The Role of Functional Postsynaptic NMDA Receptors in the Central Nucleus of the Amygdala in Opioid Dependence

Abstract

Activation of ionotropic N-methyl-D-aspartate (NMDA)-type glutamate receptors in limbic system nuclei, such as the central nucleus of the amygdala (CeA), plays an essential role in autonomic, behavioral, and affective processes that are profoundly impacted by exposure to opioids. However, the heterogeneous ultrastructural distribution of the NMDA receptor, its complex pharmacology, and the paucity of genetic models have hampered the development of linkages between functional amygdala NMDA receptors and opioid dependence. To overcome these shortcomings, high-resolution imaging and molecular pharmacology were used to (1) Identify the ultrastructural localization of the essential NMDA-NR1 receptor (NR1) subunit and its relationship to the mu-opioid receptor (μOR), the major cellular target of abused opioids like morphine, in the CeA and (2) Determine the effect of CeA NR1 deletion on the physical, and particularly, psychological aspects of opioid dependence. Combined immunogold and immuoperoxidase electron microscopic analysis showed that NR1 was prominently expressed in postsynaptic (i.e., somata, dendrites) locations of CeA neurons, where they were also frequently colocalized with the μOR. A spatial–temporal deletion of NR1 in postsynaptic sites of CeA neurons was produced by local microinjection of a neurotropic recombinant adeno-associated virus (rAAV), expressing the green fluorescent protein (GFP) reporter and Cre recombinase (rAAV–GFP–Cre), in adult “floxed” NR1 (fNR1) mice. Mice with deletion of NR1 in the CeA showed no obvious impairments in sensory, motor, or nociceptive function. In addition, when administered chronic morphine, these mice also displayed an acute physical withdrawal syndrome precipitated by naloxone. However, opioid-dependent CeA NR1 knockout mice failed to exhibit a conditioned place aversion induced by naloxone-precipitated withdrawal. These results indicate that postsynaptic NMDA receptor activity in central amygdala neurons is required for the expression of a learned affective behavior associated with opioid withdrawal. The neurogenetic dissociation of physical and psychological properties of opioid dependence demonstrates the value of combined ultrastructural analysis and molecular pharmacology in clarifying the neurobiological mechanisms subserving opioid-mediated plasticity.

I. Introduction

Opioid use can lead to an interrelated complex of adverse neural and behavioral adaptations that include tolerance, addiction, and dependence (Christie, 2008). Dependence, the focus of this paper, is typically studied within the framework of the withdrawal state produced by cessation of opioid exposure. However, dependence can also have an impact that endures beyond the immediate experience of withdrawal. One persistent effect of dependence is the result of learned associations between the noxious affective properties of withdrawal and environmental cues (O’Brien, 2008), a topic that is an important theme of this chapter. Although the neural mechanisms subserving the complex aspects of dependence are far from clear, there is reason to believe that interactions between the ionotropic N-methyl-D-aspartate (NMDA)-type glutamate receptor and the mu-opioid receptor (μOR), the major target of abused opioids like morphine, play a key role in opioid-adaptive processes. Moreover, the central nucleus of the amygdala (CeA), a critical coordinator of behavioral and emotional processes impacted by opioid exposure (LeDoux, 2000; Phelps and LeDoux, 2005), is a key anatomical substrate of NMDA-dependent forms of neural plasticity that may parallel the development of drug dependence (Hyman et al., 2006). After providing some important background information, this paper describes recent results characterizing the synaptic organization and functional interactions of NMDA and μORs in the CeA, particularly with respect to opioid withdrawal. The implications of these results are considered within the context of prior neuropharmacological data on the role of the CeA in psychological and physical aspects of opioid dependence.

II. Opioids and Dependence

Opioids have been used by humans for millennia (Brownstein, 1993; Scarborough, 1995), and their analgesic properties have made them an important part of current medical care (Inturrisi, 2002). Despite their utility, the complex relationships between opioid pharmacology, use history, genetic factors, and other physiologic and environmental variables present significant problems with respect to their rational clinical use (Kreek, 2008; O’Brien, 2008; Weaver and Schnoll, 2002).

