, Volume 180, Issue 4, pp 781–788

Effects of chronic cocaine self-administration on norepinephrine transporters in the nonhuman primate brain


  • Thomas J. R. Beveridge
    • Department of Physiology and PharmacologyWake Forest University School of Medicine
  • Hilary R. Smith
    • Department of Physiology and PharmacologyWake Forest University School of Medicine
  • Michael A. Nader
    • Department of Physiology and PharmacologyWake Forest University School of Medicine
    • Department of Physiology and PharmacologyWake Forest University School of Medicine
Original Investigation

DOI: 10.1007/s00213-005-2162-1

Cite this article as:
Beveridge, T.J.R., Smith, H.R., Nader, M.A. et al. Psychopharmacology (2005) 180: 781. doi:10.1007/s00213-005-2162-1



While cocaine blocks dopamine and serotonin transporters, considerably less emphasis has been placed on its effects following blockade of the norepinephrine transporter (NET). To date, no studies have made a systematic investigation of the effects of chronic cocaine on primate NET density.


We previously reported increases in NET density in portions of the monkey bed nucleus of stria terminalis after 100 days of cocaine self-administration. In the present study we extend these findings and assess the changes in [3H]nisoxetine binding in additional brain regions of rhesus monkeys chronically self-administrating cocaine.


[3H]Nisoxetine binding sites in the A1 noradrenergic cell group were significantly higher after 5 days of cocaine exposure. One hundred days of self-administration also induced a higher density of NET binding within the A1 cell group; however, in addition, the effects extended to the nucleus prepositus, as well as forebrain regions such as hypothalamic nuclei, basolateral amygdala, parasubiculum, and entorhinal cortex.


These data demonstrate that cocaine self-administration alters the noradrenergic system of nonhuman primates. Although cocaine affected NET binding sites in the forebrain projections of both the ventral (VNAB) and dorsal (DNAB) noradrenergic bundles, the alteration in the A1 cell group at the early time-point suggests that the VNAB appears to be more sensitive than the DNAB to the effects of cocaine. Given the role of norepinephrine in arousal and attention, as well as mediating responses to stress, long-term exposure to cocaine is likely to result in significant changes in the way in which information is perceived and processed by the central nervous system of long-term cocaine users.




Activation of the brain dopaminergic system is thought to be the primary mechanism by which cocaine exerts its reinforcing and behaviorally activating effects (de Wit and Wise 1977; Giros et al. 1996; Ritz et al. 1990; Volkow et al. 1997). Cocaine binds to the dopamine transporter (DAT), thereby blocking reuptake of dopamine into presynaptic cells and enhancing dopaminergic transmission at its target sites in the central nervous system. However, at behaviorally active doses, cocaine also binds with high affinity to both norepinephrine (NET) and serotonin transporters (Bennett et al. 1995; Ritz et al. 1990).

Although there is a vast body of work describing the function of dopamine in the etiology and sequelae of cocaine use, the role of the noradrenergic system has only begun to be appreciated, particularly with regard to its contribution to chronic self-administration. In rats trained to discriminate low and high doses of cocaine, norepinephrine (NE) reuptake inhibitors substitute only for low concentrations of cocaine, although they appear to amplify the subjective effects of the drug at a broader range of doses (Kleven and Koek 1997, 1998; Spealman 1995). In addition, the administration of the α2 adrenergic receptor agonist, clonidine, increases stress-induced cocaine reinstatement (Erb et al. 2000), whereas the α2 adrenergic receptor antagonist, yohimbine, decreases it (Lee et al. 2004). Similarly, the infusion of a β12 receptor antagonist mixture into the bed nucleus of stria terminalis (BNST) or central nucleus of the amygdala (CeA) was shown to block stress-induced, but not cocaine-induced, reinstatement of cocaine seeking (Leri et al. 2002).

