A 3 Adenosine receptors mediate oligodendrocyte death and ischemic damage to optic nerve

RESEARCH ARTICLE

A3 Adenosine Receptors MediateOligodendrocyte Death and Ischemic

Damage to Optic Nerve

Est�ıbaliz Gonz�alez-Fern�andez,1 Mar�ıa Victoria S�anchez-G�omez,1 Alberto P�erez-Samart�ın,1

Rogelio O. Arellano,1,2 and Carlos Matute1

Adenosine receptor activation is involved in myelination and in apoptotic pathways linked to neurodegenerative diseases. In thisstudy, we investigated the effects of adenosine receptor activation in the viability of oligodendrocytes of the rat optic nerve. Selec-tive activation of A3 receptors in pure cultures of oligodendrocytes caused concentration-dependent apoptotic and necrotic deathwhich was preceded by oxidative stress and mitochondrial membrane depolarization. Oligodendrocyte apoptosis induced by A3

receptor activation was caspase-dependent and caspase-independent. In addition to dissociated cultures, incubation of opticnerves ex vivo with adenosine and the A3 receptor agonist 2-CI-IB-MECA(1-[2-Chloro-6-[[(3-iodophenyl)methyl]amino]-9H-purin-9-yl]-1-deoxy-N-methyl-b-D-ribofuranuronamide)-induced caspase-3 activation, oligodendrocyte damage, and myelin loss, effectswhich were prevented by the presence of caffeine and the A3 receptor antagonist MRS 1220 (N-[9-Chloro-2-(2-furanyl)[1,2,4]-tria-zolo [1,5-c]quinazolin-5-yl]benzene acetamide). Finally, ischemia-induced injury and functional loss to the optic nerve was attenu-ated by blocking A3 receptors. Together, these results indicate that adenosine may trigger oligodendrocyte death via activation ofA3 receptors and suggest that this mechanism contributes to optic nerve and white matter ischemic damage.

GLIA 2013;00:000–000Key words: demyelinization, ischemia, mitochondria, caspases, astrocytes

Introduction

Adenosine is both a catabolic product and a precursor of

ATP (Conde et al., 2006). Although it is not considered

as a classical neurotransmitter, it plays an important role in the

homeostasis of the nervous system, and it can be released from

most cells, including neurons and glia (Ribeiro et al., 2002).

Adenosine activates A1, A2a, A2b, and A3 adenosine receptors,

which are coupled to G proteins (Fredholm et al., 2001). A1

and A3 receptors activate Gi/0 proteins, which inhibit adenylate

cyclase; A2a and A2b receptors are coupled to Gs proteins,

which activate adenylate cyclase. These receptors possess differ-

ent ranges of affinity to adenosine. The A1, A2a, and A3 recep-

tors can be activated by physiological levels (25–250 nM) of

adenosine (Dunwiddie and Masino, 2001). In contrast, the

A2b receptor requires a higher agonist concentration, which

may be present in pathophysiological states (Fredholm et al.,

2001), like hypoxia or ischemia (Latini and Pedata, 2001).

Adenosine receptors are expressed in most tissues and cells,

including the central nervous system (CNS) (Dar�e et al., 2007;

Hammarberg et al., 2004; Melani et al., 2009; Othman et al.,

2003; Ribeiro et al., 2002). In the brain, they can protect neurons

from dying or, conversely, activate apoptosis cascades (Fatokun

et al., 2008; Lauro et al., 2008). However, little is known regard-

ing adenosinergic signaling in oligodendrocytes. These cells are

damaged in many diseases including ischemia (Domercq et al.,

2010; Shibata et al., 2000; Walker and Rosenberg, 2010) and

multiple sclerosis (MS; O’Meara et al., 2011; Smith and Lass-

mann, 2002). Although adenosine receptors in oligodendroglia

modulate migration, proliferation, and differentiation (Othman

et al., 2003; Stevens et al., 2002), there are also data which pro-

pose the manipulation of adenosine receptors as a target of oligo-

dendrocyte protection after ischemic damage (Melani et al.,

2009; Trincavelli et al., 2008) and in MS (Chen et al., 2010;

Mills et al., 2008; Tsutsui et al., 2004; Wei et al., 2013).

View this article online at wileyonlinelibrary.com. DOI: 10.1002/glia.22599

Published online Month 00, 2013 in Wiley Online Library (wileyonlinelibrary.com). Received June 13, 2013, Accepted for publication Oct 28, 2013.

Address correspondence to Dr. Carlos Matute; Departamento de Neurociencias, Universidad del Pa�ıs Vasco, E-48940 Leioa, Spain. Email: [email protected]

From the 1CIBERNED, Achucarro Basque Center for Neuroscience and Departamento de Neurociencias, Universidad del Pa�ıs Vasco (UPV/EHU), E-48940 Leioa,

Spain; 2Departamento de Neurobiolog�ıa Celular y Molecular, Instituto de Neurobiolog�ıa, Universidad Nacional Aut�onoma de M�exico, Juriquilla Quer�etaro, M�exico.

Additional Supporting Information may be found in the online version of this article.