A pernicious complication of opioid consumption, dependence reflects the development of opioid-mediated adaptations that are expressed after termination of drug action by drug withdrawal-induced counter-adaptations, which are typically contrary to the acute effects of the opioid (Christie, 2008; Koob, 2009; Williams et al., 2001). Although it has been reported that only a single opioid exposure may be required (Eisenberg and Sparber, 1979), repetitive use results in a more pronounced degree of dependence (Way et al., 1969). The development of dependence on the prototypic opioid morphine requires functional μORs (Matthes et al., 1996), and its severity is significantly related to genetic factors (Kest et al., 2002; Korostynski et al., 2007).

Well characterized in humans, dependence has been studied extensively in rodent models (Martin et al., 1963). Opioid antagonist-precipitated withdrawal is the most common means of studying this phenomenon in animals, and will be the method discussed throughout the remainder of this chapter. Symptoms of opioid withdrawal can be elicited in dependent animals upon peripheral administration of general opioid receptor antagonists, such as naloxone or naltrexone (Hamlin et al., 2001). However, intraventricular (Maldonado et al., 1992) or intracranial (Stinus et al., 1990) microinjection of opioid blockers can reproduce most withdrawal signs, indicating that dependence is critically mediated by diverse neural circuits. A list of some prominent withdrawal symptoms is provided in Table 8.1. As would be expected from the highly unpleasant nature of these symptoms, avoiding or relieving withdrawal is an important factor in continued drug use (Koob, 2009).

Table 8.1.

In small rodents, opioid withdrawal symptoms can be categorized into distinct functional groupings, including somatic, autonomic, endocrine, and affective (Buccafusco, 1990; Gonzalvez et al., 1994; Martin et al., 1963; Mucha, 1987)

Somatic	Escape-jumping, wet-dog shakes
Autonomic	Diarrhea, hypertension
Endocrine	ACTH and Cort release
Affective	Acute aversion
Learning	Conditioned place aversion/avoidence

Only some of the most prominent examples are listed for each category.

ACTH: adrenocorticotropic hormone; Cort: corticosterone

In addition to its acute unconditional effects, opioid withdrawal has other consequences that may endure beyond periods of detoxification and abstinence due to learned associations between the withdrawal state and environmental cues (O’Brien, 2008). Indeed, animals will learn to spend less time in a previously neutral environment paired with aversive periods of naloxone-precipitated withdrawal (Stinus et al., 1990), a phenomenon termed conditioned place aversion (CPA). CPA is a species of classical conditioning whereby a conditional stimulus, such as a particular location with salient cues (i.e., tactile, visual, or other sensory stimuli), acquires secondary negative qualities, due to its pairing with an unconditional stimulus, such as opioid withdrawal (Cunningham et al., 2006). Because the conditioned stimulus can evoke a response similar to the unconditioned stimulus, animals engage in avoidance behavior when subsequently exposed to the aversive conditioning environment.

What is notable about CPA is that it can be produced with doses of naloxone that are so low even physical withdrawal signs are not elicited (Gracy et al., 2001). The latter finding indicates that a negative affective state (Schulteis et al., 1998), rather than physical symptoms per se, is the requisite unconditioned stimulus necessary for learning. In addition, place aversion persists for weeks after training (Stinus et al., 2000), and may have a lasting influence on behavior. Learned aversive cues may also impact drug-taking behavior, as demonstrated by evidence that withdrawal-associated stimuli can prime opioid self-administration in dependent animals (Hellemans et al., 2006; Kenny et al., 2006), likely as a means of escaping the noxious experience. Moreover, the role of aversive learning in drug-seeking behavior may also have significant clinical relevance (O’Brien et al., 1975, 1998).

III. Glutamate Systems and Opioid Dependence

Functional ionotropic glutamate receptors are required for the full development and expression of dependence. In particular, it has been shown that acute or chronic administration of NMDA receptor antagonists attenuates physical withdrawal symptoms. This has proven to be a reliable outcome as NMDA receptor blockade inhibits withdrawal in opioid-dependent mice (McLemore et al., 1997), guinea pigs (Tanganelli et al., 1991), rats (Trujillo and Akil, 1991), and humans (Bisaga et al., 2001). Moreover, in rodent models, NMDA antagonists also suppress the conditioned aversive properties (i.e., CPA) of naloxone-precipitated withdrawal (Higgins et al., 1992).

