Noradrenaline-induced release of newly-synthesized accumbal dopamine: differential role of alpha- and beta-adrenoceptors

Previous studies have shown that intra-accumbens infusion of isoproterenol (ISO), a beta-adrenoceptor-agonist, and phenylephrine (PE), an alpha-adrenoceptor-agonist, increase the release of accumbal dopamine (DA). In the present study we analyzed whether the ISO-induced release of DA is sensitive to pretreatment with the DA synthesis inhibitor alpha-methyl-para-tyrosine (AMPT). Earlier studies have shown that the PE-induced release of DA is derived from DA pools that are resistant to AMPT. In addition to PE, the alpha-adrenoceptor-antagonist phentolamine (PA) was also found to increase accumbal DA release. Therefore, we investigated whether similar to the DA-increasing effect of PE, the DA increase induced by PA is resistant to AMPT. Pretreatment with AMPT prevented the ISO-induced increase of accumbal DA. The accumbal DA increase after PA was not reduced by the DA synthesis inhibitor, independently of the amount of DA released. These results show that mesolimbic beta-, but not alpha-adrenoceptors, control the release of accumbal newly-synthesized DA pools. The DA-increasing effects of PE have previously been ascribed to stimulation of presynaptic receptors located on noradrenergic terminals, whereas the DA-increasing effects of PA and ISO have been ascribed to an action of these drugs at postsynaptic receptors on dopaminergic terminals. The fact that AMPT did not affect the accumbal DA response to PE and PA, whereas it did prevent the accumbal DA increase to ISO, supports our previously reported hypothesis that the noradrenergic neurons of the nucleus accumbens containing presynaptic alpha-adrenoceptors impinge upon the dopaminergic terminals in the nucleus accumbens containing postsynaptic adrenoceptors of the alpha but not of the beta type. The putative therapeutic effects of noradrenergic agents in the treatment of DA-related disorders are shortly discussed.


INTRODUCTION
It has previously been demonstrated that the release of mesolimbic noradrenaline directs the release of mesolimbic dopamine (DA) (in vitro studies: Nurse et al., 1984;Russell et al., 1993; in vivo studies: Cools and Tuinstra, 2003;. Both, stimulation of accumbal beta-adrenoceptors by the agonist isoproterenol (ISO), and inhibition of accumbal alphaadrenoceptors by the antagonist phentolamine (PA) have been found to facilitate accumbal DA release (see Figure 1 in Verheij and Cools, 2009b). The DA increase induced by ISO and PA have previously been ascribed to the binding of these agents at accumbal postsynaptic adrenoceptors located on dopaminergic nerve terminals (see Figure 1 in Verheij and Cools, 2009b). The mainly postsynaptic action of these agents has been confirmed in more recent studies showing that intra-accumbal administration of various DA-increasing doses of either beta-adrenoceptor agonists or alpha-adrenoceptor antagonists Aono et al., 2013), did not affect accumbal noradrenaline levels (Aono et al., 2007(Aono et al., , 2013. Interestingly, phenylephrine (PE), an alpha-adrenoceptor agonist, has been found to act mainly at the presynaptic adrenoceptors of the nucleus accumbens (see Figure 1 in Verheij and Cools, 2009a). As expected, stimulation of these presynaptic receptors by PE leads to a decreased accumbal noradrenaline release (see Figure 1 in Aono et al., 2007) that in turn was found to result in an increase of the extracellular levels of accumbal DA (see Figure 1 in Verheij and Cools, 2009a). Given that a decrease of accumbal noradrenaline at the level of the DA-stimulating postsynaptic beta-adrenoceptors would have resulted in a reduction of the release of accumbal DA (see also : Mizoguchi et al., 2008), it has previously been hypothesized that those noradrenergic neurons that contain presynaptic alpha-adrenoceptors do not impinge upon dopaminergic terminals that contain postsynaptic adrenoceptors of the beta type (see Figure 1 in . Accordingly, the finding that PE increased accumbal DA release suggests that those noradrenergic neurons that contain presynaptic alpha-adrenoceptors, impinge upon dopaminergic terminals that contain postsynaptic adrenoceptors of the DAinhibitory alpha type (see Figure 1).
Accumbal DA can be released from two types of DA pools (Ewing et al., 1983;Leviel et al., 1989;Cools, 2007, 2009b). The first pool consists of newly-synthesized DA and is sensitive to the DA synthesis inhibitor AMPT, whereas the second pool consists of stored DA and is sensitive to the DA vesicle depleter reserpine (RES; for review: . Microdialysis studies have shown that RES counteracts the PA-, but not the ISO-induced accumbal DA release (Verheij and Cools, 2009b). The first aim of the present study was to investigate whether AMPT may counteract the ISO-induced increase of accumbal DA release (see Figures 1A,B). We have previously shown that the PE-induced release of DA is derived from DA pools that are insensitive to AMPT (Verheij and Cools, 2009a). The second aim of the present study was to investigate whether the accumbal DA increasing effects of the postsynaptic acting alphaagent PA are, in addition to the accumbal DA increasing effects of the presynaptic acting alpha-agent PE, resistant to AMPT (see Figures 1C,D). The third and final aim of the present study was to provide additional evidence for the abovementioned hypothesis that the presynaptic alpha-adrenoceptor of the nucleus accumbens are located on those noradrenergic nerve terminals that impinge upon dopaminergic neurons containing postsynaptic alpha-, but not beta-adrenoceptors (see Figures 1E,F).
The subjects used in the present study were low responder (LR) and high responder (HR) to novelty rats (Piazza et al., 1989;Dellu et al., 1996;Bevins et al., 1997;Cools and Gingras, 1998;Cools and Tuinstra, 2003;Kabbaj, 2004;Verheij et al., 2009;Verheij and Cools, 2011). Given that intra-accumbens administration of PA has previously been found to result in a larger release of accumbal DA in LR than in HR , these rats represent a useful tool to study whether the expected lack of effects of AMPT in PA-treated rats is (in)dependent on the amount of DA that is released.

