JOURNAL OF NEUROCHEMISTRY

| 2009 | 109 | 11181128

doi: 10.1111/j.1471-4159.2009.06040.x

, ,1 ,

, ,1 ,

, , ,

*Institut National de la Sante et de la Recherche Medicale, Unite Mixte de Recherche S975, Centre de Recherche de lInstitut du piniere, Paris 6, France Cerveau et de la Moelle E ` Universite Pierre et Marie Curie-Paris, Paris, France Centre de Recherche Pierre Fabre, Castres, France

Abstract High plasma levels of the end product of purine metabolism uric acid (UA) predict a reduced risk of developing Parkinsons disease suggesting that UA may operate as a protective factor for midbrain dopaminergic neurons. Consistent with this view, UA exerted partial but long-term protection in a culture model in which these neurons die spontaneously. The rescued neurons were functional as they accumulated dopamine, efciently. The use of the uorescent probe dihydrorhodamine-123 revealed that UA operated by an antioxidant mechanism. The iron chelating agent desferrioxamine, the H2O2 scavenger enzyme catalase and the inhibitor of lipid peroxidation Trolox mimicked the effects of UA, suggesting that UA neutralized reactive oxygen species produced via a Fenton-type chemical reaction.

UA was, however, not signicantly accumulated into neurons, which indicates that the antioxidant effect occurred probably extracellularly. Structure activity relationships among purine derivatives revealed that the antioxidant properties of UA resulted from the presence of a 8-one substituent in its chemical structure. Of interest, the stimulation of L-type Ca2+ channels by high K+-induced depolarization and the ensuing activation of extracellular signal-regulated kinases 1/2 strongly improved the neuroprotective effect of UA whereas the depolarizing signal alone had no effect. In summary, our data indicate that UA may interfere directly with the diseases pathomechanism. Keywords: calcium, dopaminergic, oxidative stress, Parkinson, uric acid. J. Neurochem. (2009) 109, 11181128.

Parkinsons disease (PD) is a debilitating neurodegenerative disorder of ageing characterized by invalidating motor symptoms resulting from the loss of dopaminergic neurons in the substantia nigra (Agid 1991; Dauer and Przedborski 2003). Different epidemiological studies indicate that uric acid (UA), the end-product of purine nucleoside catabolism may inuence the progression of the disease; (i) Individuals with high plasma urate levels have a markedly reduced risk of developing PD (de Lau et al. 2005; Weisskopf et al. 2007). (ii) Further, among patients with early PD, higher plasma or cerebrospinal uid urate levels predict a slower clinical progression (Schiess and Oh 2008; Schwarzschild et al. 2008). (iii) A prospective study also demonstrated that individuals with gout, a rheumatic disease resulting from hyperuricemia, had a signicantly reduced occurrence of PD (Alonso et al. 2007; Gao et al. 2008). Conversely, UA levels have been reported to be reduced in the substantia nigra of

PD patients (Church and Ward 1993; Fitzmaurice et al. 2003) and probenecid an agent that stimulates the excretion of the purine metabolite in urine was found to potentiate the effects of the dopaminergic toxin MPTP in mice (Petroske et al. 2001; Alvarez-Fischer et al. 2008) which further

Received January 8, 2009; revised manuscript received March 6, 2009; accepted March 10, 2009. Address correspondence and reprint requests to Patrick P. Michel, INSERM UMR S975, Centre de Recherche de lInstitut du Cerveau et de la Moelle Epiniere, Bat. Pharmacie, Hopital de la Salpetriere, 47, bd de ` ` lhopital, Paris 75013, France. E-mail: patrick-pierre.michel@upmc.fr 1 Both these authors contributed equally to this study. Abbreviations used: Ca2+cyt, cytosolic calcium; DA, dopamine; DHR123, dihydrorhodamine-123; DIV, days in vitro; ERK, extracellular signal-regulated kinases; GFAP, glial brillary acidic protein; MAP-2, microtubule-associated protein-2; PD, Parkinsons disease; ROS, reactive oxygen species; TH, tyrosine hydroxylase; UA, uric acid.

1118

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

Protection of midbrain dopaminergic neurons by the end-product of purine metabolism uric acid | 1119

suggests that UA may interfere with the diseases pathomechanism. A number of studies have shown that UA has a strong antioxidant capacity. More specically, UA was reported to neutralize reactive species such as peroxynitrites (Keller et al. 1998) and hydroxyl radicals (Davies et al. 1986) and to prevent free radical-mediated chain reactions that results in lipid peroxidation (Muraoka and Miura 2003; Allameh et al. 2008), leading to the speculation that UA could protect dopaminergic neurons through an antioxidant mechanism (de Lau et al. 2005). This is particularly pertinent if we consider that oxidative stress is probably pivotal in the diseases process (Hirsch 1992; Jenner and Olanow 1998; Bharath et al. 2002). Yet, the mechanisms by which UA could protect dopaminergic neurons from oxidative stress-mediated insults remain rather elusive. In this study, we wished therefore (i) to demonstrate that UA was able to afford long-term protection to dopaminergic neurons in a paradigm in which these neurons die as the result of low-level oxidative stress; (ii) to explore the mechanism of this neuroprotective effect; (iii) to prove that the rescued neurons were functional; (iv) to dene the structural requirements conferring neuroprotection to UA; and (v) to determine whether its protective action could be possibly improved by other pharmacological treatments.

