specific manner. Further evidence for this was seenin the food data; compared with placebo treatment, buprenorphine had no effect on the response rates for the first five food pellets (Fig. 4, bottom panels). Buprenorphine did, however, have a marked effect on responserates for the initial cocaine (0.01 mg/kg) injections. During placebo treatment, the rate of responding for 0.01 mg/kg of cocaine increased significantly by 1.4 responses/sfrom the first to third cocaine injection (P = 0.03). During buprenorphine treatment, these increasesin responserates were no longer statistically significant. Higher buprenorphine doses attenuated the rate-acceleratingeffect of cocaine even further, resulting in increasesof only 0.3 to 0.5 responses/sfrom the first injection. The responserate data for the first five cocaine injections were subjected to a parallel-line bioassay procedure to measurethe slopes of the ascending limb of the dose-responsecurves, to determine if there were potency shifts and to assesswhether the curves were parallel

Buprenorphine

Dose

I

O-1

I 0.001

0.0001

I I 0.01

I

, 0.1

Cocaine Unit Dose (mg/kg) Fig. 5. Parallel line bioassay analysis (Tallarida and Jacob, 1979)of the responserates for the first five cocaine injections. Data are from 12 consecutive sessionsduring placebo and buprenorphine treatment conditions:fRegressionlines for each treatment dose were constructed using the raw data from each monkey (n = 5) at each cocaine unit dose on the ascendinglinear portion of the dose-responsecurve. Equations for each line are shown and lines sharing the samelower caseletter (a, b, c, d) are parallel to one another (P < 0.05).

using a parallel line bioassay (Tallarida and Jacob, 1979). The maximal response rate was 2.0 responses/s. Thus, a half-maximal rate of 1.0 response/swas used to compare the individual curves. This analysis confirmed that at the half maximal responserate of 1.Oresponse/s, the 0.1 mg/kg/day dose of buprenorphine significanfly shifted the cocaine dose-responsecurve to the right by 0.976 log dose units (Fig. 5). The resultant curve during buprenorphine treatment (0.1 mg/kg/inject) remained parallel to the baseline curve obtained during placebo treatment. In contrast, the two higher doses of buprenorphine resulted in a downward shift of the cocaine dose-responsecurve and, while these two curves were parallel to one another, their slopes were significantly different from the placebo and 0.1 mg/kg/day buprenorphine treatment (P < 0.05 and 0.01, respectively). 4. Dhcussion The main findings of the present study were: (1) buprenorphine reduced cocaine intake at doses that did not appreciably alter food intake; (2) buprenorphine had a greater effect on lower cocaine dosesthan on cocaine doses that maintained the highest levels of selfadministration; (3) buprenorphine attenuated the rateaccelerating effects of cocaine early in the sessions. 4.1. Buprenorphine effects on cocaine and food selfadministration

