Inactivation of neuronal forebrain A2A receptors protects dopaminergic neurons in a mouse model of Parkinson’s disease

Authors


Address correspondence and reprint requests to Anna R Carta, PhD, Department of Toxicology, University of Cagliari, via Ospedale 72, Cagliari 09124, Italy. E-mail: acarta@unica.it

Abstract

Adenosine A2A receptors antagonists produce neuroprotective effects in animal models of Parkinson’s disease (PD). As neuroinflammation is involved in PD pathogenesis, both neuronal and glial A2A receptors might participate to neuroprotection. We employed complementary pharmacologic and genetic approaches to A2A receptor inactivation, in a multiple MPTP mouse model of PD, to investigate the cellular basis of neuroprotection by A2A antagonism. MPTP·HCl (20 mg/kg daily for 4 days) was administered in mice treated with the A2A antagonist SCH58261, or in conditional knockout mice lacking A2A receptors on forebrain neurons (fbnA2AKO mice). MPTP-induced partial loss of dopamine neurons in substantia nigra pars compacta (SNc) and striatum (Str), associated with increased astroglial and microglial immunoreactivity in these areas. Astroglia was similarly activated 1, 3, and 7 days after MPTP administration, whereas maximal microglial reactivity was detected on day 1, returning to baseline 7 days after MPTP administration. SCH58261 attenuated dopamine cell loss and gliosis in SNc and Str. Selective depletion of A2A receptors in fbnA2AKO mice completely prevented MPTP-induced dopamine neuron degeneration and gliosis in SNc, and partially counteracted gliosis in Str. Results provide evidence of a primary role played by neuronal A2A receptors in neuroprotective effects of A2A antagonists in a multiple MPTP injections model of PD. With the symptomatic antiparkinsonian potential of several A2A receptor antagonists being pursued in clinical trials, this study adds to the rationale for broader clinical benefit and use of these drugs early in the treatment of PD.

Abbreviation used:
fbn

forebrain neurons

GFAP

glial fibrillary acidic protein

IR

immunoreactivity

KO

knockout

PD

Parkinson’s disease

SNc

substantia nigra pars compacta

Str

striatum

TH

tyrosine hydroxylase

WT

wild-type

As the initial demonstration that caffeine can attenuate MPTP-induced toxicity, a neuroprotective role of A2A receptor blockade has been suggested in several experimental models of Parkinson’s disease (PD) (Chen et al. 2001; Schwarzschild et al. 2006). In rodents, different A2A receptor antagonists have been shown to counteract the loss of dopaminergic neurons in the substantia nigra pars compacta (SNc), as well as dopamine depletion in the striatum (Str), induced by acute administration of systemic MPTP or acute infusion of 6-hydroxydopamine in the medial forebrain (fb) bundle (Chen et al. 2001; Ikeda et al. 2002; Pierri et al. 2005). Moreover, loss of striatal dopamine induced by acute MPTP was attenuated by the global genetic deletion of A2A receptors in A2A knockout (KO) mice (Chen et al. 2001). In contrast to acute models, neuroprotection by A2A antagonists in the neurotoxicity induced by multiple MPTP administration was never evaluated. On the other hand, previous studies have shown that acute or multiple injections MPTP delivery result in different histopathological features and different modes of cell death (Schmidt and Ferger 2001). Interestingly, A2A receptor antagonism or gene KO has been found to be neuroprotective in different models of neurodegeneration, such as Alzheimer’s disease, Huntington’s disease, and cerebral ischemia (Chen et al. 1999; Dall’Igna et al. 2003; Melani et al. 2003; Fink et al. 2004).

A2A receptors are enriched in the Str, where they are located either post-synaptically in striatopallidal neurons or pre-synaptically in nerve terminals (Schiffmann et al. 1991; Rosin et al. 1998; Svenningsson et al. 1999; Rebola et al. 2005). Moreover, A2A receptors are expressed at low level in other fb structures, such as cortex and hippocampus, whereas little or no A2A receptor immunoreactivity (IR) has been detected in dopaminergic neurons in the SNc (Rosin et al. 1998). Besides neurons, non-neuronal cell types such as microglia and astroglia also express A2A receptors (Fiebich et al. 1996; Saura et al. 2005).

