A Role for Accumbal Glycine Receptors in Modulation of Dopamine Release by the Glycine Transporter-1 Inhibitor Org25935

Accumbal glycine modulates basal and ethanol-induced dopamine levels in the nucleus accumbens (nAc) as well as voluntary ethanol consumption. Also, systemic administration of the glycine transporter-1 inhibitor Org25935 elevates dopamine levels in nAc, prevents a further ethanol-induced dopamine elevation and robustly and dose-dependently decreases ethanol consumption in rats. Here we investigated whether Org25935 applied locally in nAc modulates dopamine release, and whether accumbal glycine receptors or NMDA receptors are involved in this tentative effect. We also addressed whether Org25935 and ethanol applied locally in nAc interact with dopamine levels, as seen after systemic administration. We used in vivo microdialysis coupled to HPLC-ED in freely moving male Wistar rats to monitor dopamine output in nAc after local perfusion of Org25935 alone, with ethanol, or Org25935-perfusion after pre-treatment with the glycine receptor antagonist strychnine or the NMDA receptor glycine site antagonist L-701.324. Local Org25935 increased extracellular dopamine levels in a subpopulation of rats. Local strychnine, but not systemic L-701.324, antagonized the dopamine-activating effect of Org25935. Ethanol failed to induce a dopamine overflow in the subpopulation responding to Org25935 with a dopamine elevation. The study supports a role for accumbal glycine receptors rather than NMDA receptor signaling in the dopamine-activating effect of Org25935. The results further indicate that the previously reported systemic Org25935–ethanol interaction with regard to accumbal dopamine is localized to the nAc. This adds to the growing evidence for the glycine receptor as an important player in the dopamine reward circuitry and in ethanol's effects within this system.

The activity of the mesolimbic dopamine system is controlled by several neurotransmitter systems that involve primary targets of alcohol, e.g., the N-methyl-d-aspartate receptor (NMDAR) and the glycine receptor (GlyR); yet the molecular mechanisms by which ethanol activates the system is not clear. Recently studies have identified the GlyR in nAc as an access point for ethanol to the brain reward system (Molander and Söderpalm, 2005a). It was further shown that glycine perfusion in the nAc increases accumbal dopamine levels and reduces ethanol intake, whereas perfusion with the GlyR antagonist strychnine instead decreases dopamine levels and increases ethanol intake (Molander et al., 2005). It was hypothesized that accumbal GlyRs provide tonic regulation of dopaminergic output via a feed-back circuitry, where GlyRs control GABAergic medium spiny neurons projecting from nAc to the VTA. By activation, these GlyRs hyperpolarize the GABAergic neurons and reduce the inhibitory tone on ventral tegmental dopamine neurons and/or on incoming cholinergic neurons; thereby allowing the dopamine neurons to fire (Söderpalm et al., 2009).
Extracellular glycine levels are regulated by high affinity glycine transporters (GlyTs) and two distinct subtypes referred to as GlyT-1 and GlyT-2 have been characterized. GlyT-2 displays IntroductIon Alcohol addiction is a major public health problem with enormous socioeconomic and negative health consequences. Alcohol addiction is described as compulsive alcohol use with chronic relapses and with an underlying enduring pathology that may begin with a dysregulation of brain hedonic systems (Koob, 2009). The mesolimbic dopamine pathway is the central substrate for reward, reinforcement, and motivational learning (Koob, 1992;Wise, 1998;Gonzales et al., 2004) and consists of cell bodies in the ventral tegmental area (VTA) that primarily project to the nucleus accumbens (nAc). There is considerable evidence from animal studies and humans that acute alcohol administration activates this system (Di Chiara and Imperato, 1988;Boileau et al., 2003;Volkow et al., 2007). Moreover, chronic exposure to alcohol is known to result in allostatic and functional changes in the VTA-nAc pathway that is linked to development of addiction (Hyman et al., 2006;Koob and Volkow, 2010).