The observeddecreasesin cocaine injections and rates

S.E. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

of responding for some cocaine doses confirms and extends our previous observations that daily buprenorphine treatment (O-237-0.7 mg/kg/day) reduces cocaine self-administration by rhesus monkeys for periods of 15 days to 4 months (Mello et al., 1989, 1990, 1992, 1993a,b,c). Single doses of buprenorphine (0.4 mg/kg) significanfly reduced cocaine self-administration for up to 48 h (Mello et al., 1993a; see Mello and Mendelson 1995for review). The daily number of cocaine injections at unit dosesof 0.01 and 0.1 mg/kg in the present study were decreasedby an averageof 64.8 and 45.9%, respectively during 0.3 mg/day buprenorphine treatment. A reduction in cocaine self-administration in animal drug self-administration models during buprenorphine treatment has also been reported by other laboratories (Carroll and Lac, 1992; Carroll et al., 1992; Winger et al., 1992). Buprenorphine also altered food-maintained responding, but the degreeof attenuation and the range of effective dose combinations was much smaller than those which altered cocaine self-administration. In the first report of buprenorphine’s effects on cocaine and food self-administration, food intake was decreasedonly 30% while cocaine intake was decreasedby up to 90% (Mello et al., 1989). It is unlikely that any potential pharmacotherapy will be completely selective for cocaine self-administration alone. However, the degree of suppression of cocaine self-administration should be evaluated against the degree of suppression of other behaviors when evaluating the utility of a potential pharmacotherapy (Mello, 1992). It is noteworthy that while higher cocaine unit dosesresulted in reduced food intake, the suppression of cocaine intake by buprenorphine did not result in a concomitant increase in the number of food pellets earned. This profile suggeststhat there is a dissociation between cocaine and food intake, at least at the higher cocaine unit doses. The downward and right shift in the full doseresponse curve (Fig. 5) indicates that the lowest and highest cocaine unit doses were more affected by buprenorphine than the intermediate cocaine unit dose (0.01 mg/kg/inject). Examination of the cocaine injection per day data (Figs. 1 and 2) and the first five response rate data (Fig. 4) suggeststhat buprenorphine decreasesthe range of reinforcing cocaine doses.During placebo treatment, three different doses of cocaine maintained high rates of responding. As the dose of buprenorphine was increased, only 1 unit dose of cocaine was reliably self-administered. Since saline and each unit doseof cocaine were presentedin an irregular order that differed among monkeys, it is unlikely that order effects contributed to the results obtained. The diurnal pattern of low dose cocaine selfadministration suggeststhat the reinforcing effects of these doses of cocaine varies during the day. This pattern of lower intake during the afternoon and evening

95

sessionshas also been previously reported in nonhuman primates for cocaine (Harrigan and Downs, 1978;Lukas et al., 1984a; Negus et al., 1995 in press) and, various PCP analogues(Lukas et al., 1984a).While this analysis was conducted to measurethe time-course of buprenorphine’s effects,the data shown in Fig. 3 suggestthat buprenorphine suppressescocaine intake during sessions in which cocaine was not self-administered at high levels. Overall, the late afternoon and early evening cocaine sessions were most sensitive to buprenorphine’s suppressiveeffects.Buprenorphine’s relatively long half-life (Lloyd-Jones et al., 1980)most likely contributed to this effect on the evening cocaine sessions.Since there was little change in the intake during the next morning session, it is impossible to determine if buprenorphine’s effects had waned. However, a prior study by our group demonstrated that buprenorphine administered once every 48 h significantly reduced cocaine selfadministration (Mello et al., 1993a). 4.2. Response rate analysis

Data from the very first injection were included in the analysis of responserates for the first five reinforcers becausethe rates were essentially the sameacross all conditions and among all animals. Buprenorphine’s attenuation of the operant response rate accelerating effects of cocaine were clearly evident early in each session (Fig. 4). During placebo treatment conditions, monkeys typically maintained response rates of 2.5 responses/sby the second or third injection of the optimal cocaine dose (0.01 mg/kg/inject); buprenorphine consistently attenuated this rate-accelerating effect of this dose of cocaine. Since responserates before the very first cocaine injection and during food-maintained responding were not significantly reduced during buprenorphine treatment, we conclude that buprenorphine does not decreasecocaine self-administration by non-specific suppression of operant responding. The apparent increase in responserates during’saline self-administration probably reflects extinction responding. However, the variability was so large that these rate increaseswere not statistically significant. If extinction responding had been more robust, then one might conclude that buprenorphine was attenuating extinction responding as well. This ‘first five’ approach to studying the reinforcing efficacy of cocaine is similar to the analysis of the distribution of inter-infusion-times (IIT) and the latency to the first-five injections of a single session used by Mansbach and Balster (1990) to compare cocaine and saline self-administration. These authors demonstrated that the mean latencies for successivesaline and 0.1 mg/kg/inject cocaine injections decreased while the latencies for injections of the 0.033 mg/kg/inject unit doseof cocaine decreased.In the present study, a similar result was obtained in that response rates for the first

