mGlu5, Dopamine D2 and Adenosine A2A Receptors in L-DOPA-induced Dyskinesias.

Patients with Parkinson's disease (PD) receiving L-3,4-dihydroxyphenylalanine (L-DOPA, the gold-standard treatment for this disease) frequently develop abnormal involuntary movements, termed L-DOPA-induced dyskinesias (LID). Glutamate overactivity is well documented in PD and LID. An approach to manage LID is to add to L-DOPA specific agents to reduce dyskinesias such as metabotropic glutamate receptor (mGlu receptor) drugs. This article reviews the contribution of mGlu type 5 (mGlu5) receptors in animal models of PD. Several mGlu5 negative allosteric modulators acutely attenuate LID in 1-methyl-4-phenyl-1,2,3,6- tetrahydropyridine (MPTP) monkeys and 6-hydroxydopamine(6-OHDA)-lesioned rats. Chronic administration of mGlu5 negative allosteric modulators to MPTP monkeys and 6-OHDA rats also attenuates LID while maintaining the antiparkinsonian effect of L-DOPA. Radioligand autoradiography shows an elevation of striatal mGlu5 receptors of dyskinetic L-DOPA-treated MPTP monkeys but not in those without LID. The brain molecular correlates of the long-term effect of mGlu5 negative allosteric modulators treatments with L-DOPA attenuating development of LID was shown to extend beyond mGlu5 receptors with normalization of glutamate activity in the basal ganglia of L-DOPA-induced changes of NMDA, AMPA, mGlu2/3 receptors and VGlut2 transporter. In the basal ganglia, mGlu5 receptor negative allosteric modulators also normalize the L-DOPA-induced changes of dopamine D2receptors, their associated signaling proteins (ERK1/2 and Akt/GSK3β) and neuropeptides (preproenkephalin, preprodynorphin) as well as the adenosine A2A receptors expression. These results show in animal models of PD reduction of LID with mGlu5 negative allosteric modulation associated with normalization of glutamate, dopamine and adenosine receptors suggesting a functional link of these receptors in chronic treatment with L-DOPA.

Introduction: Parkinson’s disease and 
L-DOPA-induced dyskinesias

Parkinson’s disease (PD) is the most common neurodegenerative movement disorder characterized by tremor, rigidity, bradykinesia and is likely to increase due to the aging populations [1]. PD involves principally the death of dopamine (DA) neurons in the substantia nigra pars compacta (SNc) but other neurotransmitters and neuro-
modulators are also affected [2]. Treatment of motor symptoms of PD with the DA precursor, L-3,4-dihydroxyphenylalanine (L-DOPA), introduced 50 years ago still remains the gold standard for PD treatment [3]. However, various complications including motor fluctuations and abnormal involuntary movements, such as L-DOPA-induced dyskinesias (LID), limit the quality of life in PD patients and can be very difficult to manage [4]. LID increases with 
the duration of L-DOPA treatment and up to 95% of PD patients become afflicted after 15 years [5, 6]. Although the mechanisms of these involuntary movements are not well understood, they are paralleled by changes in various neurotransmitter systems and intracellular signaling pathways [7].

Glutamate receptors in Parkinson’s disease and L-DOPA-induced dyskinesias

Glutamate neurotransmission is reported to be increased in the basal ganglia in PD [7] and LID [8, 9]. Amantadine, a noncompetitive antagonist at N-methyl-D-aspartate (NMDA) ionotropic glutamate receptors and, to a lesser extent clozapine, are presently the only drug used in the clinic to reduce LID in some PD patients without worsening parkinsonian symptoms [10-14]. However, the antidyskinetic effect of amantadine may be transient [15] and high doses may not be tolerated in some PD patients because of cognitive impairment, thus limiting its use [15].

On the basis of these considerations, combined with the rich distribution and diverse physiological roles of metabotropic glutamate (mGlu) receptors within the basal ganglia, recent attention has been placed on these receptors as alternative targets to modulate glutamate hyperactivity in PD and LID [16]. Studies in animal models and PD patients indicate that antagonists of group I mGlu receptor, especially mGlu5 receptor, could be considered as a suitable therapeutic approach in PD and LID.

mGlu5 receptor specific binding was reported to be increased in the basal ganglia of parkinsonian monkeys with LID and in parkinsonian patients with motor complications [17-21].