Abbreviations: GlyR, glycine receptor; GlyT-1, glycine transporter-1 protein; nAc, nucleus accumbens; NMDAR, N-methyl d-aspartate receptor; VTA, ventral tegmental area. was kindly provided by Organon Labs. Ltd. (now Merck Sharp & Dohme, Newhouse, Lanarkshire, UK). Strychnine, Org25935, and ethanol were dissolved in Ringer solution (140 mM NaCl, 1.2 mM CaCl 2 , 3.0 mM KCl, and 1.0 mM MgCl 2 ) for perfusion into nAc by reverse microdialysis. Due to insolubility in Ringer, L-701.324 was dissolved in Polysorbate 80, diluted in 0.9% NaCl to 10%, and administered i.p. in a volume of 2 ml/kg. With reverse microdialysis it is difficult to determine the exact concentration of the perfused drug in the extracellular space outside the probe. However in our laboratory the recovery of dopamine in vitro is ∼5-10% of the actual concentration outside the probe. In vivo, we estimate the recovery/excovery to be ∼1-5% depending on the molecular size and charge and the area of the active space of the probe. We therefore estimate that perfusion of 100 μM Org25935 equals a concentration of 1-5 μM in the extracellular space. The concentration of Org25935 used was selected based on data obtained from Organon Labs. Ltd. The concentrations of the GlyR antagonist strychnine (10 and 20 μM) were selected based on results obtained in previous studies (Molander and Söderpalm, 2005b;Ericson et al., 2006;Chau et al., 2010b). The dose of L-701.324, 5 mg/kg i.p., was chosen based on studies showing antagonistic effects on relapse-like drinking in rats (Vengeliene et al., 2005). Finally, to be able to compare the present results to those obtained after systemic administration of both Org25935 and ethanol we used an ethanol concentration (300 mM) producing a dopamine elevation (∼30%) comparable to that observed after 2.5 g/kg i.p. (Blomqvist et al., 1993;Ericson et al., 2003;Lof et al., 2007).

MIcrodIalysIs technIque
The microdialysis experiments were performed in awake and freely moving rats. Rats were anesthetized by isoflurane (Baxter, Sweden), mounted into a stereotaxic instrument (David Kopf Instruments, AgnTho's AB, Lidingö, Sweden) and put on a heating pad to prevent hypothermia during the surgery. Holes were drilled for placement of two anchoring screws and a dialysis probe custom made in the laboratory. The probe with 2 mm active space was lowered into the nAc monolaterally; A/P: +1.85, M/L: −1.4, relative to bregma, and D/V: −7.8 mm relative to dura (Paxinos and Watson, 2007), coordinates corresponding to a core-shell borderline region. A 2.5% ketoprofen gel was applied on the injured tissue to relieve inflammation and pain. The dialysis probe and screws were fixed to the scull with Harward cement (DAB Dental AB; Gothenburg, Sweden). Animals were given 2 ml saline i.p. post surgery to prevent dehydration and allowed to recover for 48 h prior to the experiment. On the experimental day, the sealed inlet and outlet of the probe were cut open and connected to a microperfusion pump (U-864 Syringe Pump, AgnTho's, Sweden) via a swivel allowing the animal to move freely. The probe was perfused with Ringer solution at a rate of 2 μl/min and dialysate (40 μl) was collected every 20 min. Before sampling begun, rats were perfused with Ringer solution for 1 h to obtain a balanced fluid exchange. Animals were killed immediately after the experiment and brains were removed. The positions of the probes were verified by visual examination of brain slices, cut by a vibroslicer (Campden Instruments Ltd., Leicester, UK) and visualized in Figure 1. A number of 20 rats were excluded from the study after probe verification, due to hemorrhage or incorrect probe placement. high neuronal expression in caudal CNS regions, whereas GlyT-1 is expressed throughout most regions of the CNS, mainly by glial cells (Zafra et al., 1995;Jursky and Nelson, 1996). GlyT-2 has a specialized role in reuptake of glycine into nerve terminals for vesicular storage, whereas GlyT-1 is the principal regulator of extracellular glycine levels (Gomeza et al., 2006). Within the above described feed-back mechanism, GlyT-1 inhibition may elevate extracellular glycine concentrations and possibly protect the GlyR from ethanol-induced activation. We have examined the selective GlyT-1 blocker cis-N-methyl-N-(6-methoxy-1-phenyl-1,2,3,4-tetrahydronaphtalen-2-ylmethyl)amino-methylcarboxylic acid hydrochloride (Org25935), which easily passes the blood-brain barrier and elevates extracellular glycine levels after systemic administration (Ge et al., 2001;Lidö et al., 2009). In a twobottle free choice model Org25935 produced a robust and dosedependent decrease in ethanol consumption in rats (Molander et al., 2007). It was further demonstrated that systemic Org25935 increases dopamine levels in nAc in a similar manner as local glycine and counteracts accumbal dopamine elevations provoked by systemic ethanol administration (Lidö et al., 2009). Yet the actual involvement of accumbal GlyRs in the effects of Org25395 on dopamine levels remains to be determined.