were fed every 2 days by replacing twice 350 lL of the culture medium. Control culture media and culture media supplemented with purine derivatives and/or high K+ were equilibrated for 1 h in the incubator at 37C before being added to the cultures. Other pharmacological treatments were applied directly to the cultures. Immunouorescence detection protocols The survival of dopaminergic and more generally of mesencephalic neurons was assessed by tyrosine hydroxylase (TH) and microtubule-associated protein-2 (MAP-2) immunouorescent detection respectively. The mouse anti-TH (Diasorin, Stillwater, MN, USA) and anti-MAP-2 monoclonal antibodies (clone AP-20; SigmaAldrich) were diluted 1 : 5000 and 1 : 50, respectively, in phosphate-buffered saline (PBS) containing 0.2% Triton X-100 (Guerreiro et al. 2008). Subsequent incubations were performed with a secondary anti-mouse IgG cyanin 3 conjugate (1 : 500; Sigma-Aldrich). Glial cells and more specically astrocytes were detected using vimentin (Clone V9, Sigma-Aldrich) and glial brillary acidic protein (GFAP) (Dako France, Trappes, France) antibodies, respectively, as described before (Troadec et al. 2002). A rabbit polyclonal antibody raised against UA conjugated to a protein carrier (#ab53000; Abcam, Cambridge, UK) was used to visualize UA within cultured neurons. Briey, the cultures were exposed overnight to the anti-UA antibody diluted 1 : 1000 in PBS containing 0.2% Triton X-100 followed by a detection step with an Alexa Fluor 488 goat anti-rabbit IgG. Dopamine uptake quantication The functional integrity of dopaminergic neurons was evaluated by their ability to accumulate tritiated DA (50 nM; 40 Ci/mmol) by active transport as described (Guerreiro et al. 2008). Blank values were obtained in the presence of 3 lM GBR-12909 (Sigma-Aldrich). Uptake of [methyl-3H]-thymidine [Methyl-3H]-thymidine, a marker of DNA synthesis was used to label proliferating cells (Traver et al. 2006). Mesencephalic cultures exposed for 48 h in vitro to [3H]-thymidine (40 Ci/mmol; 0.5 lCi/ well) in the presence of the test compounds were then allowed to recover until day in vitro (DIV) 5 in the same conditions of treatment but in the absence of the tritiated deoxynucleoside. The detection of TH and GFAP positive cells was performed as described before. Thymidine positive nuclei were visualized using the Hypercoat LM-1 emulsion (GE Healthcare, Orsay, France) after an incubation of 4 days at 4C. Quantication of reactive oxygen species and cytosolic-free calcium levels Cytoplasmic-free calcium and ROS levels were measured using calcium-green acetomethylester, (Guerreiro et al. 2008) and DHR123 (Traver et al. 2005) as uorescent probes respectively. Calcium measurements were performed in the early phase of the degenerative process (4 DIV) whereas ROS were quantied at a later stage when oxidative stress is at its peak (56 DIV). Measurement of intracellular uric acid Intracellular UA was assessed with the Kit PAP150 (BioMerieux, Lyon, France) using the protocol described by Duarte et al. (2005) and based on the manufacturers instructions. This procedure

Material and methods

Pharmacological agents Pharmacological reagents including UA and its structural analogues were obtained from Sigma-Aldrich (St Quentin Fallavier, France). Dihydrorhodamine-123 (DHR-123) and calcium-green acetomethylester, the cell permeant probes used for the detection of reactive oxygen species (ROS) and cytosolic calcium (Ca2+cyt), respectively, were purchased from Molecular Probes (Invitrogen, Cergy Pontoise, France). Tritiated compounds including [3H]-dopamine (DA) and [methyl-3H]-thymidine were obtained from GE Healthcare (Orsay, France). Cultures of mesencephalic dopaminergic neurons Animals were treated in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council 1996), the European Directive N886/609, and the guidelines of the local institutional committee for animal care and use. Cultures of dopaminergic neurons were prepared from the ventral mesencephalon of Wistar rat embryos dissected at embryonic day 15.5 as described previously (Troadec et al. 2002). Mesencephalic cells in suspension were plated onto polyethylenimine (1 mg/mL; SigmaAldrich) pre-coated culture plates (24 wells) and maintained in 500 lL of chemically dened serum-free medium consisting of equal volumes of Dulbeccos minimal essential medium and Hams F12 nutrient mixture (Invitrogen). This medium was supplemented with 10 lg/mL insulin, 30 mM glucose and 100 U/mL penicillin and streptomycin. To favour cell attachment, foetal calf serum (10%) was also added to the medium but only the rst hour after plating. Cultures

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

1120 | S. Guerreiro et al.

consisted in measuring H2O2 formed through the enzymatic conversion of UA into allantoin by uricase. Briey, cultured cells were washed twice with PBS and scrapped off 16 mm cultures wells using 70 lL of distilled water. Cell material from two culture wells was pooled and 100 lL of the total cell suspension was mixed with 900 lL of 50 mM Tris buffer, pH 8.0, containing 80 IU/L uricase, 200 IU/L peroxidase, 1000 IU/L ascorbate oxidase, 250 lM 4aminoantipyrine, 30 lM potassium ferrocyanate and 2 mM 3,5dichloro-2-hydroxybenzene sulfonic acid. The absorbance was measured at 520 nm using a Beckman Coulter DU 640B spectrophotometer (Fullerton, CA, USA). Western immunoblotting of extracellular signal-regulated kinases 1/2 After acute exposure (15 min) of the cultures to the test treatments, the cells were recovered in a lysis buffer, pH 8, containing 10 mM Tris/HCl, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40 (SigmaAldrich), 0.8% sodium deoxycholate, 250 IU/mL benzonase and 1% of the complete miniprotease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany). Samples were electrophoresed through a 412% acrylamide gel (NuPAGE Novex TrisAcetate Mini Gels; Invitrogen) and blotted onto a nitrocellulose membrane. The membranes were incubated with a phospho-extracellular signalregulated kinases 1/2 (ERK1/2) antibody (1 : 300; New England Biolabs, Ipswich, MA, USA) and developed with the enhanced chemiluminescence detection kit (Pierce, Rockford, IL, USA). Membranes were exposed 5 min to the Re-blot Plus Mild stripping solution (Millipore, Temicula, CA, USA), washed three times with 0.1% Tween in PBS, incubated with an anti-ERK1/2 antibody

(1 : 1000; New England Biolabs) and then developed as described above. Statistical analysis All data points presented are the mean SEM of ve to eight independent experiments. The statistical signicance of differences between groups was analysed with the SIGMASTAT 3.5 Software from Systat (San Jose, CA, USA) using one-way ANOVA followed by Dunnetts (for multiple comparisons against a single reference group) or Bonferroni (for all pairwise comparisons) post hoc tests. The Students t-test was applied when comparing a single treatment group to its corresponding control.

