Neurobiology of Disease 65 (2014) 69–81

Contents lists available at ScienceDirect

Neurobiology of Disease journal homepage: www.elsevier.com/locate/ynbdi

Progressive Parkinsonism by acute dysfunction of excitatory amino acid transporters in the rat substantia nigra Maxime Assous a, Laurence Had-Aissouni a, Paolo Gubellini a, Christophe Melon a, Imane Nafia b, Pascal Salin a, Lydia Kerkerian-Le-Goff a,⁎,1,2, Philippe Kachidian a,⁎,1 a b

Aix-Marseille Université, CNRS, IBDML, UMR7288, 13009, Case 907, Parc Scientifique de Luminy, 13009 Marseille, France Fluofarma, 2 Rue Robert Escarpit, 33607, Pessac, France

a r t i c l e

i n f o

Article history: Received 2 October 2013 Revised 10 January 2014 Accepted 14 January 2014 Available online 27 January 2014 Keywords: Animal model Excitotoxicity Glutamate transporters Neurodegeneration Neuroprotection Oxidative stress Parkinson's disease PDC (L-trans-pyrrolidine-2,4-dicarboxylate)

a b s t r a c t Parkinson's disease (PD) is characterized by the progressive degeneration of substantia nigra (SN) dopamine neurons, involving a multifactorial cascade of pathogenic events. Here we explored the hypothesis that dysfunction of excitatory amino acid transporters (EAATs) might be involved. Acutely-induced dysfunction of EAATs in the rat SN, by single unilateral injection of their substrate inhibitor L-trans-pyrrolidine-2,4-dicarboxylate (PDC), triggers a neurodegenerative process mimicking several PD features. Dopamine neurons are selectively affected, consistent with their sustained excitation by PDC measured by slice electrophysiology. The antioxidant N-acetylcysteine and the NMDA receptor antagonists ifenprodil and memantine provide neuroprotection. Besides oxidative stress and NMDA receptor-mediated excitotoxicity, glutathione depletion and neuroinflammation characterize the primary insult. Most interestingly, the degeneration progresses overtime with unilateral to bilateral and caudo-rostral evolution. Transient adaptive changes in dopamine function markers in SN and striatum accompany cell loss and axonal dystrophy, respectively. Motor deficits appear when neuron loss exceeds 50% in the most affected SN and striatal dopamine tone is dramatically reduced. These findings outline a functional link between EAAT dysfunction and several PD pathogenic mechanisms/pathological hallmarks, and provide a novel acutely-triggered model of progressive Parkinsonism. © 2014 Elsevier Inc. All rights reserved.

Introduction Parkinson's disease (PD) is a neurodegenerative movement disorder in which cell death mainly affects substantia nigra (SN) dopamine (DA) neurons. Neuronal loss is progressive and motor symptoms appear when 50–60% of nigral DA neurons are lost. A number of mechanisms have been implicated in such cell death, including oxidative stress, mitochondrial dysfunction, protein misfolding/aggregation, autophagy, neuroinflammation and excitotoxicity (Ahmed et al., 2012; Cheung and Ip, 2009; Olanow, 2007; Schapira and Jenner, 2011; Yacoubian and Standaert, 2009). Depletion of glutathione (GSH), the main brain antioxidant, is considered as an early and key component of the pathological process. Reduced levels of GSH are measured specifically in the SN of PD patients and incidental Lewy body disease (presumably representing presymptomatic PD), but not in other pathologies affecting DA neurons (Jenner et al., 1992; Pearce et al., 1997; Perry et al., 1982; Sian et al., 1994a, 1994b). Down-regulation of GSH synthesis ⁎ Corresponding authors at: IBDML, UMR7288 CNRS/Aix-Marseille Université, Case 907, Parc Scientifique de Luminy, 13009 Marseille, France. Fax: +33 4 91 26 92 44. E-mail addresses: [email protected] (L. Kerkerian-Le-Goff), [email protected] (P. Kachidian). Available online on ScienceDirect (www.sciencedirect.com). 1 These authors contributed equally to the elaboration of this work. 2 Designate to communicate with the editorial and production offices. 0969-9961/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.nbd.2014.01.011

in vivo in rodents impacts mitochondrial complex-I activity and results in nigral DA neuron degeneration associated with protein aggregation (Chinta et al., 2007; Garrido et al., 2011). Dysfunction of excitatory amino acid transporters (EAATs) might link several of these PD pathogenic mechanisms. DA neurons express the neuronal transporter EAAC1 (EAAT3) at high levels (Plaitakis and Shashidharan, 2000; Shashidharan et al., 1997), and pathological conditions leading to excessive neuronal depolarization can affect or even reverse EAAT function by altering the transport driving force (Danbolt, 2001): for example, overactive glutamatergic inputs from the subthalamic nucleus (STN) to DA neurons (Rodriguez et al., 1998), or ATP depletion and subsequent impairment of the Na+/K+-ATPase resulting from mitochondrial dysfunction. Neuroinflammation can also affect glial EAAT function (McNaught and Jenner, 2000; Tilleux and Hermans, 2007). Besides clearing extracellular glutamate, EAATs provide substrates for GSH production since they uptake not only glutamate but also cyst(e)ine (Hayes et al., 2005). EAAT dysfunction might then, in turn, sustain excitotoxicity and oxidative stress. Accordingly, oxidative stress on EAAC1 has been involved in MPTPinduced GSH depletion (Aoyama et al., 2008), and EAAC1−/− mice show an age-dependent loss of nigral DA neurons that is prevented by N-acetylcysteine (Berman et al., 2011). Interestingly, the EAAT substrate inhibitor L-trans-pyrrolidine-2,4-dicarboxylate (PDC) triggers preferential DA neuron death in vitro through a mechanism involving oxidative stress and excitotoxicity, and can induce early loss of DA neurons when

70

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

injected in the SN (Nafia et al., 2008), whereas it does not induce neuronal damage in other brain regions (Massieu et al., 1995; Montiel et al., 2005). Here we show that a single unilateral PDC injection in the rat SN initiates a self-sustaining degenerative process selectively affecting DA versus non-DA neurons, which progresses up to 120 days after the injection. This provides a novel animal model exhibiting a unique combination of PD hallmarks: i) presumed cell death mode (GSH depletion, oxidative stress, NMDA receptor-mediated excitotoxicity, neuroinflammation); ii) evolution pattern of DA neuron loss from unilateral to bilateral and with a caudo-rostral gradient; iii) development of compensatory mechanisms; iv) appearance of motor deficits. Materials and methods Experiments were performed on male Wistar Hannover rats in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and all efforts were made to minimize their number and sufferance. Protocols were approved by the local ethical committee and conformed to the ethical guidelines of the French Ministry of Agriculture and Forests (Animal Health and Protection Veterinary Service). In vitro electrophysiology Brains from 4 weeks-old rats (n = 21) were cut in coronal slices (250 μm) by a vibratome in ice-cold solution containing (in mM): 110 choline, 2.5 KCl, 1.25 NaH2PO4, 7 MgCl2, 0.5 CaCl2, 25 NaHCO3, 7 glucose, pH 7.4, bubbled with a mix of 95% O2 and 5% CO2. Slices were kept in bubbled artificial cerebrospinal fluid (ACSF) at room temperature, composed of (in mM): 126 NaCl, 2.5 KCl, 1.2 MgCl2, 1.2 NaH2PO4, 2.4 CaCl2, 11 glucose, and 25 NaHCO3, pH7.4, added with 250 μM kynurenic acid and 1 mM sodium pyruvate. Electrophysiological experiments were done at 35 °C in standard ACSF (without kynurenic acid and sodium pyruvate) flowing at ~2.5 ml/min. Whole-cell patch-clamp recordings were performed by borosilicate micropipettes (4–5 MΩ) filled with a solution containing (in mM): 125 K-gluconate, 10 NaCl, 1 CaCl2, 2 MgCl2, 0.5 1,2-bis (2-aminophenoxy) ethane-N,N,N,N-tetraacetic acid (BAPTA), 19 N-(2-hydroxyethyl)-piperazine-N-s-ethanesulphonic acid (HEPES), 0.3 guanosine triphosphate (GTP), and 1 Mg-adenosine triphosphate (Mg-ATP), pH 7.3. Alexa Fluor 568 (20 μM) was added in the micropipette solution to localize and identify off-line the recorded

