Effects of intranucleus accumbens shell administration of dopamine agonists and antagonists on cocaine-taking and cocaine-seeking behaviors in the rat
- First Online:
- Cite this article as:
- Bachtell, R.K., Whisler, K., Karanian, D. et al. Psychopharmacology (2005) 183: 41. doi:10.1007/s00213-005-0133-1
- 285 Views
Dopamine signaling in the nucleus accumbens (NAc) plays an important role in regulating drug-taking and drug-seeking behaviors, but the role of D1- and D2-like receptors in this regulation remains unclear.
Our objective was to study the role of NAc D1- and D2-like receptors in the reinstatement of cocaine-seeking behavior and the regulation of stabilized cocaine intake in rats.
Using a within-session reinstatement procedure, whereby animals self-administer cocaine (90 min) and extinguish responding (150 min) in a single session, we assessed the ability of NAc microinfusions of the D1 agonist SKF 81297 and the D2 agonist 7-OH-DPAT to reinstate extinguished cocaine seeking. The effects of the D1 antagonist SCH 23390 and the D2 antagonist eticlopride pretreatment on agonist- and cocaine-primed reinstatement were also measured. Similar agonist and antagonist treatments were tested for their ability to modulate stabilized cocaine and sucrose self-administration.
Intra-NAc infusions of either SKF 81297 (0.3–3.0 μg) or 7-OH-DPAT (1.0–10.0 μg) dose-dependently reinstated cocaine seeking with greater efficacy in the medial core than in the shell subregion and at doses that also stimulated locomotor behavior. Intra-NAc shell infusions of SCH 23390 (1.0 μg) and eticlopride (3.0–10.0 μg) blocked cocaine-primed reinstatement (2.0 mg/kg, i.v.) and indiscriminately blocked reinstatement induced by either intra-NAc D1 or D2 agonists. Doses of agonists that triggered reinstatement failed to alter stabilized cocaine intake, whereas doses of antagonists that blocked reinstatement increased cocaine intake in the shell.
Both D1 and D2 receptors in the NAc play a prominent, and perhaps cooperative, role in regulating cocaine-taking and cocaine-seeking behaviors.
KeywordsD1 receptorD2 receptorSelf-administrationCravingRelapseReinstatementRewardIncentive motivation
Dopamine D1 and D2 receptors mediate opposite effects on the reinstatement of cocaine-seeking behavior after extinction of cocaine self-administration. Thus, in rodent studies, selective stimulation of D2 receptors is sufficient to strongly induce cocaine-seeking behavior, whereas selective stimulation of D1 receptors is virtually without effect, even at equipotent locomotor-activating doses (Wise et al. 1990; Self et al. 1996; De Vries et al. 1999; Dias et al. 2004). Moreover, stimulation of D1 receptors completely attenuates cocaine's ability to reinstate cocaine seeking, whereas stimulation of D2 receptors facilitates cocaine-induced reinstatement (Self et al. 1996; Alleweireldt et al. 2002). This D1/D2 dichotomy also regulates cocaine seeking in nonhuman primates (Spealman et al. 1999; Khroyan et al. 2000) and may regulate craving responses in humans (Haney et al. 1998, 1999).
In the nucleus accumbens (NAc), D1 and D2 receptors are differentially expressed in distinct subpopulations of NAc neurons that project to different brain regions (Aubert et al. 2000; Lu et al. 1998; Steiner and Gerfen 1998), although there is evidence for colocalization in the same neurons (Aizman et al. 2000; Schwartz et al. 1998). D1 and D2 receptors also differ in their ability to couple stimulatory and inhibitory G-proteins that differentially regulate adenylyl cyclase activity (Lachowicz and Sibley 1997). These anatomical and functional differences suggest that D1 and D2 receptors in the NAc could contribute to the differential regulation of cocaine-seeking behavior found with systemic administration of D1 and D2 receptor agonists.