Results
Uric acid affords substantial protection to degenerating dopaminergic neurons Counts of mesencephalic dopaminergic neurons identied by TH immunouorescence conrmed that these neurons degenerate progressively in our culture system; 40% of the initial population was lost after 45 DIV and the degenerative process was generally complete by 8 DIV (Fig. 1a). When UA was applied to the cultures, the loss of TH+ cells was reduced in a concentration-dependent manner (Fig. 1b). The TH+ neurons rescued by UA were morphologically normal (Fig. 1c) and functional as they accumulated [3H]DA efciently via active transport (Fig. 1b). The EC50,

(a)

(c)

(b)

(d)

Fig. 1 Neuroprotection of dopaminergic neurons by UA in mesencephalic cultures. (a) Number of TH+ neurons in UA-treated and untreated cultures as a function of the days in vitro (DIV); *p < 0.05, higher than corresponding untreated cultures at the same stage of maturation. (b) Number of TH+ neurons and quantication of [3H]-DA uptake as a function of the concentrations of UA (01000 lM); *p < 0.05, increase in the test parameter compared with correspond-

ing untreated cultures. (c) Visualization of 8 DIV dopaminergic neurons exposed or not to a chronic treatment with UA (200 lM). Scale bars, 25 lm. (d) Counts of TH+ neurons in 8 DIV mesencephalic cultures exposed or not to 200 lM UA for various time intervals (open bars). The number of TH+ neurons plated initially in these cultures is represented by the lled bar; *p < 0.05, higher than untreated 8 DIV cultures.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

Protection of midbrain dopaminergic neurons by the end-product of purine metabolism uric acid | 1121

estimated graphically for both parameters, was 50 lM. Concentrations < 30 lM were ineffective. Protection against DA cell loss was not complete as in the presence of an optimal concentration of 200 lM UA only 50% of the TH+ neurons initially present in the cultures survived at 8 DIV (Fig. 1a). Interestingly, when the treatment was delayed after plating, UA still saved a signicant proportion of dopaminergic neurons, but only if added no later than day 4 (Fig. 1d). Conversely, when the cultures were treated only transiently with UA between days 0 and 4 and then left in standard culture medium, no effect of UA was observed at DIV 8. Uric acid operates by reducing intracellular oxidative stress We explored the possibility that UA could operate as an antioxidant using the ROS-sensitive probe DHR-123 as described (Traver et al. 2005). At a time when the death process is extensive, i.e. after 56 DIV, ROS levels increased dramatically in untreated cultures. In the presence of 200 lM UA, a concentration which is optimally effective for dopaminergic cell survival, ROS production was decreased by more than 50% whereas a concentration of 10 lM which has no protective effect failed to reduce the intensity of uorescent signal (Fig. 2a and b). An illustration that depicts the antioxidant effect of UA is given in Fig. 2c.

(a)

(b)

(c)

Uric acid prevents neurodegenerative changes caused by a Fenton-type reaction Neuroprotection by UA was reproduced by desferrioxamine (10 lM) which predominantly chelates ferric iron and by catalase (500 IU/mL) (Fig. 3a) which indicates that hydroxyl radicals produced via a Fenton-type reaction kill dopaminergic neurons in this culture system (Halliwell and Gutteridge 1984). Consistent with this view, neuroprotection was also achieved by Trolox (10 lM), a cell-permeable water-soluble derivative of vitamin E known to prevent hydroxyl radicalmediated lipid peroxidation. Desferrioxamine, catalase and Trolox were equally potent to UA in reducing ROS levels in degenerating neurons (Fig. 3b). Note that the survival effect of UA (200 lM) was not improved by any of these antioxidants (unshown results). Curiously, ascorbic acid (100 lM) which has been reported to be as potent as UA to prevent oxidative stress in some in vitro assays (Whiteman and Halliwell 1996) failed to reduce ROS production and to afford protection in our culture system (Fig. 3). Uric acid does not neutralize directly intracellular oxidative stress The use of a biochemical assay revealed that intracellular levels of UA were not signicantly increased in cultures exposed chronically to the purine derivative (Fig. 4a). This suggests that UA did not require to be accumulated into dopaminergic neurons to exert its neuroprotective effect. A visual conrmation of this nding was obtained with an

Fig. 2 UA reduces intracellular ROS production. (a) Number of dopaminergic neurons and (b) ROS detected with the uorochrome DHR-123 in mesencephalic cultures exposed to UA (10 and 200 lM) or no treatment. (c) Representative elds illustrating the inhibitory effects of UA on ROS production; *p < 0.05, different from untreated cultures.

antibody raised against UA conjugated to a protein carrier (Fig. 4b). Indeed, the uorescent signal was identical in control and UA-treated cultures. Note that virtually no uorescent signal was detected when the rst antibody was omitted. Neuroprotection by structural analogues of uric acid We also tested the neuroprotective effects of some of the close structural analogues of UA, xanthine, hypoxanthine and guanine, which are also its immediate precursors in the catabolic pathway of purines (Fig. 5). None of these compounds was able to reproduce the protective action of

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

1122 | S. Guerreiro et al.