Apart from being a GlyR agonist, glycine serves as an obligatory co-agonist to glutamate at the NMDAR and along with other ligand-gated ion channels, also the NMDAR is a primary substrate for ethanol in the brain (Gass and Olive, 2008). The nAc is richly innervated by glutamatergic inputs and ethanol is reported to interact with NMDAR in this brain region (Nie et al., 1993;Moghaddam and Bolinao, 1994). In the progression from substance use to addiction, a shift from a dopamine-based to a more glutamate-based behavior is described (Ross and Peselow, 2009), and glutamate is implicated in addiction processes such as relapse (Vengeliene et al., 2005). It was therefore judged of interest to look into the mechanisms by which Org25935 modulates accumbal dopamine release. To this end, we investigated if Org25935 applied locally in nAc elevates accumbal dopamine levels, and, if so, whether the effect observed involves accumbal GlyRs or NMDAR signaling. The study also addressed whether Org25935 and ethanol applied locally in the nAc interact with dopamine in a similar manner as previously observed after systemic administration.

MaterIals and Methods anIMals
A total number of 120 male Wistar rats weighing 240-260 g were purchased from Beekay (Stockholm, Sweden). The rats were housed five per cage at constant room temperature (22°C) and humidity (65%) and kept under regular light-dark conditions (lights on at 7.00 AM and off at 7.00 PM) with free access to standard rat feed (Harlan Teklad Europe, UK) and tap water. The rats were allowed minimum 1 week of adaptation before the experiment was performed. The study was approved by the Ethics Committee for Animal Experiments in Gothenburg in Sweden (diary number 5/04).

drugs
Strychnine and L-701.324 were purchased from Sigma Aldrich (Stockholm, Sweden) and ethanol was purchased from Kamet AB (Haninge, Sweden). The selective GlyT-1 inhibitor Org25935 (PLSD) post hoc test was used in order to reduce risk of obtaining type 2 errors. Due to the biological variability in the ability of Org25935 to elicit an accumbal dopamine response, separate time courses for the responding, and non-responding subgroups were analyzed in Experiment 1 and Experiment 3 in addition to the group as a whole. The criteria for being a responder was set to >10% increase in dopamine output, while remaining animals were classified as non-responders.

experIMental procedure and neurocheMIcal assay
A high-pressure liquid chromatography system with electrochemical detection was used for the separation and detection of dopamine content of the dialysate samples, as previously described in Lidö et al. (2009). An external standard containing 3.25 fmol/μl of dopamine was used to identify the dopamine peak. For all experiments, microdialysis was performed during continuous monitoring of accumbal dopamine efflux. When three stable dopamine values were obtained (±10%), the mean of these was set to 100% for each animal, the time point was set to 0 min, and all percentages throughout the experiments were related to this. In Experiment 1, Org25935 (100 μM) was perfused into nAc from time point 40 min while Ringer perfusion was maintained throughout the experiment in the control group. Experiment 2 studied the influence of strychnine (10 and 20 μM) and L-701.324 (5 mg/kg i.p.) pre-treatment on the effect of Org25935. Therefore, perfusion of strychnine was initiated at time point 0 min as was injection of L-701.324, while perfusion of Org25935 started 40 min later, at time point 40 min. In Experiment 3 we examined the influence of Org25935 on ethanol-induced dopamine increase, and Org25935 (100 μM) was perfused from time point 0 min while ethanol (300 mM) was perfused alone or co-perfused with Org25935 from time point 100 min and throughout the experiment.