VC 2013 Wiley Periodicals, Inc. 1

In vitro and in vivo experiments have shown that oligo-

dendrocytes are highly vulnerable to hypoxic and ischemic

insults (Domercq et al., 2010; Goldenberg-Cohen et al.,

2005; Yang et al., 2011). In this study, we investigated

whether adenosine was protective or deleterious to oligoden-

drocytes. First, we found that both oligodendrocyte cultures

and optic nerve expressed adenosine receptors. Then, we ana-

lyzed the effects of adenosine on the viability of oligodendro-

cyte in these preparations, and we found that selective

activation of A3 compromised oligodendrocyte survival by

altering mitochondrial function and inducing apoptosis and

necrosis. Finally, we provide evidence that adenosine receptors

are relevant to ischemia-induced oligodendrocyte death in the

rat optic nerve and that adenosine receptor antagonists ameli-

orate the loss of oligodendrocytes and myelin triggered by

ischemic injury.

Materials and Methods

AnimalsAll experiments were conducted under the supervision and with the

approval of the internal animal ethics committee of the University of the

Basque Country (UPV/EHU). Animals were handled in accordance with

the European Communities Council Directive. All possible efforts were

made to minimize animal suffering and the number of animals used.

Adenosine Receptor DrugsAdenosine receptor ligands included adenosine and antagonist caffeine

(Sigma, St. Louis, MO), adenosine receptor-specific agonists CPA (A1

receptor agonist; N6-cyclopentyladenosine), CGS 21680 (A2a receptor

agonist; 4-[2-[[6-Amino-9-(N-ethyl-b-D-ribofuranuronamidosyl)-9H-

purin-2-yl]amino]ethyl]benzene propanoic acid hydrochloride),

NECA (A2b receptor agonist; 50-N-ethylcarboxamidoadenosine), and

2-CI-IB-MECA (highly selective A3 agonist; 1-[2-Chloro-6-[[(3-iodo-

phenyl)methyl]amino]-9H-purin-9-yl]-1-deoxy-N-methyl-b-D-ribofur-

anuronamide), and selective antagonist MRS 1220 (A3 adenosine

receptor antagonist; N-[9-Chloro-2-(2-furanyl)[1,2,4]-triazolo[1,5-

c]quinazolin-5-yl]benzene acetamide) and SCH 58261 (A2a receptor

antagonist; 2- (2- Furanyl) -7 -(2-phenylethyl) - 7H-pyrazolo [4,3-e]

[1,2,4] triazolo [1,5-c] pyrimidin-5-amine) were obtained from Tocris

(Cookson, Bristol, UK). All drugs were first dissolved in water (aden-

osine, caffeine) or DMSO (CPA, CGS 21680, NECA, 2-CI-IB-

MECA, MRS 1220, and SCH 58261), and then added to the

medium to achieve the desired final concentration.

Oligodendrocyte CulturesOligodendrocyte cultures were prepared from optic nerves of 12-

day-old Sprague-Dawley rats, as described previously (S�anchez-

G�omez and Matute, 1999), with minor modifications (Alberdi et al.,

2002). Cells were seeded onto 14-mm diameter glass coverslips,

coated with 10 lg/mL poly-D-lysine, at 1 3 104 cells per coverslip,

and maintained in 24-well plates at 37�C and 5% CO2 in a chemi-

cally defined medium (Barres et al., 1992). Cells were used after 1–2

day of culturing in vitro, when cultures comprised at least 98% of

oligodendrocyte marker O41/Galactocerebroside1 (GalC) oligoden-

drocytes; the remaining cells were GFAP1 or unidentified cells. We

did not detect A2B15 oligodendrocyte progenitors or microglial cells

in these cultures (Alberdi et al., 2002).

Reverse Transcription-Polymerase Chain ReactionTotal RNA was extracted from cell cultures and from rat tissue sam-

ples with TRIzol reagent according to the manufacturer’s recommen-

dations (Invitrogen, Barcelona, Spain). After DNase I treatment

(TURBO DNAfree TM, Ambion, Austin, TX), the RNA (2 lg)

was denatured for 5 min at 65�C, then cooled to 4�C. First strand

cDNA synthesis was carried out by adding the reverse transcriptase,

SuperscriptTM III (Invitrogen), the supplied buffer, random primers,

and recombinant RNase inhibitor for a final volume of 50 lL. The

reverse transcription reaction was performed at 50�C for 1 h. PCR

reactions were performed with 1 lL of cDNA diluted in Platinum

SYBBRGreen qPCR Supermix UDG (Invitrogen) in an ABI PRISM

7000 Sequence Detection System (Applied Biosystems). Amplifica-

tion of cDNA was performed with the primers (PrimerExpress Soft-

ware, Applied Biosystems) listed in Supporting Information, Table

S1. The PCR protocol consisted of heating at 95�C for 10 min,

then 40 cycles of 95�C for 15 s, and 60�C for 1 min. PCR products

(20 lL) were loaded on a 2% agarose gel containing SYBRVR Safe,

submitted to electrophoresis for 90 min, and visualized under UV

light. We used reverse transcriptase-negative controls to verify that

the signal was not overestimated by contamination from residual

genomic DNA amplification. The quantitative PCR products were

also subjected to a dissociation protocol to ensure that a single

amplicon of the expected melting temperature was indeed obtained.