neurons. 35 DA and 14 non-DA neurons were identified on-line by infrared videomicroscopy (and confirmed off-line by the presence or not, respectively, of Alexa Fluor 568 fluorescence) and recorded by AxoPatch 200B or Multiclamp 700B amplifiers with pClamp10.2 software (Molecular Devices, USA). For slice application, drugs (TocrisCookson, UK) were dissolved at the desired concentration in the ACSF. Data were analyzed by pClamp and Prism (GraphPad, USA) software and are expressed as average ± SEM. Stereotaxic surgery Surgery was performed in 7–8 weeks old rats (180–200 g). After equithesin anesthesia (4 ml/kg), animals received unilateral (behavioral and morphological studies) or bilateral (biochemical measurements) stereotaxic injection of 300 nmol PDC (5 μl of a 60 mM PDC solution injected at the rate of 1 μl/min) or vehicle (0.9% NaCl; 5 μl at the rate of 1 μl/min) at a point between SN pars compacta (SNc) and pars reticulata (SNr) (coordinates in mm: AP + 2.2, L ±2.0, and DV +3.3, according to de Groot (De Groot, 1959); see Figs. 1A,B). Animals for microdialysis were stereotaxically implanted in the ipsilateral DA-lesioned side or the contralateral un-injected side with a cannula guide CMA/11 (Carnegie Medicine, Stockholm, Sweden), with the tip placed above the striatum (AP: +0.2 mm and L: +/−2.8 mm from bregma; DV: −3.3 mm from dura, according to the stereotaxic atlas of Paxinos and Watson, 1998). For neuroprotection study, PDC was either co-injected with memantine and/or ifenprodil (2 mM solution) or associated with N-acetylcysteine in drinking water (2 mg/ml started the day of PDC injection until 4 dpi). Animals were sacrificed at 4 days post-injection (dpi) for biochemical measurements, at 4, 15, 30, 60 or 120 dpi for histology, and at 90 dpi for in vivo microdialysis. Biochemical assays Anesthetized animals were transcardially perfused with 400 ml of cold PBS pH 7.4 containing 0.16 mg/ml heparin. Immediately after brain removal, striatum and SN were dissected out from 2 mm coronal slices performed using a rat brain matrix, and tissue samples were weighed and homogenized in cold 20 mM HEPES buffer (1/10, w/v), pH 7.2, containing 1 mM EDTA. Homogenates were centrifuged (1600 g for 10 min at 4 °C). Part of the supernatant was used to determine

Fig. 1. PDC injection site and procedure of section collection for histological studies. Photomicrograph of a toluidine blue-stained section (A) and scheme (B) (adapted from (Paxinos and Watson, 1998) showing the needle tract (white arrows on the photomicrograph) of PDC injection in the SN. Stereotaxic PDC injection was done between SNc and SNr (abbreviations: cp: cerebral peduncle; ml: medial lemniscus; mp: mammillary peduncle; VTA: ventral tegmental area; scale bar = 100 μm). (C) For each animal, 27 consecutive 50 μm-thick frontal sections (illustrated by black, dark gray and light gray bars) covering an anteroposterior extent of 1350 μm (~90% of SNc extent) of the injected and contralateral SN were carefully collected. SNc was subdivided in three equivalent 450 μm length sub-regions covering, respectively, the anterior, peri-injection and posterior parts. Sections of the same color on the drawing have been used as sample for a given labeling.

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

γ-glutamyltranspeptidase (γ-GT) activity, protein and thiobarbituric acid reactive substances (TBARS) contents. The remaining part was centrifuged (10,000 g for 15 min at 4 °C) and used to determine GSH content and activity of the other GSH-related enzymes. γ-GT activity was determined using the γ-GT reagent (Biolabo, Malzy, France) by following p-nitroaniline formation at 405 nm. Protein content was determined by the Bradford's method with the Coomassie Protein Assay Kit (Pierce, Rockford, IL, USA) and BSA as standard. Lipid peroxidation was measured as TBARS content by the method of Richard et al. (1992). Briefly, samples were incubated in thiobarbituric acid/perchloric acid. TBARS extraction was performed by butanol-1 and TBARS content was determined using a fluorescent plate reader (Fluoroskan Ascent FL; λex = 530 nm, λem = 590 nm). Malondialdehyde standards were included with each assay and TBARS contents were calculated as nmol of malondialdehyde/mg of proteins. Contents of total GSH and GSH oxidized form (GSSG) were measured using a GSH assay kit (Cayman Chemical, Spibio, Montigny, France). Samples were diluted 1/10 in MES buffer, deproteinated by adding an equal volume of 10% metaphosphoric acid (Sigma-Aldrich, St Louis, MO), centrifuged (2500 g, 2 min) and supernatant neutralized with triethanolamine (0.2 M). Part of the sample was treated with 2-vinylpyrydine (10 mM) to derivatize reduced GSH and assay only GSSG. Total GSH and GSSG contents were determined by measuring the production of 5-thio-2-nitrobenzoic acid at 405 nm after 25 min of incubation with the kit reagents. The activity of γ-glutamate cysteine ligase (GCL) was measured using the method of Seelig and Meister (1985). After sample addition to reaction buffer, enzyme activity was determined by measuring NADH consumption at 340 nm. Glutathione peroxidase (GPx) activity was determined using a GPx assay kit (Cayman Chemical, Spibio, Montigny, France) by measuring the consumption of NADPH at 340 nm. Glutathione S-transferases (GST) activity was determined using a GST assay kit (Cayman Chemical, Spibio, Montigny, France) measuring the production of 1-chloro-2,4-dinitrobenzene conjugates at 340 nm. Immunocytochemical procedures Animals were transcardially perfused with 400 ml 4% PFA in PBS 0.1 M, pH 7.3. Brains were dissected, left 24 h in the same fixative and then transferred in 30% sucrose (2 × 48 h). Serial frontal cryosections (50 μm) at the level of SNc and striatum were collected in PBS. For SN, they were distributed in three sets of adjacent sections covering the whole extent of the structure, each set being used for a different labeling (see Fig. 1C). After rinses, sections were immersed in 0.3% H2O2 (2 × 15 min), pre-incubated for 30 min in PBS containing 5% BSA, then incubated overnight with a specific primary antibody: mouse antityrosine hydroxylase (TH; 1:1000, Millipore), rabbit anti-phosphoTyr1472-NR2B (1:300, AbCys), mouse anti-glial fibrillary acidic protein (GFAP; 1:500, Sigma), rabbit anti-ionized calcium binding adaptor molecule 1 (Iba1; 1 μg/ml, Wako), goat anti-DA transporter (DAT; 1:100, Santa Cruz), and mouse anti-PHF-Tau (AT8, 5 μg/ml, Thermo Scientific). Sections were incubated 2 h with secondary antibodies (1:200) either fluorescent (Alexa Fluor goat anti-mouse 568 or donkey anti-rabbit 488; Invitrogen) or biotinylated (goat anti-rabbit, donkey anti-goat and goat anti-mouse; Jackson Immunoresearch). For immunoperoxidase detection we used avidin–biotin complex (2 h), and revelation with 0.005% 3,3′diaminobenzidine tetrahydrochloride and 0.001% H2O2. In situ hybridization for TH After decapitation, brains were quickly removed and frozen in dry ice. Frontal cryosections (14 μm) were performed across the SN, slidemounted, post-fixed for 5 min in 3% PFA and incubated in prehybridization buffer containing 2× standard saline citrate (SSC). Radioactive (35S-radiolabeled) a 3′-end-labeled synthetic oligonucleotide probe (44mer) selected on the basis of the published sequence of TH was used. Briefly, after pre-hybridization and dehydration, each section

71

was covered with 35 μl of hybridization buffer containing the probe (500,000 cpm per section) and incubated overnight at 47 °C in humid chamber. After post-hybridization, sections were coated with Amersham LM1 autoradiographic emulsion, exposed at 4 °C for 1–3 weeks, developed in Kodak D-19 (4 min at 13 °C) and counterstained with toluidine blue. Dual fluorescent GAD67 in situ hybridization and TH immunocytochemistry A 43mer synthetic oligonucleotide probe selected from the published GAD67 sequence was 3′-end-labeled by a dUTP queue coupled with Dig. After the hybridization steps, sections were incubated with anti-Dig antibody (1:500, Roche) and then with fluorescent tyramide solution. TH immunofluorescence detection was then performed as above-described. In vivo microdialysis and measurement of extracellular DA levels Microdialysis in the ipsilateral and contralateral striatum was performed in freely moving rats at 90 dpi. The microdialysis probe (CMA/11, Carnegie Medicine, Stockholm, Sweden; length of the microdialysis membrane, 3 mm; molecular weight cut-off, 20 kDa; outer diameter, 0.5 mm) was lowered through the guide cannula so that the tip of the dialyzing membrane reached DV −6.3 mm. Probes were perfused with Ringer's solution (in mM: NaCl, 147; CaCl2, 2.5; KCl, 4) at a constant flow rate (1 μl/min) using a CMA/102 microdialysis pump (Carnegie Medicine, Stockholm, Sweden). The dialysates were collected every 20 min in tubes containing 10 μM ascorbic acid and were immediately frozen at −80 °C until analysis. Sample collection started 120 min after the onset of perfusion to achieve stable levels of DA, and 4 samples were collected for 80 min as basal values before a 40 min perfusion (2 samples) with potassium enriched Ringer's solution (in mM: NaCl, 52.8; CaCl2, 2.5; KCl, 70). Return to basal values was determined by subsequent perfusion with regular Ringer's solution during 80 min (4 samples). DA was separated using HPLC and detected with the electrochemical system. The mobile phase consisted of 0.1 M sodium acetate, 0.17 mM octyl sulfate, 8% methanol, 0.7 mM EDTA, pH 4.5 (with addition of Acetic Acid 100%, 17.5 N) and was delivered through a LC-10ADvp Shimadzu pump (Kyoto, Japan) into a C-18 (ODS2, 5 μm, 4.6 × 150 mm) Spherisorb column (Waters, Milford, MA, USA). Samples of 20 μl were injected and analyzed using a Coulochem II, ESA detector (Chelmsford, MA, USA). The potential of the reference electrode was set at −50 mV and the working electrode at + 300 mV. The limit of detection was 20 fmol/sample. Extracellular DA concentration was estimated by rationing peak heights of DA to its external standard using the Empower software. The running time for each sample was 20 min. Quantification of neuronal cell loss The number of TH-IR cells within the SNc was estimated by a stereological approach using Mercator® system (Explora Nova, France) combined with an Optiphot® microscope (Nikon, France) coupled to a DXC-990P color video camera (Sony, France). A systematic random sampling of every third section centered on the track of PDC injection needle, was done from approximately −4.80 mm to −6.15 mm relative to bregma (Paxinos and Watson, 1998). For each animal the reference volume of the SNc was estimated using the Cavalieri method. The mean coefficient of error is 0.0106 with a maximum value reaching 0.0143. The optical fractionator method was used to determine the number of TH-IR cells in the SNc. Identification of the regions of interest was performed at 10× (nA = 0.30), and stereology was performed at 40 × (nA = 0.75). An unbiased counting frame was randomly positioned in the previously outlined area. For regional analysis, three 450 μm subregions (anterior, periinjection and posterior to the injection site) were considered (Fig. 1C). Effects of memantine, ifenprodil and N-acetylcysteine, were evaluated