statIstIcs All values are expressed as means ± SEM and a probability value (p) less than 0.05 was considered statistically significant. Due to the interaction design where we studied effects of drugs/drug combinations over different time intervals, ANOVA with repeated measures was performed over relevant time periods for respective drug perfusions, and not over the time period as a whole. As the study was clearly hypothesis-driven and since Org25935-and ethanol-evoked dopamine responses could be anticipated due to previous studies, the more liberal Fisher's protected least significant difference and 3A-C. Each bar represents the mean ± SEM of the % change in dopamine output during 40 min following the respective drug perfusions. Org25935 pre-treated group (paired t-test 100-120 min, p = 0.792). Conversely, dopamine output in this group decreased 40 min after start of ethanol perfusion (paired t-test 100-140 min, p = 0.037). Figure 4B displays the dopamine output after Org25935 and/or ethanol perfusion, as described above, in the Org25935 responding (n = 8) and non-responding subgroups (n = 9). Prior to ethanol perfusion, a repeated measures ANOVA (time period 0-100 min) revealed a significant group effect [F(2,21) = 19.624, p = 0 < 0.001], time effect [F(4,84) = 5.144, p = 0.001], and group by time interaction [F(8,84) = 5.393, p = 0 < 0.001]. The Fisher's PLSD test revealed a significant difference between responders and non-responders (p < 0.001) and responders and Ringerperfused animals (p < 0.001) but not between non-responders and Ringer-perfused animals (p = 0.879). As previously shown in Figure   animals, Org25935 mitigated ethanol's effect among the nonresponding animals, as also seen after systemic administration. Thus the interaction with ethanol may take place regardless of the variability in the strength of the Org25935-induced dopamine response as such. Since ethanol allosterically modulates the GlyR (Mihic et al., 1997;Perkins et al., 2008), excessive glycine availability provoking a conformational change in the GlyR may in turn obstruct the binding of ethanol and the following rise in dopamine levels.
Due to an inconsistent dopamine response provoked by Org25935, rats were divided into responders and non-responders. Org25935 increased accumbal dopamine output by 26% in all rats, consisting of seven responders, displaying a 36% increase, and six non-responders. This is comparable to systemic Org25935, which increased dopamine levels by 13% in all rats, consisting of 11 responders, displaying a 29% increase, and 10 non-responders (Lidö et al., 2009). A similar biological variability in dopamine response was observed after accumbal glycine perfusion (Molander and Söderpalm, 2005a,b). As Org25935 produced a consistent increase in glycine levels, equal among dopamine responders, and non-responders (Lidö et al., 2009), we proposed that the variability lies in the response to glycine as such, and not in glycine removal capacity. This may be linked to a recent study reporting no association of alcohol dependence with GlyT-1gene polymorphism (Koller et al., 2009). Instead, differences in, e.g., GlyR desensitization or internalization phenomena could account for the variance in dopamine response. However a GlyR expression study using genetically high versus low alcohol-preferring rats reported very few differences in mRNA expression despite of the distinct drinking behavior, and few differences after 1 month of ethanol exposure (Jonsson et al., 2009). Also, daily Org25935 injections for 19 days produced no changes in GlyR subunits mRNA nor in GlyT-1 mRNA in several brain regions including the nAc (unpublished results). Thus the GlyR does not seem to be prone to rapid compensatory alterations. 23% compared to the time point before the ethanol challenge (paired t-test 100-120 min, p = 0.038). There was a trend for an ethanol-induced dopamine elevation in the Org25935 nonresponding group (paired t-test 100-120 min, p = 0.057) whereas ethanol did not elevate dopamine levels in responders (paired t-test 100-120 min, p = 0.374). Instead, dopamine output in responders significantly decreased 40 min after ethanol perfusion (paired t-test 100-140 min, p = 0.024).