Previous studies have shown that dopamine receptors in the NAc have both necessary and sufficient roles in reinstating cocaine-seeking behavior in rats. Thus, intra-NAc microinfusions of dopamine reinstate cocaine seeking (Cornish and Kalivas 2000), and NAc infusions of a D1 antagonist block cocaine seeking induced by systemic priming injections of cocaine, but only when infused in the NAc shell, and not the core, subregion (Anderson et al. 2003). However, the relative ability of NAc D1 and D2 receptors to induce reinstatement is unknown.
Similarly, D1 and D2 dopamine receptors in the NAc may have an important role in regulating cocaine intake during self-administration when access to cocaine is relatively unrestricted (e.g., low fixed ratio reinforcement schedules). Thus, pretreatment with NAc infusions of either D1 or D2 receptor antagonists increases cocaine self-administration in a manner resembling reducing the injection dose in rats (Phillips et al. 1983; Maldonado et al. 1993; McGregor and Roberts 1993; Caine et al. 1995). Conversely, reverse dialysis of dopamine into the NAc, at concentrations similar to that achieved during cocaine self-administration, decreases cocaine intake (Hurd and Ponten 2000), but it is unclear whether local stimulation of either D1 or D2 receptors in the NAc is sufficient to produce a similar rate-decreasing effect.
We compared the ability of intra-NAc infusions of the D1-selective agonist SKF 81297 (>500-fold D1/D2; Andersen and Jansen 1990) and the D2-selective agonist 7-OH-DPAT (>1,000-fold D2–3/D1; Levesque et al. 1992; Gonzalez and Sibley 1995; Levant et al. 1996) to reinstate cocaine-seeking behavior. We utilized microinfusions in both the medial core and the shell subregion to determine whether local dopamine receptor activation would exhibit the same subregion specificity found with receptor blockade. We also compared the ability of the D1-selective antagonist SCH 23390 (>1,000-fold D1/D2; Andersen 1988) and the D2-selective antagonist eticlopride (>100,000-fold D2/D1; Hall et al. 1986) to block agonist- and cocaine-primed reinstatement of cocaine seeking when infused into the NAc shell. In addition, we determined whether doses of agonists and antagonists that were effective in modulating reinstatement would also modulate stabilized cocaine intake during self-administration when infused into the NAc shell.
Materials and methods
Animals and housing conditions
Male Sprague–Dawley rats initially weighing 275–325 g (Charles River Laboratories, Kingston, NY) were individually housed in wire cages with food and water available ad libitum, except during sucrose pellet self-administration. Experiments were conducted during the light cycle of a 12:12-h light:dark cycle (lights on at 0700 hours) in accordance with guidelines established by the National Institute of Health and the Institutional Animal Care and Use Committee at the University of Texas Southwestern Medical Center.
Surgical procedures and apparatus
To facilitate the acquisition of operant behavior, all rats were initially food-restricted to prevent weight gain and were trained to lever-press for sucrose pellets on a fixed ratio 1 (FR1) reinforcement schedule until acquisition criteria had been achieved (100 sucrose pellets for three consecutive tests). After lever-press training, animals were fed ad libitum for at least 1 day and were surgically implanted with a chronic intrajugular catheter and 26-gauge bilateral intracerebral guide cannulae, as previously described (Self et al. 1998). Guide cannulae were aimed at the NAc core (±1.5 mm lateral) or shell (±0.8 mm lateral) subregions positioned 1.7 mm anterior to bregma and −5.7 mm ventral to dura with level skull (Paxinos and Watson 1998). Dummy cannulae (33 gauge), cut to extend 1 mm beyond the guide cannulae, were left in place throughout the experiment. Animals were allowed 5–7 days to recover before experimentation. Self-administration and reinstatement testing were performed in operant test chambers (Med-Associates, East Fairfield, VT) equipped with two response levers and an infusion pump system as described (Self et al. 1998).