(a)

(b)

Fig. 4 UA does not accumulate into mesencephalic neurons. (a) Biochemical assessment of intracellular levels of UA in 3 DIV mesencephalic cultures treated or not with 200 lM UA. Results were expressed in nmol/105 living (MAP-2+) neurons. (b) Visualization of UA in the same experimental conditions. Omit: signies that the rst antibody was not added to the cultures. Scale bar, 25 lm.

Fig. 3 Comparison of the effects of UA with that of other antioxidants. (a) Survival promotion of dopaminergic neurons in mesencephalic cultures exposed or not to UA or various antioxidants including desferrioxamine (DESF; 10 lM), catalase (CAT; 500 IU/mL), Trolox (TROL; 10 lM) and vitamin C (Vit C; 100 lM). (b) Detection of ROS in the same conditions as above; *p < 0.05, higher than untreated cultures and #p < 0.05, lower than untreated cultures.

UA. Not surprisingly, they were also ineffective in reducing ROS production. At variance, the synthetic 1,7-dimethyl derivative of UA reported previously to be protective in models of brain focal ischemia in mice (Haberman et al. 2007) retained the protective effects of its parent compound. The antioxidant potential of this compound was also comparable to that of UA in our model system. The neuroprotective effect of UA is potentiated by high K+-induced depolarization The rescuing effects of UA being only partial in our model system, we wished to determine whether dopaminergic cell survival could be improved by another treatment applied concurrently. Because dopaminergic neurons survive better when they are slightly depolarized (Michel et al. 2007), we submitted UA-treated cultures to high (30 mM) K+-induced depolarization in the presence of 1 lM MK-801 to avoid secondary excitotoxic stress (Salthun-Lassalle et al. 2004). In this particular set of experiments, MK-801 was therefore used with all other culture conditions. Our results show that the effect of UA at 8 DIV was strongly enhanced by the depolarizing treatment combining high K+ and MK-801

(Fig. 6a). Importantly, the survival rate of TH+ neurons was not modied by the presence of MK-801 either in control cultures (Fig. 6a) or in cultures exposed to UA alone. In a representative experiment, the number of TH+ neurons/well in UA-treated cultures was estimated to 1670 122 and 1759 144 in the absence or presence of MK-801 respectively. There is a possibility that the increase in TH+ cell numbers could result, at least in part, from a mitogenic effect of elevated K+ on TH+ neuroblasts or their precursor cells maintained in the presence of UA. This possibility was however, excluded as we failed to demonstrate the presence of dividing TH+ cells in cultures challenged with a marker of DNA synthesis [methyl-3H]-thymidine (Fig. 6b). In fact, the tritiated label was present only in cells with a glial phenotype detectable by vimentin or GFAP immunostaining (Fig 6b). In agreement with our previous observation (Troadec et al. 2002), the number of dividing cells expressing at least one of these two markers represented < 5% of all mesencephalic cells. The comparison of the number of surviving TH+ neurons at 0 and at 8 DIV suggested that the combined application of UA and 30 mM K+ rescued virtually all dopaminergic neurons plated in these cultures. Interestingly, the depolarizing signal alone afforded no protection, at all (Fig. 6a). It is interesting to note that the protective effect of UA was not restricted to dopaminergic neurons as the purine derivative also promoted the survival of other populations of mesencephalic neurons characterized by their immunoreactivity for MAP-2 (Fig. 6c). The effect of the depolarizing signal in the presence of UA appeared, however, restricted to dopaminergic neurons (Fig. 6c).

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

Protection of midbrain dopaminergic neurons by the end-product of purine metabolism uric acid | 1123

Fig. 5 Comparison of the effect of UA to that of some of its close structural analogues. Left panel: survival of TH+ neurons and ROS production in mesencephalic cultures exposed to UA and to some of its structural analogues, including hypoxanthine (HYPO), guanine (GUA), xanthine (XAN) and 1,7-dimethyl-UA (DMUA). All molecules were

tested at 200 lM. Right panel: chemical structures of the test molecules. Arrows represent enzymatic reactions involved in the nal stages of purine metabolism. Note that 1,7-dimethyl-UA is a synthetic analogue of UA not produced via purine nucleoside metabolism; *p < 0.05, higher than untreated cultures and #p < 0.05, lower than untreated cultures.

Mechanisms underlying high K+-mediated survival promotion of dopaminergic neurons in the presence of UA The protective effect of the depolarizing signal in the presence of UA was not accompanied by a further reduction of ROS levels (Fig. 6a and d) suggesting that it was attributable to another mechanism. In fact, the survival promoting effect of high K+ was associated to a moderate increase in intracellular calcium levels (Fig. 6e). Preventing this rise with the specic blocker of L-type voltage-gated calcium channels nifedipine (20 lM) abolished the survival promoting effect of the elevation of K+ but had no effect on that provided by UA indicating that the calcium elevation was accountable for the effect of the depolarizing signal. Note that the same rise was also detected in cultures where the depolarizing signal was applied in the absence of UA. Yet, this elevation was not sufcient per se to protect dopaminergic neurons. High K+-induced depolarization also caused an increase in ERK1/2 activation (Fig. 6f). Blockade of this increase with PD98059 (20 lM) or nifedipine abolished the survival effect of the depolarizing signal without affecting that of UA (Fig.6a). Unlike nifedipine, PD98059 had no inuence on the Ca2+ elevation caused by K+-induced depolarization (Fig. 6e). This suggests that the rise in calcium and the activation of ERK1/2 occurred sequentially to promote

dopaminergic cell survival. A schematic representation of the mechanisms by which UA may protect midbrain dopaminergic neurons is given in Fig. 7.

