Escalation of cocaine self-administration does not depend on altered cocaine-induced nucleus accumbens dopamine levels

Authors


Address correspondence and reprint requests to Loren H. Parsons, Department of Neuropharmacology, CVN-7, The Scripps Research Institute, 10550 N. Torrey Pines Road, La Jolla, CA, USA. E-mail: lparsons@scripps.edu

Abstract

Previous studies showed that prolonged access to cocaine or heroin self-administration (long access, or LgA) produces an escalation in drug intake not observed with limited access to the drug (short access, or ShA). The present experiment employed in vivo microdialysis to test the role of alterations in drug pharmacokinetics and/or efficacy in increasing dopamine (DA) levels in the nucleus accumbens (NAcc) during cocaine intake escalation. In experiment 1, both ShA and LgA rats were challenged with passive intravenous administration of cocaine (0.125–1 mg/injection). Regardless of the doses tested, there was no difference between groups in the ability of cocaine to increase NAcc DA levels and no group differences in the temporal profile of dialysate cocaine levels. In experiment 2, cocaine and DA concentrations were measured during cocaine self-administration. Self-administration produced sustained increases of DA in the NAcc with LgA rats maintaining greater steady levels of DA (750% of baseline) than ShA rats (400% of baseline). The difference in the LgA versus ShA rats was not due to differences in the efficacy of cocaine to elevate DA levels, or in the rate of cocaine metabolism, but was directly related to the amount of self-administered cocaine. These findings show that changes in cocaine efficacy or pharmacokinetics do not play a critical role in cocaine intake escalation.

Abbreviations used
aCSF

artificial cerebrospinal fluid

CRF

corticotropin-releasing factor

DA

dopamine

ICSS

intracranial self-stimulation

LgA

long access

NAcc

nucleus accumbens

n.s.

not significant

OD

outside diameter

ShA

short access

The transition from drug use to drug addiction is associated with a process of escalation whereby drug intake becomes excessive and difficult to limit (Siegel 1984; Marlatt et al. 1988; Gawin and Ellinwood 1989; Leshner 1997). Based on previous work (Deneau et al. 1969; Johanson et al. 1976; Bozarth and Wise 1985), we have shown that differential access to intravenous self-administration of cocaine or heroin produces two distinct patterns of drug intake (Ahmed and Koob 1998, 1999; Ahmed et al. 2000). With 1 h of access per day (short-access or ShA rats), drug intake remains low and stable over time, while with 6 h or more of access per day (long-access or LgA rats), first hour and total drug intake gradually and dramatically escalates. Comparisons between ShA rats and LgA rats after escalation, therefore, may point to neurobiological mechanisms that underlie the transition to compulsive drug use associated with addiction.

Ample evidence indicates that cocaine-induced increases in nucleus accumbens (NAcc) dopamine (DA) levels play a major role in regulating cocaine self-administration in both animals (Koob 1992; Wise 1996) and humans (Volkow et al. 1996, 1997). In vivo microdialysis studies have confirmed that cocaine self-administration produces sustained and stable increases of DA levels in the NAcc (Pettit and Justice 1989; Di Ciano et al. 1995; Parsons et al. 1995; Hemby et al. 1997) that may correspond to an optimal ‘drug effect’ level hypothesized to be regulated by animals during self-administration (Yokel and Pickens 1973, 1974; Ahmed and Koob 1999). Wise and colleagues have accumulated strong evidence for this hypothesis by showing that the timing of drug injections within a self-administration session is correlated with temporal fluctuations in DA levels in the NAcc (Wise et al. 1995; Ranaldi et al. 1999). The regulation of NAcc DA by cocaine self-administration may explain both the tonic change in cell firing rates observed in the NAcc during cocaine self-administration (Carelli and Deadwyler 1994, 1996; Chang et al. 1998; Peoples et al. 1999; Nicola and Deadwyler 2000) and the fluctuations in firing rate associated with the timing of cocaine injections (Chang et al. 1994; Peoples and West 1996; Chang et al. 1998; Peoples et al. 1998). These observations indicate that animals regulate cocaine intake to maintain an optimal DA tone in the NAcc.

Escalation of drug use is classically explained as tolerance to the desired effects of the drug. Based on the putative regulation of NAcc DA levels by cocaine self-administration, tolerance can be envisioned to arise from distinct, though not necessarily mutually exclusive, mechanisms including altered cocaine pharmacokinetics, a decreased ability of cocaine to elevate NAcc DA levels, and/or, more speculatively, an increased optimal NAcc DA concentration preferred by the experimental subject (Ahmed and Koob 1998, 1999). This study used in vivo microdialysis to test the potential contribution of these different mechanisms in determining the different patterns of cocaine self-administration previously observed between ShA rats and LgA rats.

