OPIOID RECEPTORS IN THE BASOLATERAL AMYGDALA BUT NOT DORSAL HIPPOCAMPUS MEDIATE CONTEXT-INDUCED ALCOHOL SEEKING

Abstract

Contexts associated with the availability of alcohol can induce craving in humans and alcohol seeking in rats. The opioid antagonist naltrexone attenuates context-induced reinstatement (renewal) of alcohol seeking and suppresses neuronal activation in the basolateral amygdaloid complex and dorsal hippocampus induced by such reinstatement. The objective of this study was to determine whether pharmacological blockade of opioid receptors in the basolateral amygdala or dorsal hippocampus would attenuate the context-induced reinstatement of alcohol seeking. Rats were trained to self-administer alcohol in one context (Context A), extinguished in a distinct context (Context B) and then tested for reinstatement of alcohol seeking in A and B contexts. Prior to the test session, rats were bilaterally microinjected with 0, 333 or 1000 ng (total) naloxone methiodide into the basolateral amygdala or dorsal hippocampus. Naloxone methiodide in the amygdala, but not the hippocampus, dose-dependently suppressed context-induced reinstatement. This suggests that opioid transmission in the basolateral amygdaloid complex is an important mediator of context-induced alcohol seeking.

Introduction

The association of an otherwise neutral stimulus with a drug of abuse, such as alcohol, can render it motivationally salient. For an abuser, it can induce craving and precipitate relapse [28]. In animal models of relapse, extinguished alcohol seeking can be induced by a variety of alcohol-associated cues: response-contingent discrete cues [29, 41], discriminative cues that are not response contingent [15, 24], and context [8, 47, 49]. A context is a multimodal cue complex that can gate responding to discrete [47] or discriminative [8] conditioned stimuli, and may even be processed by a separate – albeit, likely overlapping – brain circuitry [6, 10, 17, 37].

Although the circuitry mediating cue-induced alcohol craving (humans) or seeking (nonhumans) is complex, much evidence has implicated the endogenous opioid system. In humans, naltrexone – a relatively nonselective opioid receptor antagonist – reduces craving induced by alcohol associated cues [39, 42, 45], delays the onset of relapse [2], and reduces heavy drinking days and the number of drinks consumed on those days [40]. In animal models of relapse, systemic administration of naltrexone blocks alcohol seeking in rats induced by a variety of cues [9, 24, 29], including context. In addition, the activation of specific brain sites induced by an alcohol associated cue is reversed with systemic naltrexone pre-treatment [15, 36]. The literature also suggests that the effect of naltrexone does not apply to all types of seeking behaviour; it does not affect sucrose seeking [9] or stress-induced alcohol seeking [26, 29].

To study the involvement of opioidergic mechanisms in context-induced reinstatement of alcohol seeking, we have used a “renewal” procedure [7] that has been adapted to study context-induced drug seeking [13, 14]. In this procedure, rats are trained to self-administer a drug in one context in the presence of a discrete cue paired with each drug delivery. During the extinction phase, lever pressing in the presence of the discrete cue is extinguished in a different, non-drug context. Rats are subsequently exposed to the original drug self-administration context in the presence of the discrete cue and presses on the lever previously associated with drug delivery are measured. This procedure has been demonstrated to reliably reinstate seeking for a variety of drug and natural rewards [5, 14].

Using the renewal procedure, we have previously shown that systemic administration of naltrexone and other opioid antagonists can attenuate context-induced alcohol seeking [36, 37]. Using c-fos gene mapping, we have also correlated activity in the lateral and basolateral nuclei of the amygdala (together, nuclei that form the basolateral amygdaloid complex) and the dorsal hippocampus with context-induced reinstatement and its attenuation by naltrexone [36]. As these regions have a high density of opioid receptor binding [32], they may be key brain sites where opioid receptors play a critical role in mediating context-induced reinstatement. However, direct evidence to implicate opioid receptors in these regions in the reinstatement of alcohol seeking is lacking.