Prior to intracranial infusions, dummy cannulae were removed; infusions of 0.5 μl/side were administered through bilateral 33-gauge injection cannulae extending 1 mm beyond the tip of the guide over a 100-s period; and injection cannulae were left in place for 30 s to allow for diffusion into brain tissues before removal. Each intra-NAc treatment was separated by a minimum of 2 days of training to reestablish baseline performance. Animals received a maximum of 8–10 intracranial infusions to avoid excess tissue damage at the injection site.
Cocaine self-administration procedure
Following recovery from surgery, lever-trained animals performed cocaine self-administration in daily 4-h sessions for 5–6 days/week. A single lever-press response (FR1) at the active lever produced a 0.5-mg/kg intravenous (i.v.) injection of cocaine delivered in 0.1 ml over 5 s, concurrent with illumination of a cue light located above the lever while house light was extinguished. Each injection was followed by an additional 10-s “time-out” (TO) period when the house light remained off; active lever-press responses during the injection–TO period (15 s) had no scheduled consequence. Responses on the inactive lever were recorded but also had no consequence.
After a minimum of 10 cocaine self-administration sessions, the sessions were reduced to 90 min and animals demonstrating stable self-administration baselines (injection totals varied by <10% from the mean of three consecutive sessions) were given intra-NAc infusions of D1 or D2 agonists and antagonists as pretreatments immediately prior to the test session. In these experiments, animals were assigned to various treatment groups such that the mean baseline self-administration rates in each group were similar. The D1 agonist SKF 81297 (3.0 μg/side) and the D2 agonist 7-OH-DPAT (10.0 μg/side) were tested at doses that were maximally effective in reinstatement testing. Similar effective doses of the D1 antagonist SCH 23390 (3.0 μg/side) and the D2 antagonist eticlopride (10.0 μg/side) were tested based on results from reinstatement tests. Since these antagonist doses produced positive effects on cocaine intake, lower doses of both SCH 23390 (0.5 and 1.0 μg/side) and eticlopride (1.0 and 3.0 μg/side) were used to determine dose thresholds.
After a minimum of 10 cocaine self-administration sessions, animals were trained in a daily 4-h within-session reinstatement procedure. In this procedure, cocaine (0.5 mg/kg per injection) was available for the first 90 min of the session, followed by an extinction phase lasting 150 min when drug-paired lever responding produced only the injection cue light–TO sequence. Animals were trained in this procedure until responding during the final 60 min of the session had met extinction criteria (five responses or less at either lever for three consecutive daily sessions). The ability of intra-NAc infusions of phosphate-buffered saline (PBS), the D1 agonist SKF 81297 (0.3 and 3.0 μg/side), or the D2 agonist 7-OH-DPAT (1.0 and 10.0 μg/side) to reinstate drug-paired lever responding was measured in extinguished animals by removing them from the test chambers 5 min prior to the final hour of the test session, administering the intra-NAc infusion, and returning the animals to the test chambers, where the number of nonreinforced lever-press responses was recorded. In other experiments, the ability of intra-NAc infusions of the D1 antagonist SCH 23390 (0.3 and 1.0 μg/side) or the D2 antagonist eticlopride (1.0, 3.0, and 10.0 μg/side) to block cocaine-induced reinstatement was tested by infusing the antagonist immediately prior to a priming injection of 2.0 mg/kg of cocaine administered intravenously in 0.4 ml over 20 s. In addition, the ability of D1 (SCH 23390, 1.0 μg/side) or D2 (eticlopride, 10.0 μg/side) antagonists to block D1 (SKF 81297, 3.0 μg/side) or D2 (7-OH-DPAT, 10.0 μg/side) agonist-induced reinstatement was tested by sequential intra-NAc administration of antagonists and agonist infusions (combined volume 1.0 μl). Most animals in both self-administration and reinstatement experiments were tested with at least two different agonist or antagonist treatment regimens, counterbalanced for dose order (within treatment) and test drug (between treatments) across animals such that no two experiments were composed of the same animals.