Toxicity Assays and Cell ViabilityAfter 1 day in culture, oligodendrocytes were exposed for 15 min to

adenosine receptor agonists alone, or in the presence of antagonists

added 10 min before the agonists. In some experiments, cells were pre-

treated with caspase inhibitors or a PARP-1 inhibitor, 3,4-Dihydro-

5[4-(1-piperindinyl)butoxy]-1(2H)-isoquinoline (DPQ; Calbiochem),

for 30 min before incubating with adenosine receptor agonists. The fol-

lowing caspase inhibitors were prepared in DMSO according to the

manufacturer’s instructions (Peptides International, Louisville,

KY): acetyl-aspartyl-glutamyl-valyl-aspart-1-aldehyde (Ac-DEVD-H);

acetylisoleucyl-glutamyl-threonyl-aspart-1-aldehyde (Ac-IETD-H); ace-

tylleucyl-glutamyl-histidyl-aspart-1-aldehyde (Ac-LEHD-H). Broad-

spectrum caspase inhibitor Z-VAD-FMK [N-benzyloxycarbonyl-valyl-

alanylaspart-(OMe)-fluoromethylketone] was purchased from Tocris

(Bristol, UK). After drug incubation cells were cultures for 24 h, and

oligodendrocyte viability was assessed by loading cells with 1 lM

calcein-AM (Invitrogen; C3100) for 30 min. Fluorescence was meas-

ured with a Synergy-HT fluorimeter (Bio-Tek Instruments, Inc), as

indicated by the supplier (485 nm excitation wavelength and 530 nm

emission wavelength). All experiments were performed in triplicate,

and the values indicated represent the average of at least three inde-

pendent experiments.

Intracellular Reactive Oxygen Species MeasurementOligodendrocyte cultures were exposed to adenosine or 2-CI-IB-

MECA in the presence and absence of antagonists (caffeine or MRS

2 Volume 00, No. 00

1220), as described previously. Then, accumulation of reactive oxygen

species (ROS) within cells was measured by loading cells with 10 lM

5,6-chloromethyl-2�,7�-dichlorodihydrofluorescein diacetate (CM-

H2DCFDA; Molecular Probes, Invitrogen; C6827) for 30 min at

37�C, 5% CO2, and 1 lM calcein-AM (Molecular Probes:C3100) as

a control dye. Fluorescence was measured with a Synergy-HT fluo-

rimeter (Bio-Tek Instruments). Excitation and emission wavelengths

for CM-DCFDA were 488 and 515 nm, respectively. Data were

expressed as the percentage of CM-H2DCFDA/calcein fluorescence,

normalized to the same ratio in controls (100%). All experiments

were performed in duplicate and data were plotted as the mean of at

least three independent experiments 6 SEM.

Measurement of Mitochondrial Membrane PotentialFor evaluation of the mitochondrial potential, oligodendrocytes were

exposed to adenosine or 2-CI-IB-MECA alone or with other drugs.

Then, cells were loaded with 3 lM of 5,5,6,6-tetrachloro-1,1,3,3-tet-

raethylbenzimidazolcarbocyanine iodide (JC-1) (Invitrogen) for 15

min at 37�C, and to eliminate the excess, cells were washed twice

with HBSS without phenol red. In the cytosol, the monomeric form

of JC-1 fluoresces green (emission at 527 nm when excited at 485

nm), whereas within the mitochondrial matrix, JC-1 becomes highly

concentrated and forms aggregates that fluoresce red (emission at

590 nm when excited at 485 nm). Both JC-1 monomers and aggre-

gates were detectable with a Synergy-HT fluorimeter (Bio-Tek

Instruments). Changes in mitochondrial potential were calculated as

the red/green fluorescence ratio for each condition. All experiments

were performed in triplicate, and the values represent the normalized

mean 6 SEM of at least three independent experiments.

Annexin AssayAfter 1 day in culture, oligodendrocytes were exposed to adenosine

(10 lM–1 mM) or the A3-receptor agonist, 2-CI-IB-MECA (30 or

100 lM) for 30 min. Early apoptosis was determined with an

annexin assay. Cells were incubated with Annexin-V-Fluos (1:100;

Molecular Probes, Invitrogen) for 15 min at RT in the dark. This

procedure allows the fluorophore to bind to negatively charged phos-

pholipid surfaces in a Ca21-dependent manner. This binding is

highly specific for phosphatidylserine on the outer membrane of

apoptotic cells. In all cases, oligodendrocyte nuclei were counter-

stained with Hoechst 33258 (5 lg/mL; Molecular Probes, Invitro-

gen) to simultaneously evaluate nuclear condensation. After washing,

cells were observed under a fluorescence microscope, and apoptotic

cells were identified by the presence of Annexin-V-Fluos (green sig-

nal) on the cell membrane.