96

S. E. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

five cocaine injections (0.01 mg/kg unit dose) rapidly accelerated whereas those for food only slightly increased or remained constant throughout the session (Fig. 4). Buprenorphine attenuated the cocaine-induced response rate acceleration but had no effect on responserates for food pellets. This ‘first-five’ response rate procedure appears to differentiate the effects of buprenorphine on two different reinforcers. While it is possible that buprenorphine would have had a greater effect on foodmaintained responding if the initial rates were higher, it is clear that buprenorphine’s effectswere greateston low and high cocaine dosesand had relatively less effect on the 0.03 to 0.1 mg/kg doses which maintained the highest responding. Unlike procedures that permit only one drug injection per session(Griffiths et al., 1979),our experimental procedure permits up to a maximum of 20 cocaine injections (or 25 food pellets) per session in each of four sessionsper day. Cocaine acquired at the beginning of the sessionmay reduce responserates for subsequentinjections and responserates averagedover the entire session may not accurately reflect changes in patterns across the session. Thus, under certain conditions, overall rates may have little relationship to the actual number of drug injections or food pellets earned. We have attempted to avoid these potential problems by using a second-order scheduleof reinforcement, limiting the number of injections to only 20 per session and extending the inter-session interval to at least 3 h. Since a second-order schedule elicits more behavior before a reinforcer is actually delivered, the direct effects of cocaine on responding for subsequent injections are reduced somewhat, but not completely. The introduction of a microanalysis of the responserates for the first five reinforcers is a valuable adjunct to other measures of reinforcing efficacy, especially in the presenceof another pharmacologic agent, such as buprenorphine. Furthermore, this procedure may provide a sensitive measure of the rate suppressant effects of treatment drugs. 4.3. Mechanisms of buprenorphine’s effect on cocaine

The mechanismsby which buprenorphine reducescocaine self-administration are unclear. Recent data suggest that the p agonist component of buprenorphine may be important in its interactions with cocaine (Mello et al., 1993b).When naltrexone was administered 20 min before buprenorphine, a significant naltrexone dosedependent decrease in the buprenorphine-induced reduction of cocaine self-administration was observed. Thesedata suggestthat naltrexone antagonized the partial p agonist component of buprenorphine (Mello et al., 1993b).The role of Kand 6 opioid receptors in buprenorphine’s interactions with cocaine remains to be determined (Mello and Mendelson, 1995). Several behavioral explanations for buprenorphine’s

effects on cocaine self-administration have also been suggested. One way in which buprenorphine could reduce cocaine self-administration is to make cocaine aversive. This interpretation is consistent with a report that opioid-dependent men describedcocaine’s effectsas dysphoric during buprenorphine treatment (Kosten et al., 1989b).However, an alternative notion was reported by Foltin and Fischman (1994) in a study in which they found that liking scoreswere higher when subjectswere given buprenorphine-cocaine combinations. In the present study, if cocaine was made aversive by buprenorphine, then we would have expected the responserates (especially for the first few injections) to plummet, as is the case when the racemic mixture of N-allylnormetazocine (SKF 10 047) was substituted for cocaine (Lukas et al., 1986). Instead, the rate-accelerating effects of cocaine were simply attenuated by buprenorphine, which would argue against an aversive effect. An alternative to the dysphoria hypothesis is that buprenorphine may reduce cocaine self-administration because it enhances cocaine’s reinforcing effects and, therefore, requires lesscocaine to produce a salient positive effect. In clinical studies, the subjective effects of intranasal cocaine (2 mg/kg) were initially enhanced in five patients after 3 days of buprenorphine treatment (2 mg/day), but this effect diminished after 5 days of buprenorphine treatment (Rosen et al., 1992), suggesting that there is a dissociation between the acute and chronic effects of buprenorphine. Data consistent with the enhancement hypothesis also comes from recent studies in rats and squirrel monkeys. Brown et al. (1991) reported that a combination of cocaine (1.5 mg/kg) and a low acute dose of buprenorphine (0.01 mg/kg) elicited conditioned place preference(CPP), whereasthesedoses of each drug alone did not elicit CPP. Brown et al. (1991) concluded that cocaine and buprenorphine interacted synergistically to elicit CPP. However, these observations were not replicated in subsequent studies conducted by Kosten et al. (1991). These conflicting findings may be accounted for by several procedural differences such as the buprenorphine dose, duration of treatment and pairing of the buprenorphine injection with CPP training (see Mello and Mendelson, 1995 for review). Studies with squirrel monkeys in a drug discrimination paradigm also suggest that buprenorphine may enhance the discriminative stimulus effect of cocaine under certain conditions (Kamien and Spealman, 1991; Spealman and Bergman, 1992). Pre-treatment with relatively low acute dosesof buprenorphine (0.001-0.01 mg/kg) shifted the cocaine discrimination dose-response curve threefold or more to the left in three squirrel monkeys (Kamien and Spealman, 1991).The same dosesof buprenorphine also potentiated the rate-increasing effectsof cumulative dosesof cocaine (0.03-0.3 mg/kg) on an FI 3 schedule of shock termination (Kamien and