Materials and methods

Subjects

Thirty-six male Wistar rats, weighing 245–350 g at the start of the experiment, were used (Charles River, Hollister, CA, USA). Rats were housed in groups of two or three and were maintained in a light- (12-h light-dark cycle; lights on at 10.00 h) and temperature-controlled vivarium. All behavioral testing occurred during the dark phase of the light-dark cycle. Food and water were freely available, except when specified. All procedures were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Catheter surgery

Anesthetized rats (halothane-oxygen mixture, 1.0–1.5% halothane) were prepared with catheters in the right jugular vein as described previously (Ahmed and Koob 1997). After surgery, catheters were flushed daily with 0.2 mL of a sterile antibiotic solution containing heparinized saline (30 USP units/mL) and Timentin® (100 mg/mL; SmithKline Beecham Pharmaceuticals, Philadelphia, PA, USA). The patency of the catheter was checked by administering 0.15 mL of the ultra-short-acting barbiturate anesthesic Brevital® sodium through the catheter (1% methohexital sodium, Eli Lilly, Indianapolis, IN, USA).

Intravenous cocaine self-administration

Behavioral training was conducted in standard operant chambers (described in Ahmed and Koob 1997). One week after i.v. catheterization, rats were food restricted for maintenance of 85% of the body weight obtained with free-feeding conditions and trained to press a lever for 45-mg food pellets on a continuous reinforcement schedule during six daily 30-min sessions. After food training, rats were returned to ad libitum food for the duration of the experiments. Three days after food training, rats were tested for cocaine self-administration during two consecutive phases: a screening phase (1 day) and an escalation phase (11–13 days). During the screening phase, rats were allowed to self-administer cocaine during only 1 hour on a fixed-ratio 1 schedule (250 μg/injection in a volume of 0.1 mL delivered in 4 s) after which two balanced groups with the same mean cocaine intake were formed. During the escalation phase, one group had access to cocaine self-administration for only 1 h per day (short-access or ShA rats) and the other group for 6 h per day (long-access or LgA rats). Self-administration sessions were performed 5–6 days a week.

Intracranial surgery

Intracranial surgery was performed on two successive days after the 13th (experiment 1) or the 11th (experiment 2) session of cocaine self-administration. All rats were anesthetized with halothane (1–2%) and stereotaxically implanted with a stainless steel microdialysis guide-cannula (21 gauge; Plastics One, Roanoke, VA, USA) which terminated at the dorsal surface of the left nucleus accumbens (AP, +1.7 mm; ML, +1.4 mm; V, −6.1 mm; according to the atlas of Paxinos and Watson 1986). Unilateral placement was chosen to avoid additional variation related to brain lateralization. The cannula was secured to the skull with sterilized stainless steel screws and methylmethacrylate cement. Stainless steel wire stylets cut to the length of the cannula were inserted to prevent cannula blockade.

Intracranial microdialysis

In all experiments and in all animals, microdialysis sampling was performed 48 h after the last session of the escalation phase. Approximately 9 h prior to the start of microdialysis, animals were lightly anesthetized (1–2% halothane) and microdialysis probes were inserted through the guide-cannula. Then, animals were singly housed in standard plastic cages located in a room dedicated for microdialysis sampling. Rats regained consciousness within 5 min of probe insertion. Microdialysis probes were constructed to have a 2-mm active membrane length (300 μm outside diameter [OD]; 13 000 Da cut-off; Spectrum Medical Industries, Houston, TX, USA) that extended 2 mm beyond the end of the guide-cannula. The perfusion medium consisted of an artificial cerebrospinal fluid (aCSF) solution containing 145 mm NaCl, 2.8 mm KCl, 1.2 mm MgCl2, 1.2 mm CaCl2, 0.25 mm Ascorbate, 5.4 mm Glucose (pH 7.2–7.4) and was delivered to the probe using syringe pumps (CMA/100; Bioanalytical Systems, West Lafayette, IN, USA) via a liquid swivel (Model 375 dual channel; Instech, Plymouth Meeting, PA, USA) attached to a balance arm above the animal's cage to ensure freedom of movement. After probe implantation, aCSF was perfused at a flow-rate of 0.5 μL/min until the start of the experiment.