72

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

at 4 dpi by quantifying TH+ neurons on a set of three sections covering the peri-injection region. For examining the specificity of DA neuron damage, the numbers of TH + neurons in SNc and GAD mRNAexpressing neurons in SNr were counted on 3 consecutive sections located around the injection site. Cell loss in SNc was verified at 4 and 60 dpi by counting cells on toluidine blue-labeled sections located in same region. Quantification of nigral astroglial and microglial immunostaining Quantifications were done on the set of nine equidistant sections adjacent to those used for kinetic analysis of neuronal loss. Optical density (OD) of GFAP and Iba1 fluorescent staining was measured in a region covering 0.13 mm2 in SNc with the same microscope and camera used above and the Mercator software (Explora Nova). High magnification pictures of microglial cells showing their activation after PDC injection were performed using a Zeiss Axioimager Z1 Apotome (63×). Quantification of intraneuronal TH mRNA levels in SNc For each animal, 3 consecutive emulsion-coated sections at SNc (peri-injection) level were analyzed. Sections were observed under dark-field epilumination (final magnification 1000×). Using the Visioscan image analysis software (BIOCOM), the number of silver grains per labeled cell (N 10 grains) was estimated under polarized light by measuring OD with respect to a standard curve and background subtracted. Quantification of striatal DAT immunostaining This was performed by digitized image analysis using “Densirag” analysis system (BIOCOM, France). The mean OD value was determined from four sections per animal distributed on the antero-posterior extent of the dorsal striatum after subtracting the background signal measured in a region lacking DA terminals (parietal cortex). Behavioral study Animals were placed in a Plexiglas® cylinder and video recorded during 15 min. The number of contacts made on the cylinder wall with the ipsilateral and the contralateral forepaw, either separately or together, were counted. Results are expressed as an asymmetry score, defined as the percent of contralateral contacts minus the percent of ipsilateral contacts. Hemiparkinsonian rats are expected to show a negative asymmetry score due to akinesia of the contralateral forepaw. The spontaneous motor activity was measured for a 15-min period in an open field made of clear Plexiglas® (45 × 45 cm) by measuring the distance traversed using a force-based actimetry technique (ActiVMeter, Bioseb, Vitrolles, France). Statistical analysis Statistical analysis was performed using one-way ANOVA followed by Newman–Keuls test for multiple comparison procedures for counts of GAD67-mRNA expressing neurons, TH-immunopositive or toluidine blue-stained neurons in the overall SN, for determination of intraneuronal levels of TH-mRNA in SNc and for measure of OD of DAT immunostaining. For regional analysis of cell loss, in vivo microdialysis and behavioral studies, statistical analysis was performed using twoway ANOVA followed by Bonferroni multiple comparison test. For NR2B and glial reactivities and for enzymatic activities, statistical analysis was performed using unpaired Student's t-test. Differences are considered significant for p b 0.05. Data are expressed as average values ± SEM.

Results DA and non-DA SN neurons show differential in vitro electrophysiological responses to PDC In order to study the acute effect of PDC on SN neurons and its possible action mechanisms, we first tested this drug in vitro on brain slices by means of whole-cell patch-clamp electrophysiology. DA neurons of the SNr were identified off-line by the co-localization of both Alexa Fluor 568 fluorescence (injected via the patch-clamp micropipette) and TH immunostaining (Fig. 2A), and on-line by their localization, size and membrane properties in response to hyperpolarizing and depolarizing current step (Fig. 2B) (Seutin and Engel, 2010). In control conditions, these cells showed typical action potential (AP) pacemaker activity (Fig. 2C) with an average frequency of 1.46 ± 0.25 Hz (n = 4). When PDC was bath-applied (300 μM, 3–6 min) they responded with increased AP discharge frequency reaching 2.39 ± 0.38 Hz (n = 4, p b 0.05 vs. control, Wilcoxon test), which decreased to control values (0.99 ± 0.37 Hz) after washout (Fig. 2C). At hyperpolarized sub-threshold membrane potential (obtained by negative current injection), PDC induced a sustained (up to 15 min) and reversible membrane depolarization and action potential discharge in 77% (23/30) of the recorded neurons, which was reversible after washout (Fig. 2D). Interestingly, this PDC-induced excitatory effect was strongly reduced by pre-application of either 10 μM ifenprodil or memantine (Fig. 2E; −67.6 ± 14% in 8/8 recorded neurons; data pooled together due to similar effect), respectively antagonist of NR2B- and NR2Dcontaining NMDA receptors (Kohl and Dannhardt, 2001). Altogether, these data suggest that PDC excites SNc DA neurons through a mechanism mediated by glutamate acting on NMDA receptors containing NR2B and NR2D subunits. Non-DA neurons of the SNr were identified off-line by the absence of TH immunostaining and the presence of Alexa Fluor 568 florescence (Fig. 2F). They showed typical spontaneous fast AP discharge activity (18 ± 4.27 Hz) and membrane properties in response to hyperpolarizing and depolarizing current step (Fig. 2G) (Seutin and Engel, 2010). At hyperpolarized sub-threshold membrane potential, 78% of them (7/9) responded to PDC application (up to 15 min) by depolarization and AP discharge. However, in contrast to DA neurons, such excitation was not maintained but showed a peculiar succession of short excitation and baseline recovery periods (Fig. 2H). Finally, PDC-induced excitation was completely blocked by the co-application of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and DL-2-Amino-5-phosphonopentanoic acid (AP-5) in both DA and non-DA neurons (not shown), suggesting that this effect is likely mediated by glutamate acting on AMPA and NMDA receptors. PDC injection triggers GSH depletion and specific increase of γ-GT activity As EAATs play a role in anti-oxidant defense by supplying substrates for GSH production, in a subsequent set of experiments we studied the consequences of acute PDC injection on several biochemical markers involved in GSH metabolism at 4 dpi. Total GSH levels (reduced GSH + GSSG) were significantly decreased (− 32% vs. sham) in the SN (Fig. 3A). Levels of GSSG were unchanged (Fig. 3B), suggesting selective reduction of GSH. Nigral levels of TBARS, a marker of lipid peroxidation, were significantly increased (+45% vs. sham), indicating the occurrence of oxidative stress (Fig. 3C). Activities of GCL, the enzyme responsible for the first step of GSH synthesis, and of the detoxifying enzymes GPx and GST, which use GSH in defense mechanisms against free radical-induced oxidative damage, were not modified (Fig. 3D). These data are consistent with the view that GSH availability is compromised due to low availability of substrates for its synthesis resulting from EAATs dysfunction, and not due to impaired biosynthesis capacity or increased utilization. As a possible compensatory mechanism, activity of γ-GT, the enzyme involved in GSH recycling, was significantly increased in the SN (+ 69%