In the 6-hydroxydopamine (6-OHDA)-lesioned rat model, mGlu5 receptor negative allosteric modulators 2-methyl-6-(phenylethynyl)pyridine (MPEP), 3-[(2-methyl-1,3-thiazol-4-yl)ethynyl]pyridine (MTEP) and MRZ-8676 administered acutely inhibit LID [22-26]. The mGlu5 receptor negative allosteric modulators MPEP, MTEP, fenobam and AFQ056 (mavoglurant) were found to acutely reduce LID in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-lesioned monkeys [27-29]. In the light of these acute studies, it becomes important to test the long-term effects of mGlu5 receptor negative allosteric modulators behaviourally and their associated brain molecular changes.

Sub-chronic and chronic administrations of mGlu5 receptor negative allosteric modulators in MPTP monkeys and 6-OHDA rats reduce the development of L-DOPA induced motor complications (Table 1). In previous studies, we reported that development of LID over a month of treatment were lower by overall ∼70% with addition of MPEP to the L-DOPA treatment in de novo MPTP monkeys [17] and this was associated with a normalization of glutamate receptors [18] as shown for mGlu5, NMDA containing NR2B subunit, AMPA and mGlu2/3 receptors (Table 2). In 6-OHDA rats the vesicular glutamate transporter 2 (VGlut2) was also normalized by the addition of MPEP to the chronic treatment with L-DOPA [30].

Table 1

Behavioral effects of repeated mGlu5 negative allosteric modulator treatments in 6-OHDA-lesioned rats and MPTP monkeys.

Animal Model and Treatment	mGlu5 Receptor Negative Allosteric Modulators(s)	Regimen	Dose Tested	Main Behavioral Effect	Refs.
6-OHDA-lesioned rat, Sub-chronic	MTEP	L-DOPA and MTEP administered simultaneous once daily for 7 days	5 mg/kg	•Reduced already established abnormal involuntary movement scale (AIMS) up to the 7th day of co-treatment with L-DOPA, comparable to single acute administration of MTEP	[24]
	MPEP	MPEP administered 5 min before L-DOPA once daily for 14 days	1 mg/kg	•Reduced severity of limb dyskinesia and axial dystonia over the whole duration of L-DOPA treatment	[127]
	MRZ-8676	L-DOPA and MRZ-8676 administered simultaneous once daily for 6 days	75 mg/kg	•Reduced already established AIMS •No tolerance of antidyskinetic effect following repetitive treatment	[22]
	MPEP	MPEP administered 10 min before L-DOPA twice daily for 10 days	1.5 mg/kg	•Reduced LID	[30]
6-OHDA-lesioned rat, Chronic	MTEP	L-DOPA and MTEP administered simultaneous once daily for 21 days	5 mg/kg	•Reduced AIMS compared to L-DOPA alone in drug naïve animals• Reduced already established AIMS	[26]
	MPEP	30 min before L-DOPA challenge once daily during 21 days	1.5 mg/kg	•MPEP virtually abolished AIMS	[25]
	MTEP	L-DOPA and MTEP administered simultaneous once daily for 21 days	5 mg/kg	•Attenuated dyskinesias without adverse motor effects	[128]
MPTP-lesioned monkey, Chronic	MPEP	de novo, MPEP administered 15 min prior to L-DOPA for one month	10 mg/kg	•Decreased dyskinesias and maintained anti-parkinsonian effect•No decrease of duration of the L-DOPA antiparkinsonian effect	[18]
	Fenobam	L-DOPA and Fenobam administered simultaneous for 17 days	10 mg/kg	•Attenuated development of peak-dose dyskinesias•No effect on antiparkinsonian activity of L-DOPA	[132]

Table 2

Biochemical effects of mGlu5 negative allosteric modulator treatments on glutamate neurotransmission in animal models of Parkinson’s disease.