Experimental design

Two microdialysis experiments were conducted in separate groups of rats:

Experiment 1: measurement of NAcc DA and cocaine levels following experimenter-administered i.v. cocaine in escalated rats and non-escalated rats

One to two days after intracranial surgery, rats were allowed to self-administer cocaine for 3–9 additional sessions (1 or 6 h depending on the group) before being tested for microdialysis. Four separate microdialysis test sessions were then conducted to allow for the testing of a total seven ShA and eight LgA rats (each session separated by 1–3 days). On each the first three microdialysis days, two animals from each experimental group were tested, while one ShA and two LgA rats were tested on the final microdialysis day. Each microdialysis session was divided into two successive phases: a no-net flux phase during which basal levels of extracellular DA were assessed and a cocaine challenge phase during which DA levels were measured following different doses of i.v. cocaine (0.125, 0.25 and 1 mg/injection). One hour prior to the start of the no-net flux phase, the perfusate flow-rate was increased to 1 μL/min. Following this equilibration period, four baseline samples were collected at 15-min intervals. The perfusate then was switched to aCSF containing 5 nm DA and allowed to equilibrate for 15 min. Dialysate samples then were collected for 60 min, after which time the perfusate was switched to an aCSF containing 10 nm DA. After a 15-min equilibration period, samples were collected for 60 min. After completion of the no-net flux experiment, the perfusate was switched to a normal aCSF without DA. At this time, the i.v. catheters were flushed with heparinized saline (30 USP units/mL) and were connected via the fluid swivel to syringes containing cocaine solution (2.5 mg/mL) mounted in a syringe pump (Model 22, Harvard Apparatus, So. Natuck, MA, USA). After a 60-min equilibration period, six baseline dialysate samples were collected at 10-min sampling intervals. Subsequently, three separate i.v. injections of cocaine (order: 0.25, 0.125 and 1 mg/injection) were delivered with six 10-min samples collected after each injection, except for the last injection for which nine samples were collected.

Experiment 2: measurement of NAcc DA and cocaine levels during cocaine self-administration

Three to four days after intracranial surgery, rats were allowed to self-administer cocaine for 14–17 additional sessions before microdialysis testing. During this period, rats were habituated to the microdialysis head tether during each self-administration session. Subsequently, four microdialysis test sessions were performed to allow the testing of a total of five ShA and seven LgA animals (each microdialysis test was separated by 1 day). Approximately 9 h after probe insertion, rats were transported in individual plastic cages to the self-administration room where they were left undisturbed for 1.5–2 h at a perfusate flow rate of 0.6 μL/min. After this habituation period, six baseline dialysate samples were collected at 16.5-min sampling intervals. Immediately after taking the last baseline sample, rats were placed in the self-administration boxes and were allowed to self-administer cocaine (0.25 mg/infusion; FR-1) for 180 min, during which time dialysate samples were taken at 16.5-min intervals.

HPLC methods

Dialysate samples were split for analysis of DA and cocaine concentrations by separate HPLC methods. DA concentrations were determined from 5-μL dialysate aliquots using reversed-phase HPLC coupled with electrochemical detection. Dopamine was separated on a 1 × 100-mm column (BetaBasic 3 μm C18; Keystone Analytical, Bellefonte, PA, USA) over which a mobile phase consisting of 50 mm NaH2PO4 buffer with 17% Acetonitrile (v/v), 0.2 mm Na2EDTA, 3.7 mm 1-decanesulfonic acid and 4.9 mm triethylamine (apparent pH of 4.2) was pumped at 35 μL/min using an ISCO Model 500D pulseless syringe pump (Lincoln, NE, USA). The eluent was delivered directly to a standard electrochemical cell containing two glassy carbon working electrodes (EG&G Princeton Applied Research Model MP1304; Trenton, NJ, USA) arranged in series. The first electrode in the flow path was set at +700 mV against an AG/AgCl reference electrode (Model RE4; BioAnalytical Systems) and was used to reduce the solvent front signal at the second working electrode. The second electrode was set at 0 mV and was used for the measurement of the reduction peak produced by dopamine. The electrode potential and current analyses were controlled by an EG&G Princeton Applied Research amperometric detector (Model 400). External calibration curves were generated daily from fresh standard solutions.

Dialysate cocaine and benzoylecgonine concentrations were determined by reversed-phase HPLC coupled with UV detection. Five-microliter aliquots of dialysate were injected onto a 1 × 100-mm column (BetaBasic 3 μm C18; Keystone Analytical, Bellefonte, PA, USA) over which a mobile phase consisting of 50 mm NaH2PO4 buffer with 12% acetonitrile, 11% methanol, 4.9 mm triethylamine (apparent pH 5.0) was pumped at 33 μL/min using a 1100 series pump from Agilent Technologies (Wilmington, DE, USA). The eluent was delivered to capillary flow cell of an 1100 series UV detector from Agilent Technologies. Absorbance at a wavelength of 225 nm was used for all quantitative analyses and all data capture and analysis was performed using Chemstation software (Agilent Technologies). External calibration curves were generated daily from fresh standard solutions.