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

73

Fig. 2. PDC induces prolonged excitation of DA (A–E) but intermittent activation of non-DA (F–H) neurons of the SN. In all panels except C, neurons are hyperpolarized to sub-threshold membrane potentials by negative current injection to prevent spontaneous AP firing. Voltage scale is the same in B–H. Action potentials are truncated except in B, C and G. Gray bars represent drug application. (A) Illustration of a recorded neuron identified off-line as a SNc DA neuron by co-localization of the fluorescence of Alexa Fluor 568 (red) injected through the recording micropipette and TH immunostaining (green) (scale bar 50 μm). (B) Typical membrane responses of a SNc DA neuron to hyperpolarizing (+100 pA) and depolarizing (−100 and −200 pA) current steps (baseline −50 mV): note the slow-frequency APs triggered by positive current injection, and the prominent sag (due to the hyperpolarizationactivated current Ih) followed by short rebound spikes due to the injection of negative current that disappear with smaller negative current injection. (C) Pacemaker firing activity in control condition is dramatically accelerated by the application of 300 μM PDC and recovers after washout. (D) At hyperpolarized sub-threshold membrane potential, PDC application induces membrane depolarization triggering AP discharge, which is reversible during washout (baseline −55 mV). (E) In the presence of 10 μM ifenprodil (ifen.) or 10 μM memantine (mem.), PDC-induced excitation is greatly reduced (baseline −55 mV and −65 mV, respectively). (F) SNr non-DA neuron morphologically identified by Alexa Fluor 568 fluorescence (red) and the absence of TH immunostaining (green) co-localization (scale bar 20 μm). (G) Typical membrane responses to hyperpolarizing (+50 pA) and depolarizing (−100 pA) current steps of a SNr non-DA neuron (baseline −55 mV): note the absence of the sag and the high-frequency AP discharge compared to (B). (H) PDC application results in transient and intermittent excitation of a non-DA neuron (baseline −55 mV): compare with the continuous effect observed in a DA neuron in (D).

vs. sham; Fig. 3D). On the other hand, no significant changes in any of these markers were measured in the striatum. Intranigral PDC injection induces selective death of SNc DA neurons The vulnerability to acute PDC injection of SNc DA neurons and SNr GABA neurons was comparatively examined at 4 dpi by dual fluorescent labeling, combining TH immunodetection with in situ hybridization of

GAD67 mRNA on peri-injection sections. At 4 dpi, the number of TH+ neurons and of toluidine-blue stained cells was significantly and similarly reduced (−26 ± 4.2% and − 25.3 ± 2.3% vs. sham, respectively; Figs. 4A–C), strongly suggesting that the loss of TH+ neurons actually reflects DA neuron loss and not TH down-regulation. Conversely, the number of GAD67 mRNA-expressing neurons in the SNr was not modified, showing that cell loss was restricted to DA SNc neurons (Figs. 4A–C).

74

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

Fig. 3. GSH level in the SN is reduced by PDC at 4 dpi, while TBARS and γ-GT concentration are increased. (A) Total GSH content is decreased in the SN of PDC-injected animals, while it is not significantly affected in the striatum (Str). (B) There is no increase in GSSG content after PDC injection in the SN that could account for GSH depletion. (C) Lipid peroxidation, measured as TBARS, is increased in the SN of PDC-injected animals, but not in the striatum. (D) There is no significant change in the activity of the GSH-producing (GCL) or GSH-consuming (GPx and GST) enzymes, but the activity of the GSH-recycling enzyme (γ-GT) is significantly increased (*p b 0.05 and **p b 0.01 vs. sham values using unpaired Student's t-test; n = 4–7 for each group).

PDC-induced death of SNc DA neurons involves NMDA receptor-mediated excitotoxicity and oxidative stress To test the possible role of NMDA receptor-mediated excitotoxicity in the action mechanism of PDC, suggested by electrophysiological experiments, we co-injected ifenprodil (2 mM) or memantine (2 mM) or both together with PDC in the SN. Interestingly, in these conditions no significant loss of TH + neurons vs. sham was measured at 4 dpi (Fig. 5A). Compared to PDC, significant neuroprotection was obtained by ifenprodil combined or not with memantine (Fig. 5A). Oral treatment of PDC-injected animals with the anti-oxidant N-acetylcysteine also resulted in no significant loss of TH+ neurons vs. sham at 4 dpi (Fig. 5A). The efficient neuroprotection provided by these compounds demonstrates the involvement of excitotoxicity (through NR2B- and NR2Dcontaining NDMA receptors) and oxidative stress in the PDC-triggered cell death mechanisms. The implication of NR2B-containing NMDA receptors was further supported by the increased immunolabelling of phosphorylated NR2B on Tyr1472 residue at 4 dpi in the injected SN (+214.1 ± 38.9% vs. contralateral side) (Figs. 5B,C).

Acute unilateral injection of PDC leads to a progressive and bilateral loss of DA neurons The time-course of the degenerative process triggered by PDC was examined on subgroups of PDC-injected and sham-injected animals sacrificed at 4, 15, 30, 60 or 120 dpi. The number of TH + cells was

stereologically quantified in the overall SNc for both the injected and contralateral un-injected side. Since no significant difference was measured among the different sham groups, or between ipsi- and contralateral SNc in sham animals, data were pooled in a single sham group that showed a mean of 11,135.51 ± 186.49 cells in the overall studied area. Regional analysis was further performed in the injected SNc by considering its anterior, peri-injection and posterior region. In the whole PDC-injected SNc (ipsilateral), a significant decrease in TH + cell numbers was measured as soon as 4 dpi (− 21.2 ± 4% vs. sham) (Figs. 6A,B). Afterwards, TH+ cell loss progressively increased overtime up to 120 dpi (− 31.2 ± 4.7%, − 39.3 ± 3.3%, − 46.7 ± 2.0%, − 57.4 ± 2.7% vs. sham at 15, 30, 60 and 120 dpi, respectively). In the contralateral un-injected SNc, a slight but significant loss in TH+ cells was measured at 60 dpi (−19.1 ± 5.4% vs. sham) and such loss further increased at 120 dpi (− 35.9 ± 34.6% vs. sham) (Figs. 6A, B). We verified at 60 dpi that cell loss measured on toluidine blue sections was close to the above reported values obtained on TH-stained sections, both in the injected (−46.1 ± 1.2% vs. sham) and uninjected SNc (−19.5 ± 2.1% vs. sham). No significant cell loss was observed in the VTA of the ipsilateral and contralateral side (not shown). Regional analysis of the cell loss kinetics in the PDC-injected SNc (ipsilateral) showed significant region (F(2, 78) = 50.31, p b 0.0001), time (F(5, 39) = 42.01, p b 0.0001) and interaction (F(10, 78) = 5.15, p b 0.0001) effect. Significant TH+ cell loss in the posterior and periinjection region was measured from 4 dpi in PDC-injected animals, whereas it was delayed and significantly less pronounced in the anterior region (Fig. 6C).

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

Fig. 4. PDC induces a selective loss of SNc DA but not SNr GABA neurons at 4 dpi. Representative photomicrographs of the SN on peri-injection sections of a sham (A) and a PDCinjected rat (B). Sections are processed for dual TH immunostaining (red) to label SNc DA neurons and fluorescent in situ hybridization labeling of GAD67 mRNA (green) to identify SNr GABA neurons. Note that PDC injection leads to a reduction of TH + neurons in the SNc without affecting GAD67-labeled neurons in the SNr (scale bar = 50 μm). (C) Quantitative analysis at 4 dpi showing the equivalent loss of TH+ neurons and toluidine blue-stained cells in the SNc, and the maintenance of GAD67 mRNA-expressing neurons in the SNr (**p b 0.01 vs. sham; n = 6 in each condition; one-way ANOVA followed by Newman–Keuls post-hoc test).

75

Fig. 5. PDC-induced death of DA neurons at 4 dpi involves oxidative stress and NMDA receptor-mediated excitotoxicity. (A) There is no more significant loss of TH+ neurons in the SNc when PDC is co-injected with 2 mM ifenprodil (if), 2 mM memantine (mem) or both, as well as when N-acetylcysteine (NAC; 2 mg/ml) is provided in drinking water. Moreover, compared to PDC values, the loss of TH+ neurons is significantly reversed by ifenprodil alone or combined with memantine (**p b 0.01 vs. sham; $p b 0.05 vs. PDC; n = 4–7 for each group; one-way ANOVA followed by Bonferroni's post-hoc test). (B, C) PDC leads to a 3-fold increase of Tyr1472-phosphorylated NR2B subunit immunostaining in the injected SN (identified by a star; scale bar = 200 μm) (***p b 0.001 vs. contralateral; n = 4; unpaired Student's t-test).

Adaptive changes in the nigrostriatal pathway consecutive to PDC injection The degenerative process is associated with differential time-course of astroglial and microglial reactivity Glial reactivity was analyzed in the whole injected SN up to 60 dpi, in the overall structure and in the anterior, peri-injection and posterior regions (as for TH staining). Astroglial reactivity, assessed by GFAP immunostaining, was delayed: low at 4 dpi (+ 14.8 ± 9.6% vs. sham), peaking at 15 dpi (+ 58.2 ± 12.4% vs. sham), then returning to control level at 60 dpi (Figs. 7A,B). Regional analysis shows significant GFAP reactivity in the posterior part as soon as 4 dpi, peaking at 15 dpi in all regions and remaining significant in the anterior region at 30 dpi (Fig. 7C). Microglial reactivity, assessed by Iba-1 immunostaining, was strong and maximal at 4 dpi in PDC-treated rats (+ 75.4 ± 17.0% vs. sham), then markedly decreased at 15 dpi (+40.9 ± 4.9% vs. sham) and normalized at 60 dpi (Figs. 7D,E). Regional analysis shows that Iba-1 reactivity was observed at 4 dpi in the posterior and peri-injection regions, and peaked at 15 dpi in the anterior region (Fig. 7F). Moreover, as soon as 4 dpi, microglial cells exhibited a macrophage-like form contrasting with the resting, ramified form observed in sham animals, indicating their activation (Fig. 7G).