The NMDA receptor has complex signaling properties and is a potent modulator of cellular adaptability (Dingledine et al., 1999; Kohr, 2006). The NMDA receptor is a tetrameric heteromer composed of the essential NMDA-NR1 subunit (NR1), of which there are eight splice variants. Along with NR1, glutamate-responsive NMDA receptors require some combination of NMDA-NR2 subunits (NR2), of which there are four subtypes (NR2A–D). Critical characteristics of the NMDA receptor include a voltage-dependent Mg²⁺ blockade requiring cellular depolarization for channel activation and high permeability to Ca²⁺. In addition, NMDA receptor activation can modulate numerous intracellular signaling processes including protein kinase activity (Haddad, 2005), protein transport (Shi et al., 1999), transcription factor activity, and gene expression (Yoneda et al., 2001), as well as epigenetic events (Lubin et al., 2008). These properties of NMDA receptor activation may have profound effects on cellular and behavioral plasticity (Tsien, 2000), such as long-term potentiation (LTP) and long-term depression (LTD) (Kullmann et al., 2000), in addition to whole-animal learning and memory (Shapiro and Eichenbaum, 1999; Walker and Davis, 2002), processes that play a role in opioid dependence.

IV. The Central Nucleus and Dependence

Despite the significant relationship between NMDA receptors and opioid dependence, a brain-map linking sites of functional receptor expression to specific opioid withdrawal behaviors is at best rudimentary. Given that brain motivational systems are a critical substrate of opioid plasticity, NMDA receptors in nuclei such as the CeA are likely to play an important role in dependence. The central amygdala is a critical component of neural motivation pathways that play a role in endocrine (Schulkin et al., 1994) and autonomic function (Brody, 1988), fluid (McKinley et al., 2001) and energy balance (Glass et al., 2000), as well as behavioral (Davis, 1998) and affective (Davis, 1989) processes. Many of these activities are known to be modulated by glutamate and opioids. In order to better appreciate the role of the CeA glutamate system in opioid dependence, a brief overview of its place within the larger context of brain circuitry subserving motivated behavior is illustrated in Fig. 8.1. The basic neurochemical content of the CeA will be outlined below.

Figure 8.1.

The organization of brain motivational pathways involved in opioid dependence. Brain motivational systems are composed of a highly complex network of extero- and interoceptive processing systems, integrative memory systems, as well as hypothalamic and brainstem endocrine/autonomic and motor activators that play essential roles in maintaining homeostasis (Swanson, 2000). The CeA is a critical coordinator of relevant functional inputs and outputs within this circuitry, and may play a role in coupling homeostatic need with emotional valence and immediate survival behaviors, as well as more slowly developing learned emotional responses to those challenging conditions. A highly schematic representation of neural pathways involved in motivational processes is presented as a series of illustrated coronal brain sections. The CeA and its outputs are indicated in gray, while specific targets are highlighted by red fill. Relevant CeA inputs are indicated by differently colored dashed arrows. These include glutamate afferents (green) from areas of limbic cortices (McDonald, 1998), the thalamus (Turner and Herkenham, 1991), and the BLA. The CeA is also innervated by catecholaminergic neurons (black), including dopaminergic neurons from the mesolimbic system and noradrenergic neurons from the medulla (Asan, 1998). Neuropeptides (blue) from these and other brain areas are also expressed by CeA inputs (see text for description). This array of afferents provides the CeA with a complex set of signals about current internal and external stimuli, as well as representations in memory, that are critical in organizing behavioral responses. Both NMDA-type glutamate and μORs are expressed within the central amygdala. Critical outputs of the central amygdala include areas of the extended amygdala involved in learned emotional processes such as the bed nucleus of the stria terminalis (BNST; Zahm et al., 1999), hypothalamic systems involved in endocrine (paraventricular nucleus of the hypothalamus [PVN]) and behavioral (lateral hypothalamic area [LHA]) responses (Allen and Cechetto, 1995; Gray et al., 1989; Marcilhac and Siaud 1997), midbrain areas involved in nociception (periaqueductal gray [PAG]), reward (ventral tegmental area [VTA]), and/or motor (substantia nigra [SN]) function (Schmued et al., 1989; Zahm et al., 1999), as well as medullary nuclei (nucleus of the solitary tract [NTS], ventrolateral medulla [VLM]) that mediate cardiovascular, respiratory, and/or gastrointestinal processes (Glass et al., 2002; Wallace et al., 1989). Drawings are adapted from the brain atlas of Swanson (Swanson, 1992).