Histology

Microdialysis probe locations were examined histologically after completion of the microdialysis experiments in all of the animals. Rats were killed by administering an i.v. overdose of sodium pentobarbitol and the brains were removed and rapidly frozen on dry ice. Probe placements were determined by comparing 50 μm coronal brain sections from each animal with corresponding plates in the atlas of Paxinos and Watson (1986).

Data analysis

All data were subjected to one-way or multiple-way analyses of variance (anova) with one between-factor, experimental groups (ShA and LgA groups), and none, one or two within-factors, depending of the experiment. All post–hoc comparisons for interactions were carried out by tests for simple main effects. In all experiments, dialysate DA concentrations were stable across all baseline samples (six in total) and did not significantly differ between groups. Therefore, all data characterizing cocaine effects on DA levels were expressed as percentage of the average baseline. In the course of the study, a few rats were excluded at different stages. In experiment 1 (initial n = 18) which lasted about 2 months, a total of three rats were dropped: two rats (one ShA and one LgA rat) did not acquire the self-administration behavior (less than eight injections per hour, the operant level being at 5–7 lever presses) and one ShA rat lost probe patency. All analyses were performed on the remaining rats (seven ShA and eight LgA rats). In experiment 2 (initial n = 18) which lasted about 2 months, a total of six rats were excluded: two rats (one ShA and one LgA rat) failed to acquire the self-administration behavior, two ShA rats dramatically increased their intake after intracranial cannulation (to about 150–170% of pre-surgery levels within 1–3 sessions) and, finally, one LgA rat lost probe patency and one ShA rat had its i.v. catheter disconnected from the intravenous system of injection during microdialysis. All analyses were performed on the remaining rats (five ShA and seven LgA rats).

Results

Escalation of cocaine self-administration

To appropriately compare cocaine self-administration in rats with different access time to cocaine (1 h vs. 6 h), only first-hour intakes of cocaine were considered for analysis. We have reported previously that total cocaine or heroin intake escalates over time in animals with prolonged access to drug self-administration (Ahmed and Koob 1998, 1999; Ahmed et al. 2000). In the present experiments, cocaine intake was virtually identical between groups at the beginning of the escalation phase (Fig. 1a and b). However, cocaine intake by LgA rats increased over time above the level of cocaine intake by ShA rats (group × time: exp. 1, F12,156 = 4.63, p < 0.01; exp. 2, F10,100 = 3.47, p < 0.01) (Fig. 1a and b). Post-hoc comparisons show that significant differences between groups were detected from the fifth session of escalation onward (p < 0.05). Finally, in both experiments, intake of cocaine was significantly greater in LgA rats than in ShA rats before microdialysis testing, as measured in each individual by its average intake over the last three sessions of self-administration preceding microdialysis testing (exp. 1, F1,13 = 14.43, p < 0.01; exp. 2, F1,10 = 20.23, p < 0.01) (inserts in Fig. 1a and b).

Figure 1.

Effect of access time to cocaine self-administration on first-hour cocaine intake (mean number of self-injections ± SEM). Rats had access to cocaine (0.25 mg/injection) for either 1 h (ShA rats) or 6 h (LgA rats) per day. Self-administration tests were performed 5–6 days per week. (a) Results obtained in experiment 1 (ShA rats: n = 7; LgA rats: n = 8). (b) Results obtained in experiment 2 (ShA rats: n = 5; LgA rats: n = 7). Inserts represent mean cocaine intake during the last three self-administration sessions preceding microdialysis testing. *Different from ShA rats (p < 0.05, tests of simple main effects).

Experiment 1: basal extracellular concentrations of DA

Estimates of DA concentrations in the NAcc were calculated from the mean of four samples at each perfusate concentration (0, 5, 10 nm DA). First-order regressions were used to obtain slope and intercept values, which were used to solve for the point of no-net flux (zero intercept on the y axis) (for more details, see Parsons and Justice 1992). Estimated basal extracellular DA in the NAcc did not differ between groups (ShA rats, 10.5 ± 5.2 nm; LgA rats, 9.0 ± 2.6 nm) [F1,13 < 1 not significant (n.s.)]. Estimates of in vivo DA recovery in the NAcc, as determined by the slope of the regression line, also did not vary between groups (ShA rats, 0.52 ± 0.16; LgA rats, 0.44 ± 0.10) (F1,13 < 1, n.s.).

Experiment 1: cocaine and DA levels following i.v. cocaine challenges

Cocaine injections produced a dose-dependent and time-dependent increase in cocaine concentrations in the NAcc (dose × time: F10,130 = 41.39, p < 0.01) (Fig. 2a). Increases in cocaine levels peaked at 10 min post injection and quickly decayed over time. However, regardless of the dose tested, the post-injection time-course of cocaine concentrations was almost identical between ShA rats and LgA rats (group: F1,13 < 1, n.s.; group × dose × time: F10,130 < 1, n.s.).