The level of TH mRNA in spared SNc DA neurons (Figs. 8A,B), detected by in situ hybridization, was increased in the injected side at 30 and 60 dpi (+44.7 ± 9.7% and +48.1 ± 19.7% vs. sham, respectively). No significant changes were measured in the contralateral side. In the striatum ipsilateral to PDC injection, TH immunostaining showed a transient significant decrease vs. sham at 15 dpi (− 23.3 ± 4.8%, p b 0.05, one-way ANOVA followed by Newman–Keuls post-hoc test) and no change vs. sham at the other time points (not shown). Striatal DAT immunolabelling (Figs. 8C,D) showed an initial decrease between 4 and 15 dpi (− 21.5 ± 3.9% and − 33.8 ± 5.2% vs. sham, respectively) followed by a partial recovery at 30 dpi (− 19.0 ± 6.9% vs. sham) that reached control levels at 60 dpi. Interestingly, a secondary loss appeared at 120 dpi (− 26.1 ± 3.7% vs. sham). Such complex time-courses suggest that demise of the nigrostriatal system is followed by the occurrence of compensatory mechanisms that are in part overwhelmed after 60 dpi. Demise is further supported by the presence of dystrophic neurites revealed by AT8 immunolabelling (Fig. 8E), an index of tau hyperphosphorylation at all the time points. No significant changes were measured in the contralateral side for both TH (not shown) and DAT (Figs. 8C,D) immunostaining.

76

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

Consistent with altered striatal DA function, KCl-induced DA release was significantly decreased at 90 dpi both in the ipsilateral and, to a lesser extent, the contralateral side, in comparison with the sham group (group F(2, 12) = 7.39, p b 0.01; time F(9, 108) = 21.73, p b 0.0001; interaction F(18, 108) = 5.57, p b 0.0001). Indeed, KCl perfusion in sham animals induced a 28-fold increase of DA liberation

compared to basal, whereas in PDC-treated animals this increase was only 6-fold in the ipsilateral striatum and 11-fold in the contralateral striatum (Fig. 8F) as measured after in vivo microdialysis. Behavioral analysis The possible motor deficits in PDC-injected animals were assessed by the open-field test to measure locomotor activity and the cylinder test to evaluate the symmetry/asymmetry of forelimb use. In the open-field, ANOVA analysis showed significant group (F(1, 8) = 15.18, p b 0.01) and time (F(3, 24) = 20.19, p b 0.0001) effects. While locomotor activity decreased overtime in both groups, PDC-injected animals traveled significantly less distance vs. sham at 60 and 120 dpi (Fig. 9A). In the cylinder test, significant contralateral forelimb akinesia, as evidenced by a high asymmetry score, was measured in PDC-injected animals at 120 dpi (group effect F(1, 8) = 10.04, p b 0.05; −32.9 ± 7.6 PDC vs. − 1.07 ± 4.3 sham), with a nonsignificant tendency at 15, 30 and 60 dpi (Fig. 9B). Discussion Glutamate-mediated mechanisms have been implicated in PD pathophysiology and pathogenesis (Blandini et al., 1996; Carlsson and Carlsson, 1990; Johnson et al., 2009). EAAT function is essential for glutamate homeostasis and to provide metabolic substrates for antioxidant defenses. EAAC1−/− mice have been reported to show age-dependent neurodegeneration involving oxidative stress, including loss of DA neurons, brain atrophy, and cognitive and motivational behavioral impairments but little motor deficits at 12 months of age (Aoyama et al., 2006; Berman et al., 2011). Here we show that acutely-induced dysfunction of nigral EAATs induces selective death of DA versus non-DA neurons, involving oxidative stress and NMDA-mediated excitotoxicity. The neurodegeneration progresses with a uni-to bilateral evolution and caudo-rostral pattern and leads to motor deficit. Therefore, EAATs might be implicated in PD even though their function/expression is not altered chronically. Their dysfunction provides a novel model of progressive Parkinsonism that involves at the onset several pathogenic mechanisms characterizing PD, and recapitulates main pathological features of this disease. Mechanisms of PDC-induced neurodegeneration: role of NMDA receptormediated excitotoxicity EAATs play a major role in maintaining non-toxic extracellular glutamate concentrations (Danbolt, 2001). SN neurons express NR2B and NR2D subunits of NMDA receptor (Brothwell et al., 2008; Jones and Gibb, 2005). NMDA receptors containing these subunits have been involved in excitotoxicity (Baron et al., 2010; Liu et al., 2007) due to their preferential extrasynaptic location (Hardingham and Bading,

Fig. 6. Unilateral PDC injection in the SN induces a progressive neurodegeneration with a postero-anterior gradient, evolving bilaterally at late stage. (A) Representative photomicrographs of frontal SN sections showing TH immunostaining in sham and PDC-treated animals. A progressive decrease in the number of TH-immunopositive cells is observed in the ipsilateral SN; note that the contralateral SNc is also affected at 120 dpi (scale bar = 200 μm). (B) Stereological analysis of TH-immunostained neurons in sham and PDC-injected animals (n = 3–4 and n = 5–7 for each time point, respectively). Note the early and progressive cell loss in the injected (ipsilateral) side, compared to the delayed and less pronounced, evolving loss in the un-injected SN (contralateral) side (**p b 0.01 and ***p b 0.001 vs. sham; $p b 0.05 and $$$p b 0.001 vs. 4 dpi; £p b 0.05, ££p b 0.01 and £££ p b 0.001 vs. 15 dpi; @p b 0.05 and @@@p b 0.001 vs. 30 dpi; €p b 0.05 vs. 60 dpi; oneway ANOVA followed by Newman–Keuls post-hoc test). (C) Regional analysis of TH + cell loss in the anterior, peri-injected and posterior part of the ipsilateral SN. Cell loss in the peri-injected and posterior part follows a similar time-course and is significant since 4 dpi, whereas in the anterior part it is delayed and overall significantly less marked (§p b 0.05 and §§§p b 0.001 vs. peri-injection part, two-way ANOVA followed by Bonferroni post-hoc test; ***p b 0.001 vs. sham, one-way ANOVA followed by Newman–Keuls post-hoc test).

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

2010) and/or structural specificities (Martel et al., 2012). Glutamate spillover due to EAAT dysfunction might recruit such receptors and contribute to the PDC-induced death of DA neurons. Consistently, PDC increased NR2B phosphorylation at Tyr1472, a key factor for determining NMDA receptor trafficking and localization (Prybylowski et al., 2005), and efficient neuroprotection was provided by ifenprodil or memantine, which are considered as promising neuroprotective agents in PD (Yacoubian and Standaert, 2009). Our previous observation of intranigral α-synuclein-positive intracytoplasmic deposits after PDC injection (Nafia et al., 2008) fits with the association of altered glutamate neurotransmission with α-synucleinopathies (Price et al., 2010). Interestingly, we found that non-DA nigral neurons are not vulnerable to PDC-triggered cell death. Previous electrophysiological studies showed that DA and non-DA SN neurons express similar NMDA receptor subunit composition but have different NMDA-triggered responses (DA neurons show less receptor desensitization) presumably due to differences in receptor modulation and/or trafficking (Suarez et al., 2010). Similarly, we show that prolonged PDC application on slices triggers differential electrophysiological responses in DA vs. non-DA neurons (continuous vs. intermittent excitation), although both responses are glutamate-mediated. Differential properties of NMDA receptors might thus contribute to the preferential vulnerability of DA neurons in our model. Other mechanisms might also be involved, including sensitization to excitotoxicity by oxidative stress due to the differential Ca2+ handling of DA neurons linked to their pacemaking activity (Surmeier et al., 2011).

