Cocaine withdrawal-induced trafficking of delta-opioid receptors in rat nucleus accumbens

DOR trafficking

Regions sampled for electron microscopic analysis are depicted in Figure 1A. Immunogold-silver detection for DOR was employed to examine the subcellular distribution of DOR in dendritic profiles in the NAcbC and NAcbS of both saline and cocaine-treated male and female rats that underwent two day withdrawal from drug or vehicle treatment. In chronic saline-treated rats, DOR was localized to both plasmalemmal and intracellular compartments of dendritic profiles in both the core and shell regions in both male (data not shown) and female rats (Figure 1 B,C and Figure 2 A,B). In the NAcbS of female rats undergoing cocaine withdrawal, the majority of DOR was localized to intracellular compartments (Figure 1D, 1E). Within intracellular compartments, immunogold-silver labeling for DOR was often associated with endosome-like structures. Interestingly, DOR was localized to both the plasma membrane and intracellular compartments in the NAcbC of cocaine-treated females (Figure 2 C, D). In males, cocaine withdrawal resulted in a decrease in plasma membrane DOR in both core and shell regions (Figure 3 C,D).

Semi-quantitative analysis was conducted to determine the percentage of DOR localization to the plasma membrane of dendritic profiles in the NAcbC and NAcbS of both male (n=2) and female (n=3) rats (Figure 3). In females, withdrawal from chronic cocaine resulted in a decrease in the percentage of DOR localized to the plasma membrane in NAcbS and an increase in the percentage of plasma membrane DOR in NAcbC. Thirty-five percent of DOR in the NAcbS of saline-treated females was localized to the plasma membrane (n=185 labeled profiles) compared to only 24% (n=245) in cocaine-treated rats (Figure 3A). This represents ~ 31% overall reduction in plasma membrane DOR after two days of withdrawal from chronic cocaine. Twenty percent (n=199) of DOR in the NAcbC of saline-treated females was localized to the plasma membrane compared to 26% (n=213) in cocaine-treated females representing ~ 30% overall increase in DOR localization to the plasma membrane (Figure 3B).

In males, withdrawal from chronic cocaine resulted in a decrease in plasma membrane DOR in both core and shell regions. In the NAcbS, 29% of DOR in dendritic profiles (n = 103) from saline-treated males compared to 19% of DOR (n=103) from cocaine-treated males was localized to the plasma membrane (Figure 3C). In the NAcbC, 22% of DOR (n = 114) was localized to the plasma membrane in saline-treated males compared to only 17% (n=103) in cocaine-treated males (Figure 3D). Overall this represents a ~34 % and ~23% decrease in DOR localization to the plasma membrane in core and shell regions, respectively.

Western blot analysis was used to examine DOR expression levels in the NAcb two days following saline or cocaine administration (n=4 rats per group) (Figure 4). Incubation of the primary antibody with its blocking peptide resulted in a loss of immunoreactivity for DOR at the expected molecular weight of 50 kDa (Figure 4A) (Persson et al., 2003; Persson et al., 2005). Densitometric analysis of DOR protein expression in the NAcb indicated that DOR was not different between saline-treated vs. cocaine-treated rats (Figure 4B, 4C). There was also no difference in male vs. female rats (data not shown).

Co-localization of DOR and D1R

To determine the degree of co-existence between DOR and D1R both immuno-fluorescence and electron microscopy were employed. Single labeling for D1R and DOR was present within the NAcbC and NAcbS subregions. Confocal microscopy revealed that a subset of neurons in both the NAcbS and NAcbC exhibited both DOR and D1R immunoreactivities (Figure 5A,B). These are clearly distinguishable in merged images that show singly labeled DOR and D1R neurons as well as dually labeled DOR and D1R cells (Figure 5C).