The aim of the present study is to assess the effect of opioid receptor blockade in the basolateral amygdaloid complex and dorsal hippocampus on context-induced reinstatement of alcohol seeking. Based on our examination of patterns of neuronal activation using c-fos mRNA mapping [36], we hypothesize that blocking opioid receptors in these brain sites will attenuate context-induced reinstatement and provide the first direct evidence that opioid transmission in these sites mediate alcohol seeking in an animal model of relapse.

Methods

Subjects

Male Wistar rats (Charles River, St-Constant, QC, Canada), weighing approximately 275g, were individually housed under a 12-hr light-dark cycle (lights on at 7:00 a.m.). Free access to food and water was available in the home cage and the temperature was maintained at 21±1°C. The experimental procedures followed the “Principles of laboratory animal care” (NIH publication no. 85-23, 1996) and were approved by the local animal care and use committee.

Materials and Apparatus

The operant chambers were constructed locally and were equipped with two levers, symmetrically centred on a side panel and a houselight near the ceiling that was illuminated at the start of the 1-h session and turned off at the end of the session. The Plexiglas chambers were enclosed in lightproof, sound-attenuating boxes equipped with exhaust fans. During self-administration sessions, responding on one lever (an active lever) activated an infusion pump (Razel Sci., Stamford, CT), while responding on the other lever (an inactive lever) was recorded, but had no effect. Activation of the infusion pump resulted in the delivery of 0.19 ml of alcohol into a drinking receptacle located between the two levers and initiated a 5-sec timeout period. During the time out, a key light over the active lever was illuminated, the houselight was turned off and white noise was emitted from a speaker.

Horizontal locomotor activity was monitored in stainless-steel chambers with wire mesh floors: 40 cm (L) × 25 cm (D) × 20 cm (H). Each chamber possessed two infrared photocells mounted 3 cm above the floor that divided it into three equally sized compartments. Photobeam interruptions were recorded and tabulated automatically by computer. Locomotor activity was inferred from the total number of photobeam interruptions.

Drugs

Alcohol solution was prepared by diluting 95% ethanol (Commercial Alcohols Inc., Tiverton, ON, Canada) in tap water. Naloxone methiodide (Sigma-Aldrich, St-Louis, MO, U.S.A) was dissolved in sterile saline and infused bilaterally into the basolateral amygdala or dorsal hippocampus in a volume of 0.5 μl/injection site.

Naloxone methiodide (NLX-MI) was selected as the opioid receptor antagonist because it diffuses slowly from the infusion site and does not cross the blood-brain barrier [46]. The concentrations selected, 0.7 or 2.1 mM NLX-MI were injected in 0.5 μl volume (equivalent to 167 or 500 ng NLX-MI per injection site). These doses and volume considers the affinity of this ligand [4, 31, 48] and are within the range used in behavioural studies [3, 11, 12, 21].

General Procedures

Surgery and microinjection

Surgery was performed under sodium pentobarbital anaesthesia (65 mg/kg IP). Using standard stereotaxic techniques, 23 gauge stainless-steel guide cannulae (Plastics One, Roanoke, VA, USA) were implanted bilaterally and affixed to the skull using dental acrylic and jewellers screws. The coordinates for cannulae targeting the basolateral amygdala (Experiment 1) were (from bregma): anterior/posterior: −2.5 mm; lateral/medial: ±5.0 mm; dorsal/ventral: −6.6 mm. The coordinates for cannulae targeting the dorsal hippocampus (Experiment 2) were (from bregma, using a 15° medially aimed angle): anterior/posterior: −3.2 mm; lateral/medial: ±3.5 mm; dorsal/ventral: −2.5 mm. After surgery, 30 gauge stainless-steel obturators were inserted in the cannulae to maintain patency and rats were allowed 9–14 days in their home cages to recover. To provide relief from pain following surgery and on the initial days of recovery, rats were administered 5mg/kg ketorolac, SC twice daily.