77

Mechanisms of PDC-induced neurodegeneration: role of oxidative stress Another major role of EAATs is to sustain anti-oxidant defense by supplying substrates, glutamate and cyst(e)ine, for GSH production (Had-Aissouni, 2012). Consistently, PDC-induced neurodegeneration is prevented by the anti-oxidant N-acetylcysteine. Since neuroprotection is provided by NMDA receptor antagonists or N-acetylcysteine, excitotoxicity and oxidative stress may be part of a common mechanism underlying the selective loss of DA neurons. Our model reproduces the GSH-related biochemical changes (Schulz et al., 2000; Sian et al., 1994b) and lipid peroxidation (Dexter et al., 1989) reported in PD. The reason for GSH depletion in PD is still unclear: it is not attributable to increased use for detoxification of peroxide-type oxidants, since GSSG content and GPx or GST activity are not increased. Failure in GSH synthesis has been excluded because the activity of GCL is unaffected. Thus, GSH depletion seems rather a cause than a consequence of oxidative stress. Based on the homology with our data, GSH depletion may be due to decreased availability of the substrates glutamate/cyste(i)ne. This may also account for the γ-GT upregulation in PD as in our model, which could represent a compensatory mechanism by which astrocytes provide dipeptide precursors for neurons to generate more GSH (Schulz et al., 2000). GSH depletion and oxidative stress can in turn affect glutamate homeostasis, in particular by impairing EAAT function either indirectly, through energy failure resulting from mitochondrial dysfunction, or directly. Glutamate uptake is inhibited by oxygen radicals, hydrogen peroxide and peroxynitrite (Trotti et al., 1996). Expression of peroxiredoxins is

Fig. 7. Differential time-course of astroglial (A–C) and microglial (D–G) reactivity after PDC injection. (A) Frontal sections showing GFAP immunostaining (red) of astroglia at 4 dpi in sham, and at 4 and 15 dpi in PDC-injected SN (white arrows show the needle track; scale bar = 200 μm). (B) Quantitative GFAP immunostaining (expressed as optical density, OD) in the whole SN is significantly increased at 15 and 30 dpi and returns at sham levels at 60 dpi. (C) Regional analysis shows a significant GFAP increase in the posterior part of the SN at 4 dpi, whereas it is significant in the anterior and peri-injected parts at 15 and 30 dpi (*p b 0.05 vs. sham; unpaired Student's t-test; n = 3–6 for the PDC group; n = 2–4 for the sham group). (D) Frontal sections showing Iba-1 immunostaining (green) of microglia at 4 dpi in sham, and at 4 and 15 dpi in PDC-injected SN (white arrows show the needle track; scale bar = 200 μm). (E) Quantitative Iba-1 immunostaining (OD) in the whole SN is significant and maximal at 4 dpi, then it decreases progressively and recovers to sham levels at 60 dpi. (F) Regional analysis shows a significant Iba-1 immunostaining increase in the peri-injected and posterior parts of the SN at 4 dpi followed by a normalization from 15 dpi onward, while the anterior part only shows one significant peak at 15 dpi (*p b 0.05; **p b 0.01 vs. sham; unpaired Student's t-test; n = 3–6 for the PDC group; n = 2–4 for the sham group). (G) High magnification pictures illustrating Iba-1 immunostained cells of the SN at 4 dpi in a sham (left) and a PDC-injected animal (right), showing the morphological modification of microglial cells. Note the typical highly ramified branches of the unactivated microglia in sham whereas the activated microglia after PDC injection has an enlarged cell body with shorter, stouter branches.

78

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

Fig. 7 (continued).

low in DA SNc neurons, which may contribute to their preferential vulnerability to nitro-oxidative attacks in PD (Goemaere and Knoops, 2012). Therefore, acute EAAT impairment may start a chain of pathological events sustaining their dysfunction overtime and contribute to the vicious cycle that keeps the degenerative process going. PDC being a substrate inhibitor of neuronal and glial EAATs, the question of which EAAT mediates PDC-triggered neurodegeneration arises. Although the contribution of glial EAATs cannot be excluded, a role of EAAC1 can be suggested since SNc DA neurons in EAAC1−/− mice undergo age-related neurodegeneration associated with reduced GSH levels and increased oxidative stress (Berman et al., 2011). Evolution of the degenerative process, adaptive mechanisms and functional deficits The most outstanding peculiarity of our model for its relevance to PD is the progression of neurodegeneration, especially the unilateralto-bilateral progression that was also reported recently in a genetic model (Rousseaux et al., 2012). Such progression indicates that PDC injection triggers a cascade of deleterious events involving different and complex mechanisms that remain to be identified. DA neuron loss in the injected SN shows a kinetic consistent with the exponential decay pattern of DA neurons in PD (Biju and De la FuenteFernandez, 2009; Fearnley and Lees, 1991). Its spatiotemporal pattern is

reminiscent of that occurring in PD (Damier et al., 1999) with the posterior and median (peri-injection) parts of the SN being affected earlier and more severely than the anterior part. Dystrophic neurites are observed in the ipsilateral striatum at all time-points, indicating continuous damage of striatal afferents. Adaptive mechanisms occur, as shown by the transient increase in TH mRNA expression in spared SNc neurons and the transient recovery of striatal DAT immunolabelling. These adaptive mechanisms are overwhelmed by cell death progression over a critical threshold, as indicated by the late decrease in both DAT immunolabelling and KCl-evoked DA release in the striatum. Glial reactions in PD are suggested to contribute to the degenerative process, especially via neuroinflammation (McGeer and McGeer, 2008). In our model, inflammatory processes involving microglia might contribute to the initial degeneration wave, as microglial reaction is maximal since the start or before significant cell loss is measured, then rapidly declines. Conversely, astroglial reaction always peaks after significant neurodegeneration has occurred. Given the dual neuroprotective and neurodegenerative role of astrocytes in PD (Rappold and Tieu, 2010; Vila et al., 2001), astrocyte reactivity could either contribute here to slow-down the primary degeneration or to sustain the secondary one. Contralateral DA neuron loss is observed lately after PDC injection. This excludes an effect due to PDC diffusion and further supports the hypothesis of a chain reaction triggered by PDC. Consistent with damage to the nigrostriatal system of the non-injected hemisphere,

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

79

Fig. 8. PDC intranigral injection induces progressive and adaptive changes in the SN and striatum. (A) PDC injection induces a transient increase in TH mRNA expression at 30 and 60 dpi, without any significant modification in the contralateral SN (*p b 0.05, **p b 0.01 vs. sham; one-way ANOVA followed by Newman–Keuls post-hoc test; for each time point, n = 6–10 and n = 3–4 for the PDC and sham group, respectively). (B) Representative photomicrographs showing the higher silver grain staining in the PDC-injected SN at 60 dpi compared to sham. (C) Quantitative DAT immunolabelling (OD) shows two waves of significant decreases in the side ipsilateral to PDC-injected SN compared to sham animals. The first one (4–30 dpi) peaking at 15 dpi, and the second one measured at 120 dpi following transient normalization at 60 dpi. No significant change is measured in the contralateral side at any time point (**p b 0.01, ***p b 0.001 vs. sham; one-way ANOVA followed by Newman–Keuls post-hoc test; for each time point, n = 5–7 and n = 3–4 for the PDC and sham group, respectively). (D) Representative photomicrograph showing the decrease of DAT immunostaining in the ipsilateral side (*) at 15 dpi. (E) Immunohistochemistry for AT8 reveals dystrophic neurites in the striatum (scale bar = 10 μm). (F) Extracellular DA concentration after KCl perfusion in ipsilateral (n = 6) or contralateral (n = 4) striatum, 90 dpi of PDC or vehicle. Note that KCl-induced DA liberation is severely reduced in the ipsilateral striatum, and to a lesser extent in the contralateral striatum, in comparison to sham (n = 5) animals (***p b 0.001 vs. sham; two-way ANOVA followed by Bonferroni post-hoc test).

evoked striatal DA release is significantly decreased in the same brain side. However, DAT immunostaining remains unaffected, suggesting that compensatory mechanisms are still ongoing in this hemisphere. The cellular substrates of such astonishing unilateral to bilateral

neurodegeneration need further investigation. Possibly, afferent pathways involved in the bilateral control of nigrostriatal systems (Cheramy et al., 1983, 1984; Glowinski et al., 1978; Nieoullon et al., 1977) or crossed nigrostriatal connections (Lieu and Subramanian,

80

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81

Conclusion None of the available neurotoxin-based or genetic rodent models of PD recapitulates all the hallmarks of this pathology, thus capturing as many of these hallmarks in one single model remains a major challenge (Chesselet et al., 2012; Dawson et al., 2010; Jackson-Lewis et al., 2012; Magen and Chesselet, 2010; Martinez and Greenamyre, 2012). Here we generated a novel rodent model of progressive Parkinsonism of easy implementation, based on the manipulation of endogenous glutamate-mediated mechanisms and not on DA toxins. This progressive model, which presents unique PD-relevant features, may thus provide novel insight onto the pathophysiology of early and late PD stages and may be useful for preclinical evaluation of therapeutic strategies. Finally, it also raises the poorly investigated question of the role of EAATs in PD.

Acknowledgments This work was supported by the Centre National de la Recherche Scientifique (CNRS), Aix-Marseille University, and a France Parkinson grant (to L.H.-A.). M.A. was supported by a fellowship from the French Ministry of Education and Research. The authors are grateful to Dr. Annie Daszuta for her advice and expertise in immunocytochemistry. This work was performed using France-BioImaging infrastructure supported by the Agence Nationale de la Recherche (ANR-10-INSB-0401, call “Investissements d'Avenir”). The authors declare no competing financial interests.