Figure 2.

Dose-effect of passively delivered cocaine on nucleus accumbens cocaine (a) and dopamine concentrations (b) in ShA rats (n = 7) and LgA rats (n = 8) after escalation (mean ± SEM). Forty-eight hours after the last self-administration session, both ShA rats and LgA rats were challenged with different doses of cocaine. Cocaine doses were administered by the experimenter every hour in the following order: 0.25, 0.125 and 1 mg per injection. Dialysate samples were taken every 10 min. Vertical arrows show the timing of injection and the dose of cocaine injected.

Baseline dialysate DA concentrations did not differ between groups before cocaine testing: ShA rats (2.51 ± 0.27 nm) and LgA rats (2.33 ± 0.54 nm). Cocaine injections produced a dose-dependent and time-dependent increase in DA in the NAcc (dose × time: F10,130 = 8.84, p < 0.01) (Fig. 2b). As with the cocaine levels, increases in DA levels peaked at 10 min post injection and quickly decayed over time. However, regardless of the cocaine dose tested, peak levels and the post injection profile of DA levels did not vary between ShA rats and LgA rats (group: F1,13 < 1, n.s.; group × dose × time: F10,130 < 1, n.s.). Cocaine peak effects on DA levels were expressed as the difference from DA levels measured during baseline. Net peak increases in DA levels considerably varied with the dose of cocaine (F2,26 = 11.61, p < 0.01) (Fig. 3a) but the cocaine dose-effect function on DA levels was not shifted in LgA rats after escalation (group: F1,13 < 1, n.s.; group × dose: F2,26 < 1, n.s.). Regardless of the dose tested, the ability of a fixed amount of cocaine to elevate DA levels was virtually identical between groups as measured by cocaine peak effects on DA divided by cocaine peak concentrations (group: F1,13 < 1, n.s.; group × dose: F2,26 < 1, n.s.) (Fig. 3b).

Figure 3.

Cocaine dose-effect function on nucleus accumbens DA levels in ShA rats (n = 7) and LgA rats (n = 8) after escalation. (a) Data represent peak effects on DA levels across doses. The data are expressed as the percent difference from baseline DA levels which did not differ between experimental groups. (b) Estimates of cocaine efficacy across cocaine doses (i.e. the ability of a given amount of cocaine to elevate DA levels in the NAcc). These estimates were obtained for each dose by dividing the maximal change in dialysate DA levels by the maximal dialysate cocaine concentration. In all cases, the maximal levels of both DA and cocaine occurred in the same dialysate sample.

Experiment 2: cocaine and DA levels during cocaine self-administration

During microdialysis testing, LgA rats took significantly more i.v. cocaine doses than ShA rats, especially during the first hour-and-a-half of testing where the group difference in cocaine intake was the most pronounced (group: F1,10 = 12.13, p < 0.01) (Fig. 4c).

Figure 4.

Effects of escalated levels of cocaine self-administration on nucleus accumbens cocaine (a) and dopamine levels (b) in ShA rats (n = 5) and LgA rats (n = 7) after escalation (mean ± SEM). Forty-eight hours after the last session of self-administration, both ShA rats and LgA rats were allowed to self-administer i.v. cocaine (0.25 mg/injection) during a 3-h session. The vertical dotted line indicates the onset of self-administration. (c) Data represent the self-administration behavior during microdialysis testing (mean number of cocaine self-injections/20 min). *Different from ShA rats (p < 0.05, tests of simple main effects).