S. E. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

Spealman, 1991). Subsequent studies in three squirrel monkeys confirmed that an acute dose of buprenorphine (3.0 and 5.6 &kg), as well as a series of p agonists, shifted the cocaine discrimination dose-response curve to the left (Spealman and Bergman, 1992). We have argued previously that it seemedunlikely that a cocaine reinforcement-enhancement hypothesis could account for the effects of chronic buprenorphine treatment on cocaine self-administration by rhesusmonkeys becausemonkeys often took no cocaine injections during buprenorphine treatment (Mello et al., 1990, 1992; Mello and Mendelson, 1995). Re-examination of individual data from our previous studies (Mello et al., 1990, 1992)revealed that monkeys took zero cocaine injections on 22 to 43% of the total number of buprenorphine treatment days studied (Mello and Mendelson, 1992).Moreover, monkeys took fewer than 6 injections per day on 64% of the buprenorphine treatment days at a dose of 0.237 mglkg/day, 38% of treatment days at a dose of 0.40 mg/kg/day and 73% of treatment days at a dose of 0.70 mg/kg/day (Mello and Mendelson, 1992). In conclusion, the complex pharmacological profile of buprenorphine is clearly evident in its effectson cocainemaintained responding. However, there are clinical data which suggestthat lower dosesof buprenorphine may be more effective than higher buprenorphine doses in reducing cocaine intake (Gastfriend et al., 1992, 1993) and in normalizing sleepparameters (Lukas et al., 1992, 1995 submitted). In addition, a low dose of buprenorphine (2 mg/day s.c.) normalized EEG alpha activity in methadone-maintained patients (Lukas et al., 1984b). However, higher doses of buprenorphine appear to be more effective in reducing cocaine self-administration by polydrug abusers(Schottenfeld et al., 1993).The concordance between clinical and primate studies of buprenorphine’s effects on cocaine self-administration illustrates the potential value of this primate model for evaluating new pharmacotherapies (Mello, 1991, 1992). Data from these clinical and preclinical studies suggest that buprenorphine (and possibly many other medications) may not uniformly reduce the reinforcing effects of all doses of cocaine. Acknowkdgments This research was supported in part by grants DA 00064,DA 00101, DA 00115, DA 02519 and DA 04059 from the National Institute on Drug Abuse, NIH. We are grateful to Nicolas Diaz-Migoyo, Michael Samale and Leo Spinner for excellent technical assistance in data collection, Mark Dobrosielski for computer assistance and to Eleanor DeRubeis, Loreta Carvelli and Carol Buchanan for administrative assistance.We thank Jonathan B. Kamien, Ph.D. for assistancein planning this study and Michelle Sholar, Michael Fortin and David Albeck for assistancewith the data analysis. We