SUBJECTS
Adult male LR (n = 51) and HR (n = 51) rats of 180-220 g were selected from the outbred strain of Nijmegen Wistar rats (see Section Open-Field Selection Procedure). All rats were reared and housed in Macrolon cages (42 × 26 × 15 cm; n = 3-4 per cage) under a fixed 12/12 h light/dark cycle (lights on: 07.00 a.m.) in a temperature-controlled room (21 ± 1.7 • C). Water and food FIGURE 1 | Noradrenaline-(NA)-induced changes in accumbal DA release. Postsynaptic beta-(β)-adrenoceptors stimulate (+) the release of accumbal DA (A), whereas postsynaptic alpha-(α)-adrenoceptors inhibit (−) the release of accumbal DA (C). The first aim of the study was to investigate whether the postsynaptic acting beta-adrenoceptor agonist ISO increases DA release from AMPT sensitive pools of DA (B). The second aim of the study was to investigate whether the postsynaptic acting alpha-adrenoceptor antagonist PA increases DA release from AMPT-resistant pools of DA (D). Presynaptic alpha-(α)-adrenoceptors regulate accumbal NA levels (E). The presynaptic acting alpha-adrenoceptor agonist PE is known to reduce accumbal NA release (F). The third aim of this study was to investigate whether the noradrenergic neurons that contain presynaptic alpha-adrenoceptors impinge upon dopaminergic neurons that contain postsynaptic alpha-, but not beta-adrenoceptors (F).
pellets (Ssniff, Soest, Germany) were available ad libitum, except during the testing periods. All experiments were performed in accordance with institutional, national and international guidelines for animal care and welfare (see NRC 2003 guidelines). Every effort was made to minimize the number of animals used and their suffering.

SELECTION OF LR AND HR TO NOVELTY
Rats were individually housed 3 days before the open-field selection procedure Verheij and Cools, 2007). Testing took place between 09.00 h and 17.00 h in a room illuminated by white light of 170 Lux at the middle of the openfield. Rats were placed on a black square table with no walls (160 × 160 cm) for a period of 30 min. This open-field was 95 cm elevated above the floor and surrounded by a white neutral background (270 × 270 × 270 cm). As described by Cools et al. (1990), behavior was recorded with a computerized tracking system. Both ambulation and habituation time were used to select LR and HR. Ambulation was defined as the overall distance (cm) traveled in 30 min. Habituation time was defined as the duration of the period(s) that started as soon as the rat began to explore the open-field and ended as soon as the locomotor activity stopped for at least 90 s. Rats that habituated in less than 480 s and walked less than 4800 cm in 30 min were labeled LR, whereas rats that habituated after 840 s and walked more than 6000 cm in 30 min were labeled HR Verheij and Cools, 2007). Habituation time in addition to ambulation was used as selection criterion, because traveled distance per se is not always a reliable criterion (Cools et al., 1997;Saigusa et al., 1999). Rats that did not fulfill the criteria were excluded from this study. Efforts were made to use these rats in other studies.