Baseline dialysate DA concentrations did not differ between groups before cocaine self-administration testing: ShA rats (0.94 ± 0.37 nm) and LgA rats (1.18 ± 0.22 nm). As expected, i.v. cocaine self-administration produced time-dependent increases of DA (time: F13,130 = 16.88, p < 0.01) and cocaine (time: F13,130 = 28.11, p < 0.01) in the NAcc (Fig. 4a and b). In both groups, DA and cocaine reached maximum levels after about 30 min of cocaine self-administration and then remained above baseline for the duration of the testing session. Self-maintained increases in NAcc DA levels were significantly greater in LgA rats than in ShA rats (group × time: F13,130 = 2.08, p < 0.05). Post-hoc comparisons revealed that LgA rats maintained significantly higher steady levels of DA in the NAcc (about 750% of baseline) than ShA rats (about 400% of baseline) during the interval of 33–115 min during cocaine self-administration (p < 0.05) (Fig. 4b). After that period of time, increased DA levels in LgA rats remained above increased DA levels in ShA rats, but this difference was no longer significant. During the same period of time, dialysate cocaine was more elevated in LgA rats than in ShA rats, but this difference was not statistically significant (group × time: F13,130 = 1.55, n.s.). To obtain estimates of cocaine efficacy in elevating interstitial DA levels in the NAcc during cocaine self-administration, DA levels were divided by cocaine levels over each time interval (Fig. 5a). In both ShA rats and LgA rats, similar amounts of cocaine produced similar elevations in interstitial DA levels, indicating that the pharmacodynamic effect of cocaine on DA remained unaltered after cocaine intake escalation (group: F1,10 < 1, n.s.; group × time: F10,100 = 1.24, n.s.). Finally, to estimate cocaine metabolism rates that lead to the formation of benzoylecgonine, dialysate benzoylecgonine levels were divided by dialysate cocaine levels over each time interval (Pettit and Pettit 1994). Results show that groups did not differ in cocaine metabolism rates during cocaine self-administration (group: F1,10 < 1, n.s.; group × time: F10,100 = 1.47, n.s.) (Fig. 5b). The increase over time of the benzoylecgonine/cocaine ratio (time: F10,100 = 9.72, p < 0.01) primarily reflects the accumulation of benzoylecgonine in the NAcc during cocaine self-administration.

Figure 5.

Cocaine efficacy (a) and cocaine metabolism (b) during cocaine self-administration in ShA rats (n = 5) and LgA rats (n = 7) (mean ± SEM; for other details, see legend of Fig. 4). Estimates of cocaine efficacy were obtained by dividing DA levels by cocaine concentrations over each time interval. Estimates of the rate of cocaine metabolism were obtained by dividing benzoylecgonine concentrations by cocaine concentrations over each time interval.

Histology

In all cases, the active portion of the dialysis membranes was located in the nucleus accumbens core/shell just medial to the anterior commissure between 2.20 and 1.20 mm anterior to bregma (Fig. 6).

Figure 6.

Placement of microdialysis probes from experiments 1 and 2. In all cases, the active portion of the dialysis membranes was located in the rostral pole of the nucleus accumbens with the active dialysis area transecting the shell and core subdivisions.

Discussion

As previously reported (Ahmed and Koob 1998, 1999), repeated prolonged access to cocaine self-administration (LgA rats) produced an escalation in cocaine intake not observed with limited access to the drug (ShA rats). Assessment of the temporal profile of microdialysate cocaine levels from the NAcc (i.e. peak concentrations and elimination rates) following passive administration of different doses of i.v. cocaine revealed no difference between ShA and LgA rats. In addition, the cocaine dose-effect function on DA levels was not shifted after cocaine intake escalation, indicating that efficacy of cocaine to increase interstitial DA levels in the NAcc remained unaltered in animals with escalated cocaine use. Finally, assessment of NAcc DA levels during cocaine self-administration showed that LgA rats maintained significantly higher steady levels of DA (about 750% of baseline) than ShA rats (about 400% of baseline). The sustained increase in NAcc DA levels observed in LgA rats was not associated with a decrease in cocaine efficacy or with an alteration in the rate of cocaine metabolism during cocaine self-administration. In this regard, however, it is noteworthy that the dialysis probes in both of the present experiments were in the rostral pole of the NAcc (+1.2 to 2.2 mm from bregma; see Fig. 6) with the active dialysis area transecting the shell and core subdivisions (Heimer et al. 1997; Zahm 2000). In view of the functional heterogeneity of the NAcc, along both its medial-lateral (Kelley 1999) and anterior-posterior axis (Reynolds and Berridge 2001; Martin et al. 2002), it is possible that alterations in cocaine efficacy would have been detected if probes had been more systematically positioned in selective sub-regions of the NAcc (for example, see Cadoni et al. 2000). Further work is necessary to determine whether the escalation of cocaine self-administration is associated with alterations in cocaine pharmacodynamics in specific NAcc sub-territories. With this caveat in mind, however, the present data suggest that neither alterations in the efficacy of cocaine to increase NAcc DA levels nor alterations in the post-injection temporal profile of cocaine levels in brain play a critical role in cocaine intake escalation.