91

also thank S. StevensNegus, Ph.D. of the Alcohol and Drug Abuse Research Center for his helpful comments on the manuscript. Preliminary data were presented to the Committee on Problems of Drug Dependence in 1992. References Brown, E.E., Finlay, J.M., Wong, J.T.F., Damsa, G. and Fibiger, H.C. (1991) Behavioral and neurochemical interactions between cocaineand buprenorphine: Implications for the pharmacotherapy of cocaine abuse. J. Pharmacol. Exp. Ther. 256, 119-126. Carroll, M.E. and Lac, ST. (1992) Effects of buprenorphine on selfadministration of cocaine and a nondrug reinforcer in rats. Psychopharmacology 106,439-446. Carroll, M.E., Cannona, G.N., May, S.A., Buzalsky, S. and Larson, C. (1992) Buprenorphine’s effects on self-administration of smoked cocaine base and orally delivered phencyclidine, ethanol and saccharinin rhesusmonkeys. J. Pharmacol. Exp. Ther. 261,26-37. Dykstra, L.A., Gmerek, D., Winger, G. and Woods, J.H. (1987) Kappa opioids in rhesusmonkeys. I. Diuresis, sedation, analgesia and discriminative stimulus effects. J. Pharmacol. Exp. Ther. 242, 413-420. Foltin, R.W. and Fischman, M.W. (1994)Effects of buprenorphine on the self-administration of cocaine by humans. Behav. Pharmacol. 5, 79-89. Fudala, P.J., Johnson, R.E. and Jaffe, J.H. (1991) Outpatient comparison of buprenorphine maintenance. II. Effects on cocaine usage,retention time in study and missedclinic visits. In: Problems of Drug Dependence 1990,NIDA Research Monograph No. 105 (Harris, L.S., ed.), U.S. Government Printing Office, Washington, DC. Gastfriend, D.R., Mendelson, J.H., Mello, N.K. and Teoh S.K. (1992) Preliminary results of an open trial of buprenorphine in the outpatient treatment of combined heroin and cocaine dependence.In: Problems of Drug Dependence 1991,NIDA Research Monograph No. 119 (Harris, L.S., ed.). U.S. Government Printing Office, Washington, DC. Gastfriend, D., Mendelson, J.H., MeIIo, N.K., Teoh, SK. and Reif, S. (1993) Buprenorphine pharmacotherapy for concurrent heroin and cocaine dependence.Am. J. Addict. 2, 269-278. Grifliths, R.R., Bradford, L.D. and Brady, J.V. (1979) Progressive ratio and fixed ratio schedulesof cocaine-maintained responding in baboons. Psychopharmacology 65, 125-136. Harrigan, S.E. and Downs, D.A. (1978) Continuous intravenous naltrexone effects on morphine self-administration in rhesusmonkeys. J. Pharmacol. Exp. Ther. 204, 481-486. Johnson, R.E., Jaffe, J.H. and Fudala, P.J. (1992) A controlled trial of buprenorphine treatment for opioid dependence.J. Am. Med. Assoc. 267, 2750-2752. Kamien, J.B. and Spealman, R.D. (1991) Modulation of the discriminative stimulus effects of cocaine by buprenorphine in squirrel monkeys. Behav. Pharmacol. 2, 517-520 Kosten, T.R., KIeber, H.D. and Morgan, C. (198r~, Role of opioid antagonists in treating intravenous cocaine i L’X. Life Sci. 44, 887-892. Kosten, T.R., Kleber, H.D. and Morgan, C. (1989b)Treatment of cocaine abuse with buprenorphine. Biol. Psychiatr. 26, 170-172. Kosten, T.A., Marby, D.W. and Nestler, E.J. (1991) Cocaine conditioned place preference is attenuated by chronic buprenorphine treatment. Life Sci. 49, PL-201-PL-206. Lewis, J.W. (1974) Ring C-bridged derivatives of thebaine and oripavine. In: Narcotic Antagonists. Advances in Biochemical Psychopharmacology (Braude, M.C., Harris, L.C., May, E.L., Smith, J.P. and Villarreal, J.E., eds.), pp. l23- 136. Raven Press, New York.