Immunocytochemistry in Oligodendrocyte CulturesAfter oligodendrocyte stimulation, cells were loaded with Mito-

tracker OrangeVR (Molecular Probes; Invitrogen), to identify mito-

chondria, according with manufacturer instructions. Cells were then

fixed with 4% paraformaldehyde, permeabilized in 0.1% Triton

X-100 in phosphate-buffered saline (PBS) and processed for immu-

nofluorescence with anti-Bax-NT antibody (1:50; Upstate). After

washing with PBS, cells were incubated with Alexa FluorVR

488-conjugated secondary antibody (1:50, Molecular Probes) washed

in PBS, mounted with Glycergel (Dako, Glostrup, Denmark), and

visualized with a laser scanning confocal microscope (Olympus Fluo-

view FV500) at the Analytic and High Resolution Microscopy Facil-

ity in the University of the Basque Country. No staining was

detected in control samples run in parallel without primary

antibodies.

Adenosine Receptor Activation in Isolated RatOptic NervesOptic nerves of 12-day-old Sprague-Dawley rats were dissected out

in cold oxygen-saturated artificial CSF containing 126 mM NaCl, 3

mM KCl, 2 mM MgSO4, 26 mM NaHCO3, 1.25 mM NaH2PO4,

2 mM CaCl2, and 10 mM glucose, as previously described (S�anchez-

G�omez et al., 2003). Subsequently, the nerves were placed in a 48-

well plate containing artificial CSF and incubated with adenosine

(100 lM) or 2-CI-IB-MECA (30 lM) for 1 h at 37�C and 5% CO2

in the presence or absence of caffeine (1 mM), MRS 1220 (10 lM),

or vehicle. Following incubation, media was replaced with fresh

oxygen-saturated artificial CSF and incubated for another 2 h at

37�C. Tissue damage was analyzed by measuring lactate dehydrogen-

ase (LDH) release in the incubation media at 1 and 2 h poststimulus,

using the CytoTox 96VR assay (Promega Biotech Ib�erica, Spain). At

the end of the experiment, the optic nerves were homogenized in ice-

cold RIPA buffer supplemented with HaltTM protease and phospha-

tase inhibitor cocktail (Thermo Fisher Scientific, Spain), and sub-

jected to centrifugation (8000 rpm at 4�C for 10 min) to remove

insoluble material. Experiments were performed in duplicate, normal-

ized as a function of protein content, and LDH values expressed as

the percentage of LDH release by control nerves within each assay.

To evaluate tissue damage by histology, we used the double

transgenic mice (hGFAP-EGFP/PLP-DsRed) (generous gift from Dr.

Frank Kirchhoff, University of Saarland, Homburg, Germany).

Optic nerves were extracted and incubated with adenosine receptor

agonists and antagonists, as described above. At the end of experi-

ments, optic nerves were fixed for 1 h in 4% paraformaldehyde,

washed, mounted with ProlLongVR Gold antifade (Invitrogen; molec-

ular probes), and analyzed with a laser scanning confocal microscope

(Olympus Fluoview FV500) at the Analytic and High Resolution

Microscopy Facility in the University of the Basque Country. The

number of z-sections were optimized with Fluo View software, sec-

tions were <0.75 lm thickness, and approximately 10 z-sections for

each optic nerve. The number of cells was quantified by Image J

software.

Oxygen and Glucose Deprivation in Intact OpticNerveOptic nerves from 12-day-old Sprague-Dawley rats were dissected in

oxygen-saturated CSF, as previously described (Domercq et al.,

2010). To simulate ischemia, we replaced external O2 with N2 and

external glucose with sucrose (10 mM) and added 20 lM iodoace-

tate to block glycolysis. After 1 h of ischemia at 37�C, the medium

was replaced with fresh solution containing glucose and equilibrated

with 95% O2, 5% CO2 and the nerves were incubated for another

3 h at 37�C to simulate reperfusion. Tissue damage was analyzed by

measuring LDH release in the incubation medium at different

Gonz�alez-Fern�andez et al.: Adenosine Role in Oligodendrocyte Death

Month 2013 3

poststimulus points, as described above. Data were normalized to

LDH release by control nerves and to protein content.

To analyze cellular damage to oligodendrocytes and astrocytes

in optic nerves under ischemia conditions, we used the double trans-

genic mice (hGFAP-EGFP/PLP-DsRed). After the ischemic insult,

optic nerves were fixed, mounted, analyzed, and quantified as

described in the preceding section.

Oxygen and Glucose Deprivation in AstrocyteCulturesPrimary cultures of cerebral cortical astrocytes were prepared from

newborn P0-P2 Sprague-Dawley rats as described previously (McCar-

thy and de Vellis, 1980) with minor modifications (Alberdi et al.,

2013). For OGD experiments, cells were detached with trypsin and

seeded at 105 cells/well onto poly-D-lysine-coated coverslips. Cultures

were maintained in DMEM with 10% FBS at 37�C in a humidified

atmosphere with 5% CO2. To induce OGD, we replaced external O2

with N2 and external glucose with sucrose (10 mM) and added 20-

lM iodoacetate to block glycolysis. After 1 h in OGD conditions,

astrocytes were returned to fresh medium which was collected at dif-

ferent poststimulus times (3–24 h). The toxicity of these media to

oligodendrocytes was quantified by calcein-AM assay 24 h after incu-

bation. In addition, OGD-induced astrocyte damage was monitored

at different times by measuring LDH release to the culture media.