SURGERY
One day after the open-field selection procedure took place, LR and HR were unilaterally implanted with a stainless steel guide cannula (length: 5.5 mm, outer diameter: 0.65 mm, inner diameter: 0.3 mm) directed to the right nucleus accumbens according to previously described procedures Verheij and Cools, 2007). Rats, anesthetized with pentobarbital (60 mg/kg, i.p.), were placed in a stereotactic apparatus and the following coordinates were used according to the atlas of Paxinos and Watson (1986): anterior: +10.6 mm (relative to the interaural line) and lateral: −1.5 mm (relative to the midline suture). The guide cannula was lowered 5.5 mm relative to the dura surface resulting in a vertical coordinate of +3.5 mm for the cannula tip. Finally, the cannula was angled 10 • laterally to the right side. The rats were allowed to recover from surgery for the next 7-10 days in Plexiglas dialysis cages (25 × 25 × 35 cm) covered with sawdust on the floor. On 3 consecutive days just prior to the start of the microdialysis experiment, each rat was gently picked up in order to habituate to the procedure assessed on the day when the concentration of accumbal DA was measured. This handling procedure was repeated three times per day.

MICRODIALYSIS
As previously described, a dialysis probe (type A-I-8-02, outer diameter: 0.22 mm, 50,000-molecular-weight cut-off, Eicom, Tokyo, Japan) was carefully inserted into the brain of a conscious rat and secured to the guide cannula with a screw Verheij and Cools, 2007). The tip of the dialysis probe protruded 2 mm below the distal end of the guide cannula into the nucleus accumbens. The probes had an in vitro recovery of 10-12% for DA. The inlet and outlet of the probe were connected to a swivel allowing the rat to move undisturbed. Accumbal dialysates were analyzed for DA (pg/40 µl) according to previously described procedures Verheij and Cools, 2007). Briefly, the probe was perfused at a rate of 2.0 µl/min with modified Ringer solution (see Section Compounds) and the outflow was collected in a sample loop and injected, once every 20 min, into a high performance liquid chromatography (HPLC) system. DA was separated from the remaining neurotransmitters by means of reversed phase, ion-paring, liquid chromatography and the concentration was measured using electrochemical detection (ECD). The HPLC-ECD system (HTEC-500: Eicom, Tokyo, Japan) was calibrated with a standard DA solution before and after each experiment. The detection limit was 500 fg per sample.
All rats were randomly divided over the different treatment groups. The noradrenergic drugs were administered 100 min after AMPT, because at this time AMPT still reduced the levels of accumbal DA (see Section Results and Watanabe et al., 2005), whereas the amount of accumbal noradrenaline has previously been found to be returned to baseline levels (Verheij and Cools, 2009a). The dose of PA and ISO were chosen because these doses have been shown to increase extracellular levels of DA in the nucleus accumbens in a receptor-specific manner Verheij and Cools, 2009b

HISTOLOGY
At the end of each experiment the rat was deeply anesthetized with an overdose of sodium-pentobarbital (60 mg, i.p.) and intracardially perfused with 60 ml 4% paraformaldehyde solution. Vibratome sections (100 µm, Leica VT1000F; Leica, Rijswijk, The Netherlands) were cut to determine the exact location of the microdialysis probe.

COMPOUNDS
The following compounds and solutions were used Verheij and Cools, 2009b): (1)

STATISTICAL ANALYSIS
The drug-induced changes of accumbal DA were expressed as a percentage of the concentration of DA that was measured in the 20-min-lasting period just before these drugs were infused. All data are expressed as the mean ± SEM. A three-way ANOVA with the factors rat type, treatment, and time (for repeated measures) was assessed followed by a post-hoc Student's t-test where appropriate. A probability level of p < 0.05 was taken as significant. SPSS for Windows (Release 12) was used to statistically analyze the data.