The mechanism through which A2A receptor antagonists achieve neuroprotection in PD models has not been elucidated. Given the cellular distribution of A2A receptors in different cell types in brain, neuroprotection by A2A receptor blockade may be achieved through an action on receptors located either on neurons or on glial cells. Attenuation of gliosis by A2A receptor antagonists in the Str and SNc of mice acutely treated with MPTP, suggests that reduction of neuroinflammation may be involved (Ikeda et al. 2002; Pierri et al. 2005). Moreover, it was recently suggested that A2A receptors located on glial cells may play a role in neuroprotection mediated by A2A antagonists against an acute MPTP-induced striatal dopamine depletion (Yu et al. 2008).

To get insight into the mechanism by which A2A receptor antagonists induce neuroprotection, this study evaluated (i) the neuroprotective activity of an A2A antagonist in a mouse model of PD obtained by a multiple injections MPTP delivery, (ii) the role of neuronal A2A receptors on this effect, and (iii) whether neuronal A2A receptor blockade might affect glial response to MPTP. To this aim, we treated mice with MPTP plus the A2A antagonist SCH58261, or administered MPTP to genetically manipulated mice selectively lacking A2A receptors in fb neurons (fbnA2AKO mice) (Bastia et al. 2005; Shen et al. 2008). Neuronal damage was evaluated in the SNc and Str using tyrosine hydroxylase (TH) immunohistochemistry, whereas the inflammatory response in the SNc and Str was studied through analysis of glial fibrillary acidic protein (GFAP) and CD11b-IR as markers of astroglial and microglial cells, respectively.

This study provides evidence of a primary role played by fbn A2A receptors in the neuroprotection that A2A receptor antagonists confer against dopaminergic neuron degeneration and glial activation induced by repeated MPTP.

Materials and methods

Animals and treatments

Male C57BL/6J mice (Charles River, Milano, Italy) were used for experiments involving pharmacological treatments with the A2A receptor antagonist. Mice were housed five per cage, with a 12 : 12 h light/dark cycle and with food and water ad libitum. Experiments were conducted in accordance with the guidelines for care and use of experimental animals of the European Communities Council Directive of November 24, 1986 (86/609/EEC) and the National Institutes of Health.

Pharmacological treatment

Male C57BL/6J mice (25–30 g, 3-months old), received a multiple injections treatment with vehicle (n = 15), MPTP·HCl (20 mg/kg i.p.) once a day for 4 days (n = 18), or the A2A antagonist SCH58261 (0.5 mg/kg i.p.) twice a day plus MPTP (20 mg/kg i.p.) once a day for 4 days (n = 23). SCH58261 was injected half an hour before MPTP administration. After MPTP treatment discontinuation SCH58261 treatment continued once a day, until killing, which occurred 1, 3, or 7 days after MPTP treatment. Injections were made at 8 am and 8 pm.

Postnatal conditional fbn A2AKO mice were generated using the Cre/loxP system based on the specificity of CaMKIIα promoter (Bastia et al. 2005; Yu et al. 2008; see also Appendix S1). FbnA2AKO mice and littermate controls were treated with repeated MPTP (n = 18 or 15) or vehicle (n = 12 or 10) as described in Pharmacological treatment section. Experiments with the A2A antagonist (see Results section) showed that loss of TH-immunoreactive cells was stable 1, 3, and 7 days after MPTP treatment and that CD11b-IR was greatest 1 and 3 days post-MPTP treatment. Based on these considerations, neuroprotection and glial reactivity parameters in transgenic mice were evaluated 3 days after MPTP treatment.

Drugs

MPTP·HCl (Sigma-Aldrich, Milano, Italy or St. Louis, MO, USA) was dissolved in 0.9% saline in a volume of 0.1 mL/10 g; SCH58261 (kindly provided Prof. Baraldi, Ferrara) was suspended in 0.5% of methylcellulose.