Electrophysiology RecordingsPropagated compound action potentials (CAPs) in 12 and 30-day-

old Sprague-Dawley rats were evoked using a bipolar silver electrode

placed on one end of the optic nerve. Stimulus pulse (30-ls dura-

tion delivered every 15 s) strength was adjusted to evoke supramaxi-

mal stimulation (Tekkok et al., 2007). CAPs were recorded at 37�C

with a suction electrode connected at the opposite end of the optic

nerve and back-filled with CSF. Optic nerves were perfused (1 mL/

min) with CSF continuously bubbled with 95% O2/5% CO2 or

95%N2/5% CO2 during normoxia and ischemia, respectively. The

signal was amplified 5000 X and filtered at 30 kHz. After recording

control CAPs for 30 min, ischemia was induced for 1 h. CAP recov-

ery after ischemia was recorded 1 h after reoxygenation and reperfu-

sion with glucose. Drugs were applied during all experiments. TTX

(1 lM) was applied at the end of the experiment to obtain the stim-

ulus artifact, which was subtracted from all the records. Area under

the curve of the CAP profile was measured for comparison among

control, ischemia, and ischemia-treated optic nerves.

Western Blotting AnalysisOligodendrocyte cultures or isolated optic nerves were lysed by

homogenation in RIPA buffer supplemented with HaltTM protease

and phosphatase inhibitor cocktail (Thermo Fisher Scientific, Spain).

Samples were boiled in sample buffer for 10 min and subjected to

SDS-PAGE in 4–20% polyacrylamide gels. After electroblotting on

nitrocellulose membranes (Bio-Rad), proteins were visualized using

rabbit anti-Bax-NT (1:1000; Upstate), rabbit anti-Puma (1:500; Cell

signaling), mouse anti-CNPase (1:1000; Sigma), mouse anti-MBP

(1:2000; Millipore), rabbit anti-caspase-3 (1:500; Santa Cruz Tech-

nologies); rabbit anti-GFAP (1:500; Sigma) and rabbit anti-actin

(1:5000; Sigma). After washing, the blots were developed using

HRP-conjugated anti-IgG secondary antibodies (1:5000) and an

enhanced chemiluminescence detection kit, according to the manu-

facturer’s instructions (Supersignal; Thermo Scientific). Images were

acquired with a ChemiDoc MP system (BioRad) and quantified with

Image J Software.

Statistical AnalysisAll data are expressed as the mean 6 S.E.M. (n), where n refers to

the number of cultures assayed. Statistical analysis was carried out

with the Student’s two-tailed t-test except for multiple comparisons,

which were performed with one-way analysis of variance (ANOVA)

and Bonferroni post hoc test. Significance was determined at P <

0.05 and statistical P values are reported in figure legends. Data fit-

ting was performed with the GraphPad Prism 4.0 program.

Results

Oligodendrocytes Express Adenosine Receptors,Enzymes, and TransportersAdenosine exerts important effects in the CNS, but the

expression and function of the adenosine signaling system in

oligodendrocytes is not well-established. We initially used RT-

PCR to analyze the expression of adenosine receptors, the

enzymes adenosine deaminase, (ADA) and adenosine kinase

(ADK), and equilibrative nucleoside transporters, ENT1 and

ENT2. The results demonstrated the expression of all known

adenosine receptors, ADA, ADK, ENT1, and ENT2 in the

FIGURE 1: Expression of adenosine receptors, transporters, andenzymes in cultured oligodendrocytes derived from rat opticnerve. RT-PCR detection of mRNAs for adenosine receptors (A1,A2a, A2b, A3), adenosine transporters (ENT1 and ENT2), andenzymes, adenosine deaminase (Ada) and adenosine kinase(Adk) in (A) rat brain (positive control), (B) in optic nerve, and (C)cultured oligodendrocytes from optic nerve of 12-day-old rats.

4 Volume 00, No. 00

brain, as expected, as well as in optic nerves and in cultures

of oligodendrocytes (98% of O41/GalC1cells; Fig. 1A–C).

Adenosine Receptor Activation InducesOligodendrocyte Death via A3 ReceptorAdenosine receptors have been detected in different tissues,

and their activation is responsible for diverse biological

responses in various cellular types (Melani et al., 2009; Miya-

zaki et al., 2008). To analyze the sensitivity of our cultures to

adenosine, oligodendrocytes were incubated for 15 min with

increasing concentrations of adenosine (from 10 lM to 1

mM) and cell viability was measured after 24 h. We found

that, at concentrations above 10 lM, adenosine was toxic to

oligodendrocytes in a concentration-dependent manner (Fig.