OPEN-FIELD SELECTION PROCEDURE
The open-field selection procedure revealed 25% LR and 26% HR. The average distance traveled in 30 min (±SEM) was 3425 ± 123.6 cm and 8273 ± 233.0 cm in LR and HR, respectively. The average habituation time (±SEM) was 353 ± 20.8 s in LR and 1322 ± 58.2 s in HR.

HISTOLOGY
Following histological verification, 1 LR treated with the solvent of AMPT and PA had to be excluded from the analysis due to an incorrect placement of the microdialysis probe (number of rats included in data analysis: 51 − 1 = 50 LR and 51 − 0 = 51 HR). The coronal region of the nucleus accumbens in which all correctly placed microdialysis probe tracks were located is shown in Figure 2.

EFFECTS OF ALPHA-METHYL-PARA-TYROSINE
As previously reported (Saigusa et al., 1999;Verheij and Cools, 2007;, accumbal DA levels increased less in novelty-challenged LR than in novelty-challenged HR (Figure 3: rat type × time effect: F (7,336) = 4.304, p < 0.001). The noveltyinduced increase of DA lasted 60 min in LR ( Figure 3A: one FIGURE 2 | (A) Example of three unilateral microdialysis probe tracks located in the right nucleus accumbens. The probe protrudes 2 mm below the distal end of the guide cannula. (B-D) Schematic illustration of coronal brain sections containing the nucleus accumbens. The brain region in which correctly placed probes were found is indicated as a gray oval. IA corresponds to the distance (mm) from the interaural line according to Paxinos and Watson (1986), NAC = nucleus accumbens, CPU = caudate putamen, CC = corpus callosum. Both PA and ISO were infused 100 min after AMPT was given. This infusion point was labeled basal sample b4. It is important to note that the DA levels of the control animals were returned to baseline when PA and ISO were infused, whereas the DA levels in AMPT-treated animals were still reduced. Data are expressed as mean percentage ± SEM. The solid horizontal line represents the infusion time of AMPT. # Significant DA increase after novelty (DA levels were compared to baseline levels using one sample t-tests). * Significant DA decrease after AMPT (DA levels of AMPT-treated rats were compared to DA levels of control (Ringer) animals using Student's t-tests).

RATS ARE AT REST DURING THE INFUSION OF ISOPROTERENOL AND PHENTOLAMINE
As mentioned above, the DA levels of the control rats did not differ anymore from the baseline levels of DA at t = 100 min after novelty (Figure 3: one sample t-test at t = 100 min: LR: ns, HR: ns). In addition, the accumbal DA levels of these rats did not change over time during the period after t = 100 min (Figures 4, 6: time effect: LR: ns, HR: ns). These data demonstrate that both LR and HR were not anymore challenged at the time of the infusion of ISO or PA.

EFFECTS OF ISOPROTERENOL IN ALPHA-METHYL-PARA-TYROSINE-TREATED RATS
The dose of 0.001 mM of the beta-adrenoceptor-agonist ISO equally increased accumbal DA levels in non-challenged LR and non-challenged HR that were treated with the solvent of AMPT (Figure 4: treatment × time effect: F (5,150) = 12.477, p < 0.001; rat type × treatment (× time) effect: ns). AMPT equally reduced this ISO-induced DA increase in both type of rats ( Figure 5: treatment × time effect: F (5,140) = 9.045, p < 0.001; rat type × treatment (× time) effect: ns). ISO did not increase anymore accumbal DA levels in either AMPT-treated LR or AMPT-treated HR (Figure 5: treatment (× time) effect: ns).

EFFECTS OF PHENTOLAMINE IN ALPHA-METHYL-PARA-TYROSINE-TREATED RATS
The dose of 0.01 mM of the alpha-adrenoceptor-antagonist PA increased accumbal DA levels in both LR and HR that were treated with the solvent of AMPT (Figure 6). The PA-induced increase of accumbal DA was larger in non-challenged LR than in nonchallenged HR (Figure 6: rat type × treatment × time effect:  In fact, the DA-increasing effects of PA in AMPT-treated rats were not different from the DA-increasing effects of PA in rats that were treated with AMPT solvent (Figure 7: treatment (× time) effect: ns).