Electron microscopy allowed for further analysis of DOR and D1R co-localization. At the ultrastructural level, immunogold-silver labeling for D1R and DOR was detected within the NAcbC and NAcbS in both single and dual-labeled profiles. Black deposits indicative of gold-silver immunolabeling were distinguishable from the electron dense peroxidase reaction product. To circumvent potential caveats caused by the visualization of these two different labeling techniques, immunomarkers were reversed in a separate set of tissue run simultaneously to ensure that the localization of each antigen was not thwarted by the detection method used. Results were similar irrespective of the labeling approach used.

In both the NAcbC and NAcbS, single-labeled profiles were visualized within similar portions of the neuropil. Often, single-labeled D1R and DOR somatodendritic processes were in close proximity to one another. In some cases, single-labeled profiles were apposed to one another. Localization of D1R and DOR to the same cellular profiles was evident in both the shell and core regions (Figure 6A, B). Clusters of immunoperoxidase labeling for D1R could be identified within dendrites (Figure 6A) and somata that also contained gold-silver immunoreactivity for DOR. Furthermore, co-localization was also present when labels were reversed (Figure 6B).

Semi-quantitative analysis was conducted to determine the percentage of co-existence between DOR and D1R in the NAcb (Figure 6C). In the NAcbS, the percentage of co-existence was similar with 33% of D1R labeled profiles (n=215) containing DOR immunolabeling and 34% of DOR labeled profiles (n=211) containing D1R immunoreactivity. In the NAcbC, D1R labeled profiles (n=316) contained DOR immunoreactivity in 24% of sampled neurons. Interestingly, a greater percentage of total DOR labeled profiles (n=198) showed immunolabeling for D1R (n=75, 38%). As both of these receptors were primarily localized to somatodendritic processes, co-existence was predominantly seen in these cellular profiles.

Methodological Considerations

A potential limitation known to be associated with the pre-embedding immunolabeling technique is attributed to penetration of immunoreagents in thick Vibratome sections (~40 µm) (Chan et al., 1990). To circumvent this caveat, analysis of ultrathin sections was carried out exclusively on sections near the tissue/plastic interface where penetration is maximal. Further, immunolabeling of antigens was reversed to ensure that patterns of immunolabeling were not attributed to differences in the detection methods (see below).

The specificity of the D1R and DOR primary antibodies are discussed in the section entitled “Experimental Procedure” and we have performed the requisite control experiments to ensure antibody specificity here and in our previous studies (Commons et al., 2001; Ambrose et al., 2006). Our findings are in agreement with others that used the same D1R primary antibody and reported a primarily postsynaptic distribution of D1R along the plasma membrane of neurons in both subregions of the NAcb (Hara and Pickel, 2005). Furthermore, our results are consistent with localization of D1R in medium-sized spiny neurons (Yung et al., 1995; Hara and Pickel, 2005; Dumartin et al., 2007; Hara and Pickel, 2007).

Both receptors were primarily localized to somatodentritic processes in both the core and shell of the NAcb. In both the core and shell region of NAcb, DOR was predominantly intracellular as previously shown using immunohistochemistry in other CNS regions (Cheng et al., 1995; Zhang et al., 1996). This pattern of prominent somatodendritic labeling for the DOR is consistent with previous studies using antibodies specific for this region of the N-terminus (amino acids 3–17 on the cloned mouse DOR1) (Cahill et al., 2001; Lucido et al., 2005). For example, Lucido et al. demonstrated DOR membrane recruitment in NAcb dendritic profiles following morphine treatment and reported that most of DOR immunoreactivity was localized to neuronal somata and dendrites, with low to moderate labeling of synaptic terminals. However, studies using antibodies raised to different sequences of the receptor have reported different patterns of labeling. For example, Svingos and colleagues (using an antibody raised to amino acids 34–47 of the N-terminal domain) reported that 60% of DOR-immunoreactive profiles in the NAcbS were axon terminals and small unmyelinated axons with 35% of DOR profiles characterized as dendrites and dendritic spines (Svingos et al., 1998). This discrepancy could be due to multiple factors. It is possible that the inclusion criteria used in the present analysis for immunogold labeling may have led to an under-representation of presynaptic DOR labeling. In addition, variations in antibody recognition could cause differences in detection. Although this caveat remains, the present study is consistent with previous reports and extends these findings by showing cocaine withdrawal-induced trafficking of DOR and co-existence of D1R and DOR in the NAcb.