The CeA has a diverse compliment of glutamate afferents that arise from areas of brain motivational systems. These include limbic cortices (McDonald, 1998), the thalamus (Turner and Herkenham, 1991), and the basolateral nucleus of the amygdala (BLA) (Pitkanen et al., 1997). The CeA is notable for its abundant expression of glutamate receptors (Petralia and Wenthold, 1992; Petralia et al., 1994; Sato et al., 1993). In particular, neurons in the CeA express the NR1 gene (Sato et al., 1995) and protein (Petralia et al., 1994), as well as NMDA ligand-binding sites (Monaghan and Cotman, 1985).

In addition to glutamate, CeA neurons have a rich array of signaling molecules. Many central amygdala neurons express the enzyme glutamic acid decarboxylase 67 kDa, which is responsible for synthesizing the inhibitory transmitter γ-aminobutyric acid (GABA) (Carta et al., 2008). In addition to GABA, there are also endogenous CeA neurons and/or axons from extrinsic sources that express a variety of neuroactive peptides as well as receptors for these modulators. These include corticotropin-releasing factor (Beyer et al., 1988; Swanson et al., 1983), components of the renin–angiotensin system (Brown and Gray, 1988; Brownfield et al., 1982; Lavoie et al., 2004), neuropeptide Y (Gray et al., 1986; Heilig et al., 1993), orexin (Baldo et al., 2003), oxytocin (Lee et al., 2005; Roozendaal et al., 1993), and vasopressin (Francis et al., 2002; Roozendaal et al., 1993). In addition, the CeA is also innervated by catecholamine releasing axonal varicosities (Asan, 1998) and contains neurons that express adrenergic receptors (Glass et al., 2002).

Among the peptide systems active in the central amygdala, the opioids are notably abundant. Enkephalin (Cassell et al., 1986; Petrovich et al., 2000; Wray and Hoffman, 1983), dynorphin (Fallon and Leslie, 1986; Reyes et al., 2008; Zardetto-Smith et al., 1988), and β-endorphin (Gray et al., 1984) are each present in the CeA. Moreover, opioid receptors, including the μOR, are expressed in the central amygdala (Mansour et al., 1988, 1994). Deciphering the synaptic organization and functional relationships between NMDA receptors and opioid receptors is a critical issue in understanding of central amygdala function and its relationship to opioid dependence.

V. The Synaptic Relationship Between NMDA and μ-Opioid Receptors in the CeA

The cellular relationship between NMDA receptors and μORs in the CeA has only been inferred by the results of electrophysiological studies, and is a matter of contention (Zhu and Pan, 2004, 2005). Because of its high spatial resolution, immunoelectron microscopy employing gold and peroxidase markers is a valuable tool for examining the fine structural location of functionally interacting proteins in areas of brain motivational pathways (Gracy and Pickel, 1995; Van Bockstaele et al., 2000). Using immunoelectron microscopic analysis, the NR1 subunit was shown to be highly enriched in postsynaptic (i.e., dendrites) sites of central amygdala neurons (Fig. 8.2A and B) (Glass et al., 2009). Labeled dendritic profiles were typically small to intermediate in size (0.5–1 μm cross-sectional area). NR1 was frequently present in near intracellular endomembranous organelles, reflecting sites of protein storage and trafficking. However, this protein was also present on the plasma membrane, the primary location of functional receptors. Furthermore, many of these dendritic profiles were contacted by axon terminals forming asymmetric excitatory-type synapses, typical of those made by glutamatergic axon terminals (Peters et al., 1991). In summary, NR1 was mainly localized to dendritic (i.e., postsynaptic) structures where it was frequently present in vesicular organelles linked to protein transport, as well as the extrasynaptic and synaptic plasma membrane contacted by excitatory-type axon terminals.

Figure 8.2.