2A). In contrast, when oligodendrocytes were pretreated with

the adenosine receptor antagonist, caffeine, adenosine toxicity

was prevented (Fig. 2B,F). These results strongly suggested

that extracellular adenosine led to oligodendroglial death

FIGURE 2: Adenosine induces oligodendrocyte death. Oligodendrocyte viability was measured 24 h after the treatment, by fluorimetrywith calcein-AM (1 lM). (A) Optic nerve oligodendrocytes exposed (15 min) to increasing adenosine concentrations (10 lM–1 mM)caused concentration-dependent oligodendrocyte death. The percentage of dead cells was compared to that in control untreated cells.(B) Blocking adenosine receptors with caffeine (300 lM) significantly reduced cell death caused by adenosine (10 lM–1 mM). (C) Forsko-lin (50 lM), an adenylate cyclase activator, reduced oligodendrocyte death when added prior to adenosine stimulation (10 lM–1 mM).(D) Specific adenosine receptor agonists (100 lM or 1 mM) showed that toxicity was mainly mediated by A3 receptors. (E) The selectiveblockade of adenosine A3 receptor with specific antagonist MRS 1220 (10 lM) diminished significantly cell death induced by adenosine(100 lM). Bars represent the mean 6 SEM (n � 3) (*P < 0.05, **P < 0.01, ***P < 0.001 compared to control, in figure A and D, and com-pared to adenosine alone, in figures B, C, and E; ANOVA test). (F) Representative images of oligodendrocyte viability as assessed withcalcein (green fluorescence) 24 h after activation of A3 receptors (adenosine 100 lM or 2 CI-IB-MECA 30 lM) in the presence or absenceof antagonists (caffeine 300 lM, or MRS1220 10 lM). Scale bar, 40 lm. [Color figure can be viewed in the online issue, which is availableat wileyonlinelibrary.com.]

Gonz�alez-Fern�andez et al.: Adenosine Role in Oligodendrocyte Death

Month 2013 5

through activation of specific receptors, which were function-

ally expressed in these primary oligodendrocytes derived from

rat optic nerve.

We next determined which receptor subtypes mediated

the observed toxicity. Because A1 and A3 adenosine receptors

inhibit adenylate cyclase (Fredholm et al., 2000), we tested

whether forskolin, an adenylate cyclase activator, was protec-

tive. Indeed, adenosine-induced oligodendrocyte death was

totally abolished when cells were pretreated with forskolin

(50 lM; 30 min) before adding adenosine, at all concentra-

tions tested (Fig. 2C). These results suggested that A1 and/or

A3 receptors-mediated adenosine-induced oligodendrocyte

death.

To further assess the adenosine receptors involved in

adenosine injury to oligodendrocytes, we treated them with

receptor-specific agonists (at 100 lM and 1 mM) for each

adenosine receptor subtype. Selective activation of the A1,

A2a, or A2b receptor showed very low toxicity (Fig. 2D); how-

ever, A3 receptor activation with 2-CI-IB-MECA elicited

robust oligodendrocyte death (44.6 6 4.3% and 73.9 6

4.5% at 100 lM and 1 mM, respectively; Fig. 2D,F). Fur-

thermore, adenosine-mediated toxicity was prevented in pres-

ence of MRS 1220 (10 lM), a specific A3 receptor

antagonist (Fig. 2E,F).

In line with the above observations, A3 receptor agonist

2-CI-IB-MECA caused concentration-dependent oligodendro-

cyte death (Fig. 3A). Oligodendrocytes also express adenosine

transporters and both adenosine and agonists of adenosine

receptors can be transported into the cell and act intracellu-

larly. However, 2-CI-IB-MECA toxicity was not caused via

this mechanism since dipyridamole (10 lM; 10 min), an

adenosine transporter inhibitor (Meester et al., 1998), was

not protective, whereas MRS 1220 was (Fig. 3B).

Taken together, these results clearly indicated that A3

receptor was the main receptor subtype involved in oligoden-

drocyte death induced by adenosine.

Adenosine Receptor Activation Causes OxidativeStress and Disrupts Mitochondrial MembranePotential in OligodendrocytesExtracellular adenosine can cause oxidative stress and altera-

tions in the mitochondrial membrane potential, which may

lead to mitochondrial-dependent cell death (Galluzzi et al.,

2009; Ijima et al., 2006; Sai et al., 2006; Xu et al., 2005).

Therefore, we examined whether adenosine perturbed mito-

chondrial functions in oligodendrocytes. To that end, we eval-

uated levels of intracellular ROS with DCFDA and

mitochondrial membrane depolarization with the JC-1.

When adenosine was applied for 15 min, a concentration-

dependent increase in intracellular ROS was observed in oli-

godendrocytes. This increase was abolished when oligoden-

drocytes were stimulated in the presence of the antagonist,

caffeine (300 lM, 10 min; Fig. 4A). Consistent with the tox-

icity assays, we found that the A3 receptor agonist, 2-CI-IB-

MECA (30 lM, 15 min) also increased ROS levels (Fig. 4B).

Again, this increase was completely abolished in the presence

of the A3 receptor-specific antagonist, MRS 1220 (10 lM;

pretreated for 10 min). Moreover, adenosine (100 lM or 1

mM, 15 min), depolarized the mitochondrial membrane

potential, shown by the reduction in JC-1 fluorescence in oli-

godendrocytes (Fig. 4C). This effect was mimicked by the

specific activation of A3 receptors with 2-CI-IB-MECA (30

lM; 15 min). The A3 receptor activation provoked a signifi-

cant depolarization that was prevented in the presence of the

A3 receptor-specific antagonist MRS 1220 (10 lM) (Fig.