DISCUSSION
We have previously demonstrated that binding of the betaadrenoceptor agonist ISO to the postsynaptic beta-adrenoceptors of the nucleus accumbens increases accumbal DA release (Figure 1). The first aim of the present study was to investigate whether this ISO-induced DA release is derived from AMPTsensitive pools of DA (see also Figure 1). In addition to the presynaptic acting alpha-adrenoceptor-agonist PE, the postsynaptic acting alpha-adrenoceptor-antagonist PA was also found to increase accumbal DA release (Figure 1). The second aim of this study was to investigate whether the DA-increasing effects of PA were like the DA-increasing effects of PE resistant to AMPT (see also Figure 1).

DUAL ROLE OF NORADRENALINE IN MEDIATING DOPAMINE RELEASE
As expected, intra-accumbens infusion of PA and ISO increased accumbal DA levels in rats treated with the solvent of AMPT (Figures 4, 6). This accumbal DA increase may very well explain the previously reported increase in locomotor activity observed following the inhibition and stimulation of accumbal alpha-and beta-adrenoceptors, respectively (Verheij et al., accepted). It has been demonstrated previously that the DA increase following intra-accumbens administration of PA and ISO are dose-dependent and receptor-specific . The present study also confirms the previously reported finding that the effects of ISO do not differ between non-challenged LR and HR , whereas the effects of PA are larger in the former than in the latter . These results have previously been ascribed to individual differences in the noradrenergic activity at accumbal postsynaptic alpha-, but not beta-adrenoceptors in non-challenged LR and HR rats (for details: .  Verheij and Cools, 2009b). (C-D) An increase of NA at the level of the postsynaptic alpha-, but not beta-adrenoceptors of the nucleus accumbens inhibits (-) accumbal DA release from RES sensitive vesicular DA pools (see: Verheij and Cools, 2009b) that are resistant to AMPT (present study). (E-F) The presynaptic alpha-adrenoceptors of the nucleus accumbens (see: Verheij and Cools, 2009a) are known to inhibit (-) accumbal NA release (see: Aono et al., 2007). The present study indicates that this reduction of accumbal NA leads to an increase of accumbal DA (see also: Verheij and Cools, 2009a) because the noradrenergic neurons containing presynaptic alpha-adrenoceptors impinge upon dopaminergic neurons equipped with postsynaptic alpha-, but not beta-receptors.

EFFECTS OF ALPHA-METHYL-PARA-TYROSINE ON THE DOPAMINE RELEASE AFTER ISOPROTERENOL OR PHENTOLAMINE
The finding that AMPT completely prevented the ISO-induced increase of accumbal DA levels in both LR ( Figure 5A) and HR ( Figure 5B) shows that ISO increases DA release from AMPT-sensitive DA pools. Importantly, the local infusion of AMPT completely depleted the newly-synthesized pools of DA in both rat types (Figure 5). Despite the fact that the newlysynthesized pools of DA were shown to be completely empty (Figure 5), the local infusion of PA led to an increase of accumbal DA levels in both AMPT-treated LR ( Figure 7A) and AMPTtreated HR (Figure 7B). In fact, the PA-induced DA release in these AMPT-treated rats did not differ from the PA-induced DA release in rats treated with AMPT-solvent (Figure 7). The finding that not only the small PA-induced DA increase in HR, but also the large PA-induced DA increase in LR is insensitive to AMPT (Figure 7) illustrates that the PA-mediated DA release is independent on AMPT-sensitive pools of DA even when the release of DA is high.