Discussion
We showed here that UA the nal product of purine nucleoside metabolism in humans and higher primates exerts partial but long-term protective effects in a culture model of spontaneous dopaminergic cell death. UA prevented the activation of a death pathway involving ROS by a mechanism that did not require its accumulation into neurons. Structureactivity relationship revealed that the 8-one moiety in its chemical structure was crucial for neuroprotection. Interestingly, UA-mediated neuroprotection was improved substantially by high K+-induced depolarization through a mechanism involving Ca2+ elevation and subsequently ERK1/2 activation. Uric acid affords robust but partial protection to dopamine neurons Mesencephalic dopaminergic neurons die spontaneously when maintained in a serum-free environment that contains trace amounts of iron (Troadec et al. 2002). This study showed that a signicant proportion of these neurons could

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

1124 | S. Guerreiro et al.

(a)

(d)

(b)

(c)

(e)

(f)

Fig. 6 The survival promoting effects of UA are improved by high K+induced depolarization. (a) Number of dopaminergic neurons immediately after plating and after 8 DIV in cultures maintained chronically with 200 lM UA and 30 mM K+ applied, separately or in combination. The normal concentration of K+ in the culture medium was 4 mM. Evaluation of the effects of nifedipine (NIF; 20 lM) or PD98059 (PD; 20 lM) on the treatment combining UA and 30 mM K+ or on the treatment with only UA. (b) Upper: TH+ neuron (red label) excluding the DNA marker [3H]-thymidine (blue label); Lower: GFAP+ astrocyte (red label) exhibiting tritiated thymidine in its nucleus. (c) Number of MAP-2+ neurons in 8 DIV control cultures and in cultures exposed to

UA alone or in the presence of 30 mM K+. (d) DHR-123 and (e) calcium-green-1 uorescence levels in the conditions of treatment described in the gure. (f) Western blot analysis of phospho-ERK1/2 (activated) and ERK1/2 (total) in the same treatment conditions. In all the experiments depicted in this gure, the culture medium was supplemented with MK-801 (1 lM) to avoid secondary excitotoxic stress generated by high K+-induced depolarization (Salthun-Lassalle et al. 2004); *p < 0.05, higher than control cultures at the same stage of maturation; **p < 0.05, higher than cultures exposed to UA alone; # p < 0.05, reduced compared with cultures exposed to UA and 30 mM K+; and ##p < 0.05, reduced compared with control cultures.

be rescued in a functional state when UA was applied chronically to the cultures. Of importance, UA was also protective if added when the loss of dopaminergic neurons was already ongoing. To be effective, however, the treatment with UA had to be initiated before 4 DIV which indicates that beyond a certain time-point, dopaminergic neurons become engaged in a cell death pathway that cannot be blocked any longer by UA. Du et al. (2007) reported previously that astrocytes intervene crucially in the neuroprotective effect of UA in a model system of spinal cord neurons dying by excitotoxic stress. Obviously, the effect of UA on dopaminergic neurons did not depend on astrocytes as the density of glial cells was very low in our culture paradigm. Accordingly, the protective effect of UA persisted

when glial cells were totally removed from the cultures by a treatment with the antimitotic cytarabine (data not shown). We may therefore surmise that the protective mechanism of UA varied with the nature of the lethal insults applied to the cultures and with the phenotypes of neurons as well. Uric acid protects dopamine neurons by reducing ROS production Several observations support the idea that UA protected DA neurons from death by reducing oxidative stress; (i) Dopaminergic cell demise in our preparation resulted from the presence of trace amounts of iron, a transition metal that catalyses the formation of hydroxyl radicals via a Fenton-type reaction (Halliwell and Gutteridge 1984). (ii) Neuroprotection

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

Protection of midbrain dopaminergic neurons by the end-product of purine metabolism uric acid | 1125

Fig. 7 Schematic representation illustrating how UA may protect midbrain dopaminergic neurons. Soluble oxygen gives rise spontaneously to superoxide anions which by dismutation produce H2O2. In the presence of H2O2, the transition metal catalyst iron triggers a Fenton-type reaction that generates hydroxyl radicals. Hydroxyl radicals initiate lipid peroxidation at the plasma membrane level which stimulates in turn intracellular ROS production by a mechanism that is probably indirect. UA prevents the propagation of oxidative stress from

the extracellular to the intracellular milieu by preserving the integrity of the plasma membrane at the lipid-aqueous interface boundary. Note that UA is unlikely to penetrate into the cytoplamic compartment. High K+-induced depolarization amplies neuroprotection provided by UA through a mechanism involving Ca2+ elevation and ERK1/2 activation. CaVL: L-type voltage-gated calcium channel; CAT: catalase; DESF: desferrioxamine; TROL: Trolox.

by UA was mimicked by conventional antioxidants including the chain breaking antioxidant Trolox and the hydroxyl radical detoxifying enzyme catalase. Accordingly, ROS levels were reduced substantially in UA-treated mesencephalic cultures. These observations are also consistent with studies showing that UA can operate as a potent antioxidant either in acellular preparations (Ames et al. 1981; Davies et al. 1986) or in a number of paradigms where neuronal cell death is caused by excitotoxic stress (Yu et al. 1998; Du et al. 2007) or metabolic impairment (Duan et al. 2002). Importantly, the antioxidant effect provided by UA was not sufcient in itself to rescue all cultured dopaminergic neurons, suggesting that some of these neurons die either by a mechanism unrelated to oxidative stress or alternatively that the antioxidant effect is not sufcient in itself to provide protection to this subpopulation of dopaminergic neurons. It is also worth noting that UA failed to prevent dopaminergic cell death induced by the nitric oxide donor diethylamine NONOate (unshown results) which conrms that the protective action of the purine derivative did not extend to reactive nitrogen species or their derivatives (Rodrguez-Martn et al. 2002). Mechanism underlying the antioxidant effect of uric acid Given that UA has the potential to form stable 1 : 1 and 2 : 1 coordination complexes with either Fe3+ or Fe2+ cations