4D). Overall, these results suggested that the integrity of

FIGURE 3: Selective activation of A3 receptors is toxic to oligo-dendrocytes. Oligodendrocyte viability was measured 24 h afterthe treatment, by fluorimetry with calcein-AM (1 lM). (A)Concentration-dependent oligodendrocyte death caused byexposure to the A3 receptor agonist, 2-CI-IB-MECA (10 lM–1mM; 15 min). Bars represent the mean 6 SEM (n � 3). (**P <0.01, ***P < 0.001 compared to untreated control cells). (B) Tox-icity was blocked by A3 receptor antagonist, MRS 1220 (10 lM),but not by adenosine transporter inhibitor, dipyridamole (10lM). Bars represent the mean 6 SEM (n � 3). (*P < 0.05, com-pared to 2-CI-IB-MECA alone; ANOVA test).

6 Volume 00, No. 00

oligodendroglial mitochondria was compromised by extracel-

lular adenosine and that A3 receptors mediated the mitochon-

drial damage detected.

Adenosine Induces Oligodendrocyte Death throughCaspase-dependent and Caspase-independentPathwaysBecause mitochondrial alterations constitute a crucial step in

apoptosis, we next examined whether adenosine toxicity caused

oligodendrocyte apoptosis. Apoptosis is characterized by mor-

phological features, including the loss of plasma membrane

asymmetry, externalization of phosphatidylserine (PS), conden-

sation of the nucleus, and fragmentation of DNA. When oli-

godendrocytes were exposed to adenosine (10 lM, 100 lM,

or 1 mM) or 2-CI-IB-MECA (30 or 100 lM), they showed

early signs of apoptosis. One sign was Annexin V-FITC bind-

ing to PS residues on the outer leaflet of the lipid bilayer (Fig.

5A). In all cases, cells were counterstained with Hoechst 33258

to simultaneously evaluate nuclear condensation.

Next, we evaluated cell death in the presence of differ-

ent caspase inhibitors that block caspase proteolytic activity.

The broad-spectrum caspase inhibitor, ZVAD (50 lM),

reduced oligodendrocyte death induced by adenosine (13.6 6

2% for adenosine 100 lM vs. 7.9 6 2.2% in the presence of

ZVAD; 12.5 6 1.5% for adenosine 1 mM vs. 4.3 6 2.9%

in the presence of inhibitor). Moreover, selective inhibition of

caspase-8 with IETD (100 lM) significantly reduced cell

death caused by 10 lM adenosine, but it did not prevent tox-

icity at higher concentrations of adenosine (Fig. 5B). That

result suggested that the concentration of adenosine deter-

mined the activation of different apoptotic death pathways.

In contrast, oligodendroglial damage by adenosine was signifi-

cantly prevented, or even abolished, by the mitochondria-

associated caspase-9 inhibitor, LEHD (100 lM) (Fig. 5B).

Together, these findings suggested that adenosine toxicity

involved the activation of extrinsic and/or intrinsic apoptotic

pathways, depending on the intensity of the stimulus.

On the other hand, death caused by selective activation

of adenosine A3 receptors with a low concentration of 2-CI-

IB-MECA (30 lM) could be significantly diminished by

selective inhibition of caspase-3 (DEVD; 100 lM) or

mitochondrial-associated caspase-9 (LEHD; 100 lM), but

not by inhibition of caspase-8 with IETD (Fig 5C). In con-

trast, death caused by activation of adenosine A3 receptors

FIGURE 4: Adenosine receptor activation triggers oxidative stress and depolarization of the mitochondrial membrane. (A, B) ROS levelswere quantified with CM-H2DCFDA (20 lM), immediately after cultured oligodendrocytes were incubated with agonists. (A) ROS produc-tion induced by adenosine (100 lM–1 mM; 15 min) was blocked in the presence of caffeine. (B) ROS production induced by the adenosineA3 receptor-specific agonist (2-CI-IB-MECA) was blocked in the presence of adenosine-specific antagonist (MRS 1220). (C, D) Mitochondrialmembrane potential was detected by fluorimetry of cultured cells loaded with JC-1 (3 lM) immediately after agonist stimulation. Mito-chondrial membrane potential in control cell cultures was taken as 100% (dashed line). (C) Concentration-dependent depolarization ofmitochondrial membrane with adenosine stimulation. (D) Depolarization of the mitochondrial membrane induced by adenosine A3 recep-tor agonist, 2-CI-IB-MECA, was blocked in the presence of A3 antagonist, MRS 1220. Bars represent the mean 6 SEM (n � 3). *P < 0.05,**P < 0.01, ***P < 0.001, compared to non treated cells; #P < 0.05, ##P < 0.01, ###P < 0.001, compared to agonist alone.

Gonz�alez-Fern�andez et al.: Adenosine Role in Oligodendrocyte Death

Month 2013 7

with high concentrations of 2-CI-IB-MECA (100 lM) could

not be prevented by any of the caspase inhibitors (Fig 5D).