Animal Model	mGlu5 Receptor Negative Allosteric Modulator(s)	Dose(s) Tested and Treatment Regimen	Main Biochemical Effects	Refs.
MPTP-lesioned monkey	MPEP	10 mg/kg, once daily for one month (de novo), 15 min prior to L-DOPA	•Prevented the increase by L-DOPA of striatal mGlu5 receptors density measured with [3H]ABP688 specific binding •Maintained at control levels striatal binding affinity of [3H]ABP688 to mGlu5 receptors •Maintained at control levels striatal mGlu5 receptor mRNA levels •Prevented the increase by L-DOPA of [3H]Ro-25-6981 specific binding to NMDA receptors containing NR1/NR2B subunits in the caudate nucleus and putamen•Prevented the increase by L-DOPA of [3H]Ro-48-8587 specific binding to AMPA receptors in the caudate nucleus and putamen•Prevented the decrease by L-DOPA of [3H]LY341495 specific binding to mGlu2/3 receptor in the caudate nucleus and putamen	[18]
6-OHDA-lesioned rat	MPEP	MPEP (1.5 mg/kg) administered 10 min before L-DOPA twice daily for 10 days	•MPEP with L-DOPA decreased the levels of VGlut2 in the striatum ipsilateral to the lesion	[30]

mGlu5 receptor negative allosteric modulators, mavoglurant and ADX-48621 (dipraglurant), were shown to reduce LID in parkinsonian patients and were well tolerated without worsening motor symptoms [31-33]. However, recent studies with mavoglurant did not show reduced LID in PD patients [34, 35].

Dopamine receptors in Parkinson’s disease and L-DOPA-induced dyskinesias

The mechanisms underlying the development of LID remain unknown, but evidence suggests that LID is the result of maladaptive plasticity at striatal synapses [7, 36] and in an altered activity of dopaminergic neurotransmission in the basal ganglia [37]. DA binds to five different subtypes of G protein-coupled DA receptors divided in two classes, the D1 (D1 and D5) and the D2 class (D2, D3 and D4) of receptors [38]. D1 receptors were shown to be expressed in neurons containing substance P and dynorphin, projecting to the substantia nigra pars reticulata (SNr) and to the internal globus pallidus (GPi), which constitute the direct striatal output pathway [39]. D2 receptors are predominantly localized in neurons expressing enkephalin, projecting to the external globus pallidus (GPe), constituting the indirect pathway [39]. D2 receptors are found on postsynaptic and pre-synaptic nigrostriatal dopaminergic terminals, of the substantia nigra neurons and of presynaptic corticostriatal terminals where they can inhibit striatal glutamate release [40].

Denervation-induced supersensitivity of D1 and D2 receptors was initially recognized as a plausible mechanism of LID [41, 42]. Numerous studies measured the density of D1 and D2 receptors in the brain of human and animal models, but no general consensus emerged; the wide methodological discrepancies, time to sacrifice, post-mortem delays, etc, may account for this lack of consensus. Post-mortem studies have shown that striatal DA receptors particularly the D2 subtype were increased in PD patients [42, 43] or unchanged [44, 45], while both D1 and D2 receptor subtypes were increased in MPTP monkeys [46-48]. Administration of L-DOPA was shown to reverse these increases in PD patients [42, 43] and primates in many studies [47-49]. No general consensus also emerged for DA receptor mRNA. Hence, D1 mRNA levels are reported to remain unchanged [50-52] or to be reduced after MPTP lesion in monkeys [51-53]. This decrease was corrected with L-DOPA [52, 53]. D2 receptors mRNA levels were reported to be increased concomitantly with its corresponding protein levels in striatum with MPTP [51-54] and returned to control levels when L-DOPA was administered [52]. These reports support that LID are more complex than hypersensitivity due to a simple increase in the density of striatal DA receptors and its mRNA.

Numerous interactions between mGlu5 receptor and 
D1, D2, NMDA, A2A adenosine receptors suggest that these receptors may function together as closely associated signaling partners in the development of LID [55, 56].

A chronic treatment with MPEP in MPTP monkeys treated with L-DOPA that prevented the development of LID was shown to normalize changes produced by L-DOPA on D2 receptor and its mRNA, preproenkephalin (PPE) mRNA, preprodynorphin (PPD) mRNA, phosphorylated extracellular signal-regulated kinase 1 and 2 (ERK1/2) and phosphorylated Akt/GSK3β signaling proteins, but not D1 and its mRNA (Table 3). Similar findings for PPE and PPD mRNA levels in 6-OHDA rats were observed with chronic MPEP or MTEP treatment with L-DOPA associated with the prevention of development of abnormal involuntary movement scale (AIMS) (Table 3). MTEP was also shown to normalize phosphorylated ERK1/2 and phospho-mitogen- and stress-activated protein kinase-1 (MSK-1) in L-DOPA treated 6-OHDA rats (Table 3). While both MPEP and MTEP have off-target activities these are at higher doses [57-61] and at the doses shown in Table 3 the results reported are more likely due to mGlu5 receptor negative allosteric modulator activity.