Immunohistochemistry

Animals were anesthetized with chloral hydrate (400 mg/kg i.p.) or Avertin (0.1 mL/10 g i.p.) prior to transcardial perfusion with 20 mL of saline and 60 mL of 4%p-formaldehyde. Brains were removed and post-fixed for 2 h. Adjacent coronal sections (50 μm) from the Str and SNc were cut on a vibratome and stored at −20°C in an antifreeze medium until use (Schintu et al. 2009). For TH, GFAP, and CD11b immunostaining, adjacent sections were processed as described (Schintu et al. 2009; see also Appendix S1).

Analysis and statistics

Images were digitized (videocamera Pixelink PL-A686, Ottawa, ON, Canada) under constant light conditions to standardize the measurements. Immunostained sections containing left and right SNc were captured at 10× magnification (the entire SNc, corresponding to three frames, was digitized for the analysis). Immunostained sections of bilateral striata were captured at 20× magnification. One portion from the dorsolateral Str and one from the ventromedial Str (520 × 380 μm) were analyzed. For each animal, three sections corresponding to rostral (within −2.90/−3.20 mm from bregma), medial (−3.20/−3.50), and caudal (−3.50/−3.80) SNc levels, and three sections corresponding to rostral (within 1.20/0.90 mm from bregma), medial (0.90/0.60), and caudal (0.60/0.30) Str (accordingly to Mouse Brain Atlas, Paxinos and Franklin 2001) were analyzed for each protein marker evaluated in the study.

Tyrosine hydroxylase and glial fibrillary acidic protein analyses

As the number of cells was different in the three SNc and Str levels analyzed, for each mouse the number of TH- or GFAP-positive cells/level was first normalized with respect to the vehicle. Individual values from the three levels were then averaged to generate a mean. Adjacent SNc sections were Nissl-stained to confirm cell loss in this area.

The SNc from three mice randomly selected from each experimental group were counterstained with cresyl violet for evaluation by unbiased stereological counting, and the mean of TH-IR and Nissl-stained neurons per mm3 ± SEM was calculated (see Appendix S1).

CD11b analysis

Images were digitized in a gray scale, and CD11b immunostaining was evaluated with the analysis program scion Image (Scion Corporation, Frederick, MD, USA). A threshold, the value of which was set above the mean ± SEM of the background, was applied for background-correction. Inside each frame, the area occupied by gray values above the threshold was automatically calculated. For each level of SNc or Str, the obtained value was first normalized with respect to vehicle; the individual values from the three levels were then averaged to generate a mean.

Results from mice treated with MPTP plus SCH58261, or from fbnA2AKO mice were statistically compared with a two-factors anova, followed by Tukey’s post hoc test, for comparison between experimental groups.

Results

A2A receptor antagonist prevents dopaminergic cell loss in the SNc

In C57BL/6J mice, MPTP treatment induced a partial dopaminergic neurons degeneration in the SNc (Fig. 1a and b and Table 1). Counting of Nissl-stained cells confirmed this result (Table 1). Dopaminergic neurons loss was statistically significant 1 day after MPTP treatment (n = 5) and remained significant after 3 (n = 5) and 7 days time points (n = 8). Combined treatment with A2A antagonist SCH58261 plus MPTP, attenuated TH-positive neurons loss in the SNc at 1 (n = 5), 3 (n = 5), and 7 days (n = 13) (Fig. 1a and b and Table 1), as confirmed by Nissl-staining (Table 1). Two-factors anova showed a significant effect of treatment (see Table S1 for corresponding F and p values). In the Str, MPTP-induced decrease in TH-IR was significantly attenuated by treatment with the A2AR antagonist, as measured 3 days after treatment (Fig. 2).

Figure 1.