Locomotor activity procedure
Locomotor activity was conducted in eight cocaine-trained animals 1 week after their last reinstatement test. Locomotor activity was assessed in a circular corridor (a 12-cm-wide runway) equipped with four photocell detectors located equidistantly at 90° intervals around the 1.95-m circumference. Each 6-h locomotor activity test session was divided into three 120-min phases (habituation, vehicle infusion, and dopamine agonist infusion). Following the habituation phase, animals were removed from the test chambers, given intra-NAc infusions of PBS, and returned to the chambers. Locomotor activity was then recorded for 120 min. After 2 h, animals were removed again, given intra-NAc infusions of either the D1 (SKF 81297, 3.0 μg/side) or D2 (7-OH-DPAT, 10.0 μg/side) agonist at a dose that was determined to be maximally effective in reinstatement tests, and returned to the chambers for the final 120 min of testing. Each animal was given both the D1 and D2 agonists on consecutive days in a counterbalanced order.
Sucrose self-administration procedure
Animals were surgically implanted with only the intra-NAc guide cannulae as described above and subsequently maintained on a restricted diet (fed after each daily test session) to prevent weight gain over the course of testing. Following acquisition of lever-press behavior (FR1) as described above, animals self-administered 50 sucrose pellets in daily sessions and were tested with intra-NAc infusions after response rates had stabilized (the time to self-administer 50 pellets varied by <20% of the mean of three consecutive tests). On the test day, animals were given bilateral intra-NAc infusions of SKF 81297 (3.0 μg/side), 7-OH-DPAT (10.0 μg/side), SCH 23390 (0.5, 1.0, and 3.0 μg/side), or eticlopride (1.0, 3.0, and 10.0 μg/side), corresponding to doses tested in cocaine self-administration experiments. Doses and test drugs were counterbalanced across animals as described above.
Nucleus accumbens infusion sites were determined in chloral-hydrate-anesthetized (400 mg/kg, i.p.) animals by infusing 0.5 μl of cresyl violet dye through bilateral injection cannulae as described above. Animals were immediately decapitated, brains were removed, and 0.8-mm-thick coronal slices were collected throughout the forebrain and analyzed under a dissecting microscope for the location of infusion sites according to the stereotaxic coordinates of Paxinos and Watson (1998). Micrographs for core and shell infusion sites were taken from cresyl-violet-stained brain sections (40 um) in animals that received nine NAc infusions of PBS, SKF 81297 (2×0.3 and 3.0 μg/side), and 7-OH-DPAT (2×1.0 and 10.0 μg/side), at least 2 days apart, prior to perfusion with formalin.
Self-administration and reinstatement data were analyzed with between-subject statistical analyses, since some animals exhibited catheter failure or failed to recover baseline response rates before completing all test doses within a treatment condition. Reinstatement data for each agonist in the core and shell subregions were analyzed using a three-factor region × dose × lever analysis of variance (ANOVA) with repeated measures on lever (drug-paired vs inactive). Subsequent reinstatement experiments in the NAc shell used a two-factor dose/drug × lever ANOVA with repeated measures on lever. Cocaine and sucrose self-administration data were analyzed with a two-factor dose/drug × treatment ANOVA with repeated measures on treatment (baseline vs treatment). Locomotor time course data following PBS and agonist infusions were pooled into 1-h bins (expressed as a percentage of baseline activity; photocell counts in the 1-h period preceding each infusion) and analyzed by two-factor region × time bin ANOVA with repeated measures on time bin. Interactive effects were followed by simple main effects analyses and post hoc tests with Dunnett's test (compared to PBS or the lowest inactive dose), Fisher's least significant difference (LSD) test (when multiple control groups were present), or Student's paired t tests, corrected for multiple comparisons (drug-paired vs inactive lever or baseline vs test day) where appropriate. Statistical significance was preset at p<0.05.