Regulation of DA receptor function and cocaine intake escalation

As outlined in the introduction, several lines of evidence suggest that cocaine self-administration is a self-regulatory behavior aimed at maintaining an optimal DA tone in the NAcc and perhaps in other forebrain regions innervated by DA neurons (Hurd et al. 1997). Within this regulatory framework, the maintenance of higher NAcc DA levels during cocaine self-administration by LgA versus ShA rats may reflect an effort to surmount a decrease in the post-synaptic effects of DA on NAcc neurons. Dopamine acts on two families of receptors, the D1 class (D1, D5) and D2 class (D2, D3 and D4), with each class possessing different molecular, anatomical and pharmacological properties (Hartman and Civelli 1997; Missale et al. 1998). Previous pharmacological studies have shown that both types of DA receptors contribute to cocaine self-administration. For instance, in a continuous reinforcement schedule such as that used in the present study, blockade of D1 and/or D2 receptors by selective DA antagonists increase the rate of cocaine self-injections (Koob et al. 1987; Corrigall and Coen 1991; Hubner and Moreton 1991; Caine and Koob 1994), a compensatory behavioral phenomenon that is thought to reflect a decrease in the rewarding effects of cocaine. Similar effects on cocaine self-administration are produced when DA antagonists are directly injected into the NAcc (Maldonado et al. 1993; McGregor and Roberts 1993; Caine et al. 1995) and other anatomically and functionally related forebrain regions (Caine et al. 1995; Hurd et al. 1997; Epping-Jordan et al. 1998a). Thus, what animals appear to regulate during cocaine self-administration is not DA levels per se but rather an overall level of dopaminergic neurotransmission (including the downstream effects of post-synaptic DA receptor stimulation). It follows that increased steady levels of DA observed in animals with escalated cocaine use could result from a decrease in DA receptor function induced by the chronic exposure to prolonged cocaine self-administration.

Several recent reports converge to show that both D1-like and D2-like receptors are decreased after long-term, heavy exposure to passive administration or self-administration of cocaine or other stimulant drugs in rats (Tsukada et al. 1996; Graziella de Montis et al. 1998; Maggos et al. 1998), non-human primates (Moore et al. 1998a,b; Nader et al. 2002) and humans (Volkow et al. 1993, 2001). For instance, in one study, rats were exposed to a binge pattern of passive cocaine administration for 14 days, a treatment schedule that approximates the LgA condition of the present study. Using positron emission tomography, this treatment was shown to produce an apparent decrease in the in vivo affinity of both D1-like and D2-like receptors (Tsukada et al. 1996). Interestingly, dopamine D2-like receptors returned to normal levels after 21 days of withdrawal from chronic cocaine (Maggos et al. 1998), a time course of recovery that parallels the time course of recovery from escalated levels of cocaine intake seen in LgA rats after an equivalent period of withdrawal (Ahmed and Koob 1998). In another study, both D1-like receptor number and dopamine-stimulated adenylyl cyclase activity were decreased in rats exposed to prolonged cocaine self-administration for 30 days (6 h per day) (Graziella de Montis et al. 1998). In addition to reductions in the number or affinity of dopamine receptors, it is also possible that long-term cocaine exposure disrupts the coupling between DA receptors and their respective G-proteins (Nestler et al. 1990; Self et al. 1994) and/or produces alterations in intracellular signal-transduction pathways (Berhow et al. 1996; Fienberg et al. 1998; Self et al. 1998). The net result of these alterations may be a reduction in the efficacy of dopamine neurotransmission in the nucleus accumbens. Taken together, these findings suggest that LgA rats may consume more cocaine and sustain greater interstitial DA levels than ShA rats, to surmount a relative decrease in NAcc DA receptor function. The validity of this interpretation is supported by a study showing that blockade of D1-like receptors by SCH23390 concomitantly produces increases in cocaine intake and in DA-related electrochemical signals in the NAcc during self-administration (Egilmez et al. 1995).

In contrast to the findings described above, several electrophysiological studies have demonstrated that the net response of accumbal neurons to DA is enhanced following repeated exposure to either non-contingently administered (Henry et al. 1989; Henry and White 1991, 1995) or self-administered cocaine (White et al. 1995a; Peoples et al. 1999). The amount of cocaine exposure in these latter studies was similar to that given to the ShA group of the present study. However, the increased sensitivity to DA reported in these studies is relative to that observed in either cocaine-naïve subjects (Henry et al. 1989; Henry and White 1991, 1995; White et al. 1995b) or relative to the response observed during the second or third exposure to cocaine (Peoples et al. 1999). It is presently not known whether the responsiveness of accumbal neurons to DA is differentially altered by short versus long daily exposure to cocaine.

In addition to cocaine-induced changes in dopaminergic function, there is growing evidence that extended cocaine exposure alters the function of several other neurotransmitter systems known to be involved in modulating the behavioral effects of cocaine. For example, chronic cocaine exposure has been shown to alter glutamatergic function (White et al. 1995a; Pierce et al. 1996; Zhang et al. 1997; White and Kalivas 1998; Swanson et al. 2001; Xi et al. 2002), corticotropin-releasing factor (CRF) function (Goeders et al. 1990; Richter et al. 1995; Zhou et al. 1996; Ambrosio et al. 1997; Richter and Weiss 1999) and serotonergic function (Parsons and Justice 1993; Parsons et al. 1995; Simms and Gallagher et al. 1996; Perret et al. 1998; Mash et al. 2000). Cocaine-induced alterations in molecular mechanisms have also been well documented (Hope 1996; Nestler 2001). Although each of these factors likely contributes to an altered behavioral response to cocaine following chronic cocaine exposure, the relative involvement of these processes in the escalation of cocaine self-administration is presently unknown.