Ultrastructural distribution of the NMDA receptor and its relationship to the μOR in the central amygdala. (A) A dendritic profile (NR-1-d) expressed diffuse immunoperoxidase reaction product for NR1. This profile exhibited a spine neck (sp-n) and was contacted by an unlabeled axon terminal (ut-1) that did not appear to form a synapse. (B) When labeled by either immunoperoxidase (ABC) or immunogold (Gold) secondary markers, the majority of NR1-labeled processes were dendrites. (C) A dendritic profile (NR-1 and μOR-d) expressed diffuse immunoperoxidase reaction product for NR1 and immunogold labeling (small arrows) for the μOR. This dual labeled profile received distinct asymmetric excitatory-type synapses (arrow heads) from two unlabeled axon terminals (ut-1 and ut-2). (D) Of all dual labeled neuronal profiles, the majority were dendrites. This was the case when NR1 was labeled by immunoperoxidase (ABC) and the μOR by immunogold (Gold) markers, and when secondary antisera were reversed (i.e., NR1 labeled by gold and the μOR by peroxidase). Scale bars: 0.5 μm. See Glass et al. (2009) for details.

In terms of its subcellular location, the μOR was found near intracellular membranous organelles and the plasma membrane, a pattern similar to that of NR1. However, relative to the glutamate receptor subunit, the μOR had a more heterogeneous distribution in neuronal compartments. Although frequently present in somata and dendrites, there were also many instances of μOR expressing axons and axon terminals (for details see Glass et al., 2009).

There were numerous instances of neuronal profiles in the CeA that expressed labeling for both NR1 and the μOR. Despite the prominent dendritic distribution of NR1 and the mixed dendritic and axonal localization for the μOR, dual labeling for these proteins was most commonly found in dendrites (Fig. 8.2C and D). Like the single-labeled NR1 and μOR containing dendrites, structures that expressed immunoreactivity for both proteins were small to medium in size. Labeling for each protein was present near vesicular organelles characteristic of those involved in protein transport, and was also found on the surface membrane. Synapses formed on these dual labeled dendrites were typically of the asymmetric excitatory kind. In sum, these results indicate that NMDA receptors and μORs are strategically positioned for postsynaptic comodulation of glutamate signaling in dendrites of CeA neurons.

VI. Deletion of Postsynaptic NR1 in Central Amygdala Neurons Attenuates Opioid Withdrawal-Induced Place Aversion

Establishing relationships between functional postsynaptic NMDA receptor expression and opioid dependence is not feasible by traditional pharmacological approaches. Currently available NMDA receptor antagonists cannot discriminate between neuronal dendrites, axon terminals, or other ultrastructural compartments. In addition, interpreting neuropharmacological studies involving NMDA receptor antagonists is problematic given that central amygdala blockade of NMDA receptors can have rewarding or aversive properties, depending on the particular agent used (Watanabe et al., 2002), as well as other effects likely to confound performance of learned behaviors including alterations in motor function (Andrzejewski et al., 2004).

To examine the relationship between opioid dependence and functional postsynaptic central amygdala NMDA receptors a spatial–temporal gene deletion strategy employing Cre-lox technology was used. The NR1 subunit was deleted by local CeA microinjection of a neurotropic replication deficient recombinant rAAV that expressed a fusion protein of the enzyme Cre recombinase (Cre) and a reporter, GFP, termed “rAAV–GFP–Cre” (South et al., 2003). Injections were made in adult male transgenic loxP knockin mice that have strategically placed loxP sites in the NR1 gene (i.e., “floxed NR1” (fNR1) mice) flanking exons that encode for the four membrane domains and the C-terminus (South et al., 2003).

Direct microinjection of rAAV–GFP–Cre into the CeA of adult fNR1 mice resulted in recombination specifically in neurons (Glass et al., 2008). Moreover, there was a significant reduction in NR1 expression in the target area (Fig. 8.3). This reduction occurred in somata and dendritic profiles (i.e., postsynaptic). There were no concomitant reductions of presynaptic NR1 or somatodendritic NR2 immunolabeling, and no effects on local cell number or cellular morphology (Glass et al., 2008).

Figure 8.3.