SKF 81297 [(±)-6-chloro-7,8-dihydroxy-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine hydrobromide], 7-OH-DPAT [(+)-7-hydroxy-N,N-di-n-propyl-2-aminotetralin], SCH 23390 [R-(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine], and eticlopride [S(−)-3-chloro-5-ethyl-N-[(1-ethyl-2-pyrrolidinyl)methyl]-6-hydroxy-2-methoxybenzamide hydrochloride] were purchased from Sigma (St. Louis, MO). All drugs were dissolved in PBS (pH 5.5) vehicle. Cocaine hydrochloride was obtained from the National Institute on Drug Abuse (Research Triangle Park, NC) and was dissolved in sterile-filtered physiological saline.
NAc infusions of D1 or D2 receptor agonists reinstate cocaine-seeking behavior
Mean number of cocaine self-injections during the 90-min self-administration phase of the within-session reinstatement procedure
Dopamine agonist reinstatement
Previous day mean (±SEM)
Treatment day mean (±SEM)
0.3 μg of SKF 81297
3.0 μg of SKF 81297
1.0 μg of 7-OH-DPAT
10.0 μg of 7-OH-DPAT
0.3 μg of SKF 81297
3.0 μg of SKF 81297
1.0 μg of 7-OH-DPAT
10.0 μg of 7-OH-DPAT
Dopamine antagonist–cocaine reinstatement
Saline mean (±SEM)
Cocaine mean (±SEM)
0.3 μg of SCH 23390
1.0 μg of SCH 23390
1.0 μg of eticlopride
3.0 μg of eticlopride
10.0 μg of eticlopride
Dopamine antagonist–agonist reinstatement
3.0 SKF 81297 mean (± SEM)
10.0 7-OH-DPAT mean (± SEM)
3.0 μg of SCH 23390
10.0 μg of eticlopride
The highest dose of 7-OH-DPAT increased inactive lever responding above PBS control levels in the shell (F2,32=3.32, p<0.05) and in the core (F2,33=3.65, p<0.05). However, drug-paired lever responding exceeded inactive lever responding at both doses of 1.0 μg/side (t12=2.79, p<0.05) and 10.0 μg/side (t9=4.79, p<0.01) in the shell and at the lower 1.0-μg/side dose in the core (t11=3.23, p<0.05). In contrast, SKF 81297 did not increase inactive lever responding above PBS-infused controls, and drug-paired lever responding exceeded inactive lever responding at the higher 3.0-μg/side dose in the shell (t19=3.88, p<0.01) and at both doses of 0.3 μg/side (t12=2.92, p<0.05) and 3.0 μg/side (t13=3.86, p<0.01) in the core.
The temporal profile of reinstatement produced by the D1 and D2 agonists in the NAc shell paralleled the profile of locomotor stimulation (Fig. 2e). Thus, infusions of SKF 81297 (3.0 μg/side) produced a low but sustained drug-paired lever responding throughout the 1-h reinstatement test session, whereas 7-OH-DPAT (10.0 μg/side) produced a transient but more robust drug-paired lever responding that peaked 15–30 min postinfusion. In the core, infusion of SKF 81297 induced transient increases in reinstatement that peaked after 15–30 min while locomotor stimulation endured, whereas 7-OH-DPAT transiently increased both reinstatement and locomotor activity (Fig. 2f).
NAc shell infusions of D1 or D2 receptor antagonists block cocaine-induced reinstatement
Crossblockade of D1- and D2-agonist-induced reinstatement by coinfusion with either D1 or D2 antagonists in the NAc shell
NAc shell infusions of D1 or D2 receptor agonists fail to alter stabilized cocaine self-administration
NAc shell infusions of D1 or D2 receptor antagonists increase cocaine self-administration
NAc infusions of the D2 antagonist eticlopride also produced a significant pretreatment × dose interaction (F2,19=6.57, p<0.01) and a main effect when compared to self-administration baselines (F1,19=11.88, p<0.01). Eticlopride pretreatments increased cocaine intake at the 10.0-μg/side dose (t6=2.86, p<0.01), but not at 1.0- or 3.0-μg/side dose, consistent with the dose threshold for the blockade of reinstatement. Eticlopride dose-dependently increased cocaine self-administration throughout the 90-min test session without disrupting the regularity of self-administration patterns (Fig. 6b and f). These doses of eticlopride produced no effect on sucrose pellet self-administration rates (n=6 animals/dose; data not shown).