dIscussIon
This study shows that local administration of the GlyT-1 blocker Org25935 in nAc increases accumbal dopamine output in the rat. The effect was reversed by the GlyR antagonist strychnine but not by antagonism of the glycine site on the NMDAR, suggesting that the GlyT-1 blocker modulates the accumbal dopamine tone via accumbal GlyRs rather than via NMDAR signaling. This is in line with previous findings, where accumbal glycine perfusion elevated dopamine levels by ∼50% (Molander and Söderpalm, 2005b), and where taurine and β-alanine, also endogenous GlyR agonists, increased dopamine levels in the same manner (Ericson et al., 2006. The dopamine responses provoked by glycine, taurine, and β-alanine were reversed by strychnine, supporting the GlyR involvement also in the dopamine response to Org25935. Next, the study reports that Org25935 prevents ethanol from increasing accumbal dopamine levels, thus underlining the interaction with ethanol seen after systemic administration (Lidö et al., 2009). That the drugs were perfused locally in nAc and yet produced results almost identical to those observed after systemic administration, suggests that the Org25935-ethanol interaction takes place in the nAc. This corroborates previous results pointing to the GlyR in nAc as not only a modulator of basal dopamine release in nAc, but as an access point for ethanol to the mesolimbic dopamine system (Molander and Söderpalm, 2005a;Molander et al., 2005). In addition to a robust prevention of ethanol-induced dopamine output in Org25935-responding FiGure 4 | (A) Accumbal dopamine levels as measured by in vivo microdialysis after perfusion of 300 mM ethanol (n = 7), Ringer (n = 8), and after Org25935 perfusion followed by ethanol co-perfusion (n = 17), shown as means ± SEM. Org25935 perfusion started at time point 0 min and ethanol perfusion started at time point 100 min, as indicated by arrows. Basal dopamine concentrations (mean ± SEM) for ethanol, ethanol-Org25935, and Org25935 were 2.91 ± 0.64, 2.14 ± 0.24, and 2.66 ± 0.66 nM, respectively. For statistics: See Section "Results, Experiment 3. " (B) Accumbal dopamine levels as measured by in vivo microdialysis after perfusion of 300 mM ethanol (n = 7) and after Org25935 perfusion followed by ethanol co-perfusion in the responding (n = 8) and non-responding (n = 9) subgroup, shown as means ± SEM. Org25935 perfusion started at time point 0 min and ethanol perfusion started at time point 100 min, as indicated by arrows. Basal dopamine concentrations (mean ± SEM) for ethanol, drug responders and non-responders were 2.91 ± 0.64, 2.13 ± 1.4, and 2.14 ± 0.22 nM, respectively. For statistics: See Section "Results, Experiment 3" .
internalization (Nong et al., 2003). In contrast, different GlyR agonists consistently elevate dopamine whereas the GlyR antagonist strychnine concentration-dependently reduces accumbal dopamine levels (Ericson et al., 2006. As a parallel also acamprosate, a drug structurally related to taurine and in clinical use for treating alcohol dependence, also promotes dopamine release in nAc via the GlyR rather than the NMDAR (Chau et al., 2010a,b). In summary, most evidence support a GlyR mechanism rather than an NMDAR mechanism being involved in alcohol's interference with accumbal dopamine levels. We therefore suggest that the Org25935-induced dopamine response here observed primarily involves accumbal GlyR. As also other neurotransmitters and receptors are implied in alcohol intake behavior, activation of both NMDARs and GlyRs may be important for the alcohol intake-reducing effect of the GlyT-1 blocker. This view is reinforced by recent findings that Org25935 treatment persistently reduces relapse-like drinking behavior, and that this durable effect may result from restoration of both glycinergic and glutamatergic signaling (Vengeliene et al., 2010).
Substances that stimulate dopamine release in nAc may be associated with risk of abuse. Org25935 elevates dopamine levels in nAc after systemic administration but several points argue against an addictive potential. Org25935 has a mild and indirect mode of increasing dopamine levels that probably rather affects tonic dopamine release. This contrasts to the rapid increase in dopamine levels that is associated with a reported "high" in humans, and which in turn is considered an important feature of a potentially addictive drug working via the dopamine system (Volkow et al., 1999). Org25936 may have ethanol-like effects on dopamine, but the brain perceives ethanol through several known primary targets and not only through the GlyR, and the interoceptive cues of Org25935 and ethanol in the brain are most likely not the same. Preclinical information obtained from Organon Labs. Ltd. (now MSD) reported no hedonic profile of Org25935 on intra-cranial self administration and no effect on locomotion. Org25935 has previously been tested in clinical trials with no reportage of a positive hedonic profile in man. In reports studying the antipsychotic effects of glycine, glycine agonists, and GlyT-1 inhibitors