Postsynaptic NR1 deletion in central amygdala neurons. (A) Unilateral microinjection of rAAV–GFP–Cre (indicated by arrow) in the CeA produced a localized knockout of NR1 as shown by in situ hybridization. (B) Gene deletion corresponded to expression of the GFP reporter as seen in a serial section. GFP expressing neurons can be seen at a higher magnification in the inset. (C, D) Higher magnification views of the CeA as seen in Fig. 8.3A. Note the significantly diminished NR1 expression in the injected hemisphere (C) compared to the uninjected (D) side. (E) Compared to unilateral CeA injection of the control GFP vector, injection of Cre produced significant reductions in NR1 labeling in somatodendritic sites in the injected compared to the uninjected hemisphere as measured by immunoelectron microscopy (NR1 immuno). This corresponded to reductions in NR1 gene expression (NR1 in situ) in the CeA (Glass et al., 2008). *p < 0.05 compared to GFP. Scale bars: 1 mm.

Bilateral knockout of the NR1 gene in the CeA did not produce obvious basal behavioral deficits. In particular, deletion of CeA NR1 did not impact locomotor activity, body weight, sensory-motor coordination, thermal nociception, or somatosensation (Glass et al., 2008).

In regard to opioid dependence, CeA NR1 knockout mice were chronically exposed to morphine by subcutaneously implanted morphine pellets and then administered naloxone to precipitate withdrawal. The CeA NR1 knockouts did not differ from control animals with respect to somatic signs, such as escape-jumping and wet-dog shakes, or autonomic symptoms, notably diarrhea and weight loss (Fig. 8.4A). Contrary to what was found with physical symptoms, central amygdala NR1 gene deletion did impair naloxone withdrawal-induced place aversion (Fig. 8.4B). Therefore, functional CeA NMDA receptors were not necessary for the induction of many major physical withdrawal symptoms, but were required for the production of a learned withdrawal-induced negative affective state.

Figure 8.4.

Postsynaptic deletion of NR1 in central amygdala neurons and opioid dependence. (A) Separate groups of fNR1 mice, including those given no intracranial injection (No-inj), or bilaterally microinjected with either rAAV–GFP (GFP) or rAAV–GFP–Cre (Cre), were chronically administered morphine by subcutaneous implantation of a morphine pellet (25 mg), which was replaced by a fresh pellet every fourth day as needed. Dependence was determined by visual observations of physical withdrawal signs including diarrhea, wet-dog shakes, and escape jumping after an acute injection of naloxone (1 mg/kg, i.p.). There were no differences between any of the groups in the number of somatic (wet-dog shakes + jumping) or autonomic (diarrhea; not shown) symptoms. (B). Place aversion training took place in an apparatus consisting of a two-chamber box, each with distinct tactile and visual cues, that was inserted into an automated activity monitor. After measuring baseline selection, during which time animals were allowed to freely explore each chamber, animals began place aversion training. On alternate training days morphine dependent mice were injected with saline (0.9%, i.p.) or naloxone (1 mg/kg, i.p.) and then restricted to the respective chamber for 30 min. On the test day, subjects were allowed to freely explore both chambers for 30 min as during baseline. The difference in time spent in the naloxone-paired chamber during the testing and preconditioning phases served as the measure of place aversion. In distinction to physical withdrawal, there were significant between-group differences in CPA. Unlike animals in both control groups, mice bilaterally microinjected with Cre in the CeA did not spend less time in the withdrawal-paired chamber, indicating that CeA NR1 knockout interfered with a conditioned place aversion in response to opioid withdrawal. Reductions in NR1 mRNA were seen selectively in the animals injected with rAAV–GFP–Cre as previously reported (Glass et al., 2008).

The finding that CeA NR1 deletion impaired withdrawal-induced CPA is consistent with prior reports that NMDA receptor activity in the CeA plays an important role in aversive learning and memory (Goosens and Maren, 2003). The precise role of the central amygdala in emotional learning has been controversial. Traditionally, the CeA has been considered less as a sight of learning and more as a modulator of other areas that encode new associations, however, emerging evidence contradicts this view. For example, recent findings have shown that the CeA is required for the normal acquisition and consolidation of conditioned fear (Rabinak and Maren, 2008; Wilensky et al., 2006). These neurological results are supported by neurophysiological data. It has been reported that neurons in the CeA exhibit a form of cellular plasticity (i.e., LTP) that occurs in established neuroanatomical substrates of learning and memory, such as the hippocampus and lateral amygdala (Samson and Paré, 2005). Moreover, CeA LTP is also sensitive to withdrawal from drugs of abuse in an NMDA receptor-dependent manner (Pollandt et al., 2006).