In addition to the findings linking NE to the pharmacological and behavioral effects of cocaine, several studies indicate that the noradrenergic system may be particularly important in withdrawal from cocaine and opiates. For example, Delfs et al. (2000) demonstrated that NE release in the BNST, a major noradrenergic projection site, is vital for opiate-induced withdrawal-associated aversion in the rat, and discontinuation of chronic cocaine use in humans has been demonstrated to induce heightened noradrenergic tone (McDougle et al. 1994). Moreover, hyperactivity in rat locus coeruleus (LC) neurons has been demonstrated following withdrawal from cocaine (Harris and Williams 1992), supporting an association between enhanced noradrenergic tone and drug withdrawal. Therefore, the involvement of norepinephrine and its transporter in the pharmacological and behavioral effects of cocaine is of potential importance.

Despite this, however, few studies have examined the response of the norepinephrine transporter (NET) to cocaine administration. The rodent NET appears to be fairly unresponsive to exposure to cocaine (Belej et al. 1996; Benmansour et al. 1992), and in HEK-293 cells transfected with human NET cDNA, the NET was unchanged after 3 days of incubation with cocaine (Zhu et al. 2000). However, NET mRNA was shown to increase in response to cocaine administration in rodents (Burchett and Bannon 1997), although Arroyo and colleagues reported no change in expression in a rodent model of the early and post-acquisition stages of cocaine self-administration (Arroyo et al. 2000). In a study of nonhuman primates, binding to the NET transporter was found to be higher within the subdivisions of the BNST following chronic cocaine exposure (Macey et al. 2003). However, there have been no systematic evaluations of the NET across brain regions in cocaine abusers or nonhuman primates exposed to cocaine. The purpose of this study, therefore, was to investigate the effects of cocaine exposure on the distribution of NET binding in nonhuman primates by using the relatively selective ligand [3H]nisoxetine.

Materials and methods


Fourteen adult male rhesus monkeys (Macaca mulatta) served as subjects. All procedures were performed in accordance with the National Institute of Health Guidelines for Proper Care and Use of Laboratory Animals, and all protocols were reviewed by the Wake Forest University Animal Care and Use Committee. All behavioral, surgical, and self-administration procedures used in the present study have been previously described in detail (Nader et al. 2002). Briefly, monkeys were individually housed in stainless steel cages with water ad libitum and their body weight was maintained at ∼90–95% of free-feeding weight (7.5–13.0 kg) by supplemental feeding of Purina Monkey Chow. Additionally, animals received fresh fruit and/or peanuts at least 3 days per week. Cocaine self-administration and food-reinforced responding occurred in standard operating chambers (0.915×0.74×0.76 m; Med Associates, East Fairfield, VT), which were designed to accommodate a primate chair (Model R001, Primate Products, Redwood, CA).


Details of surgical and self-administration procedures for this group of monkeys have been previously described (Nader et al. 2002). Briefly, monkeys were initially trained to respond on one of two levers under a fixed-interval (FI) 3-min schedule for 1-g banana-flavored pellets. Daily training sessions terminated after 30 reinforcers were obtained and a minimum of 20 sessions with stable performance (±20% of the mean for three consecutive sessions) satisfied training requirements. Upon achieving response stability, the feeder was unplugged and the effects of extinction on responding were examined for five consecutive sessions. Following extinction, food-maintained responding was reinstated for all monkeys. Indwelling intravenous catheters were implanted into the femoral vein and subcutaneous ports were positioned in the lower back in all monkeys. After surgery, food-reinforced responding was re-established, after which monkeys were divided into three groups. Control monkeys (n=6) continued responding under the food-reinforced FI 3-min schedule for the duration of the study (n=3 for 5 days, n=3 for 100 days). NET binding from these two control groups was essentially the same, therefore the results were pooled. Experimental monkeys (n=4 per group) began self-administration of cocaine (0.3 mg/kg per injection) for a period of 5 or 100 days. Daily sessions ended after 30 reinforcers were obtained. Detailed analyses of the behavioral data have been reported previously (Nader et al. 2002). Total daily cocaine intake per session was 9.0 mg/kg, and total lifetime intake of cocaine was 45 mg/kg for the initial phase group and 900 mg/kg for the chronic phase group.