Table 3

Biochemical effects of mGlu5 negative allosteric modulator treatments on dopamine receptors, their associated signaling proteins and neuropeptides in animal models of Parkinson’s disease.

Animal Model	mGlu5 Receptor Negative Allosteric Modulator(s)	Dose(s) Tested and Treatment Regimen	Main Biochemical Effects	Refs.
MPTP-lesioned monkey	MPEP	10 mg/kg, once daily for one month (de novo), 15 min prior to L-DOPA	•[3H]SCH23390 specific binding to D1 DA receptors remained low in the striatum and globus pallidus of MPTP, MPTP+L-DOPA, and L-DOPA+MPEP-treated monkeys while no changes were observed compared to control in the levels of D1 receptor mRNA.•[3H]raclopride specific binding to D2 DA receptor and its corresponding mRNA levels in the striatum were elevated in L-DOPA+MPEP-treated monkeys compared to L-DOPA alone and was comparable to MPTP-treated monkeys•Prevented the increase by L-DOPA of striatal preproenkephalin and preprodynorphin mRNA levels •Prevented the increase by L-DOPA of striatal phosphorylated ERK1 and ERK2 •Prevented the increase by L-DOPA of striatal phosphorylated forms of Akt (Ser473) and GSK3b (Ser9)	[109]
6-OHDA-lesioned rat	MTEP	L-DOPA and MTEP (0.25, 1.25 and 6.25 mg/kg) administered simultaneous (acute)	•Prevented the increase by L-DOPA of striatal prodynorphin mRNA with MTEP 1.25 and 6.25 mg/kg•MTEP combined with L-DOPA did not modify up-regulation of striatal preproenkephalin mRNA-induced by DA denervation•MTEP alone did not modify striatal prodynorphin and preproenkephalin mRNA expression compared to vehicle-treated 6-OHDA rats.	[26]
6-OHDA-lesioned rat	MTEP	L-DOPA and MTEP (5 mg/kg) administered simultaneous once daily for 21 days	•Up-regulation of striatal prodynorphin on the lesion side of L-DOPA+MTEP-treated rats less pronounced that with L-DOPA alone treatment •MTEP alone reversed the lesion-induced down-regulation of striatal prodynorphin•MTEP with L-DOPA partially blocked the additional up-regulation of striatal preproenkephalin induced by L-DOPA alone•MTEP alone did not reverse the lesion-induced up-regulation of striatal preproenkephalin	[26]
6-OHDA-lesioned rat	MPEP	MPEP (1 mg/kg) administered once daily 5 min before L-DOPA for 14 days	•MPEP reduced the increases in preprodynorphin mRNA levels in striatonigral neurons ipsilateral to the lesion	[127]
6-OHDA-lesioned rat	MPEP	MPEP (1.5 mg/kg) 30 min before L-DOPA challenge once daily for 21 days	•MPEP combined with L-DOPA reduced dramatically the increased striatal expression of FosB/Delta FosB induced by L-DOPA	[25]
6-OHDA-lesioned rat	MTEP	L-DOPA and MTEP (1.25 and 6. 25 mg/kg, acute; 5 mg/kg once daily for 21 days) administered simultaneous	•Acute administration of MTEP with L-DOPA reduced striatal L-DOPA-induced phospho-ERK1/2 and phospho-MSK-1 expression at the two doses tested•Chronic administration of MTEP with L-DOPA blocked the up-regulation of striatal prodynorphin mRNA induced by L-DOPA	[128]

Adenosine receptors in Parkinson’s disease and L-DOPA-induced dyskinesias

Adenosine, a purinergic messenger, plays a crucial role in many physiological processes and is released by many cells including neurons and glia [62, 63]. In the basal ganglia, adenosine interacts closely with DA and is involved in the function of striatal GABAergic striatopallidal neurons projecting from the caudate nucleus and the putamen, mainly to the GPe [64-69]. The implication of adenosine to regulate the excessive glutamate neurotransmission observed in PD and LID is also demonstrated [37, 70, 71]. Hence, adenosine has received increasing attention because of its interaction with DA and glutamate receptors; this could have major implications for the development of new pharmacological targets for the treatment of PD and LID.