References

Fig. 9. Motor deficits after PDC treatment. (A) Open-field test: PDC-injected animals travel less distance compared to sham from 60 dpi. (B) Cylinder test: PDC-injected rats present a forelimb-use asymmetry (akinesia of the contralateral forelimb) that reaches significance at 120 dpi. (*p b 0.05 and **p b 0.01 vs. sham values; n = 5 in each group; two-way ANOVA followed by Bonferroni post-hoc test).

2012) could be implicated. A working hypothesis is that interhemispheric imbalance produced by unilateral PDC leads to an overactivity of glutamate projections to the contralateral SNc resulting in the long term in damage to the DA neurons. Significant motor deficits, i.e. hypolocomotion and forelimb akinesia, are detected from 60 and 120 dpi, respectively. It is noteworthy that the onset of forelimb akinesia is unilateral, whereas, at 120 dpi, the damage to the nigrostriatal system is bilateral although asymmetrical. This is reminiscent of the characteristic asymmetrical presentation of PD symptoms at early stages. It has been previously suggested from comparisons of asymmetrical versus symmetrical damage to the nigrostriatal DA system in rat and monkey that the interdependent regulation of the two nigrostriatal systems may provide some compensatory support for the function and behavioral performance of the lesioned side via the normal unlesioned side (Blesa et al., 2011; Roedter et al., 2001). In other words, lower DA neuron loss will lead to symptoms in bilateral versus unilateral lesion conditions. Such hypothesis might explain why significant motor deficits are detected in our model when only 50–60% of DA neurons are lost in the most affected side but the degeneration has reached the contralateral SNc, whereas it is known that more than 80% loss of DA neurons is required to trigger such deficits in strictly unilateral lesion model.

Ahmed, I., Liang, Y., Schools, S., Dawson, V.L., Dawson, T.M., Savitt, J.M., 2012. Development and characterization of a new Parkinson's disease model resulting from impaired autophagy. J. Neurosci. 32, 16503–16509. Aoyama, K., Suh, S.W., Hamby, A.M., Liu, J., Chan, W.Y., Chen, Y., Swanson, R.A., 2006. Neuronal glutathione deficiency and age-dependent neurodegeneration in the EAAC1 deficient mouse. Nat. Neurosci. 9, 119–126. Aoyama, K., Matsumura, N., Watabe, M., Nakaki, T., 2008. Oxidative stress on EAAC1 is involved in MPTP-induced glutathione depletion and motor dysfunction. Eur. J. Neurosci. 27, 20–30. Baron, A., Montagne, A., Casse, F., Launay, S., Maubert, E., Ali, C., Vivien, D., 2010. NR2Dcontaining NMDA receptors mediate tissue plasminogen activator-promoted neuronal excitotoxicity. Cell Death Differ. 17, 860–871. Berman, A.E., Chan, W.Y., Brennan, A.M., Reyes, R.C., Adler, B.L., Suh, S.W., Kauppinen, T.M., Edling, Y., Swanson, R.A., 2011. N-acetylcysteine prevents loss of dopaminergic neurons in the EAAC1−/− mouse. Ann. Neurol. 69, 509–520. Biju, G., De la Fuente-Fernandez, R., 2009. Dopaminergic function and progression of Parkinson's disease: PET findings. Parkinsonism Relat. Disord. 15 (Suppl. 4), S38–S40. Blandini, F., Porter, R.H., Greenamyre, J.T., 1996. Glutamate and Parkinson's disease. Mol. Neurobiol. 12, 73–94. Blesa, J., Juri, C., Garcia-Cabezas, M.A., Adanez, R., Sanchez-Gonzalez, M.A., Cavada, C., Obeso, J.A., 2011. Inter-hemispheric asymmetry of nigrostriatal dopaminergic lesion: a possible compensatory mechanism in Parkinson's disease. Front. Syst. Neurosci. 5, 92-. Brothwell, S.L., Barber, J.L., Monaghan, D.T., Jane, D.E., Gibb, A.J., Jones, S., 2008. NR2B- and NR2D-containing synaptic NMDA receptors in developing rat substantia nigra pars compacta dopaminergic neurones. J. Physiol. 586, 739–750. Carlsson, M., Carlsson, A., 1990. Interactions between glutamatergic and monoaminergic systems within the basal ganglia—implications for schizophrenia and Parkinson's disease. Trends Neurosci. 13, 272–276. Cheramy, A., Chesselet, M.F., Romo, R., Leviel, V., Glowinski, J., 1983. Effects of unilateral electrical stimulation of various thalamic nuclei on the release of dopamine from dendrites and nerve terminals of neurons of the two nigrostriatal dopaminergic pathways. Neuroscience 8, 767–780. Cheramy, A., Romo, R., Godeheu, G., Glowinski, J., 1984. Effects of electrical stimulation of various midline thalamic nuclei on the bilateral release of dopamine from dendrites and nerve terminals of neurons in the nigro-striatal dopaminergic pathways. Neurosci. Lett. 44, 193–198. Chesselet, M.F., Richter, F., Zhu, C., Magen, I., Watson, M.B., Subramaniam, S.R., 2012. A progressive mouse model of Parkinson's disease: the Thy1-aSyn (“Line 61”) mice. Neurotherapeutics 9, 297–314. Cheung, Z.H., Ip, N.Y., 2009. The emerging role of autophagy in Parkinson's disease. Mol. Brain 2, 29. Chinta, S.J., Kumar, M.J., Hsu, M., Rajagopalan, S., Kaur, D., Rane, A., Nicholls, D.G., Choi, J., Andersen, J.K., 2007. Inducible alterations of glutathione levels in adult dopaminergic midbrain neurons result in nigrostriatal degeneration. J. Neurosci. 27, 13997–14006.

M. Assous et al. / Neurobiology of Disease 65 (2014) 69–81 Damier, P., Hirsch, E.C., Agid, Y., Graybiel, A.M., 1999. The substantia nigra of the human brain. II. Patterns of loss of dopamine-containing neurons in Parkinson's disease. Brain 122 (Pt 8), 1437–1448. Danbolt, N.C., 2001. Glutamate uptake. Prog. Neurobiol. 65, 1–105. Dawson, T.M., Ko, H.S., Dawson, V.L., 2010. Genetic animal models of Parkinson's disease. Neuron 66, 646–661. De Groot, J., 1959. The Rat Forebrain in Stereotaxic Coordinates. Dexter, D.T., Carter, C.J., Wells, F.R., Javoy-Agid, F., Agid, Y., Lees, A., Jenner, P., Marsden, C.D., 1989. Basal lipid peroxidation in substantia nigra is increased in Parkinson's disease. J. Neurochem. 52, 381–389. Fearnley, J.M., Lees, A.J., 1991. Ageing and Parkinson's disease: substantia nigra regional selectivity. Brain 114 (Pt 5), 2283–2301. Garrido, M., Tereshchenko, Y., Zhevtsova, Z., Taschenberger, G., Bahr, M., Kugler, S., 2011. Glutathione depletion and overproduction both initiate degeneration of nigral dopaminergic neurons. Acta Neuropathol. 121, 475–485. Glowinski, J., Niecoullon, A., Cheramy, A., 1978. Regulations of the activity of the nigrostriatal dopaminergic pathways by cortical, cerebellar, and sensory neuronal afferences. Adv. Biochem. Psychopharmacol. 19, 75–87. Goemaere, J., Knoops, B., 2012. Peroxiredoxin distribution in the mouse brain with emphasis on neuronal populations affected in neurodegenerative disorders. J. Comp. Neurol. 520, 258–280. Had-Aissouni, L., 2012. Toward a new role for plasma membrane sodium-dependent glutamate transporters of astrocytes: maintenance of antioxidant defenses beyond extracellular glutamate clearance. Amino Acids 42, 181–197. Hardingham, G.E., Bading, H., 2010. Synaptic versus extrasynaptic NMDA receptor signalling: implications for neurodegenerative disorders. Nat. Rev. Neurosci. 11, 682–696. Hayes, D., Wiessner, M., Rauen, T., McBean, G.J., 2005. Transport of L-[14C]cystine and L-[14C]cysteine by subtypes of high affinity glutamate transporters over-expressed in HEK cells. Neurochem. Int. 46, 585–594. Jackson-Lewis, V., Blesa, J., Przedborski, S., 2012. Animal models of Parkinson's disease. Parkinsonism Relat. Disord. 18 (Suppl. 1), S183–S185. Jenner, P., Dexter, D.T., Sian, J., Schapira, A.H., Marsden, C.D., 1992. Oxidative stress as a cause of nigral cell death in Parkinson's disease and incidental Lewy body disease. The Royal Kings and Queens Parkinson's Disease Research Group. Ann. Neurol. 32, S82–S87 (Suppl.). Johnson, K.A., Conn, P.J., Niswender, C.M., 2009. Glutamate receptors as therapeutic targets for Parkinson's disease. CNS. Neurol. Disord. Drug Targets. 8, 475–491. Jones, S., Gibb, A.J., 2005. Functional NR2B- and NR2D-containing NMDA receptor channels in rat substantia nigra dopaminergic neurones. J. Physiol. 569, 209–221. Kohl, B.K., Dannhardt, G., 2001. The NMDA receptor complex: a promising target for novel antiepileptic strategies. Curr. Med. Chem. 8, 1275–1289. Lieu, C.A., Subramanian, T., 2012. The interhemispheric connections of the striatum: implications for Parkinson's disease and drug-induced dyskinesias. Brain Res. Bull. 87, 1–9. Liu, Y., Wong, T.P., Aarts, M., Rooyakkers, A., Liu, L., Lai, T.W., Wu, D.C., Lu, J., Tymianski, M., Craig, A.M., Wang, Y.T., 2007. NMDA receptor subunits have differential roles in mediating excitotoxic neuronal death both in vitro and in vivo. J. Neurosci. 27, 2846–2857. Magen, I., Chesselet, M.F., 2010. Genetic mouse models of Parkinson's disease. The state of the art. Prog. Brain Res. 184, 53–87. Martel, M.A., Ryan, T.J., Bell, K.F., Fowler, J.H., McMahon, A., Al-Mubarak, B., Komiyama, N.H., Horsburgh, K., Kind, P.C., Grant, S.G., Wyllie, D.J., Hardingham, G.E., 2012. The subtype of GluN2 C-terminal domain determines the response to excitotoxic insults. Neuron 74, 543–556. Martinez, T.N., Greenamyre, J.T., 2012. Toxin models of mitochondrial dysfunction in Parkinson's disease. Antioxid. Redox Signal. 16, 920–934. Massieu, L., Morales-Villagran, A., Tapia, R., 1995. Accumulation of extracellular glutamate by inhibition of its uptake is not sufficient for inducing neuronal damage: an in vivo microdialysis study. J. Neurochem. 64, 2262–2272. McGeer, P.L., McGeer, E.G., 2008. Glial reactions in Parkinson's disease. Mov. Disord. 23, 474–483. McNaught, K.S., Jenner, P., 2000. Extracellular accumulation of nitric oxide, hydrogen peroxide, and glutamate in astrocytic cultures following glutathione depletion, complex I inhibition, and/or lipopolysaccharide-induced activation. Biochem. Pharmacol. 60, 979–988. Montiel, T., Camacho, A., Estrada-Sanchez, A.M., Massieu, L., 2005. Differential effects of the substrate inhibitor L-trans-pyrrolidine-2,4-dicarboxylate (PDC) and the nonsubstrate inhibitor DL-threo-beta-benzyloxyaspartate (DL-TBOA) of glutamate transporters on neuronal damage and extracellular amino acid levels in rat brain in vivo. Neuroscience 133, 667–678. Nafia, I., Re, D.B., Masmejean, F., Melon, C., Kachidian, P., Kerkerian-Le, G.L., Nieoullon, A., Had-Aissouni, L., 2008. Preferential vulnerability of mesencephalic dopamine neurons to glutamate transporter dysfunction. J. Neurochem. 105, 484–496.