Intracranial infusions were administered with a 5 μl Hamilton microsyringe connected via polyethylene tubing (PE20) to 31 gauge injectors (Plastics One, Roanoke, VA, USA) that extended 2 mm (amygdala) or 1 mm (hippocampus) below the guide cannulae. Infusions occurred over the course of 120 s and injectors remained in place for an additional 60 s. The test session began 5-min following the termination of the infusion.

At the end of testing rats were euthanized with CO₂. Targets were verified by infusing 0.5 μl of 0.1% bromophenol blue (Sigma-Aldrich) in dH₂O, then extracting and snap freezing brains in isopentane (2-methylbutane). Sites of interest were sliced into 20 um coronal sections using a cryostat. Rats with infusion sites outside of the lateral or basolateral nuclei (Experiment 1), or the dorsal hippocampus (Experiment 2) were excluded from the final analyses.

Training

Naïve rats were first given a choice of alcohol solution or tap water in Richter tubes for 30 min/day in drinking cages. The alcohol solution was provided in escalating concentrations: 3% (w/v) for the first 5 days, 6% (w/v) for the next 5 days and 12% (w/v) for the next 10 days. Subsequently, rats were trained to operantly self-administer 12% alcohol (w/v) in one of two counterbalanced contexts distinguishable on a number of sensory properties. One context had 10 μl peppermint extract (McCormick, London, ON, Canada) applied to gauze at the top of the cage, a white house light, wallpaper with vertical bars, fans off with the enclosure door slightly ajar, a wire-grid floor and testing that occurred between 8:00 and 10:00 am. The other context was unscented, had a green house light, clear walls, fans on, a bar floor and testing that occurred between 10:00 am and 12:00 pm. Sessions were 1-hr/day, 5d/week and began on a fixed ratio (FR) -1 schedule of reinforcement for 2 weeks, followed by a FR-2 schedule for 1 week, and a FR-3 schedule for 2 weeks. In cases where alcohol was left over in the drinking receptacle after a self-administration session, it was measured and subtracted in the calculation of intake. Rats underwent surgery and recovery over a three week period and were re-introduced to alcohol self-administration on a FR-3 schedule for a minimum of seven days. Once rats showed stable alcohol self-administration at FR-3 (SD < 20% of the mean for two consecutive days for each rat), they received 1-h/day extinction sessions in the alternate (B) context. Except for the shift in context and the inactivation of alcohol infusion pumps, the extinction procedure resembled that of self-administration. All rats were habituated to the infusion procedure (handled in the same manner but not infused) on the four days preceding the initial test day, and every non-test day thereafter. The extinction criterion was at least two consecutive days with 15 or fewer lever presses.

Experiment 1: Role of Opioid Receptors in the Basolateral Amygdala

Thirty-five rats with extinguished lever pressing were matched for alcohol intake on the last three days of self-administration and separated into three groups (n = 11–12 per group). On the test days, rats were pre-treated with bilateral infusions of 0, 167 or 500 ng NLX-M in 0.5 μl sterile saline per hemisphere through cannulae aimed at the basolateral amygdala and tested for reinstatement in the context where they previously self-administered alcohol (Context A, ABA) and the extinction context (Context B, ABB) in counterbalanced order with the extinction criterion met between tests.

A separate cohort of rats (n = 9) were habituated to locomotor chambers (1h/day), and handling associated with infusion procedures (without actual infusions) for five days, and then tested for locomotor activity following bilateral infusions of 0, 167 or 500 ng NLX-M in 0.5 μl sterile saline per hemisphere in a counterbalanced repeated-measures design separated by two habituation days between tests.

Experiment 2: Role of Opioid Receptors in the Dorsal Hippocampus

Thirty-one rats with extinguished lever pressing were matched for alcohol intake on the last three days of self-administration and separated into three groups (n = 10–11 per group). On the test days, rats were pre-treated with bilateral infusions of 0, 167 or 500 ng NLX-M in 0.5 μl sterile saline per hemisphere through cannulae aimed at the dorsal hippocampus and tested for reinstatement in Context A (ABA) and Context B (ABB) in counterbalanced order with the extinction criterion met between tests.