Adenosine binds to four classes of specific G-protein-coupled receptor subtypes named A1, A2A, A2B and A3 [72]. Adenosine receptor subtypes A1 and A2A are mainly localized in the basal ganglia, more precisely in the striatum, whereas A2B and A3 receptors are widely distributed in the brain [73, 74].

In the basal ganglia, A1 receptors are mainly present in the striatonigral GABAergic and corticostriatal glutamatergic neurons [75]. The stimulation of A1 receptors generally leads to inhibition of neurotransmitter release, such as the inhibition of glutamate release at corticostriatal terminals in the striatum [62, 76].

The A2A receptor is co-expressed at postsynaptic sites of the same medium spiny neurons as those bearing DA D2 receptors [77, 78] and projecting to the GPe of the indirect pathway. The restricted and specific distribution of A2A receptors to the indirect pathway of the basal ganglia provide specificity that could lead to a reduced incidence of adverse effects. Presynaptic and glial A2A receptors are also localized in the caudate nucleus and the putamen [73]. A2A receptors are found on presynaptic glutamatergic corticostriatal terminals and these receptors can modulate positively the glutamatergic cortical input by stimulating glutamate release [73].

Adenosine A2A antagonists can reduce the excessive striatopallidal and subthalamic neuronal activity and bring a new target in PD therapy [37, 65, 67, 79]. Several behavioural analyses have shown the potential efficacy of adenosine receptor modulators in the treatment of PD and LID, such as A2A receptor antagonists [79-83]. The A2A receptor antagonists Preladenant and KW-6002 (istradefylline), are reported 
to extend the duration of the antiparkinsonian effect of 
L-DOPA and enhance the parkinsonian activity of low doses of DA agonists; this is without worsening dyskinesias showing the DOPA-sparing activity of A2A receptor antagonists [81, 84-87]; reviewed in: [88-91]. Moreover a recent paper showed that the A2A receptor antagonist istradefylline did not increase LID of MPTP-treated marmosets in a chronic treatment and at days 21 and 28 it slightly but significantly reduced dyskinesias [92].

Presynaptically, A2A receptors can colocalize with A1 receptors in corticostriatal afferents, where they act together to modulate and regulate glutamate release [93]. Experiments using isolated striatal nerve terminal preparations have shown that most of the striatal glutamatergic terminals contain both A1 and A2A receptors [93]. In response to variation in adenosine concentrations, functional studies in striatal glutamatergic terminals have shown that the A1–A2A adenosine receptor heteromer provides a “switch mechanism” that can produce opposite effects on glutamate release [93-95].

Functional interactions between DA and adenosine receptors are supported by the anatomic localization of these receptors in striatal projection neurons [67]. A1 receptors are mainly co-expressed with D1 DA receptors on striatal neurons that project to the GPi and the substantia nigra [56]. D1 and adenosine A1 receptors are known to form functionally interacting complexes, the heteromer D1-A1, in cortical neurons and basal ganglia [96, 97].

Using co-immunoprecipitation and colocalization, as well as bioluminescence resonance energy transfer (BRET) and fluorescence resonance energy transfer (FRET) techniques, D2 and adenosine A2A receptors were shown to form functional homo- and hetero-oligomers [96, 98, 99]. Moreover, in cell lines cotransfected with plasmids containing D2 and A2A receptors and in primary culture of striatal neurons, long-term administration of A2A or D2 agonists induces an internalization and desensitization of the D2-A2A complex [98, 99], whereas D2 antagonists trigger an increase in D2 and A2A immunoreactivity [100]. Also, in SH-SY5Y neuroblastoma cells cotransfected with the D2 receptor, A2A antagonists enhance striatal D2 receptor signaling and block increased A2A receptor signaling [101]. Interestingly, a recent report in rats and monkeys has shown that D2-A2A-cannabinoid (CB)1 receptor heteromers are present in the striatum of intact and hemiparkinsonian 6-OHDA-lesioned rat as well as in intact and MPTP parkinsonian monkeys [102, 103] and that this heteromer expression is altered in lesioned striatum. Also, acute or chronic L-DOPA treatment induced disruption of D2-A2A-CB1 heteromers [102, 103]. These authors suggested that LID could be induced, at least in part, by an imbalance in the indirect pathway due to absence of cross-talk between D2-A2A-CB1 receptor heteromers functional components [102, 103].