Tissue processing

Monkeys utilized in this study were also used for 2-deoxy-d-[14C] glucose (2-DG) analysis (Beveridge et al. 2004; Letchworth et al. 2001; Macey et al. 2003; Nader et al. 2002; Porrino et al. 2002, 2004). The 2-DG experiment was initiated within 2 min of the final reinforcer on the last self-administration session (either food or cocaine reinforcement at 5 or 100 day time points), involved sampling of arterial blood, and lasted ∼45 min. Immediately after the 2-DG procedure, animals were euthanized with sodium pentobarbital (100 mg/kg, i.v.). Brains were immediately removed, blocked, frozen in isopentane (−35 to −80°C), and then stored at −80°C. Coronal sections (20 μm) were cut on a cryostat, thaw-mounted onto chrome–alum/gelatin-subbed or electrostatically charged slides, desiccated, and stored at −80°C until processing.

[3H]Nisoxetine autoradiography

Procedures for [3H]nisoxetine autoradiography were adapted from Tejani-Butt (1992). Tissue sections were preincubated at room temperature in buffer (50 mM Tris, 300 mM NaCl, 5 mM KCl, pH 7.4) for 20 min to remove any residual cocaine and 2-DG. Sections were then incubated for 4 h at 4°C in buffer containing 3 nM [3H]nisoxetine (80 Ci/mmol; Perkin-Elmer Life Sciences, Boston, MA) in the presence (nonspecific binding) or absence (total binding) of 1 μM mazindol. Sections were rinsed three times (5 min each) in buffer at 4°C, with a final 10-s rinse in ice-cold water. Sections were immediately dried under a stream of cold air and placed on Hyperfilm-3H (Amersham, Arlington Heights, IL) for 2–6 weeks in the presence of [3H] standards (Amersham). After appropriate exposure times, films were developed with Kodak GBX developer, fixed, and rinsed.

Densitometry and data analysis

Brain regions were identified using Nissl-stained sections adjacent to those used for autoradiography. The areas analyzed were chosen based on their afferents from either the ventral noradrenergic bundle (VNAB) or dorsal noradrenergic bundle (DNAB), as well as the cells of origin of these two pathways. Nomenclature and identification of structures were determined according to the atlas of Paxinos et al. (2000). Analyses of autoradiograms were conducted by quantitative densitometry with a computerized image processing system (MCID, Imaging Research, St. Catherines, Ontario, Canada). Tissue equivalent values (in fmol/mg wet weight tissue) were determined from optical densities using a calibration curve obtained by densitometric analysis of the autoradiograms of [3H] standards. Specific binding was determined by subtracting nonspecific binding values from the total binding values, measured in adjacent sections. Statistical analysis was performed for each brain structure measured via a one-way analysis of variance, followed by multiple comparisons least significant difference (LSD) comparing the drug treatment group to controls.


Animals in which responding was maintained by food presentation (controls) showed a heterogeneous distribution of [3H]nisoxetine binding. Within the areas measured, the highest densities were observed in the noradrenergic cell bodies of the LC and parabrachial nucleus. More moderate densities were found in hypothalamic nuclei and the central nucleus of the amygdala, and the lowest density in the striatum and hippocampal formation (Table 1).
Table 1

Binding of [3H]nisoxetine in selected nonhuman primate brain regions

Brain region


Initial cocaine (5 days)

% Change

Chronic cocaine (100 days)

% Change

Amygdala and striatum

Bed nucleus of stria terminalisa






Central amygdala






Basolateral amygdala






Caudate nucleus






Putamen nucleus






Nucleus accumbens core






Nucleus accumbens shell







Supraoptic nucleus






Medial preoptic






Lateral preoptic






Periventricular nucleus






Paraventricular nucleus






Dorsomedial hypothalamus






Ventromedial hypothalamus






Lateral hypothalamus






Arcuate nucleus







Anterior cingulate






Area 25






Area 32






Posterior Cingulate






Entorhinal cortex






Hippocampal formation




















Locus coeruleus






Nucleus of tractus solitarius






A1 nucleus






Parabrachial nucleus






Nucleus prepositus






Data represent means±SEM (fmol/mg tissue)