Statistical Analyses

All results are expressed as mean ± SEM. Reinstatement of alcohol seeking was analysed using a three-way mixed ANOVA with drug condition as the between factor and context (ABB vs. ABA) and lever (active vs. inactive) as the within factors. The dependent variable was total number of lever presses in the 1-h test session. Locomotor activity was analysed using a one-way repeated measures ANOVA with drug condition as the repeated factor. The dependent variable was total photo-beam interruptions in the 1-h test session. Multiple comparisons were performed post hoc using the Holm-Sidak method. Significance for all tests was regarded at p < 0.05.

Results

Table 1 shows the average total numbers of active and inactive lever presses, alcohol reinforcements delivered and g/kg ethanol consumed on the last three days of self-administration prior to extinction for rats included in the final analyses. There were no significant differences between groups.

Table 1.

Operant and drinking behaviour during the last three days of self administration

	n	Active lever presses	Inactive lever presses	Reinforcements delivered	Ethanol intake (g/kg)
Experiment 1
vehicle	11	65.9 ± 8.7	4.2 ± 0.9	19.2 ± 2.5	0.71 ± 0.09
333 ng NLX-MI	10	53.9 ± 6.4	3.7 ± 1.4	16.1 ± 2.1	0.67 ± 0.09
1000 ng NLX-MI	10	64.7 ± 12.5	1.2 ± 0.5	17.8 ± 3.1	0.71 ± 0.11
Experiment 2
Vehicle	8	57.3 ± 5.9	1.9 ± 0.8	17.9 ± 1.7	0.90 ± 0.06
333 ng NLX-MI	10	59.4 ± 5.7	3.7 ± 2.0	17.7 ± 1.5	0.80 ± 0.11
1000 ng NLX-MI	7	53.0 ± 3.9	1.2 ± 0.6	16.5 ± 1.1	0.92 ± 0.06

Experiment 1: Role of Opioid Receptors in the Basolateral Amygdala

For rats infused in the amygdala (see Fig. 1), there was a main effect for test condition (ABA >ABB; F_1,28 = 18.64, p < 0.01), drug dose (1000 > 0 μg; F_2,28 = 3.54, p < 0.05), and lever (active > inactive; F_1,28 = 37.74, p < 0.01). There was also a significant three-way interaction (F_2,28 = 4.28; p < 0.05), showing that vehicle- and 333 ng NLX-MI-treated rats pressed the active lever significantly more in the ABA condition than the ABB condition (p < 0.05). This indicated a reinstatement of alcohol seeking. However, reinstatement was not observed with rats infused with 1000 ng NLX-MI. Within the ABA condition alone, there was a significant difference in active lever presses between rats infused with vehicle and 1000 ng NLX-MI, but not 333 ng NLX-MI (p < 0.05), indicating a dose-dependent attenuation of reinstatement by NLX-MI. There were no group differences for inactive lever presses.

Figure 1.

(A) A representative photomicrograph of the injection site (A) and illustrations of bilateral injection sites for all the animals (B) with intra-amygdala infusions. Active (C) and inactive (D) lever presses on the reinstatement test days in the extinction (ABB) and self-administration (ABA) contexts following bilateral intra-amygdala infusions of 0, 333 and 1000 ng total naloxone methiodide (NLX-MI) per rat. *denotes a significant difference from the ABB condition (p < 0.05), †denotes a significant difference from the vehicle treated group in the ABA condition (p < 0.05).

To test whether the effect of NLX-MI on context-induced reinstatement was due to a non-specific suppression in responding, a separate cohort of rats (n = 8) was administered intra-amygdala infusions of vehicle, 333 ng NLX-MI and 1000 ng NLX-MI per rat and tested in locomotor boxes in counterbalanced order. Total photocell beam counts were 280 ± 61 (vehicle), 298 ± 56 (333ng NLX-MI) and 341 ± 78 (1000 ng NLX-MI). There was no significant difference between treatments on total photobeam interruptions (F_2,14 = 1.32, p > 0.05).