CONCLUSIONS
It has recently been shown that the increase of accumbal DA induced by ISO is due to the simultaneous stimulation of accumbal beta-1 and beta-2 receptors (Aono et al., 2013). The present finding that the ISO-induced increase of accumbal DA is sensitive to AMPT (see above) shows that the postsynaptic beta1/2-adrenoceptors of the nucleus accumbens mediate (i.e., stimulate) accumbal DA release from newly-synthesized pools of DA (Figure 8). In a previous study we have shown that the ISO-induced increase of accumbal DA release is not sensitive to reserpine (RES; Verheij and Cools, 2009b). This indicates that the AMPT-sensitive and newly-synthesized DA pools that are controlled by postsynaptic beta1/2-adrenoceptors are not vesicular (Figure 8). The PA-induced increase of accumbal DA is due to inhibition of alpha-1 but not alpha-2 receptors (Ihalainen et al., 2004;Saigusa et al., 2012). In contrast to the ISO-induced increase of accumbal DA, the PA-induced increase of accumbal DA was found to be sensitive to RES (Verheij and Cools, 2009b) demonstrating that the postsynaptic alpha-1-adrenoceptors of the nucleus accumbens mediate (i.e., inhibit) accumbal DA release from vesicular pools of DA (Figure 8). The present finding that the PA-induced increase of accumbal DA release is not sensitive to AMPT (see above) indicates that the RES-sensitive and vesicular DA pools controlled by alpha-1-adrenoceptors do not contain newly synthesized DA (Figure 8). The fact that the DA increase mediated by postsynaptic beta-1/2-adrenoceptors is sensitive to AMPT (present study), whereas the DA increase mediated by presynaptic alpha-1-adrenoceptors is not (Verheij and Cools, 2009a), indicates that the presynaptic alpha-1-adrenoceptors of the nucleus accumbens are not located on the noradrenergic nerve-terminals that impinge upon dopaminergic neurons expressing postsynaptic beta-1/2-receptors (Figure 8). The finding that both the DA increase mediated by presynaptic alpha-1-adrenoceptors (Verheij and Cools, 2009a) and the DA increase mediated by postsynaptic alpha-1-adrenoceptors (present study) are insensitive to AMPT confirms our hypothesis (see Section Introduction) that presynaptic alpha-1-adrenoceptors in the nucleus accumbens are located on those noradrenergic nerveterminals that impinge upon dopaminergic neurons equipped with post-synaptic alpha-1-receptors (Figure 8).

IMPACT
Changes in accumbal DA are generally accepted to play a key role in addiction (Koob, 1992;Everitt and Robbins, 2005), schizophrenia (Gray et al., 1991;Gray, 1995;Carlsson et al., 2001;Kuepper et al., 2012) and Parkinson's disease (Hornykiewicz, 1998;van Oosten et al., 2005). This explains why dopaminergic manipulations are widely used to treat these disorders. However, serious side effects of drugs that directly act at the dopaminergic receptors of the brain have been reported (Platt et al., 2002;Serretti et al., 2004;Ahlskog, 2011;Stowe et al., 2011;Meltzer, 2013). Partial DA agonists, however, are predicted to lead to fewer side effects because these drugs do not result in maximum stimulation of DA receptors (Jenner, 2002;Platt et al., 2002;Ohlsen and Pilowsky, 2005). Growing evidence indicates that moderate changes of endogenous DA levels in only a limited number of dopaminergic neurons, or at a limited number of dopaminergic receptors, may represent very powerful tools to treat dopaminergic diseases without producing severe side effects (Withers et al., 1995;Dixon et al., 1999;Müller, 2001;Mailman and Murthy, 2010;Diana, 2011). The fact that alpha-and beta-adrenoceptors control the release of DA from different types of dopaminergic pools (Figure 8, see also: Cools, 2008, 2009b), which are in turn believed to control their own class of dopaminergic receptors (for review: , indicates that by choosing the appropriate noradrenergic treatment it may very well be possible to change the release of endogenous DA only from those pools that are involved in the therapeutic effects of DA, but not from the pools that induce DA-related side effects. Whether an alpha-or beta-adrenergic agent has therapeutic potential depends on the disease to be treated. We have previously hypothesized that alpha-adrenoceptor-induced changes in vesicular DA release may be beneficial in the treatment of cocaine addiction and Parkinson's disease, whereas beta-adrenoceptorinduced changes in the release of newly-synthesized DA may be beneficial in the treatment of amphetamine addiction and the positive symptoms of schizophrenia (for references see: . Indeed, it has recently been found that alpha-adrenoceptor antagonists reduce both the behavioral and accumbal DA response to cocaine (Mitrano et al., 2012), presumably by acting on the pre-synaptic alpha-1-adreneceptors of the nucleus accumbens (Verheij et al., 2013). In the future, additional animal studies will be needed to further investigate the beneficial effects of noradrenergic drugs as treatment for DArelated disorders.