% change Percentage difference from control values

*p<0.05, compared to controls

**p<0.01, compared to controls

***p<0.001, compared to controls

aData previously reported (Macey et al. 2003)

After 5 days of cocaine self-administration, [3H]nisoxetine binding sites were significantly higher within the A1 cell group (43%, p<0.05), as opposed to the other noradrenergic cell bodies in the brainstem, such as the locus coeruleus (LC), parabrachial nucleus and nucleus tractus solitarius (NTS), none of which showed significant effects (Table 1) compared to controls. No other cell body, or target brain region of the noradrenergic system, showed significant increases at this initial time point.

Animals self-administering cocaine for 100 days, however, showed substantial increases in NET binding throughout the brain. These alterations were within discrete regions of the brainstem, hypothalamus, extended amygdala, cortex, and hippocampal formation. In addition to the A1 cell body (48%; p<0.05), chronic cocaine self-administration also induced a significantly higher [3H]nisoxetine binding in another brainstem region, the nucleus prepositus (65%; p<0.01), although the LC, NTS, and parabrachial nucleus again failed to change significantly. In other areas receiving ventral noradrenergic projections, such as the hypothalamus, widespread increases in [3H]nisoxetine binding density were observed at both anterior and posterior levels, with large significant increases in the periventricular (63%; p<0.05) and medial preoptic nuclei (59% p<0.01), whereas the paraventricular (35%; p<0.05) and supraoptic (36%; p<0.05) nuclei exhibited more moderate increases (Figs. 1, 2). Other hypothalamic areas, such as the dorsomedial and lateral hypothalamus, as well as the lateral preoptic and arcuate nuclei all showed higher levels of NET binding compared to controls, although they failed to reach significance (Table 1). In addition, 100 days of cocaine self-administration also induced significant increases in NET binding sites within the BNST (45% p<0.01; Macey et al. 2003), although the CeA, another region shown to receive dense ventral noradrenergic projections (Roder and Ciriello 1993; 1994), failed to show significant effects.
Fig. 1

Schematic diagram showing brain regions measured, and degree of increase in [3H]nisoxetine binding following chronic cocaine exposure, at anterior and posterior hypothalamus and amygdala. Yellow indicates an increase in norepinephrine transporter binding of 20–34%, orange an increase of 35–49%, and red an increase of more than 50%, compared to controls. GP globus pallidus, ic internal capsule, BNST bed nucleus of stria terminalis, ac anterior commissure, ot olfactory tubercle, Amy amygdala, MPOA medial preoptic area, LPOA lateral preoptic area, SO supraoptic area, Pa paraventricular hypothalamus, L lateral hypothalamus, RTN reticular thalamic nucleus, DM dorsomedial hypothalamus, Pe periventricular hypothalamus, VM ventromedial hypothalamus, Arc arcuate nucleus, CeA central amygdala, BLA basolateral amygdala, ER entorhinal cortex
Fig. 2

Pseudocolor enhanced autoradiogram of hypothalamus and amygdala of control monkey (a) and monkey self-administering cocaine for 100 days (b). CeA central amygdala, MeA medial amygdala, LA lateral amygdala, BLA basolateral amygdala, AB accessory basal amygdala, VM ventromedial hypothalamus, Arc arcuate nucleus, Pa paraventricular hypothalamus, Pe periventricular hypothalamus, L lateral hypothalamus

Regions receiving noradrenergic projections via the dorsal bundle, such as the basolateral amygdala (69%; p<0.05; Fig. 2), parasubiculum (89%; p<0.05), and entorhinal cortex (69%; p<0.05), also showed significantly higher NET binding following chronic cocaine exposure (Table 1). However, there were no significant differences in other areas, such as the anterior and posterior cingulate cortex, Brodmann’s areas 25 and 32, hippocampal formation, striatum, and nucleus accumbens (Table 1).