in man, the question of tentative addictive properties of GlyT-1 blockers was never raised (Shim et al., 2008). Chronic elevation of extracellular glycine in humans appears safe in the clinic (Patel et al., 1990;Shoham et al., 2001;Yang and Svensson, 2008) and overall, the abuse liability risk associated with Org25935 is low. This is the first study to demonstrate a role for accumbal GlyRs in the dopamine-modulating effects of a GlyT-1 blocker, and it appears that the GlyR rather than NMDAR signaling is involved in the Org25935-evoked dopamine elevation. The study suggests that the Org25935-ethanol interaction takes place in nAc, since Org25935 locally in nAc increases basal dopamine levels and counteracts ethanol-induced activation of dopamine. We propose a partial substitution and antagonism mechanism of Org25935. The GlyT-1 blocker elevates and stabilizes extracellular glycine levels, which via the GlyR elevate dopamine levels (in responders) or preserve stable dopamine levels (in non-responders), which in turn prevents additional ethanol-mediated GlyR activation and dopamine elevation. This partial agonistic mechanism may offer Contrary to the inconsistent dopamine response, Org25935 produced a consistent anti-alcohol intake effect in ethanol highand medium-preferring rats (Molander et al., 2007). The consistent decrease in voluntary ethanol consumption in all rats contrasts to the effect of glycine perfusion, which decreased voluntary ethanol consumption in animals responding to glycine with a dopamine elevation, but not in the non-responders (Molander et al., 2005). Moreover, an overrepresentation of glycine-dopamine responders in ethanol high-preferring rats compared to naïve rats was reported, suggesting that a powerful glycine-dopamine interaction may be linked to high ethanol preference. Similarly, the selected rats with a high ethanol liking used in the alcohol consumption study with Org25935 may constitute a group of mostly dopamine responsive subjects. Alternatively, as sarcosine-based GlyT-1 inhibitors like Org25935 exhibit a non-competitive mode of action, i.e., independent of glycine concentration (Mezler et al., 2008), this independency of variations in endogenous glycine levels may underlie the consistent anti-alcohol intake effect of Org25935. Nevertheless, both the behavioral and neurochemical results suggest that Org25935 employs a combined substitution and antagonism mechanism, where GlyT-1 inhibition elevates and/or stabilizes extracellular glycine levels, which in turn elevate dopamine in Org25935 responders or preserve/stabilize dopamine levels in non-responding animals. This results in a blocking/ dampening mechanism preventing additional ethanol-induced GlyR activation and dopamine elevation. These events are possibly mediated via the proposed GABAergic feed-back circuitry (reviewed in detail in Söderpalm et al., 2009). Given that extracellular glycine modulates ethanol consumption and that Org25935 displays excellent anti-alcohol drinking effects in experimental animals (Molander et al., 2007;Vengeliene et al., 2010), the present results further supports GlyT-1 inhibition as a new concept for pharmacotherapy of alcohol addiction.
Besides serving as a GlyR agonist glycine is also an essential co-agonist to glutamate at the NMDAR, and glycine uptake inhibitors display NMDAR potentiating effects (Shim et al., 2008). Glutamate signaling is implicated in alcohol addiction and withdrawal from chronic alcohol is characterized by elevated glutamate levels in nAc (Dahchour and De Witte, 1999). A negative NMDAR modulation is proposed to be useful for alcohol addiction pharmacotherapy (Nagy, 2004) and various NMDAR antagonists decreases ethanol intake (Vengeliene et al., 2005;Gass and Olive, 2008). This view is not in concordance with a tentative NMDAR mechanism by Org25935, which potentiates instead of antagonizes the NMDAR. The present result with L-701.324 suggests that NMDAR signaling is not involved in the dopamine response to Org25935. Unfortunately, due to solubility problems L-701.324 could not be perfused via the dialysis probe, which weakens the comparison with strychnine in this respect. L-701.324 in the present dose prevents relapse-like drinking behavior (Vengeliene et al., 2005) and ethanol-induced conditioned place preference (Biala and Kotlinska, 1999). However, NMDAR effects on mesolimbic dopamine are not clear-cut and the precise NMDAR contribution to dopamine release in nAc is not established (Spanagel, 2009). Moreover, low doses of glycine (5-20 μM) potentiates NMDAR currents, whereas glycine in the 100 μM range or higher have been found to prime NMDARs for great potential for GlyT-1 inhibition as a new concept for decreasing alcohol intake, and GlyT-1 inhibition as a tentative "anti-alcohol drinking therapy" is emphasized by Org25935's excellent ability to decrease ethanol consumption in experimental animals. The study adds to the growing evidence for the GlyR as an important player in the dopamine reward circuitry and in ethanol's effects within this system.