 Adenosine A2Areceptor antagonist SCH58261 prevents dopaminergic cell loss in the SNc. (a) Shows representative sections immunostained for TH from SNc of mice killed 3 days after MPTP treatment. Left inset shows TH-positive cells at higher magnification and right inset shows cresyl violet-stained sections; scale bar, 50 μm. Mice were treated with MPTP·HCl (20 mg/kg once a day for 4 days), plus SCH58261 (0.5 mg/kg) or vehicle (twice a day during MPTP treatment and once daily thereafter until killing), and killed 1, 3, and 7 days after MPTP treatment. (b) Shows analysis of TH immunostaining at 1, 3, and 7 days after MPTP, reported as a percentage of TH-positive cells when compared with vehicle-treated mice; *p < 0.001 versus vehicle and #p < 0.001 versus MPTP group, by Tukey’s post hoc test. Scale bar, 50 μm.

Table 1.   Unbiased evaluation of TH-IR and Nissl-stained neurons by stereological analysis in the SNc of mice following pharmacological blockade of A2ARs with the antagonist SCH58261 or A2AR genetic depletion
TreatmentnDensity of TH-IR neurons/mm3Density of Nissl-IR neurons/mm3
  1. TH, tyrosine hydroxylase; KO, knockout; WT, wild-type; IR, immunoreactivity; *p < 0.05 versus vehicle group and ^p < 0.05 versus the corresponding MPTP-treated group. Stereological evaluation of TH-IR and Nissl-stained neurons in the substantia nigra pars compacta.

Pharmacological blockade of A2AR
Vehicle329 725.89 ± 439.5637 702.70 ± 2520,89
MPTP315 211.79 ± 1300.06*24 827.52 ± 1689.72*
SCH + MPTP323 003.66 ± 495.76*^32 817.67 ± 1225.20^
FbnA2AR depletion
WT vehicle327 447.34 ± 1365.0832 198.28 ± 3341,77
WT MPTP314 278.85 ± 2493.30*21 784.92 ± 2266.61*
KO vehicle323 654.36 ± 2277.7430 903.52 ± 4264.73
KO MPTP324 267.96 ± 2277.83^29 832.48 ± 883.17^
Figure 2.

 Adenosine A2A receptor antagonist SCH58261 attenuates degeneration of dopaminergic terminals in the Str. (a) Representative sections immunostained for TH, from the Str of mice killed 3 days after MPTP treatment. (b) Results from pharmacological blockade with SCH58261 or genetic A2AR depletion are presented. In the left column, − and + indicate the administration of vehicle or SCH58261to MPTP-treated mice. In the right column + and − indicate fbnA2AWT and fbnA2AKO mice, respectively.

A2A receptor antagonist inhibits astroglia and microglia activation

In brain sections from vehicle-treated mice, few GFAP-positive cells (Fig. 3), and a low CD11b immunostaining (Fig 4), were detected in the SNc and Str. Repeated MPTP treatment induced an increase in GFAP and CD11b-positive cells in both the SNc and Str of C57BL/6J mice.

Figure 3.

 Adenosine A2A receptor antagonist SCH58261 counteracts astroglia activation in the SNc and Str. (a) Shows representative sections immunostained for GFAP, from SNc (upper images) and Str (lower images) of mice killed 3 days after MPTP treatment. Mice were treated as described in Fig 1. (b) Shows analysis of GFAP immunostaining 1, 3, and 7 days after MPTP, reported as percentage of GFAP-positive cells when compared with vehicle-treated mice in the SNc (left graph) and in the Str (right graph); *p < 0.001 versus corresponding vehicle and MPTP + SCH58261 groups and #p < 0.001 versus corresponding MPTP group, by Tukey’s post hoc test. Scale bar, 50 μm.

Figure 4.

 Adenosine A2Areceptor antagonist SCH58261 counteracts microglia activation. (a) Representative images from the SNc immunostained for CD11b as a marker of microglia activation. Mice were treated as described in Fig. 1 and killed 1, 3, and 7 days after MPTP treatment. (b) CD11b analysis in SNc and Str was performed in gray scale digitized images. The area occupied by gray values above a threshold was calculated and expressed as square pixels and as percentage of vehicle-treated mice. Tukey’s post hoc test: *p < 0.001 versus vehicle and MPTP + SCH58261 group; #p < 0.001 versus MPTP group; ^p < 0.05, ^^p < 0.001 versus the indicated time point. Scale bar, 50 μm.