In addition to adenosine A2A and DA D2 receptors, mGlu5 receptors are also known to interact and colocalize postsynaptically in the striatopallidal GABAergic efferent neurons [16, 71, 104]. Experiments using optical sectioning techniques found that A2A and mGlu5 receptors are also co-localized in rat striatal cultures [105]. In addition, double labelling electron microscopy colocalization experiments in the putamen of monkey have shown a substantial degree of A2A and mGlu5 receptor colocalization mainly in postsynaptic elements [104]. In fact, 60-70% of A2A receptors immuno-
reactive dendrites or spines in the monkey putamen co-express mGlu5 receptors [104]. Accordingly, A2A receptors activation can increase the phosphorylation of DARPP-32 at Thr-34 via an ERK pathway and the induction of c-fos expression is also increased in striatopallidal neurons when A2A and mGlu5 receptors are co-activated [71, 106]. An important increase in adenosine release was observed when glutamatergic neurotransmission becomes overactive [71, 107]. Functional interaction of A2A and mGlu5 receptors have also been demonstrated at the behavioural level, a synergistic effect on the induction of motor activity being observed following acute combined treatment with a low doses of mGlu5 negative allosteric modulator, MPEP, and A2A antagonist KW-6002 [108]. Hence, the mGlu5 receptor-mediated effect in the striatum was abolished by the blockade of A2A receptors [70]. This colocalization provides a structural framework for the existence of multiple functional interactions of A2A, D2 and mGlu5 receptors [16]. A better understanding of the functional interactions that exist between these receptor heteromers in normal and lesioned striatum as well as following acute or chronic L-DOPA treatment foreshadow a potential therapy targeting mGlu5-D2-A2A receptor heteromers in PD patients.

As presented in the previous section of this review, in de novo MPTP monkeys, development of LID was shown to be lower with addition of MPEP to the L-DOPA treatment [17] and to be associated with a normalization of glutamate [18] and DA receptors [18, 109]. This antidyskinetic activity of MPEP could also extend to A2A receptors. Therefore, we investigated A2A receptors in the brain of these de novo MPTP monkeys treated with L-DOPA and MPEP where motor behaviour [17], glutamate [18] and DA receptors [109] were previously reported. A2A receptors were measured by in situ hybridization of their mRNA levels using oligonucleotides probes corresponding to bases 593–637 and 714–757 of human A2A receptors cDNA [110, 111] under conditions we previously described for MPTP monkeys [112] and human post-mortem brains [80].

Fig. shows representative A2A receptor mRNA levels in brain slices of control and MPTP monkeys treated with vehicle, L-DOPA, and L-DOPA+ MPEP. High expression of this receptor in control monkeys is observed in the caudate nucleus and putamen but not in the GPi and GPe as we previously reported [112].

In the posterior caudate nucleus analyzed (Fig. ), no significant effect of lesion and treatment was measured on A2A receptor mRNA levels (one-way analysis of variance (ANOVA) followed by post-hoc pairwise comparisons 
with Fisher’s least significant difference test, (Fisher’s test) F3,13 = 2.026, p = 0.1600) consistent with our previous findings in another group of MPTP monkeys [112]. By contrast, significant changes of A2A receptor mRNA levels were measured in the lateral (ANOVA and Fisher’s test, 
F3,13 = 6.140 p = 0.0078) and medial (ANOVA and Fisher’s test, F3,14 = 3.740, p = 0.0389) putamen. The putamen showed an increase of A2A receptor mRNA levels with the MPTP lesion in its lateral and medial parts (+61% and +43% in the lateral and medial parts respectively compared to intact-saline-treated monkeys) (Fig. 2). A2A receptor mRNA levels remained elevated in MPTP monkeys treated with L-DOPA, those animals displaying dyskinesias. In MPTP monkeys treated with L-DOPA and MPEP that developed less dyskinesias than those treated with L-DOPA [17], A2A receptor mRNA levels were lower and this was significant compared to MPTP saline treated animals in the lateral putamen a sub-region associated with motor control whereas in the medial part the difference was intermediate, neither significant compared to controls nor MPTP + L-DOPA treated MPTP monkeys. These results are consistent with our previous findings where A2A receptor mRNA levels were elevated in MPTP and dyskinetic MPTP + L-DOPA treated MPTP monkeys and lower in MPTP monkeys treated with L-DOPA and CI-1041, an NMDA glutamate receptor antagonist, that prevented the development of dyskinesias [112]. However, we cannot rule out the possibility that MPEP or CI-1041 treatments could have a direct effect on the expression of A2A receptors thereby associating the change in levels of these receptors to the antagonist treatment themselves rather than to the dyskinetic condition. Indeed in mice, striatal binding of the A2A receptor antagonist ligand [125I]ZM241385 was reduced in brain slices pre-incubated with MPEP compared to untreated animals and this was attributed to a possible change of conformation of the mGlu5-A2A heteromers and consequently the binding affinity of [125I]ZM241385 for A2A receptors [113].