The present study compared NAcbC and NAcbS but these regions can be further subdivided medially and laterally. Although the majority of samples were taken from the medial aspect of the core and shell of the NAcb, we did not investigate subdivisions of the core and shell here. Anatomical differences between medial and lateral accumbens shell have been reported in regards to both neuronal morphology and connectivity (Meredith et al., 1992; Voorn et al., 2004; Ikemoto, 2007). Additional studies focusing on medial and lateral divisions of the core and shell will be required to determine if any differences exist.

Lastly, we used both males and females in the present analysis and uncovered some potentially important sex differences. Sex differences in response to cocaine and other drugs of abuse are emerging [for reviews, see (Lynch et al., 2002; Carroll et al., 2004; Becker and Hu, 2007)]. In the NAcb, there are sex differences in dopaminergic function as well as regulation of cell signaling in both naïve and cocaine-treated animals (Becker, 1999; Lynch et al., 2006). The present study did not investigate the influence of hormonal variations associated with the estrous cycle on DOR trafficking. It is possible that varying hormone levels may influence DOR modulation in female rats. Future studies investigating therapeutic agents targeting dopamine and opioid systems will need to consider potential sex differences.

DOR trafficking in NAcbS and NAcbC during early cocaine withdrawal

Receptor trafficking can regulate the number of receptors available at the cell surface and can therefore influence cell reactivity to neurotransmitters and drugs. Under normal conditions, only a small percentage of DOR are available at the plasma membrane. The majority of DOR is localized intracellularly (Arvidsson et al., 1995; Cheng et al., 1995; Cheng et al., 1997; Cahill et al., 2001; Commons et al., 2001). Consistent with previous evidence, 20–35% of DOR was localized to the plasma membrane of saline-treated controls. In the NAcbC of males and the NAcbS of males and females, early cocaine withdrawal-induced internalization of DOR resulted in a reduction in the number of receptors at the cell surface. This condition did not lead to a decrease in the overall expression of receptors since Western blot analysis did not reveal any change in expression levels. The present findings are consistent with previous evidence reporting no change in DOR receptor number using quantitative autoradiography on tissue obtained immediately following 14 day binge-pattern cocaine administration (Unterwald et al., 1994). These data suggest that the difference in DOR during early cocaine withdrawal is more likely due to redistribution of DOR and may not result from changes in protein expression.

Desensitization of DOR has been reported immediately following chronic cocaine and during early withdrawal (Unterwald et al., 1993; Unterwald and Cuntapay, 2000; Perrine et al., 2008). The results of the present study demonstrate cocaine withdrawal-induced trafficking of DOR which may contribute to the increased anxiety and depression associated with withdrawal from chronic cocaine. Although the reported difference in subcellular localization of DOR did not achieve conventional statistical significance, it is likely that the change in receptor localization at the plasma membrane represents an important change in the number of receptors at the cell surface available for agonist binding. Decrease in cell surface DOR could lead to alterations in emotional state as DOR has been shown to be involved in anxiety and depression, conditions that are both linked to cocaine withdrawal (Filliol et al., 2000; Perrine et al., 2008). Since the NAcb communicates with many limbic regions, cocaine-withdrawal induced neuroadaptations in this region have an influence on behavioral state. The NAcbS is most associated with limbic functions and therefore, it is possible that alterations in DOR in the NAcbS could modulate emotional behavior. Interestingly, there was a slight increase in the localization of DOR to the plasma membrane in the NAcbC of female rats. The core is known to be more closely related to motor functions. Future studies are needed to determine if cocaine-induced changes in this region may be involved in the underlying mechanisms involved in the increased behavioral sensitization to cocaine exhibited by female rats (Hu and Becker, 2003).