Glial fibrillary acidic protein-positive cells displayed a highly branched morphology with tiny processes and a small body in control sections, and became hypertrophic in response to MPTP treatment (Fig. 3a); moreover, CD11b-positive cells were ramified at baseline, but took on an ameboid aspect after MPTP treatment (Fig. 4a), indicative of astroglial and microglial activation. In both the SNc and Str, GFAP immunolabeling was of similar intensity 1, 3, and 7 days post-MPTP treatment (Fig. 3b). Combined SCH58261 plus MPTP treatment, completely prevented the increase in GFAP-IR in the SNc and partially prevented it in the Str, at 1, 3, and 7 days after MPTP treatment (Fig. 3b). Two-factors anova for GFAP analysis showed a significant effect of the treatment at all time points analyzed in the SNc and in the Str (see Table S1).

CD11b immunolabeling was highest 1 day after MPTP, gradually declining to basal levels after 7 days in both SNc and Str (Fig. 4a and b). Combined SCH58261 plus MPTP treatment completely prevented the increase in CD11b-IR in the SNc 1 and 3 days after MPTP treatment (Fig. 4a and b). In the Str, SCH58261 partially prevented MPTP-induced increase in CD11b at 1 day, and totally prevented it at 3 days after MPTP treatment. Two-factors anova and post hoc analysis for CD11b revealed a significant effect of treatment, time, and a treatment/time interaction, 1 and 3 days after MPTP administration in the SNc and in the Str (see Table S1).

Dopaminergic cell loss is prevented in fbnA2AKO mice

In fbnA2A wild-type (WT) mice, repeated MPTP treatment induced a significant loss of dopamine neurons in the SNc (Fig. 5a and b and Table 1). This result was confirmed by a reduction in Nissl-stained cells (Table 1 and see right inset in Fig. 5a). In contrast, in fbnA2AKO mice, repeated MPTP treatment did not result in a decrease in TH-immunolabeling in the SNc (Fig. 5a and b and Table 1). Two-factors anova revealed a significant effect of treatment and genotype (see Table S1 for corresponding F and p values). In the Str of fbnA2AKO mice repeated MPTP reduced TH immunolabeling to a lesser extent when compared with fbnA2AWT controls (Fig. 2).

Figure 5.

 FbnA2A KO mice are protected against MPTP-induced loss of dopaminergic cells in the SNc. (a) Shows representative sections from SNc immunostained for TH. Insets show higher magnification of TH-labeled (left) and cresyl violet-labeled (right) cells. Mice were treated with MPTP (20 mg/kg once a day for 4 days) or vehicle. (b) Shows analysis of TH immunostaining in fbnA2AKO mice, reported as a percentage of TH-positive cells when compared with vehicle-treated mice. Tukey’s post hoc test: *p < 0.05 versus vehicle group and #p < 0.05 versus WT MPTP group. Scale bar, 50 μm.

Astroglial and microglial activation is attenuated in fbnA2AKO mice

Glial fibrillary acidic protein-IR in the SNc and Str was significantly enhanced in MPTP-treated fbnA2AWT mice (Fig. 6a and b). In contrast, in fbnA2AKO mice, MPTP-induced increase in GFAP-IR was totally prevented in the SNc, and partially prevented in Str (Fig. 6a and b). In both brain regions two-factors anova revealed a significant effect of treatment, genotype and a significant treatment/genotype interaction (see Table S1).

Figure 6.

 Astroglia activation is attenuated in SNc and Str of fbnA2AKO mice. (a) Representative images from the SNc immunostained for GFAP, as a marker of astroglial cells. (b) Graphs show the analysis of GFAP immunostaining in SNc and Str, in fbnWT and fbnA2AKO mice treated with vehicle or MPTP. Tukey’s post hoc test: *p < 0.001 versus vehicle group and #p < 0.001 versus WT MPTP group. Scale bar, 50 μm.