Nevertheless, these findings in MPTP monkeys model well the human condition where we showed that A2A receptor mRNA levels were elevated in PD patients with dyskinesias compared to PD patients without dyskinesias and controls in the putamen while no significant effect was measured in the caudate nucleus [80]. In this regard, post-mortem studies have shown an increase of basal ganglia A2A adenosine receptors and its mRNA levels in PD patients with LID as compared to PD patients without LID and controls [80]. As for post-mortem human brains, A2A receptor levels were only increased in dyskinetic monkeys as compared to untreated monkeys [112]. These results in PD patients were in accordance with a report by Varani et al. (2010) showing elevated A2A receptor mRNA levels (5.2 fold increase) and an overexpression of A2A protein levels (280%), measured by Western blot in the putamen of PD patients, when compared to healthy subjects [114]. In the present study and in our previous ones in MPTP monkeys [112] and PD patients [80] similar changes were observed for A2A receptor mRNA levels and [3H]-SCH 58261 specific binding to A2A receptors supporting that changes of expression of this receptor in PD is the result of altered transcriptional activity and are functionally associated with changes of this receptor protein levels. A2A receptors availability in PD patients with and without LID was evaluated by PET studies and higher levels of A2A receptor binding in PD patients with LID compared to those without LID were reported [115, 116]. This suggests that LID might be the result of altered adenosine transmission and that A2A receptor antagonists could produce potential beneficial antidyskinetic activity. The molecular mechanisms involving A2A receptors in the induction of dyskinesia is not yet fully elucidated. However, it is well known that activation of A2A receptors disrupts the inhibitory actions of D2 receptors within the indirect pathway resulting in a facilitation of DA transmission [117]. In the light of various biochemical and behavioural results showing the functional interaction between A2A and D2 receptors, it was postulated that altered striatal D2-A2A functional interaction in PD creates an imbalance in the activity of the two efferent pathways of the basal ganglia that could be associated with motor alterations inherent to dyskinesia [102, 117]. These changes suggest that both striatal A2A receptors and their mRNA levels are modulated by chronic L-DOPA treatment and might be related to the close interactions with DA and glutamate systems in the basal ganglia. These findings also suggest that the striatal mGlu5–mGlu5 dimeric receptor complexes on the striato-pallidal efferents neurons [104] as well as the putative mGlu5-D2-A2A complexes on the same neuronal population, might be involved in striatal plasticity and hence could be relevant for the management of PD and LID [67, 71, 105].

Discussion

The mechanisms involved in the occurrence of LID are complex and have been investigated in numerous studies using animal models and PD patients. Multiple changes in the basal ganglia dopaminergic systems in D1 and D2 receptors and their signaling pathways have been observed such as modulation in the expression and the activity of subtypes of DA receptors, G proteins, effectors, protein kinases, transcription factors, etc [118]. Abnormal adaptation in the striatum leading to LID may also involve faulty interaction between glutamate and DA inputs and DA-glutamate signaling in the nigrostriatal pathway [119].

Glutamate is the brain most abundant excitatory neuro-
transmitter mediating as much as 70% of synaptic transmission in the central nervous system and its overactivity is well documented in PD and LID. An attractive strategy to treat LID is to use anti-glutamatergic adjunct drugs that can modulate basal ganglia dopaminergic neurotransmission [120-122]. The present review showed that a chronic treatment with mGlu5 negative allosteric modulators in 
6-OHDA rats and MPTP monkeys treated with L-DOPA normalized changes produced by L-DOPA on D2 receptor and its mRNA, PPE mRNA, PPD mRNA, ERK1/2 and Akt/GSK3β signaling proteins, but not D1 and its mRNA.