These data clearly demonstrate that cocaine self-administration alters the noradrenergic system of nonhuman primates, producing significant changes in the density of the NET throughout the brain including the cells of origin of forebrain noradrenergic projections, portions of the extended amygdala, hypothalamus, and hippocampus. These changes begin early in the course of experience with cocaine self-administration with increases in NET binding in the A1 region of the brainstem, the source of noradrenergic neurons in the ventral noradrenergic bundle, apparent after only 5 days of exposure to cocaine. As cocaine exposure becomes more chronic, elevations in the density of the NET were observed throughout the forebrain, suggesting a widespread compensatory response to the increased levels of norepinephrine that result from NET blockade by cocaine. Therefore, in addition to the well-documented effects of cocaine self-administration on dopaminergic systems in human cocaine abusers and animal models of cocaine self-administration (Letchworth et al. 2001; Little et al. 1993; Mash et al. 2002), these data show significant alterations in both the dorsal noradrenergic projections of the locus coeruleus and the ventral system of projections to the hypothalamus and other forebrain regions. These data also extend previous findings from this laboratory demonstrating alterations in NET binding of the bed nucleus of the stria terminalis (Macey et al. 2003) and functional activity in brain regions regulating cardiovascular function (Beveridge et al. 2004) as a result of chronic cocaine exposure in nonhuman primates.

Although the concentration of NET binding sites in the forebrain projections of both the VNAB and DNAB were both affected by cocaine self-administration, the VNAB appears to be more sensitive to the effects of cocaine than the DNAB. The cells of origin of the VNAB show significant increases in NET concentration suggesting changes in autoreceptor regulation of the firing of these cells, whereas the density of NET binding in the cells of origin of the DNAB, the LC, was largely unchanged by cocaine. Furthermore, the only brain region to show significant changes in NET binding following initial cocaine exposure was the A1 cell group. Given that compensatory changes occurred at such an early time point, these data suggests that the A1 cell group may be one of the first regions to change in response to cocaine self-administration and may subsequently affect downstream structures in a progressive manner (through its extensive ventral projections).

In addition, the primary projections of the VNAB, that include the BNST, and hypothalamus, were the brain regions that showed some of the largest changes in NET density. Involvement of the hypothalamus in the endocrine and neurobehavioral effects of cocaine has also been demonstrated, largely through its participation in stress-related responses as a component of the HPA axis, in which it is the source of corticotrophin-releasing factor (CRF) (for review see Goeders 2002). There is evidence to suggest, for example, that acute cocaine activates the HPA axis (Broadbear et al. 1999; Forman and Estilow 1988; Levy et al. 1991; Sarnyai et al. 1996) and that this activation is due to the enhanced release of CRF by the hypothalamus (Calogero et al. 1989; Rivier and Lee 1994; Teoh et al. 1994). In addition, microinjection of CRF into the BNST, but not the CeA, reinstates cocaine seeking (Erb and Stewart 1999), demonstrating a differential contribution of CRF in these brain regions. Thus heightened noradrenergic tone elicited by chronic cocaine, resulting in a stimulatory effect on CRF release from the hypothalamus via excitation of CRF-releasing cells, may contribute to the dysregulation of HPA axis activity.

Norepinephrine in the BNST, which also receives projections from the A1 cell group (Aston-Jones et al. 1999; Roder and Ciriello 1994; Terenzi and Ingram 1995), has been shown to be a critical component in the etiology of withdrawal from drugs of abuse (Aston-Jones et al. 1999; Delfs et al. 2000). Heightened noradrenergic tone in this region may be critical for the aversive properties of drug withdrawal (Aston-Jones et al. 1999; Leri et al. 2002).