CD11b immunolabeling in the SNc and Str was significantly enhanced in MPTP-treated fbnA2AWT mice (Fig. 7a and b). In fbnA2AKO mice CD11b activation was totally prevented in the SNc and Str (Fig. 7a and b). Two-factors anova revealed a significant effect of treatment, genotype and a treatment/genotype interaction (see Table S1).

Figure 7.

 Microglia activation is prevented in SNc and Str of fbnA2AKO mice. (a) Representative images from the SNc immunostained for CD11b, as a marker of microglia activation. (b) Graphs show the analysis of CD11b immunostaining in SNc and Str, in fbnWT and fbnA2AKO mice treated with vehicle or MPTP. Tukey’s post hoc test: *p < 0.001 versus vehicle group and #p < 0.001 versus WT MPTP group. Scale bar, 50 μm.

Discussion

Antagonism of adenosine A2A receptors or their selective deletion in fb neurons produced similar protection of TH-positive nigral neurons in a multiple MPTP injections mouse model of PD. These complementary pharmacological and genetic means of A2A receptor disruption also attenuated the neurotoxin-triggered activations of astroglia and microglia along the nigrostriatal pathway. Together these data provide evidence that therapeutically accessible A2A receptors located on fb neurons play a critical role in nigral dopaminergic neuron degeneration and inflammatory processes in the multiple MPTP injections mouse model of PD.

Pharmacological blockade of A2A receptors prevents MPTP-induced dopaminergic neuron degeneration and glial activation

Systemic administration of the A2A receptor antagonist SCH58261 prevented the degeneration of nigrostriatal TH-positive neurons induced by repeated MPTP exposure in mice. Changes in number of TH-positive neurons correlated with changes in Nissl-stained (cresyl violet-positive) cells, indicating that MPTP treatment resulted in actual loss of dopaminergic neurons, which were rescued by SCH58261.

A neuroprotective effect of A2A receptor antagonists was previously observed upon acute administration of high MPTP doses in mice (Chen et al. 2001; Ikeda et al. 2002; Pierri et al. 2005; Yu et al. 2008). Here, we report that neuroprotection with A2A antagonism can also be achieved upon multiple low-doses MPTP exposure. A number of studies have provided evidence that a repeated daily MPTP administration protocol similar to the one used here, presents histopathological features that more closely reproduce the human PD neuropathology, including apoptotic death of dopaminergic neurons (Jackson-Lewis et al. 1995; Tatton and Kish 1997). Therefore, this study further substantiates the neuroprotective potential of A2A antagonism in PD. To this regard, it is noteworthy that A2AR antagonists were shown to inhibit apoptotic neuronal death in hippocampal neurons (Silva et al. 2007).

Based on their differential location in the Str or other brain regions, A2A receptors may hold different levels of expression and intracellular signaling, reflecting A2A receptor multiple functions (Kull et al. 2000; Pedata et al. 2003; Rosin et al. 2003; Rebola et al. 2005; Shen et al. 2008). According to such varied roles of the A2A receptor, diverse effects have been attributed to A2A antagonists, ranging from symptomatic antiparkinsonian actions to neuroprotection in various neurodegenerative conditions (Alfinito et al. 2003; Blum et al. 2003; Chen et al. 1999; Melani et al. 2003; Popoli et al. 2002). Motor effects of A2A receptor antagonists are likely mediated by A2A receptors located on striatal neurons projecting to globus pallidus, whereas several mechanisms have been hypothesized for their neuroprotective effects, involving either neuronal or glial A2A receptors, though no single mechanism has yet been proven to prevail (Popoli et al. 1995; Chen et al. 2001; Carta et al. 2003; Melani et al. 2003; Pedata et al. 2003; Huang et al. 2006; Schwarzschild et al. 2006; Yu et al. 2008). Noticeably, in this study neuroprotection by SCH58261 was achieved at doses similar to those effective in other neurodegenerative conditions, but several times lower than doses displaying a symptomatic efficacy in PD (Chen et al. 2001; Dall’Igna et al. 2003; Melani et al. 2003; Pinna et al. 2007), supporting the concept that different mechanisms might account for A2A-mediated neuroprotection or symptomatic effects.

