Role of Noradrenergic Signaling by the Nucleus Tractus Solitarius in Mediating Opiate Reward

See allHide authors and affiliations

Science  17 Feb 2006:
Vol. 311, Issue 5763, pp. 1017-1020
DOI: 10.1126/science.1119311


Norepinephrine (NE) is widely implicated in opiate withdrawal, but much less is known about its role in opiate-induced locomotion and reward. In mice lacking dopamine β-hydroxylase (DBH), an enzyme critical for NE synthesis, we found that NE was necessary for morphine-induced conditioned place preference (CPP; a measure of reward) and locomotion. These deficits were rescued by systemic NE restoration. Viral restoration of DBH expression in the nucleus tractus solitarius, but not in the locus coeruleus, restored CPP for morphine. Morphine-induced locomotion was partially restored by DBH expression in either brain region. These data suggest that NE signaling by the nucleus tractus solitarius is necessary for morphine reward.

The pleasurable experiences associated with taking a drug encourage repeated usage, which can lead to neurochemical changes that promote addiction. Seminal studies of the neurotransmitters underlying opiate reward implicated the catecholamines NE and dopamine (DA), but the role of DA in this process has received the most attention (1, 2). Although many studies implicate NE in the adverse effects of opiate withdrawal (3), its role in mediating the rewarding and stimulatory effects of opiates remains equivocal. We examined opiate-mediated reward and locomotion in dopamine β-hydroxylase knockout (DBH-KO) mice, which cannot synthesize NE (4, 5). This genetic lesion is specific and complete, but it is conditional in that NE synthesis can be restored by administration of a pseudo-substrate (4).

To assess opiate reward in DBH-KO mice, we used a balanced and unbiased conditioned place preference (CPP) paradigm (6). Rodents can learn to associate the pleasurable aspects of a stimulus with the environments in which they experience that stimulus, and they preferentially seek such environments (referred to as CPP) while avoiding environments where unpleasant stimuli are encountered (conditioned place aversion, CPA). Quantification of time spent in a drug-paired environment after conditioning can be used as a measure of the rewarding properties of that drug (6). DBH-KO mice showed no preference for the morphine-paired environment at any dose tested, but they showed aversion to the drug-paired side at a dose of 25 mg/kg intraperitoneally (ip) (Fig. 1B). In contrast, littermate controls showed a typical CPP to morphine at doses of 15, 20, and 25 mg/kg (Fig. 1A). Noradrenergic neurotrans-mission was restored in DBH-KO mice by administering l-3,4-dihydroxyphenylserine (DOPS), which can be converted to NE in the absence of DBH (4). Treatment with DOPS rescued CPP for morphine in DBH-KO mice but did not affect control mice (Fig. 1, A and B). To ensure that this deficit in CPP was due to lack of morphine reward rather than a learning or attention deficit, we assessed the ability of DBH-KO mice to form a CPP for food; DBH-KO mice showed a robust preference for a food-paired environment (Fig. 1C). Other studies have shown that DBH-KO mice can form a CPP to cocaine (7).

Fig. 1.

CPP and locomotor responses of DBH-KO and control mice to morphine. (A) Control mice treated with various doses of morphine during the conditioning phase develop CPP at intermediate doses (mg/kg ip); administration of DOPS had no effect (n = 7 to 15 mice per group). (B) DBH-KO mice treated with various doses of morphine during the conditioning phase of the CPP paradigm fail to manifest CPP; administration of DOPS before the conditioning with morphine at 20 mg/kg restored CPP (n = 7 to 13 mice per group). Analysis revealed a significant genotype effect but no genotype × dose interaction. (C) Both DBH-KO and control mice form a CPP for food (n = 6 or 7 mice per group). (D) Control mice show increased locomotion in response to all morphine doses tested, whereas the response of DBH-KO mice is significantly blunted (n = 6 to 10 mice per group). CPP data are expressed as means ± SEM of time spent on the drug-paired side before and after conditioning. Locomotor data are expressed as average total beam breaks ± SEM during a 2-hour session. *P < 0.05, **P < 0.01.