Experiment 2: Role of Opioid Receptors in the Dorsal Hippocampus

For rats infused in the dorsal hippocampus (see Fig. 2), there was a main effect for test condition (ABA >ABB; F_1,22 = 45.08, p < 0.01) and lever (active > inactive; F_1,22 = 84.99, p < 0.01), and a two-way interaction between these (F_1,22 = 4.28, p < 0.01). An analysis of this interaction revealed that active lever presses were significantly greater in the ABA than ABB condition regardless of drug dose, but no differences existed for inactive lever presses. This demonstrates that the ABA condition induced a reinstatement of alcohol seeking across groups and this was unaffected by NLX-MI. There were no other main effects or interactions.

Figure 2.

(A) A representative photomicrograph of the injection site (A) and illustrations of bilateral injection sites for all the animals (B) with intra-hippocampal infusions. Active (C) and inactive (D) lever presses on the reinstatement test days in the extinction (ABB) and self-administration (ABA) contexts following bilateral intra-dorsal hippocampus infusions of 0, 333 and 1000 ng total naloxone methiodide (NLX-MI) per rat. *denotes a significant difference from the ABB condition (p < 0.05).

Discussion

The main finding in this study is that opioid receptor blockade within the basolateral amygdaloid complex significantly reduces context-induced reinstatement of alcohol seeking. This effect was not due to alterations in motoric activity since it had no effect on locomotor behaviour or on the inactive lever. In contrast, opioid receptor blockade in the dorsal hippocampus did not produce a significant attenuation of alcohol seeking. This is the first direct evidence supporting that opioid receptors in a specific brain region mediate alcohol seeking in a model of relapse.

Substantial evidence suggests that the endogenous opioid system plays a major role in mediating alcohol seeking induced by a variety of cues in laboratory animals. For example, systemic administration of naltrexone can attenuate or block alcohol seeking induced by response-contingent discrete cues [29], non-response-contingent discriminatory cues [24] and contexts [8] that signal alcohol availability. While it also attenuates cue-induced reinstatement of amphetamine seeking [1], naltrexone only modestly and at higher doses attenuates reinstatement for cocaine [9] and nicotine [30]. As naltrexone has no effect on stress-induced reinstatement of alcohol seeking [26, 29] or cue-induced reinstatement of sucrose seeking [9], opioidergic mechanisms appear to play a role that is particularly important for seeking elicited by alcohol-conditioned stimuli. The present and preceding [36] studies from our laboratory are aimed at identifying where in the brain opioids mediate these effects, particularly in context-induced reinstatement.

The present findings correspond to patterns of neuronal activation in the amygdala induced by re-exposure to a context previously associated with alcohol self-administration. Using c-fos mRNA expression as a marker of neuronal activation, we previously found that a context associated with the availability of alcohol reinstated alcohol seeking and activated cells residing in the lateral and basolateral nuclei of the amygdala [36]. Other studies have similarly demonstrated that context-induced reinstatement of seeking for sucrose [18], beer [19] and cocaine [20] elevated c-fos in the basolateral or lateral amygdala. Moreover, we demonstrated that treatment with 1 mg/kg naltrexone SC substantially diminished neuronal activation in the lateral nucleus. Based on the correlation of neuronal activation with seeking and opioid blockade, we reasoned that opioid receptors in this region regulated neuronal activity mediating context-induced reinstatement. The basolateral and lateral nuclei are rich in mu-opioid (MOP) and delta-opioid (DOP) receptors, respectively [43]. The present findings provide direct evidence that these receptors regulate context-induced alcohol seeking.

The present findings do not support the involvement of the dorsal hippocampus in mediating context-induced reinstatement. NLX-MI administered directly in the hippocampus did not significantly alter alcohol seeking, despite a dense expression of opioid receptors in the pyramidal layer [32]. Moreover, both context and discriminative cues respectively induced naltrexone-reversible c-fos mRNA and c-fos protein expression in the dorsal hippocampus, particularly in the CA3 subregion [15, 36]. The present findings indicate that naltrexone, when administered systemically, reduces context-induced reinstatement of alcohol seeking by blocking opioid receptors in extrahippocampal regions, such as the basolateral amygdala. They likewise suggest that the effects of naltrexone on c-fos mRNA induced by an alcohol-associated context occur through an action on extrahippocampal opioid receptors.