SCH58261 fully prevented astroglia and microglia activation in the SNc, while only partially inhibiting astroglia and microglia reactivity in the Str, in line with a partial protection of dopaminergic terminals. Noteworthy, A2A receptor antagonism prevented both astroglia and microglia activation in the SNc and Str at all time points evaluated, in accordance with blockade of neurodegeneration.

Although the mechanism through which A2A receptor blockade produces neuroprotective effects in PD models is unclear, the modulation of neuroinflammation has been proposed as a likely target for neuroprotection (Hunot and Hirsch 2003). Several findings have suggested that neuroinflammation may play an active role in the pathogenesis of neurodegeneration in PD, as focal inflammation has been described in the SNc of PD patients and MPTP-treated primates (McGeer et al. 1988; Barcia et al. 2004). Intriguingly, blockade of microglia reactivity in mice rescued dopamine neurons from acute MPTP toxicity (Wu et al. 2002). Moreover, in mice acutely treated with MPTP, dopamine neuron neuroprotection by pre-treatment with an A2A antagonist was associated with an attenuation of astroglia and microglia activation in SNc and Str (Ikeda et al. 2002; Pierri et al. 2005), consistent with a causal relation between the two events.

Selective deletion of neuron-specific forebrain A2A receptors prevents MPTP-induced dopamine neuron degeneration and glial activation

Previous studies evaluating A2A receptor-mediated neuroprotection have hypothesized several mechanisms that might underlie this process. To determine the role of neuronal versus glial A2A receptors in neuroprotection of dopamine neurons, we exploited genetically modified mice with selective depletion of A2A receptors from fb neurons (Bastia et al. 2005; Shen et al. 2008). Importantly, in the fbnA2AKO mice, deletion of A2A receptor is not only spatially restricted (to fbn A2A receptor) but is also temporally limited to postnatal A2A receptors (Bastia et al. 2005; Yu et al. 2008), thus avoiding potential confounds of compensatory responses to A2A receptor gene disruption during development as might occur in constitutive A2A KO mice.

Our results revealed that selective deletion of A2A receptors from fb neurons totally prevented dopaminergic neuron loss in the SNc following multiple MPTP injections, while partially preventing damage to striatal dopaminergic terminals. A2AR deletion provided a greater protection of SNc neurons than the A2AR antagonist, as expected from a permanent when compared with a temporal pharmacological blockade of the receptor, in line with the reported half-life of SCH58261 of 2–3 h.

A2A receptors are located both pre- and post-synaptically in striatal and cortical neurons, and are expressed in microglial as well as astroglial cells (Küst et al. 1999; Cunha 2001; Rosin et al. 2003; Nishizaki 2004;Rebola et al. 2005). Positive modulation of post-synaptic signaling as well as of pre-synaptic release of neurotransmitters as glutamate and acetylcholine by A2A receptors have been described (Popoli et al. 1995; Marchi et al. 2002; Fredholm et al. 2003; Fuxe et al. 2003; Schwarzschild et al. 2006; Schiffmann et al. 2007). In addition, A2A receptors interfere with glia-mediated synthesis and release of neurotoxic factors such as cyclooxygenase-2, prostaglandins, nitric oxide, and glutamate, which have been hypothesized to play central roles in inflammatory processes and neuronal damage (Fiebich et al. 1996; Li et al. 2001; Saura et al. 2005).