One interesting exception to the robust changes in the projection fields of the VNAB is the central nucleus of the amygdala. [3H]Nisoxetine binding in the CeA was unaffected by chronic cocaine exposure despite dense basal levels of NET. This result paralleled an earlier finding of a lack of effect of cocaine on local cerebral glucose utilization in the same animals, at the same time point (Beveridge et al. 2004). The CeA, as the caudal extent of the lateral division of the extended amygdala, shares many neurochemical, cytoarchitectural, and connectivity characteristics with the lateral division of the BNST (Alheid et al. 1998; Alheid and Heimer 1988; Martin et al. 1991). Given this close association with the BNST, the lack of effect of cocaine in the CeA is somewhat surprising, in view of the role this nucleus has been shown to play in mediating reinstatement of cocaine self-administration following abstinence (Erb et al. 2001; Leri et al. 2002).

Although they were not as widespread as the changes in NET density in the projections of the VNAB, areas innervated by the DNAB also exhibited changes in the concentration of the NET. For example, there was a large effect of cocaine on NET binding density in the basolateral amygdala (BLA), which receives the majority of its projections from the LC, as well as within the hippocampal formation, including the parasubiculum and entorhinal cortex. Connections between the basolateral amygdala and parasubiculum have been previously demonstrated (van Groen and Wyss 1990). In fact, the entorhinal cortex receives its most dense inputs from the basolateral amygdala (Room and Groenewegen 1986), and has strong reciprocal connections with the parasubiculum (van Groen and Wyss 1990). Indeed, the interactions of the BLA and entorhinal cortex with the hippocampus in memory formation were recently linked to norepinephrine neurotransmission (Ferry et al. 1999; Roozendaal et al. 1999). The changes in the noradrenergic system within these brain regions, then, could have significant influences on learning and memory functions and contribute to the cognitive deficits that have been associated with cocaine addiction (Rogers and Robbins 2001).

One observation to come from these results is that, although the increases in NET that we see here in monkeys are substantial, there have been no analogous findings in the rodent, where studies have largely reported decreases. For example, Belej et al. (1996) found decreases in rat [3H]nisoxetine binding densities in the inferior olive, lateral parabrachial area, and the ventral BNST after 30 days of oral cocaine. Other studies have reported that various cocaine administration paradigms have had no effect on NET binding sites (Benmansour et al. 1992), NET mRNA (Arroyo et al. 2000), or brain NE levels (Karoum et al. 1990; Yeh and De Souza 1991). On the other hand, Burchett and Bannon (1997) reported a small but significant upregulation of NET mRNA in the locus coeruleus after 14 days of cocaine administration. However, it should be noted that these studies, with the exception of Arroyo et al. (2000), administered cocaine noncontingently, in contrast to the present study (which utilized self-administration). Although there are other differences such as duration of exposure, length of withdrawal, etc., the conflicting results may indicate significant differences across species. Regardless, care should be taken in extrapolating from rat to human cocaine abuse.

In summary, these data have demonstrated considerable adaptations within the noradrenergic system of monkeys as a result of exposure to cocaine self-administration. There is a significant upregulation of norepinephrine transporters within the hypothalamus, bed nucleus of the stria terminalis, as well as portions of the amygdala and hippocampal formation. Although the noradrenergic system does not appear to mediate the rewarding properties of cocaine directly (de Wit and Wise 1977; Sora et al. 2001), its alteration may be an important consequence of chronic cocaine exposure, contributing to many of the changes that cocaine users undergo during the process of addiction. In particular, these changes may play an important role in withdrawal, in view of the effectiveness of the β adrenoceptor antagonist propranolol in decreasing cocaine withdrawal symptoms in human cocaine addicts, especially in those reporting the most severe symptoms of withdrawal (Kampman et al. 2001). Furthermore, given the role of norepinephrine in arousal and attention, as well as mediating responses to stress, long-term exposure to cocaine is likely to result in significant changes in the way in which information is perceived and processed by the central nervous system of long-term cocaine users.


This study was supported by a grant (DA09085) from the National Institute on Drug Abuse.

Copyright information

© Springer-Verlag 2005