98

S.E. Lukas et al. /Drug and Alcohol Dependence 40 (1995) 87-98

Lewis, J.W., Rance, M.J. and Sanger, D.J. (1983)The pharmacology and abusepotential of buprenorphine: A new antagonist analgesic. In: Advances in SubstanceAbuse: Behavioral and Biological Research(Mello, N.K., ed.), pp. 103-154, JAI Press,Greenwich, CT. Lloyd-Jones, J.G., Robinson, P., Henson, R., Biggs, S.R. and Taylor, T. (1980)Plasmaconcentrations and disposition of buprenorphine after intravenous and intramuscular doses to baboons. Eur. J. Drug Metab. Pharmacokinet. 5, 233-239. Lukas, S.E., Griffiths, R.R., Brady, J.V. and Wurster, R.M. (1984a) Phencyclidine-analogueself-injection by the baboon. Psychopharmacology 83, 316-320. Lukas, S.E., Jasinski, D.R. and Johnson, R.E. (1984b) Electroencephalographic and behavioral correlates of buprenorphine administration. Clin. Pharmacol. Ther. 36, 127-132. Lukas, SE., Brady, J.V. and Griffiths, R.R. (1986) Comparison of opioid self-injection and disruption of schedule-controlled performance in the baboon. J. Pharmacol. Exp. Ther. 238, 924-931. Lukas, S.E., Mello, N.K., Bree, M.P. and Mendelson, J.H. (1988)Differential tolerance development to buprenorphine-, diprenorphineand heroin-induced disruption of food-maintained responding in macaque monkeys. Pharmacol. Biochem. Behav. 30, 977-982. Lukas, S.E., Dorsey, C.M., Abdulali, A., Fortin, M., Abdulali, S., Mendelson, J.H., Mello, N.K. and Cunningham, S. (1992) EEG sleep architecture in cocaine- and heroin-dependent subjects during buprenorphine maintenance. In: Problems of Drug Dependence 1991,NIDA Research Monograph No. I I9 (Harris, L.S., ed.). U.S. Government Printing OfIice, Washington, DC. Lukas, SE., Dorsey, C.M., Mello, N.K., Mendelson, J.H., Sholar, M., Cunningham, S.L. and Lundahl, L.H. (1995)Reversal of sleep disturbances in cocaine- and heroin-dependent subjects during chronic buprenorphine treatment (submitted for publication). Mansbach, R.S. and Balster, R.L. (1990) Temporal determinants of intravenous cocaine self-administration in rhesusmonkeys. Behav. Pharmacol. I, 317-326. Mello, N.K. (1991) Pre-clinical evaluation of the effects of buprenorphine, naltrexone and desipramine on cocaine self-administration. In: Problems of Drug Dependence 1990, NIDA Research Monograph No. I05 (Harris, L.S., ed.). U.S. Government Printing ORice, Washington, DC. Mello, N.K. (1992) Behavioral strategies for the evaluation of new pharmacotherapies for drug abuse treatment. In: Problems of Drug Dependence 1991, NIDA Research Monograph No. II9 (Harris, L.S., ed.). U.S. Government Printing Office, Washington, DC. Mello, N.K. and Mendelson, J.H. (1980) Buprenorphine suppresses heroin use by heroin addicts. Science27, 657-659. Mello, N.K. and Mendelson, J.H. (1992)Primate studiesof behavioral pharmacology of buprenorphine. In: Buprenorphine: An Alternative Treatment for Opioid Dependence, NIDA Research Monograph No. I21 (Blaine, J.D. and Snyder, M., eds.). U.S. Government Printing Office, Washington, DC. Mello, N.K. and Mendelson, J.H. (1995) Buprenorphine treatment of cocaine and heroin abuse. In: Buprenorphine: Combatting Drug Abuse With a Unique Opioid (Cowan, A., and Lewis, J., eds.), pp. 241-287. John Wiley & Sons, Inc., New York. Mello, N.K., Mendelson, J.H. and Kuehnle, J.C. (1982) Buprenor-