Histological verification of infusion sites
The role of NAc dopamine receptors in cocaine seeking
Intra-NAc infusions of either D1- or D2-selective agonists dose-dependently reinstated cocaine-seeking behavior. The ability of the D2 agonist 7-OH-DPAT to reinstate cocaine seeking is consistent with the effects of systemic administration, whereas reinstatement induced by the D1 agonist SKF 81297 is opposite to the effects of systemic administration (Alleweireldt et al. 2002; Self et al. 1996). These differences are not attributable to differential psychomotor activation with systemic and intra-NAc infusions, since both routes of administration modulate cocaine-seeking behavior at doses that also stimulate locomotor activity. Instead, these results indicate that selective activation of either D1 or D2 receptors in the NAc is sufficient to trigger a cocaine-seeking response. Therefore, while D2 receptors in the NAc could contribute to reinstatement induced by systemic D2 agonist administration, it seems unlikely that NAc D1 receptors would contribute to the attenuation of cocaine seeking by systemic D1 agonist administration (but see discussion below).
Given that systemic cocaine injections preferentially increase dopamine levels in the NAc shell of rats (Hedou et al. 1999; Pontieri et al. 1995), and that blockade of D1 receptors in the shell, but not the core, has been shown to block cocaine-primed reinstatement of cocaine seeking (Anderson et al. 2003), we compared the ability of D1 and D2 receptor antagonists to block cocaine-induced reinstatement in the NAc shell. We found that blockade of either D1 or D2 receptors in the NAc shell attenuates cocaine seeking elicited by i.v. injections of cocaine. These results are similar to the effects of systemic D1 or D2 antagonist administration in rats (Alleweireldt et al. 2002; Weissenborn et al. 1996) and monkeys (Khroyan et al. 2000), indicating that D1 and D2 receptors in the NAc shell each have a critical role in cocaine-induced reinstatement of cocaine seeking. Interestingly, both D1- and D2-selective antagonists completely and indiscriminately attenuated reinstatement induced by either the D1- or D2-selective agonist when agonist/antagonist combinations were coinfused into the NAc shell. These results suggest that some degree of endogenous dopamine tone at both D1 and D2 receptors is required to enable either receptor to mediate the reinstatement of cocaine-seeking behavior. This notion is supported by well-established cooperative interactions between D1 and D2 receptors in the NAc on neuronal firing (Hopf et al. 2003; White 1987), locomotor activity (Dreher and Jackson 1989; Essman et al. 1993; Koshikawa et al. 1996), and reinforcement mechanisms (Ikemoto et al. 1997), and is consistent with the effects of systemic administration of D1 and D2 receptor ligands (Waddington and Daly 1993).
Differential regulation of cocaine intake by intra-NAc agonist and antagonist infusions
In contrast to their effects in reinstatement, neither the D1 nor the D2 agonist altered stabilized cocaine self-administration when infused into the NAc shell and thus failed to recapitulate the rate-decreasing or additive-like interaction with cocaine (prolonging the postinjection pause) typically found with systemic administration in rats (Caine et al. 1999; Hubner and Koob 1990; Pulvirenti and Koob 1994; Self et al. 1996, 2000). In contrast, both the D1 antagonist SCH 23390 and the D2 antagonist eticlopride increased stabilized cocaine intake when infused into the NAc shell, similar to the effect of systemic administration (e.g., Britton et al. 1991; Corrigall and Coen, 1991). These findings suggest that D1 and D2 receptors in the NAc shell are saturated with dopamine during cocaine self-administration, and receptor blockade reveals their inhibitory regulation of cocaine intake by shortening the pause between successive self-injections. However, given that reverse dialysis of dopamine into the NAc reduces cocaine intake (Hurd and Ponten 2000), coactivation of both D1 and D2 receptors within the NAc may be required for such localized effects. Alternatively, agonist-mediated inhibition of cocaine intake could require a more widespread activation of dopamine receptors within the NAc, or a concurrent activation of dopamine receptors in other brain regions. In this regard, dopamine antagonists increase cocaine intake when infused into several other brain regions, including the ventral tegmental area, prefrontal cortex, amygdala, and bed nucleus of the stria terminalis (Caine et al. 1995; Epping-Jordan et al. 1998; Hurd et al. 1997; McGregor and Roberts 1993, 1995; Ranaldi and Wise 2001; Sun and Rebec 2005).