It may be argued that the difference in the self-administration rate between ShA rats and LgA rats simply could result from the extended experience of the LgA rats with the operant behavior and not from the extended exposure to the drug per se. To test this hypothesis, two groups of rats were exposed to cocaine 6-h per day. The first group (or SA group) was allowed to self-administer cocaine during the whole 6-h testing session. The second group (or YOKED group) was allowed to self-administer cocaine only during the first hour of each daily testing session. For the remaining 5 h of the session, the operant lever was retracted and these animals could not self-administer the drug. Instead each animal was yoked to a SA rat and thus received a non-contingent cocaine infusion each time the SA rat pressed on the lever. Thus the only difference between YOKED and SA animals was the level of experience with the operant behavior. It was found that in both groups the first-hour cocaine intake escalated at the same rate between-days and reached the same maximum level (unpublished findings). This indicates that the escalation of cocaine self-administration presently observed in the LgA group is not the result of greater experience with the operant behavior.

Deficit in brain reward function and cocaine intake escalation

The hypothesis that animals seek to regulate dopaminergic neurotransmission at some optimal level raises important questions about the effects of sustained increases in interstitial DA concentrations on NAcc processing of hedonic stimuli. One of the best-documented effects of cocaine is to lower the threshold for intracranial self-stimulation (ICSS) reward (Esposito et al. 1978; Kornetsky and Esposito 1981; Bauco and Wise 1997), an effect mediated by stimulation of DA receptors (Kita et al. 1999). As, under normal conditions, brain reward pathways are activated by environmental stimuli, the lowering of ICSS reward thresholds by cocaine may reflect a facilitated activation of these systems by otherwise ineffective environmental stimuli, a phenomenon that may suffice to explain the rewarding properties of drugs of abuse. Thus, animals may seek to sustain increased DA neurotransmission to increase the functional reactivity of brain reward pathways to the environment. As proposed recently, however, repeated prolonged exposure to cocaine self-administration compromises such regulatory processes by the establishment of a chronic deficit in the basal functioning of brain reward pathways (Koob and Le Moal 1997, 2001). This hypothesis is based on numerous studies showing that acute withdrawal from a single binge-like exposure to drugs of abuse produces a transient elevation in ICSS reward thresholds that opposes the effect ordinarily produced by the drug (Leith and Barrett 1976; Markou and Koob 1991; Schulteis et al. 1995; Wise and Munn 1995; Epping-Jordan et al. 1998b). More recent data show that this transient counter-adaptive reaction fails to return to normal levels between repeated, prolonged exposure to cocaine self-administration, creating a greater and greater elevation in ICSS thresholds that persists for several days and that is highly correlated with cocaine intake escalation (Ahmed et al. 2002). The establishment of a new basal ICSS threshold, above initial pre-escalation levels, suggests that LgA rats increase their cocaine intake in an effort to maintain a comparable activation of the reward system as experienced prior to escalation. Future studies are required to determine whether this compensatory phenomenon is driven by a decrease in DA receptor function in the NAcc, as discussed here, or via other alterations in the function of the reward pathway.

Summary

The present findings show that animals with escalated cocaine intake maintain higher steady levels of DA during self-administration (about two times higher) than do animals that have stable and moderate levels of intake. This large difference between groups is not associated with an altered temporal profile of brain cocaine levels or an altered effect of cocaine on interstitial DA levels in the NAcc. It is hypothesized that animals with escalating cocaine use need to maintain increased steady levels of DA in the NAcc to counteract a chronic deficit in brain reward function. Whether or not this reward deficit directly results from alterations in DA receptor function remains to be determined. A critical challenge for future experimentation will be to understand how the continued presence of increased interstitial levels of DA during cocaine self-administration specifically alters NAcc processing of reward-related stimuli from the environment and how this processing can be compromised in animals with escalated drug use.

Acknowledgements

This work was supported by the CNRS, France (SHA) and by the NIH, USA (DA04398, GFK; DA11004, LHP). The authors would like to thank Robert Lintz and David Stouffer for their invaluable technical assistance and Mike Arends for assistance with manuscript preparation. This is manuscript number 13354-NP from The Scripps Research Institute.

Ancillary