(Davies et al. 1986), one may speculate that UA reproduced the effect of desferrioxamine and thus operated by sequestration of iron. As ferric iron is present at a concentration of 1.5 lM in the culture medium, a protective effect by sequestration would be expected to occur at concentrations on the order of 3 lM. This was not the case as UA promoted dopaminergic cell survival when its concentrations reached at least 3050 lM. This means that the chelating properties of UA were not responsible for its protective action in our model system. Alternatively, UA was also reported to operate as an inhibitor of lipid peroxidation (Davies et al. 1986; Muraoka and Miura 2003), a mechanism which is more likely to explain the effects of UA in the present paradigm as the concentrations of UA protective in mesencephalic cultures (i.e. 30500 lM) were very similar to those reported effective against lipid peroxidation in liposomal or microsomal preparations (Davies et al. 1986; Muraoka and Miura 2003). Still consistent with this hypothesis, the inhibitor of lipid peroxidation Trolox mimicked the protective effect of UA for DA neurons in our model system. The fact that the treatment with UA caused a decrease in intracellular ROS production may signify that UA operated after being accumulated into neurons. Specic UA transporters exist but essentially in renal proximal tubule cells (Endou and Anzai 2008) and passive transport is unlikely

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

1126 | S. Guerreiro et al.

owing to the hydrophilic character of UA (Muraoka and Miura 2003). Accordingly, UA levels were not increased in neuronal cells chronically exposed to this compound. This means that UA exerted probably its antioxidant effect extracellularly. Consistent with this hypothesis, the initial source of ROS was also extracellular in our model system as catalase which does not diffuse through plasma membranes (Buettner 1993), provided nevertheless robust protection to dopaminergic neurons. Therefore, we may assume that UA operated by preventing the propagation of oxidative stress from the extracellular to the intracellular milieu, possibly by preserving the integrity of the plasma membrane. Supporting this view, UA has been reported to operate at the lipidaqueous boundary of membranes in microsomal preparations (Muraoka and Miura 2003). Interestingly, the lowest protective concentrations of UA in mesencephalic cultures, i.e. 30 50 lM, were physiologically relevant as they corresponded to levels of the purine metabolite detected extracellularly in the striatum (Osborne and Hashimoto 2006). What structural features confer antioxidant effects to uric acid? To determine what structural features conferred antioxidant properties to UA, we tested the neuroprotective ability of some of its close structural analogues xanthine, hypoxanthine and guanine which are also its precursors in the catabolic pathway of purine nucleosides (Rodwell 1993). Xanthine the immediate precursor of UA, which is produced via the degradation pathway of adenosine or that of guanosine, failed to afford neuroprotection. Likewise, hypoxanthine and guanine the two purines bases which give rise to xanthine via catabolism of adenosine and guanosine, respectively, were ineffective too. In line with this observation, xanthine, hypoxanthine and guanine failed to reduce intracellular oxidative stress. Interestingly, Muraoka and Miura (2003) reported very weak antioxidant properties of xanthine and hypoxanthine in microsomal preparations in comparison to UA. Given that UA differs from xanthine only by a 8-one moiety (Rodwell 1993), we may surmise that this chemical group greatly contributed to the antioxidant activity of UA. In line of this observation, the 1,7-dimethyl derivative analogue of UA which retains the 8-one moiety in its structure was as potent as UA itself. This nding also suggested that N-methylation on nitrogens at positions 1 and 7 did not weaken the antioxidant potential of the purine structure. K+-induced depolarization strongly potentiates the effect of UA by a mechanism that does not involve a reduction in ROS production We demonstrated in this study that UA provided robust protection to dopaminergic neurons but only to a fraction of them. Yet, neuroprotection by UA was strongly enhanced when the cultures were exposed to a low-level

depolarizing stimulus produced by elevation of extracellular concentrations of K+. The increase in TH+ neurons was not because of a possible mitogenic effect on precursor cells (Daadi and Weiss 1999) as tritiated thymidine a marker of DNA synthesis was absent from the nucleus of these neurons. The induction of the TH enzyme in neurons that did not express originally detectable levels of the enzyme as reported in other model systems (Traver et al. 2006) was also excluded for the reason that the number TH+ cells exposed to UA in the presence of elevated concentrations of K+ remained identical or slightly below the number of TH+ neuroblasts detectable immediately after plating. Therefore, the effect of the depolarizing signal reected most likely a true neuroprotective effect. It is worth noting that the elevation of K+ did not further reduce ROS production in UA-treated cultures which indicates that the depolarizing signal did not operate itself via an antioxidant mechanism. This may explain why elevated K+ failed to provide protection to dopaminergic neurons in the absence of UA. The rescuing effect of K+-induced depolarization is mediated by Ca2+-dependent ERK1/2 activation The potentiation of the effect of UA by elevated K+ resulted most likely from a small increase in Ca2+cyt as preventing this increase with the L-type voltage-gated calcium channel blocker nifedipine prevented the rescuing effect of the depolarizing signal. Note that these observations are somehow paradoxical if we refer to data showing that UA and its analogues were reported previously to limit the effects of ischemic insults and excitotoxic stress (Yu et al. 1998; Du et al. 2007; Haberman et al. 2007), i.e. pathological conditions causing calcium overload. These results are, however, in line with the concept that calcium has to be maintained slightly above control levels in DA neurons to prevent their demise (Salthun-Lassalle et al. 2004; Guerreiro et al. 2008). Next, we explored the possibility that the effect of elevated Ca2+cyt in UA-treated cultures could result from the activation of the ERK1/2 kinases, which are involved in neuronal survival in a number of experimental paradigms (Troadec et al. 2002) although this remains controversial (Zhuang and Schnellmann 2006). Blocking the activation of mitogenactivated protein kinase kinase the direct upstream kinase of ERK1/2 with PD98059 had no effect on the elevation of Ca2+cyt caused by high K+-induced depolarization but it prevented the rescuing effect of the depolarizing signal in UA-treated cultures. Nifedipine, however, prevented both the inux of calcium and the activation of ERK1/2 elicited by the depolarizing signal which indicates that Ca2+cyt elevation and ERK1/2 activation were involved sequentially in the survival of the population of dopaminergic neurons for which UA alone was ineffective. Interestingly, neither nifedipine nor PD98059 inhibited the effects of UA suggesting that ERK1/2 activation by calcium was solely responsible for the effect of