In the basal ganglia of post-mortem PD patients with motor complications, changes of glutamate neurotransmission and its receptors such as NMDA, AMPA, mGlu2 and mGlu5 receptors were reported [8, 20, 123]. In these same patients, A2A receptor levels were measured and they were increased in the basal ganglia and associated with LID [80]. These observations in human brains do not provide a causal link between glutamate and adenosine receptors in motor complications. Nevertheless, they show altered adenosine and glutamate receptors in human PD and that normalization of glutamatergic transmission in MPTP monkeys with a mGlu5 receptor negative allosteric modulator normalizes also A2A receptors suggesting possible interactions between these receptors. Moreover, in MPTP monkeys, a chronic treatment with L-DOPA and a NMDA receptor antagonist CI-1041, that prevented LID, was shown to normalize basal ganglia A2A receptors as well as NMDA receptors suggesting the close link between these neurotransmitters in dyskinesias [112].

The interactions of A2A receptors, with dopaminergic and glutamatergic receptors represent an interesting area of research. In the light of theses interactions, new therapeutic approaches could include the combination of an A2A receptor antagonist with dopaminomimetic drugs or with anti-
glutamatergic drug. This combination could lead to the use of lower doses of each drug, especially L-DOPA, and could have an impact on PD symptoms and the development of motor complication. For example, A2A receptor antagonists could be combined to mGlu5 receptor negative allosteric modulators and to lower doses of L-DOPA or DA agonists to increase the antiparkinsonian activity of the dopaminergic drugs without increasing or even decreasing dyskinesias [124]. The mGlu5 receptor subtype is highly expressed in striatal medium spiny neurons [16] and plays a key role in modulating the responses mediated by NMDA receptors and L-type calcium channels [125]. In addition, an antagonistic interaction between the D2 receptor and mGlu5 receptors is reported [55]. The response of the basal ganglia to a chronic treatment with a mGlu5 receptor negative allosteric modulators in rodent models of PD show that striatal molecular changes relevant to LID are reversed by MPEP or MTEP, including delta FosB protein [126], prodynorphin mRNA [26], glutamic acid decarboxylase (GAD65 and GAD67) mRNA [127] and phospho-ERK1/2 protein levels [128]. We showed that MPEP prevented dyskinesias and the effects of chronic L-DOPA on various ionotropic and metabotropic glutamate receptors in the basal ganglia of MPTP monkeys thus showing the widespread activity of mGlu5 receptor negative allosteric modulators in the basal ganglia [18]. MPEP did not affect D1 receptor levels and its mRNA but was associated with an increased in D2 receptors levels, its mRNA and its associated signaling proteins.

Conclusion

In rodent and primate models of PD and LID, mGlu5 receptor negative allosteric modulators are well documented to inhibit dyskinesias and have an extensive range of beneficial biochemical effects. In parkinsonian primate and rodent treated with L-DOPA and mGlu5 negative allosteric modulators, crosstalk between mGlu5-D2-A2A receptors seems to be present whereas recent studies reported that L-DOPA disrupts D2-A2A-CB1 receptor heteromers crosstalk in the striatum of hemiparkinsonian rats [103] and primates in the caudate nucleus but not the putamen (two levels were investigated, pre and post commissural) [102]. The mGlu5 negative allosteric modulator treatment in parkinsonian primates and rats affects not only glutamate receptors but also DA and A2A receptors. This may be indirect by restoring glutamate neurotransmission but could also involve the direct interactions of the trio mGlu5-D2-A2A receptor crosstalk [96, 129]. Moreover, supporting a close mGlu5-A2A receptor interaction, MTEP treatment was reported to decrease mice brain A2A receptor specific binding and regulate the conditioned effects of cocaine [113].

Thus, mGlu5 receptor negative allosteric modulators preventing the development of LID and inhibiting the expression of already developed LID not only affects several glutamate receptors but also D2 and A2A receptors, neuropeptides and Akt/GSK3β and ERK1/2 signaling. An abundant literature mainly from rodent models of PD supports the striatal overactivation of ERK1/2 via DA D1 receptors to be associated with dyskinetic behaviors [130]; reviewed in [131] that is critically modulated by striatal mGlu5 receptors. Thus mGlu5 receptor negative allosteric modulators affect various markers of both the striatal direct and indirect output pathways. The implication of these receptor interactions in mental and neurodegenerative diseases and more specifically in the development and expression of PD symptoms and LID needs to be considered and further investigated to find novel targets and ultimately, novel pharmacological treatments.