The role of opioids in the circuitry mediating contextual learning and its expression has yet to be properly addressed, despite an extensive learning literature. A large body of evidence from the conditioning literature has pointed to the basolateral amygdaloid complex and the hippocampus, which are reciprocally connected [44], as being critically involved in reinstatement induced by re-exposure to conditioned cues associated with drugs or other stimuli. Inactivation or lesions of the basolateral amygdala and dorsal hippocampus have been demonstrated to impair the expression of fear to a shock-conditioned context [25, 33, 34] and prevent the context-induced reinstatement of cocaine seeking [17]. However, the basolateral amygdala, but not the hippocampus, has also been implicated in the reinstatement of cocaine seeking [17] or conditioned fear [25, 33, 35] induced by discrete cues. The present findings demonstrate that opioid receptor activation in the basolateral amygdaloid complex play a role in this circuitry, at least as it pertains to context-induced alcohol seeking. We did not test here whether opioid receptors in the basolateral amygdala mediate reinstatement induced by discrete cues alone. However, this would be an interesting question to explore considering that 1) systemically administered naltrexone effectively attenuates alcohol seeking induced by discrete cues [29] and 2) the activation of neurons residing in the basolateral amygdala have been implicated in conditioned responses to discrete cues paired with both appetitive and aversive stimuli [17, 25, 33]. It is also important to consider that the neuronal mechanisms that underlie context-induced reinstatement may not be the same for all classes of drug. For example, others have associated context-induced alcohol seeking with the activation of hypocretin (orexin) neurons in lateral hypothalamus, but not c-fos induction in ventral medial prefrontal cortex, while the opposite is true for cocaine seeking [19, 20]. Therefore, the generalizability of our findings on the role of opioid receptors in context-conditioning, as it would pertain to other conditioned or unconditioned stimuli, remains to be elucidated.

The limited anatomical specificity of targeting these complex structures and their functionally distinct subregions may be a factor in our results. Although NLX-MI is less lipophilic than naloxone and does not cross the blood brain barrier [46], it is likely that the spread of the drug from the injection site to adjacent nuclei – including the central nucleus – may have contributed to our findings. Although opioid receptors are not abundant in the central amygdala [43], studies have shown that their blockade here can inhibit alcohol self-administration [16, 21]. Also, the basolateral amygdala itself can be divided into rostral and caudal subregions, with the rostral subsection likely playing a more significant role in cue-induced drug seeking [23, 38]. This was not dissociated in the present study nor were the cornu Ammonis (CA) subregions of the dorsal hippocampus. Our analysis of neuronal activation specifically implicated the CA3 subregion of the hippocampus; however, others have implicated other hippocampal subregions in reinstatement [15] and spread of NLX-MI likely reached pyramidal cells of CA3 [46].

A final consideration is the role of opioid receptor subtypes. This study does not differentiate between opioid receptor subtypes. NLX-MI demonstrates a much greater affinity for MOP than DOP receptors [27]. However, using systemically delivered antagonists, we have more strongly implicated DOP receptor activation in mediating context-induced alcohol seeking than MOP receptors, although both may play a role [37]. Others have shown that both DOP and MOP receptors in the basolateral amygdala mediate alcohol self-administration in Wistar rats [22]. Future studies should be aimed at determining the individual contributions of opioid receptor subtypes in this region on the context-induced reinstatement of alcohol seeking.

In conclusion, the present findings demonstrate that opioid receptors in the basolateral amygdaloid complex play a role in mediating context-induced reinstatement of alcohol seeking. On the other hand, opioid receptors in the dorsal hippocampus are not involved in the reinstatement effect. To our knowledge, this is the first evidence implicating opioid receptors in a specific brain region in a relapse model of alcohol seeking and supports further exploration into the use of opioid antagonists as a treatment aid for alcohol relapse.