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

Protection of midbrain dopaminergic neurons by the end-product of purine metabolism uric acid | 1127

the depolarizing signal. The nature of the downstream targets of ERK1/2, remains, however, to be established. Ultimately, these results demonstrate that UA acted in concert with a ERK1/2-dependent mechanism elicited by high K+-induced depolarization to rescue dopaminergic neurons. Yet, it appears that the two mechanisms were not mutually substitutable. In summary, we demonstrate here that UA can exert potent survival promoting effects for dopaminergic neurons through a direct antioxidant mechanism that can be possibly amplied (but not mimicked) by low-level depolarization. On this basis, we may speculate that reduced levels of UA in the substantia nigra of PD patients (Church and Ward 1993; Fitzmaurice et al. 2003) may sensitize dopaminergic neurons to oxidative stress-mediated insults whereas high levels of the purine metabolite may exert protective effects against the diseases process.

Acknowledgments
AP performed this work in partial fullment of a Master degree in Biology of Aging at University Pierre et Marie Curie. This work was supported by INSERM and UPMC. PPM was supported by Laboratoire Pierre Fabre.

References
Agid Y. (1991) Parkinsons disease: pathophysiology. Lancet 337, 1321 1324. Allameh A., Maleklou N., Zargari M. and Sanati M. H. (2008) The inuence of uric acid treatments on liver glutathione system prevent oxidative damages in experimental autoimmune encephalomyelitis mice. Neurosci. Lett. 439, 111115. Alonso A., Rodrguez L. A., Logroscino G. and Hernan M. A. (2007) Gout and risk of Parkinson disease: a prospective study. Neurology 69, 16961700. Alvarez-Fischer D., Guerreiro S., Hunot S., Saurini F., Marien M., Sokoloff P., Hirsch E. C., Hartmann A. and Michel P. P. (2008) Modelling Parkinson-like neurodegeneration via osmotic minipump delivery of MPTP and probenecid. J. Neurochem. 107, 701 711. Ames B. N., Cathcart R., Schwiers E. and Hochstein P. (1981) Uric acid provides an antioxidant defense in humans against oxidant- and radical-caused aging and cancer: a hypothesis. Proc. Natl Acad. Sci. USA 78, 68586862. Bharath S., Hsu M., Kaur D., Rajagopalan S. and Andersen J. K. (2002) Glutathione, iron and Parkinsons disease. Biochem. Pharmacol. 64, 10371048. Buettner G. R. (1993) The pecking order of free radicals and antioxidants: lipid peroxidation, alpha-tocopherol, and ascorbate. Arch. Biochem. Biophys. 300, 535543. Church W. H. and Ward V. L. (1993) Uric acid is reduced in the substantia nigra in Parkinsons disease: effect on dopamine oxidation. Brain Res. Bull. 33, 419425. Daadi M. M. and Weiss S. (1999) Generation of tyrosine hydroxylaseproducing neurons from precursors of the embryonic and adult forebrain. J. Neurosci. 19, 44844497. Dauer W. and Przedborski S. (2003) Parkinsons disease: mechanisms and models. Neuron 39, 889909.