phine effects on human heroin self-administration: An operant analysis. J. Pharmacol. Exp. Ther. 223, 30-39. Mello, N.K., Mendelson, J.H., Bree, M.P. and Lukas, SE. (1989) Buprenorphine suppressescocaine self-administration by rhesusmonkeys. Science245, 859-862. Mello, N.K., Mendelson, J.H., Bree, M.P. and Lukas, S.E. (1990) Buprenorphine and naltrexone effects on cocaine self-administration by rhesus monkey. J. PharmacoI. Exp. Ther. 254, 926-939. Mello, N.K., Lukas, SE., Kamien, J.B., Mendelson, J.H., Drieze, J. and Cone, E.J. (1992)The effects of chronic buprenorphine treatment on cocaine and food self-administration by rhesusmonkeys. J. Pharmacol. Exp. Ther. 260, 1185-l 193. Mello, N.K., Kamien, J.B., Lukas, S.E., Mendelson, J.H., Drieze, J.M. and Sholar, J.W. (1993a) Effects of intermittent buprenorphine administration on cocaine self-administration by rhesus monkeys. J. Pharmacol. Exp. Ther. 264, 530-541. Mello, N.K., Lukas, SE., Mendelson, J.H. and Drieze, J. (1993b) Naltrexone-buprenorphine interactions: Effects on cocaine selfadministration. Neuropsychopharmacology 9, 2 I l-224. Mello, N.K., Mendelson, J.H., Lukas, SE., Gastfriend, D., Teoh, S.K. and Holman, B.L. (1993c)Buprenorphine treatment of opiate and cocaine abuse: Clinical and preclinical studies. Harvard Rev. Psychiatr. I, 168-183. Mendelson, J.H., Mello, N.K., Teoh, SK., Kuehnle, J., Sintavanarong, P. and Dooley-Coufos, K. (1991) Buprenorphine treatment for concurrent heroin and cocaine dependence:Phase I study. In: Problems of Drug Dependence 1990, NIDA Research Monograph No. I05 (Harris, L.S., ed.). U.S. Government Printing Oflice, Washington, DC. Negus, S.S. and Dykstra, L.A. (1988) Kappa antagonist properties of buprenorphine in the shock titration procedure. Eur. J. Pharmacol. 56, 77-86. Negus, S.S., Picker, M.J. and Dykstra, L.A. (1989) Kappa antagonist properties of buprenorphine in non-tolerant and morphinetolerant rats. Psychopharmacology 98, 141-143. Negus, S.S., Mello, N.K., Lukas, S.E. and Mendelson, J.H. (1995) Diurnal patterns of cocaine and heroin self-administration in rhesus monkeys responding under a schedule of multiple daily sessions. Behav. Pharmacol. (in press). Rosen, M.I. and Kosten, T.R. (1991) Buprenorphine: Beyond methadone? Hosp. Commun. Psychiatr. 42, 347-349. Rosen, M.K., Pearsall, H.R., McDougle, C.J., Price, L.H., Woods, S.W. and Kosten, T.R. (1992) Effects of acute buprenorphine on responsesto intranasal cocaine. (Personal commuication). Schottenfeld, R.S., Pakes, J., Ziedonis, D. and Kosten, T.R. (1993) Buprenorphine: Dose-related effects on cocaine and opioid use in cocaine-abusing opioid-dependent humans. Biol. Psychiatr. 34, 66-74. Spealman, R.D. and Bergman, J. (1992) Modulation of the discriminative-stimulus effects of cocaine by p and Y opioids. J. Pharmacol. Exp. Ther. 261, 607-615. Tallarida, R.J. and Jacob, L.S. (1979)The Dose-ResponseRelation in Pharmacology, Springer Verlag, New York. Winger, G., Skjoldager, P. and Woods, J.H. (1992) Effects of buprenorphine and other opioid agonists and antagonists on alfentanil- and cocaine-reinforced responding in rhesus monkeys. J. Pharmacol. Exp. Ther. 261, 311-317.