Acute administration of morphine also stimulates locomotion in rodents (811). At all doses tested, morphine increased locomotion in control mice; however, DBH-KO mice were significantly less active than controls at each dose of morphine, but not with saline (Fig. 1D). DOPS administration before morphine (20 mg/kg) restored morphine-induced locomotion in DBH-KO mice but did not affect control mice (Fig. 1D).

Two noradrenergic nuclei, the locus coeruleus (LC) and the nucleus tractus solitarius (NTS), have been implicated in the aversive effects of opiate withdrawal (3, 12). We hypothesized that they may also play a role in the rewarding properties of opiates. Therefore, we selectively restored DBH expression in either the LC or the NTS of DBH-KO mice with the use of adeno-associated virus (AAV) encoding DBH and the fluorescent reporter protein DsRed2 (AAV1-DBH-DsRed2, Fig. 2A); we also used a control virus expressing luciferase (AAV1-Luc). Bilateral injection of AAV into the NTS or LC resulted in DsRed expression in the respective nuclei (Fig. 2, B and C) but not in other noradrenergic cell groups (13). Most of the DsRed-positive neurons were noradrenergic (but in the NTS, also possibly adrenergic), as indicated by costaining with antibodies to tyrosine hydroxylase (TH) (Fig. 2, D and E). After AAV1-DBH-DsRed2 injection into the LC, expression of DBH was observed in neuronal processes in the anterior cingulate area of the prefrontal cortex (PFC), a brain region that receives noradrenergic innervation solely from the LC (14) (Fig. 2F). When AAV1-DBH-DsRed2 was injected into the NTS, DBH expression was not observed in the PFC (Fig. 2G), but it was found in neuronal processes in the nucleus accumbens (NAc) (Fig. 2I) and the bed nucleus of the stria terminalis (BNST, fig. S1), which receive the majority of noradrenergic input from the NTS rather than from the LC (12, 14, 15). These latter areas were sparsely labeled after AAV1-DBH-DsRed2 transduction of the LC (Fig. 2H).

Fig. 2.

Restoration of DBH expression in selected noradrenergic nuclei by viral transduction. (A) Schematic of AAV1-DBH-DsRed2 construct that was used to make AAV for injection. (B) Representative coronal section through the brain of an animal injected with AAV1-DBH-DsRed2 bilaterally into the LC. Red cells are neurons expressing the DsRed transgene. (D) The same section labeled with an antibody to TH, a marker of noradrenergic neurons. (C and E) Representative sections through the brain of an animal injected with AAV1-DBH-DsRed2 bilaterally into the NTS with visualization of DsRed (C) and DsRed and TH (E). (F to I) Immunohistochemistry with antibodies to DBH in PFC (anterior cingulate region) (F) or shell of NAc (H) after injection of AAV1-DBH-DsRed2 into the LC; similar sections and staining of PFC (G) and NAc (I) after injection of AAV1-DBH-DsRed2 into the NTS. AC, anterior commissure; Co, NAc core; Sh, NAc shell. Scale bars, 100 μm.

Expression of DBH in selected noradrenergic nuclei restored behavioral responses to morphine. Bilateral injection of AAV1-DBH-DsRed2 into the NTS completely rescued CPP for morphine, whereas injection of AAV1-Luc had no effect (Fig. 3A). In contrast, injections of either AAV1-DBH-DsRed2 or AAV1-Luc into the LC failed to rescue CPP for morphine (Fig. 3B). Injections of AAV1-DBH-DsRed2 into either the LC or the NTS partially restored morphine-induced locomotion, whereas injections of AAV1-Luc into either region had no effect (Fig. 3C). Viral injections into either the LC or the NTS had no effect on locomotor response to saline injection (fig. S2).

Fig. 3.