Davies K. J., Sevanian A., Muakkassah-Kelly S. F. and Hochstein P. (1986) Uric acid-iron ion complexes. A new aspect of the antioxidant functions of uric acid. Biochem. J. 235, 747754. Du Y., Chen C. P., Tseng C. Y., Eisenberg Y. and Firestein B. L. (2007) Astroglia-mediated effects of uric acid to protect spinal cord neurons from glutamate toxicity. Glia 55, 463472. Duan W., Ladenheim B., Cutler R. G., Kruman I. I., Cadet J. L. and Mattson M. P. (2002) Dietary folate deciency and elevated homocysteine levels endanger dopaminergic neurons in models of Parkinsons disease. J. Neurochem. 80, 101110. Duarte A. I., Santos M. S., Oliveira C. R. and Rego A. C. (2005) Insulin neuroprotection against oxidative stress in cortical neurons: involvement of uric acid and glutathione antioxidant defenses. Free Radic. Biol. Med. 39, 876889. Endou H. and Anzai N. (2008) Urate transport across the apical membrane of renal proximal tubules. Nucleosides Nucleotides Nucleic Acids 27, 578584. Fitzmaurice P. S., Ang L., Guttman M., Rajput A. H., Furukawa Y. and Kish S. J. (2003) Nigral glutathione deciency is not specic for idiopathic Parkinsons disease. Mov. Disord. 18, 969976. Gao X., Chen H., Choi H. K., Curhan G., Schwarzschild M. A. and Ascherio A. (2008) Diet, urate, and Parkinsons disease risk in men. Am. J. Epidemiol. 167, 831838. Guerreiro S., Toulorge D., Hirsch E., Marien M., Sokoloff P. and Michel P. P. (2008) Paraxanthine, the primary metabolite of caffeine, provides protection against dopaminergic cell death via stimulation of ryanodine receptor channels. Mol. Pharmacol. 74, 980989. Haberman F., Tang S. C., Arumugam T. V., Hyun D. H., Yu Q. S., Cutler R. G., Guo Z., Holloway H. W., Greig N. H. and Mattson M. P. (2007) Soluble neuroprotective antioxidant uric acid analogs ameliorate ischemic brain injury in mice. Neuromol. Med. 9, 315 323. Halliwell B. and Gutteridge J. M. (1984) Oxygen toxicity, oxygen radicals, transition metals and disease. Biochem. J. 219, 114. Hirsch E. C. (1992) Why are nigral catecholaminergic neurons more vulnerable than other cells in Parkinsons disease? Ann. Neurol. 32, S88S93. Jenner P. and Olanow C. W. (1998) Understanding cell death in Parkinsons disease. Ann. Neurol. 44, S72S84. Keller J. N., Kindy M. S., Holtsberg F. W., St Clair D. K., Yen H. C., Germeyer A., Steiner S. M., Bruce-Keller A. J., Hutchins J. B. and Mattson M. P. (1998) Mitochondrial manganese superoxide dismutase prevents neural apoptosis and reduces ischemic brain injury: suppression of peroxynitrite production, lipid peroxidation, and mitochondrial dysfunction. J. Neurosci. 18, 687697. de Lau L. M., Koudstaal P. J., Hofman A. and Breteler M. M. (2005) Serum uric acid levels and the risk of Parkinson disease. Ann. Neurol. 58, 797800. Michel P. P., Alvarez-Fischer D., Guerreiro S., Hild A., Hartmann A. and Hirsch E. C. (2007) Role of activity-dependent mechanisms in the control of dopaminergic neuron survival. J. Neurochem. 101, 289 297. Muraoka S. and Miura T. (2003) Inhibition by uric acid of free radicals that damage biological molecules. Pharmacol. Toxicol. 93, 284 289. Osborne P. G. and Hashimoto M. (2006) Brain antioxidant levels in hamsters during hibernation, arousal and cenothermia. Behav. Brain Res. 168, 208214. Petroske E., Meredith G. E., Callen S., Totterdell S. and Lau Y. S. (2001) Mouse model of Parkinsonism: a comparison between subacute MPTP and chronic MPTP/probenecid treatment. Neuroscience 106, 589601. Rodrguez-Martn E., Casarejos M. J., Canals S., De Bernardo S. and Mena M. A. (2002) Thiolic antioxidants protect from nitric

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128

1128 | S. Guerreiro et al.

oxide-induced toxicity in fetal midbrain cultures. Neuropharmacology 43, 877888. Rodwell V. W. (1993) Metabolism of purines and pyrimidine nucleotides, in Harpers Biochemistry (Murray R. K., Granner D. K., Mayes P. A. and Rodwell V W., eds), pp., 363377. Appleton and Lange, London. Salthun-Lassalle B., Hirsch E. C., Wolfart J., Ruberg M. and Michel P. P. (2004) Rescue of mesencephalic dopaminergic neurons in culture by low-level stimulation of voltage-gated sodium channels. J. Neurosci. 24, 59225930. Schiess M. and Oh I. (2008) Serum uric acid and clinical progression in Parkinson disease: potential biomarker for nigrostriatal failure. Arch. Neurol. 65, 698699. Schwarzschild M. A., Schwid S. R., Marek K. et al.; Parkinson Study Group PRECEPT Investigators. (2008) Serum urate as a predictor of clinical and radiographic progression in Parkinsons disease. Arch. Neurol. 65, 716723. Traver S., Salthun-Lassalle B., Marien M., Hirsch E. C., Colpaert F. and Michel P. P. (2005) The neurotransmitter noradrenaline rescues septal cholinergic neurons in culture from degeneration caused by low-level oxidative stress. Mol. Pharmacol. 67, 18821891. Traver S., Marien M., Martin E., Hirsch E. C. and Michel P. P. (2006) The phenotypic differentiation of locus ceruleus noradrenergic

neurons mediated by brain-derived neurotrophic factor is enhanced by corticotropin releasing factor through the activation of a cAMPdependent signaling pathway. Mol. Pharmacol. 70, 3040. Troadec J. D., Marien M., Mourlevat S., Debeir T., Ruberg M., Colpaert F. and Michel P. P. (2002) Activation of the mitogen-activated protein kinase (ERK1/2) signaling pathway by cyclic AMP potentiates the neuroprotective effect of the neurotransmitter noradrenaline on dopaminergic neurons. Mol. Pharmacol. 62, 10431052. Weisskopf M. G., OReilly E., Chen H., Schwarzschild M. A. and Ascherio A. (2007) Plasma urate and risk of Parkinsons disease. Am. J. Epidemiol. 166, 561567. Whiteman M. and Halliwell B. (1996) Protection against peroxynitritedependent tyrosine nitration and alpha 1-antiproteinase inactivation by ascorbic acid. A comparison with other biological antioxidants. Free Radic. Res. 25, 275283. Yu Z. F., Bruce-Keller A. J., Goodman Y. and Mattson M. P. (1998) Uric acid protects neurons against excitotoxic and metabolic insults in cell culture, and against focal ischemic brain injury in vivo. J. Neurosci. Res. 53, 613625. Zhuang S. and Schnellmann R. G. (2006) A death-promoting role for extracellular signal-regulated kinase. J. Pharmacol. Exp. Ther. 319, 991997.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2009) 109, 11181128