The behavioral effects of D1 and D2 agonists and antagonists are not attributable to generalized effects on response rates, since they differentially altered lever-press responding during cocaine and sucrose self-administration and in reinstatement experiments. Thus, both D1 and D2 agonists increased responding in reinstatement and had no effect on cocaine self-administration; the D2 agonist decreased, rather than increased, responding for sucrose pellets, similar to the effect of systemic administration (Franklin and Tang 1995; Hoffman et al. 1990; McQuade et al. 2003). Conversely, the D1 and D2 antagonists decreased cocaine-induced responding in reinstatement, but increased cocaine self-administration while having no effect on sucrose self-administration. Therefore, changes in lever-press rates in self-administration and reinstatement experiments are likely due to motivational, rather than performance-related, influences.
Dopaminergic involvement of both core and shell in cocaine seeking
The D1 and D2 receptor agonists reinstated cocaine-seeking behavior when infused into either the core or the shell subregion, although more robust effects were found with core infusions. The ability of core infusions to induce greater cocaine seeking cannot be attributed to greater psychomotor activation, since absolute locomotor activities were similar following core and shell infusions. These results suggest that dopamine receptor activation in both subregions is sufficient to trigger cocaine-seeking behavior, whereas receptor activation in the shell, but not the core, is necessary for cocaine-induced reinstatement (Anderson et al. 2003). While this latter finding agrees with the ability of cocaine to increase dopamine levels selectively in the shell, exposure to unanticipated cocaine-associated cues increases dopamine selectively in the core (Ito et al. 2000). Thus, it is possible that dopaminergic neurotransmission in the shell and core differentially mediates cocaine- and cue-elicited relapse to cocaine seeking, respectively.
It is important to note that we obtained positive effects with dopamine agonist infusions in the medial core, whereas Anderson et al. (2003) obtained negative effects with antagonist infusions in the lateral core. Although we cannot rule out the potential for diffusion between subregions, the fact that agonist-induced locomotor activation had similar time courses in both the core and the shell suggests that the behavioral effects were mediated by receptors in the vicinity of the infusion site. In addition, the transient reinstatement produced by intra-NAc infusion of the D2 agonist is consistent with local receptor activity in either the core or the shell.
The shell subregion receives limbic innervation that may be more pertinent to the reinforcing actions of cocaine, while the core is thought to represent a major component of the motor circuit, but the distribution of D1 and D2 dopamine receptors does not differ substantially between subregions (Heimer et al. 1991). The role of each subregion in the reinstatement of cocaine-seeking behavior is less clear, since dopamine antagonists block cocaine-induced reinstatement in the shell, but γ-aminobutyric acid-receptor-mediated inhibition of the core also attenuates reinstatement (McFarland and Kalivas 2001). Our results suggest that dopaminergic input to both the core and shell is important. The fact that 7-OH-DPAT produced greater cocaine-seeking responses in the core is preliminary evidence that the D2 subtype of the D2 receptor family mediates this effect, since D3 receptors are not expressed in the core and D4 receptors are expressed at low levels in the NAc (Bouthenet et al. 1991; Van Tol et al. 1991; Seeman and Van Tol 1994; Diaz et al. 1995). Similarly, the D1—rather than the D5—subtype of the D1 receptor family could mediate the effects of SKF 81297, since D5 receptor messenger ribonucleic acid is absent in the rat NAc (Tiberi et al. 1991). However, conclusive evidence will require further work with subtype-selective ligands, and these data do not rule out the possible actions of D3, D4, or D5 receptors in other brain regions on the reinstatement of cocaine-seeking behavior.