Behavioral consequences of restoring DBH expression to the NTS or LC. (A) Injection of AAV1-DBH-DsRed2 into the NTS of DBH-KO mice restored morphine CPP, whereas AAV1-Luc did not (n = 7 to 13 mice per group). (B) Injection of AAV1-DBH-DsRed2 into the LC of DBH-KO mice did not restore CPP for morphine, and neither did AAV1-Luc (n = 7 to 13 mice per group). (C) Injection of AAV1-DBH-DsRed2 into either the NTS or the LC of DBH-KO mice partially rescued morphine-induced locomotion, whereas AAV1-Luc had no effect (n = 5 to 11 mice per group). (D) Morphine administration (20 mg/kg) increased the number of cFos-positive cells in the NTS, including those that are TH-positive and TH-negative (n = 4 or 5 mice per group). *P < 0.05.

To explore how opiates affect NTS neurons, we measured cFos activation in the NTS of control mice given a dose of morphine (20 mg/kg) that induces optimal CPP. We observed a significant increase in cFos labeling of both TH-positive and TH-negative NTS neurons relative to saline-injected controls, as previously documented in rats (16) (Fig. 3D) (fig. S3).

We conclude that the inability of DBH-KO mice to form a normal CPP to morphine is due to an inability to experience morphine reward, not to an impairment of their ability to learn a place association or to a general deficit in reward. These mice also manifest a blunted locomotor response to morphine. Neither of these deficits is due to chronic compensatory changes in response to loss of NE, because acute restoration of NE via administration of DOPS rescues these phenotypes. Our data suggest that NTS neurons are activated by morphine and that NE signaling by this nucleus is critical to opiate reward, whereas NE release from both the LC and the NTS contributes to morphine-induced locomotion.

A role for NE in opiate reward was suggested three decades ago (2, 16), in part because inhibition of DBH activity attenuated opiate reward (1). However, subsequent studies of the LC yielded mixed results, and the role of other noradrenergic cell groups was not studied (16). Some recent studies have implicated NE in opiate reward, but the results are hard to interpret. Adrenergic receptor agonists and antagonists have been shown to alter CPP for morphine (1719); however, the results are conflicting (1719) and they are confounded by the fact that some noradrenergic drugs induce CPP or CPA on their own (6, 17) or are not selective for adrenergic receptors (20). The finding that adrenergic receptor (Adra1b)–null mice have attenuated CPP at one dose of morphine supports our results (10); however, these mice have other learning deficits and compensatory changes that confound interpretation of their inability to establish CPP for morphine (2123). A recent study showed that 6-hydroxydopamine lesions of noradrenergic fibers in the PFC, which arise from the LC (14), attenuate morphine CPP (8). The apparent discrepancy with our findings may be explained by differences in technique: Ablation of noradrenergic fibers in the PFC would eliminate not only NE signaling but also that of cotransmitters released from LC neurons. In addition, disruption of noradrenergic signaling in the PFC could lead to deficits in attention or memory (24) that would impair an animal's ability to learn a place association without affecting other forms of memory, thereby disrupting morphine CPP without affecting opiate reward. Because the LC and NTS are extensively interconnected, it is also possible that interactions between these two nuclei regulate morphine CPP in normal mice.

How noradrenergic neurotransmission from the NTS regulates opiate reward remains to be explored. Our cFos data and that of others (25) suggest that morphine increases NE release, and possibly epinephrine release, by NTS neurons. The extended amygdala (EA) is extensively innervated by the NTS (12, 14, 15), and it has been implicated in the aversive properties of opiate withdrawal and in opiate-seeking behaviors (12, 26, 27). It has been proposed that NE derived from the NTS, rather than the LC, mediates these effects (12, 28), which raises the possibility that NE signaling within the EA may also mediate the rewarding aspects of morphine. Which adrenergic receptors are involved and which neuronal circuits are affected by NE signaling from the NTS remain to be identified.

The partial rescue of morphine-induced locomotion that we observed after AAV injection into the LC is probably due to NE release in the PFC, where it is known to regulate opiate-induced locomotion (9). However, our results indicate that NE release from the NTS also contributes to locomotion. Both the LC and the NTS project to the ventral tegmental area, where NE may regulate DA neuron activity (14, 29) and thereby facilitate locomotion and reward (6).

Supporting Online Material

Materials and Methods

Figs. S1 to S3

References and Notes

View Abstract

Navigate This Article