Our data indicate that alterations in DOR signaling following withdrawal from cocaine are not readily explained by changes in receptor expression levels as Western blot analysis did not show any difference in DOR expression levels when comparing cocaine withdrawn rats to saline-treated controls. Interpretation of these data should consider, however, that both the core and shell subregions of the NAcb were sampled together and that regional differences may exist that were not readily identified in this study. Therefore, in situ hybridization analysis of mRNA for DOR should be conducted in future studies to unequivocally establish the presence or lack of DOR receptor downregulation.

Co-existence of D1R and DOR in NAcbS and NAcbC

Previous work from our laboratory has demonstrated co-existence of D1R and DOR in a subset of neuronal profiles in the dorsolateral striatum (Ambrose et al., 2006). Since the NAcb is a major target for cocaine, the present study sought to address whether a similar organization exists in the NAcb core and shell regions. Co-localization of DOR and D1R in accumbal neurons is a first step in elucidating the cellular mechanisms involved in the heterologous desensitization of DOR. This effect is mediated by D1R as chronic administration of a D1R agonist also attenuates DOR-induced inhibition of adenylyl cyclase activity (Unterwald and Cuntapay, 2000). It is interesting to note that although the NAcbS had an almost equal percentage of DOR profiles containing D1R and D1R profiles containing DOR, this was not the case in the NAcbC. In addition, although the co-expression of these two receptors was evident in the ventral striatum, it occurs to a lesser extent compared to the dorsolateral striatum which demonstrated about 50% co-localization (Ambrose et al., 2006). A recent study investigating the subcellular localization of D1R in the dorsal striatum, NAcbC and NAcbS also reported differences among these regions (Dumartin et al., 2007). Potentially, this regional heterogeneity could affect the contribution of each of these regions to the interaction between the dopaminergic and opioidergic systems following cocaine.

Further investigation is essential in order to determine other factors that may be involved and the degree to which this co-existence plays a role in the interaction between these two receptor systems. In a recent study utilizing a novel strategy where the nuclear translocation pathway was adapted to visualize opioid-dopamine receptor hetero-oligomers, data suggested that DOR and D1R did not form hetero-dimers while µ-opioid receptor (MOR) and D1R did form hetero-dimers (Juhasz et al., 2007). Although this does not support a direct interaction between DOR and D1R, there is evidence demonstrating the ability of MOR and DOR to form functional hetero-dimers (George et al., 2000; Gomes et al., 2000; Gomes et al., 2004) at the cell surface (Law et al., 2005) This could be a mechanism by which D1R can directly interact with both MOR and DOR receptors. Electron microscopy demonstrates co-localization of DOR and MOR in the striatum (Wang and Pickel, 2001) but dual-labeling in the NAcb has not been examined. However, separate studies using single labeling for MOR and DOR have shown these receptors to be localized to accumbal neurons [present results; (Svingos et al., 1996; Svingos et al., 1998)].

DOR agonists as potential therapeutic agents during cocaine withdrawal

Both desensitization and trafficking of DOR are possible mechanisms underlying anxiety and following abstinence from chronic cocaine treatment. Receptor trafficking can regulate the number of receptors available at the cell surface therefore influencing the availability of receptor for stimulation by ligands. Interestingly, results from manipulation of the DOR system suggest a modulatory role for DOR in anxiety (Filliol et al., 2000; Perrine et al., 2005; Saitoh et al., 2005). Moreover, a recent study demonstrated that cocaine-induced anxiety- and depression-like behaviors during early cocaine withdrawal accompanied DOR desensitization (Perrine et al., 2008). Administration of DOR agonist SNC-80 was able to ameliorate these cocaine-induced behaviors (Perrine et al., 2008). The present data report a decrease in plasma membrane DOR available for binding, however, it is possible that DOR agonists are able to activate the functional DOR remaining on the plasma membrane.