Opposite effects of systemic and intra-NAc D1 agonist administration on reinstatement
One of the most intriguing findings is that intra-NAc infusions of the D1 agonist induce cocaine-seeking behavior, whereas systemic administration inhibits cocaine seeking. The most straightforward explanation for this difference is that D1 receptors in other brain regions mediate the inhibitory effects of systemic D1 agonist administration on reinstatement. However, we found that intra-ventral-tegmental-area infusions of SKF 81297 also reinstate cocaine seeking in rats (K. -H. Choi and D. W. Self, unpublished observations), an effect possibly mediated by releasing glutamate from afferent fibers onto dopamine neurons (Adell and Artigas 2004). In the dorsal medial (but not prelimbic) prefrontal cortex and amygdala, dopamine receptor antagonists have been shown to block cocaine seeking induced by systemic cocaine injections or cocaine cues (Capriles et al. 2003; McFarland and Kalivas 2001; Park et al. 2002; See et al. 2001; Sun and Rebec 2005; Weiss et al. 2000), indicating that dopamine receptors in these regions also play a necessary role in the reinstatement of cocaine-seeking behavior. Given that infusions of D1 receptor antagonists into these brain regions also increase cocaine self-administration, D1 receptors in multiple regions could contribute to inhibitory regulation, or satiation, of cocaine seeking during self-administration.
Thus, it is possible that the satiating effects of systemic D1 agonist administration require a simultaneous stimulation of D1 receptors in multiple brain regions. It is also possible that D1 receptors on NAc neurons have different neuromodulatory effects when D1 agonists are focally applied than when active throughout the brain. For example, D1-receptor-mediated activation of prefrontal cortical afferents to NAc neurons alters the response to local D1 receptor activation, which produces opposing effects depending on whether NAc neurons are in “up” or “down” states (O'Donnell 2003). If so, then such highly localized D1 receptor activation would not recapitulate the neuromodulatory effects of systemic D1 agonist administration or mesolimbic dopamine activity throughout the brain. Another caveat is that benzazepine-based D1 receptor agonists are reported to cause an amphetamine-like elevation in dopamine levels when microinjected into striatal brain regions (Tomiyama et al. 1995). This dopamine-releasing action is not found with systemic D1 agonist administration (Zetterstrom et al. 1986) and is not mediated by D1 receptors, but could underlie the prolonged locomotor activation found with intra-NAc infusions of SKF 81297 relative to 7-OH-DPAT. Reinstatement induced by such an amphetamine-like action in the NAc would be sensitive to blockade by either D1 or D2 receptor antagonists, since both antagonists blocked cocaine-induced reinstatement.
In summary, we found that microinfusion of either D1 or D2 receptor agonists into the medial core and shell of the NAc triggers cocaine-seeking behavior, and these effects are codependent on both D1 and D2 receptor activity. These results firmly establish a role for NAc D2 receptors in the induction of cocaine-seeking behavior and suggest that NAc D1 receptors may cooperate with D2 receptors in mediating cocaine seeking, while the inhibitory effect of systemic D1 agonist administration may involve D1 receptors in different or multiple brain regions. In contrast, stabilized cocaine self-administration is sensitive to either D1 or D2 receptor blockade in the NAc, but not D1 or D2 receptor stimulation, suggesting that NAc dopamine receptors are saturated during cocaine self-administration. These findings suggest that D1 and D2 receptors in the NAc play a major role in regulating both drug-taking and drug-seeking behaviors.
This work was supported by United States Public Health Services grants DA 10460, DA 08227, and DA 18481 (R.K.B.), and the Wesley Gilliland Professorship in Biomedical Research.