81

Nieoullon, A., Cheramy, A., Glowinski, J., 1977. Interdependence of the nigrostriatal dopaminergic systems on the two sides of the brain in the cat. Science 198, 416–418. Olanow, C.W., 2007. The pathogenesis of cell death in Parkinson's disease—2007. Mov. Disord. 22 (Suppl. 17), S335–S342. Paxinos, G., Watson, C., 1998. The Rat Brain in Stereotaxic Coordinates. Pearce, R.K., Owen, A., Daniel, S., Jenner, P., Marsden, C.D., 1997. Alterations in the distribution of glutathione in the substantia nigra in Parkinson's disease. J. Neural Transm. 104, 661–677. Perry, T.L., Godin, D.V., Hansen, S., 1982. Parkinson's disease: a disorder due to nigral glutathione deficiency? Neurosci. Lett. 33, 305–310. Plaitakis, A., Shashidharan, P., 2000. Glutamate transport and metabolism in dopaminergic neurons of substantia nigra: implications for the pathogenesis of Parkinson's disease. J. Neurol. 247 (Suppl. 2), II25-II35. Price, D.L., Rockenstein, E., Ubhi, K., Phung, V., Lean-Lewis, N., Askay, D., Cartier, A., Spencer, B., Patrick, C., Desplats, P., Ellisman, M.H., Masliah, E., 2010. Alterations in mGluR5 expression and signaling in Lewy body disease and in transgenic models of alpha-synucleinopathy—implications for excitotoxicity. PLoS One 5, e14020-. Prybylowski, K., Chang, K., Sans, N., Kan, L., Vicini, S., Wenthold, R.J., 2005. The synaptic localization of NR2B-containing NMDA receptors is controlled by interactions with PDZ proteins and AP-2. Neuron 47, 845–857. Rappold, P.M., Tieu, K., 2010. Astrocytes and therapeutics for Parkinson's disease. Neurotherapeutics 7, 413–423. Richard, M.J., Portal, B., Meo, J., Coudray, C., Hadjian, A., Favier, A., 1992. Malondialdehyde kit evaluated for determining plasma and lipoprotein fractions that react with thiobarbituric acid. Clin. Chem. 38, 704–709. Rodriguez, M.C., Obeso, J.A., Olanow, C.W., 1998. Subthalamic nucleus-mediated excitotoxicity in Parkinson's disease: a target for neuroprotection. Ann. Neurol. 44, S175–S188. Roedter, A., Winkler, C., Samii, M., Walter, G.F., Brandis, A., Nikkhah, G., 2001. Comparison of unilateral and bilateral intrastriatal 6-hydroxydopamine-induced axon terminal lesions: evidence for interhemispheric functional coupling of the two nigrostriatal pathways. J. Comp. Neurol. 432, 217–229. Rousseaux, M.W., Marcogliese, P.C., Qu, D., Hewitt, S.J., Seang, S., Kim, R.H., Slack, R.S., Schlossmacher, M.G., Lagace, D.C., Mak, T.W., Park, D.S., 2012. Progressive dopaminergic cell loss with unilateral-to-bilateral progression in a genetic model of Parkinson disease. Proc Natl Acad Sci USA 109, 15918–15923. Schapira, A.H., Jenner, P., 2011. Etiology and pathogenesis of Parkinson's disease. Mov. Disord. 26, 1049–1055. Schulz, J.B., Lindenau, J., Seyfried, J., Dichgans, J., 2000. Glutathione, oxidative stress and neurodegeneration. Eur. J. Biochem. 267, 4904–4911. Seelig, G.F., Meister, A., 1985. Glutathione biosynthesis; gamma-glutamylcysteine synthetase from rat kidney. Methods Enzymol. 113, 379–390. Seutin, V., Engel, D., 2010. Differences in Na+ conductance density and Na+ channel functional properties between dopamine and GABA neurons of the rat substantia nigra. J. Neurophysiol. 103, 3099–3114. Shashidharan, P., Huntley, G.W., Murray, J.M., Buku, A., Moran, T., Walsh, M.J., Morrison, J.H., Plaitakis, A., 1997. Immunohistochemical localization of the neuron-specific glutamate transporter EAAC1 (EAAT3) in rat brain and spinal cord revealed by a novel monoclonal antibody. Brain Res. 773, 139–148. Sian, J., Dexter, D.T., Lees, A.J., Daniel, S., Agid, Y., Javoy-Agid, F., Jenner, P., Marsden, C.D., 1994a. Alterations in glutathione levels in Parkinson's disease and other neurodegenerative disorders affecting basal ganglia. Ann. Neurol. 36, 348–355. Sian, J., Dexter, D.T., Lees, A.J., Daniel, S., Jenner, P., Marsden, C.D., 1994b. Glutathionerelated enzymes in brain in Parkinson's disease. Ann. Neurol. 36, 356–361. Suarez, F., Zhao, Q., Monaghan, D.T., Jane, D.E., Jones, S., Gibb, A.J., 2010. Functional heterogeneity of NMDA receptors in rat substantia nigra pars compacta and reticulata neurones. Eur. J. Neurosci. 32, 359–367. Surmeier, D.J., Guzman, J.N., Sanchez-Padilla, J., Schumacker, P.T., 2011. The role of calcium and mitochondrial oxidant stress in the loss of substantia nigra pars compacta dopaminergic neurons in Parkinson's disease. Neuroscience 198, 221–231. Tilleux, S., Hermans, E., 2007. Neuroinflammation and regulation of glial glutamate uptake in neurological disorders. J. Neurosci. Res. 85, 2059–2070. Trotti, D., Rossi, D., Gjesdal, O., Levy, L.M., Racagni, G., Danbolt, N.C., Volterra, A., 1996. Peroxynitrite inhibits glutamate transporter subtypes. J. Biol. Chem. 271, 5976–5979. Vila, M., Jackson-Lewis, V., Guegan, C., Wu, D.C., Teismann, P., Choi, D.K., Tieu, K., Przedborski, S., 2001. The role of glial cells in Parkinson's disease. Curr. Opin. Neurol. 14, 483–489. Yacoubian, T.A., Standaert, D.G., 2009. Targets for neuroprotection in Parkinson's disease. Biochim. Biophys. Acta 1792, 676–687.