Since the NAcb plays a key role in the rewarding aspects of drugs of abuse, continued investigations of dopamine–opioid interactions may elucidate novel targets for the treatment of addiction. Clinical observations indicate that cocaine addicts use opiates to reduce cocaine-induced anxiety and irritability (Kosten et al., 1986; O'Brian, 2001). Moreover, DOR agonists are reported to have fewer negative side effects and lower abuse potential compared to MOR agonists such as morphine (Jutkiewicz, 2006). DOR agonists may be beneficial as a therapeutic for both addiction and other opioid-related disease states (i.e. pain).

Conclusions

These results support potential sex and regional differences in cocaine withdrawal-induced trafficking of DOR in the NAcb. In addition, these results support the existence of a subset of NAcb neurons co-expressing D1R and DOR. Based on these data, it is tempting to speculate that co-existence of opioid and dopamine receptors in common neurons may mediate DOR trafficking during cocaine withdrawal.

Animals

All animal care and procedures were conducted in compliance with the guidelines set forth by the Institutional Animal Care and Use Committee of Thomas Jefferson University and the National Institutes of Health Guide for the Care and Use of Laboratory Animals. In addition, adequate measures were taken to minimize pain or discomfort. Nine male and fourteen female Sprague-Dawley rats (Harlan Laboratories, Indianapolis, IN; age ~ 60 days) were used in this study. Animals were housed three to a cage and maintained on a 12 hour light/dark cycle (lights on at 7 am) with free access to food and water.

Drug Administration

Cocaine HCl (generously provided by NIH/NIDA) was dissolved in saline (0.9%) and injected in a volume of 1 ml/kg body weight. Male and female Sprague-Dawley rats were administered binge-pattern (3 injections/day at 1hour intervals) cocaine (15 mg/kg i.p.) or saline (1ml/kg i.p.) for 14 days followed by a 48 hour withdrawal period. In humans, addicts generally self-administer cocaine repeatedly over a short time period, “binging”, typically in the early evening hours followed by a period of abstinence (Gawin, 1991). This dosing schedule in animals mimics this pattern in terms of temporal pattern and also in relation to circadian rhythm. Following the 48 hours of withdrawal from drug administration, animals were euthanized and brains were processed for either immunohistochemistry or Western blot analysis.

Tissue Preparation for Immunohistochemistry

Data Analysis

For electron microscopy, the classification of neuronal profiles was based on that of Peters et al. (Peters et al., 1991). A profile was defined as immunogold-silver labeled when two or more gold-silver particles were detected (Ambrose et al., 2004). To determine spurious gold-silver deposits in the neuropil, blood vessels and the myelin sheath of myelinated axons, structures not expected to exhibit immunoreactivity for dopamine or opioid receptors, were used as an indicator of the degree of spurious labeling. Of fifty randomly sampled micrographs from tissue processed for the co-localization study, only three randomly sampled myelinated axons contained a gold particle. This demonstrates that minimal spurious gold-silver labeling was present in the tissue used for ultrastructural analysis. A profile was defined as immunolabeled with peroxidase when electron dense deposits were detected in the cytoplasm or along the plasma membranes. These were more electron dense than that of morphologically similar profiles observed within adjacent portions of the neuropil. Gold particles were identified as plasmalemmal if they were associated with the plasma membrane, and cytoplasmic if they were not in contact with the plasma membrane.

Four different treatment groups were used for the analysis of DOR trafficking: Male Cocaine, Male Saline, Female Cocaine, and Female Saline. Each treatment group included 2–3 rats and both core and shell regions of the nucleus accumbens were analyzed. At least forty labeled dendritic profiles were quantified from at least two grids from each brain region of each animal. Unpaired t-tests were used for statistical analysis. For co-localization studies, tissues from three naïve male rats were analyzed. At least 2 grids from each brain region were examined.