This study, by showing that selective deletion of A2A receptors from fb neurons protects dopaminergic neurons from MPTP toxicity, endorses a primary role of neuronal receptors in mediating neuroprotection in this multiple injections MPTP model of PD. As very low levels of A2A receptors are expressed by dopaminergic neurons in the SNc, it is unlikely that a direct action at this level might mediate neuroprotection from MPTP toxicity in the SNc. Rather, an indirect effect at the pre-synaptic level, through an inhibition of A2A-mediated glutamate release, which contribute to neuronal damage, could be envisaged (Aguirre et al. 2005; Battaglia et al. 2004; Monopoli et al., 1998; Popoli et al. 2002). Interestingly, recent studies have reported a tight cross-talk between adenosine and glial cell line-derived neurotrophic factor receptors, resulting in a fine modulation of glutamate and dopamine release (Gomes et al. 2006, 2009). In the Str, A2AR blockade would impair glial cell line-derived neurotrophic factor-stimulated increase of corticostriatal glutamate release, thus providing a beneficial effect on neurodegeneration. In addition, A2A receptor antagonism on striatopallidal or subthalamic neurons might be protective from MPTP toxicity by modulating excessive activation of subthalamic nucleus, thereby reducing excitotoxic glutamate efflux to SNc neurons (Wallace et al. 2007).

A study by Yu et al. (2008) reported that fbnA2AKO mice from the same line as used here, were not protected from striatal dopamine loss in response to acute MPTP exposure (in a single dose or multiple doses over 4 h). These results open to several compelling interpretations. First, the cellular basis of A2A receptor-dependence of MPTP toxicity might vary depending on the duration of toxin exposure, in line with the different type of neurotoxicity produced by acute when compared with subchronic MPTP. Moreover, it should be taken into account the different parameters were used to evaluate nigrostriatal neurons damage. The drop of striatal dopamine levels assessed by Yu et al. may reflect functional injury to dopaminergic terminals, whereas the measure of dopaminergic nigral neuron employed in our study may reflect an underlying neurodegenerative process in this area. All together results support the concept that A2A receptors display complex actions related to the duration of insult, cellular elements and brain areas targeted by neurodegenerative processes.

Lack of a glial reaction in fbnA2AKO mice, when compared with the robust astroglia and microglia activation in MPTP-treated control mice, indicates that deletion of neuronal A2A receptors may indirectly inhibit the inflammatory response. Glutamate is a main contributor to the complex neuron–glia crosstalk engaged by pathological events, which trigger both microglia and astroglia activation. For instance, by an action on NMDA receptors, glutamate release stimulates mitogen-activated protein kinases. Neuronal as well as glial p38 mitogen-activated protein kinase activation has been involved in cell suffering and apoptotic death, being activated and inducing several inflammatory mediators (Kawasaki et al. 1997; Irving et al. 2000; Piao et al. 2003; Gianfriddo et al. 2004). Therefore, A2A antagonists indirectly through a reduction of glutamate release might counteract glial reactivity and neuroinflammation in both SNc and Str (Melani et al. 2003, 2006). Activated glial cells, by the release of several toxic species as cytokines, free radicals, glutamate, is known to contribute to neuronal damage, and has been suggested to sustain a self-amplifying cycle which perpetuates MPTP toxicity (Hunot and Hirsch 2003). Hence, interruption of such a detrimental vicious cycle might indirectly contribute to A2A receptor-dependent neuroprotection. Accordingly results suggest that in our model, operational glial A2A receptors did not contribute to MPTP toxicity when A2A fb neuronal receptors where deleted. All together, our results suggest that A2A antagonists, by blocking neuronal A2A receptors, might act upstream in the cascade of toxic events and lead to an attenuation of dopamine neuron degeneration in PD.

Acknowledgements

This study was supported by Ministero dell’Università e della Ricerca Scientifica e Tecnologica, project FIRB [Grant number RBNE03YA3L-2005] and National Institute of Health [Grant numbers ES010804, S54978] and U.S. Army Medical Research Acquisition Activity [Grant number W81XWH-04-1-0881]. The authors thank Prof P. G. Baraldi and Dr M. A. Tabrizi, Department of Pharmaceutical Sciences, University of Ferrara, for providing SCH58261; David H. Gutmann and M. Livia Bajenaru for providing GFAP-cre mice; and Yuehang Xu for expert technical assistance.

Ancillary