VII. Does the CeA Selectively Participate in the Conditioned Aversive Properties of Opioid Withdrawal?

The finding that CeA NR1 deletion selectively interferes with CPA has interesting parallels with other reports in the literature. For example, in opioid-dependent animals, withdrawal induced by CeA microinjection of an opioid antagonist preferentially produces CPA (Stinus et al., 1990). In addition, it has been shown that lesioning the ventral noradrenergic bundle, which provides an important source of norepinephrine containing axons in the CeA and related areas of the extended amygdala, inhibits opioid withdrawal place aversive (Delfs et al., 2000). The latter effect is paralleled by the ability of CeA administered alpha-2-adrenergic agonists to inhibit only affective signs of withdrawal (Taylor et al., 1998). Based on this evidence, one may conclude that the CeA has a privileged role in mediating the conditioned emotional properties of withdrawal.

The notion that the CeA has a special role in affective learning processes related to dependence, however, may be too simplistic. For example, it has been reported that blockade of non-NMDA type glutamate receptors in the CeA of dependent animals reduces somatic signs of opioid withdrawal (Taylor et al., 1998). This result is similar to other findings in the literature showing that blockade of CeA CRF receptors also inhibits physical symptoms of withdrawal from opioids (McNally and Akil, 2002).

It appears that manipulating particular receptor systems in the CeA can produce distinct outcomes with respect to the acute physical and psychological features of opioid withdrawal. One reasonable explanation for these complex outcomes is that particular manifestations of dependence may be coded by distinct neurochemical signals acting within circumscribed central amygdala circuits. One hypothetical scheme detailing how such a system may be organized is illustrated in Fig. 8.5.

Figure 8.5.

Schematic representation illustrating hypothetical model of central amygdala synaptic coding of physical and psychological signals linked with opioid withdrawal. One group of CeA neurons may express non-NMDA type (i.e., AMPA, kainate), glutamate (dark green stars), and CRF (blue stars) receptors, and receive preferential input from CRF expressing neurons (blue arrow). The physical withdrawal pathway may be engaged by an upstream opioid receptor-dependent pathway, presumably expressing glutamate or CRF. This system would be expected to mediate immediate opioid withdrawal symptoms. The other group (light gray shading) of CeA neurons may express NMDA (pale green stars), mu-opioid (yellow star), and noradrenergic (black star) receptors, and receive input from noradrenergic afferents (black arrow), particularly from areas of the medulla (NTS, VLM), as well as opioid peptides (not shown). This system would be expected to have a slow onset and require consolidation and long-lasting synaptic integration. Glutamate (green arrow) would be expected to activate both systems, however, whether segregation of function is mediated by separate input sources (i.e., source of glutamate), or by distinct processing capacities (i.e., glutamate receptor types) of their targets is uncertain. The role of CRF (Stinus et al., 2005) and non-NMDA (Kawasaki et al., 2005) receptors may not be limited to immediate unconditional responses, and an alternative model would incorporate these signaling systems into both immediate physical and affective/learning processing streams.

VIII. Conclusion

Both NMDA receptors and μORs are strategically positioned for the comodulation of excitatory postsynaptic signaling in CeA neurons. Moreover, genetic deletion of postsynaptic NMDA receptors impairs the conditioned aversive, but not the physical features of opioid withdrawal. Because conditioned cues may promote behaviors to avoid or escape their associated adverse state, NMDA receptor activity in the CeA may be essential in complex neural processes that engage relief-seeking behaviors, like drug taking (Koob, 2009).

The intriguing speculation that the neural substrates of particular behavioral features of dependence may involve differing synaptic organizational arrangements within central amygdala circuitry will require further exploration. This hypothesis can be addressed by, among other techniques, multilabeling immunoelectron microscopic ultrastructural analysis. This can be combined with application of spatial–temporal gene deletion methodology similar to that described in this chapter, as well as a development of this approach that incorporates phenotype-specific promoters to manipulate gene expression in selected neurochemical populations (Jasnow et al., 2009; Oh et al., 2009). Such efforts should elucidate the synaptic and molecular bases of the many manifestations of dependence, knowledge that may be critical in developing the next generation of neurobiological tools to manage the long-term consequences of opioid abuse. Moreover, this approach also has implications for other psychiatric disorders involving aversive learning, such as anxiety disorders.