The limited ability of caspase inhibitors to block

oligodendroglial damage indicated that another, caspase-

independent, apoptotic, and/or necrotic death pathway may

have contributed to the toxic effect of adenosine receptor acti-

vation. To investigate the possibility of caspase-independent

apoptosis, we tested the effects of the nuclear enzyme poly

(ADP-ribose) polymerase-1 (PARP-1), an important mediator

of apoptosis-inducing factor (AIF) released from the mito-

chondria. PARP-1 was shown to be associated with apoptotic

mechanisms independent of caspases. To that end, we treated

oligodendrocytes with DPQ (30 lM), an inhibitor of PARP-

1, before adding 100 lM and 1 mM adenosine. We observed

that DPQ reduced cell death induced by adenosine (Fig.

5E), which indicated that activation of PARP-1 contributed

to adenosine-triggered oligodendrocyte apoptosis. In contrast,

PARP-1 inhibition did not reduce oligodendroglial toxicity

elicited by the selective activation of adenosine A3 receptors

with low and high 2-CI-IB-MECA concentrations (Fig. 5F).

Taken together, these results indicated that adenosine

receptor activation killed oligodendrocytes by both caspase-

dependent and caspase-independent mechanisms. In addition,

selective A3 receptor activation may cause oligodendrocyte

death via caspase-dependent apoptosis or by necrosis, depend-

ing on the intensity of the stimulus.

FIGURE 5: Adenosine receptors mediate caspase-dependent and caspase-independent apoptosis in oligodendrocytes. (A) Annexin-V-FLUO-stained oligodendrocytes show induction of phosphatidylserine residues (green) in the plasma membrane caused by adenosine (10lM-1 mM) or 2-CI-IB-MECA (30 or 100 lM). Scale bar: 20 lm. (B–F) Oligodendrocytes were pretreated with inhibitors of caspases (100lM; 30 min) before adenosine receptor activation. Cell viability was measured by fluorimetry with calcein-AM (1 lM). The percentage ofdead cells was compared to that in control untreated cells. (B) Adenosine (10 lM–1 mM) toxicity was prevented with caspase-8 (IETD;only at low adenosine) and caspase-9 (LEHD) inhibitors. (C, D) Inhibitors of caspases-3 (DEVD) and -9 (LEHD), but not inhibition ofcaspase-8 (IETD), prevented toxicity caused by mild 2-CI-IB-MECA activation of A3 receptors (30 mM). In addition, caspase were noteffective in preventing toxicity caused by higher 2-CI-IB-MECA (100 lM). (E, F) The PARP-1 blocker, DPQ (30 lM), inhibited caspase-independent apoptosis induced by adenosine (D), but DPQ did not prevent toxicity by 2-CI-IB-MECA (E). Bars represent the mean 6

SEM. (n � 3). *P < 0.05, **P < 0.01, ***P < 0.001 compared to cells treated with agonist (ANOVA test). [Color figure can be viewed inthe online issue, which is available at wileyonlinelibrary.com.]

8 Volume 00, No. 00

A3 Receptor-Mediated Apoptosis Involves Bax andPuma ActivationThe results described above suggest that milder A3 receptor

activation leads to oligodendrocyte apoptosis via

mitochondria-associated pathways. Because of that, we inves-

tigated the role of proapoptotic members of the Bcl-2 protein

family, which plays a key role in mitochondria-dependent

intrinsic pathways (Green and Kroemer, 2004). Previous data

obtained in our laboratory illustrated that proapoptotic pro-

tein Bax mediates oligodendrocyte apoptosis triggered by exci-

totoxicity after AMPA and kainate receptor activation

(S�anchez-G�omez et al., 2011). To test, if this were the case

after A3 receptor activation, we analyzed changes in the levels

of activated-Bax protein and of Puma, a proapoptotic BH3-

only member of Bcl-2 family, which can help Bax in execut-

ing apoptosis (Gallenne et al., 2009; Steckley et al., 2007).

Indeed, Western blot analysis showed an early and sustained

(10 min to several hours) increase in the levels of both activated-

Bax and Puma after selective activation of A3 receptors (2-CI-IB-

MECA, 30 lM; 10 min; Fig. 6A–C) in cultured oligodendro-

cytes. These effects were prevented in presence of adenosine A3

receptor antagonist, MRS 1220 (Fig. 6E,F; 10 lM).

Bax is a cytosolic protein that, on activation by apopto-

tic stimuli, translocates into the mitochondrial membrane. To

test this phenomenon, oligodendrocytes were treated with 2-

CI-IB-MECA 30 lM, for 15 min, in the absence or presence

of antagonist MRS 1220 (10 lM) and loaded with Mito-

tracker Orange to identify mitochondria (in red). Under these

conditions, immunofluorescence labeling of activated-Bax in

oligodendrocytes treated with the A3 receptor agonist showed

a robust elevation of this proapoptotic protein along with its

translocation into mitochondria, which was prevented by

MRS1220 (Fig. 6D).

Taken together, these results suggest that Bcl-2 proapop-

totic Bax and Puma are involved in the induction of oligoden-

droglial apoptosis triggered by adenosine A3 receptor activation.

A3 Receptor Activation Damages Oligodendrocytesand Myelin in the Optic NerveTo further investigate the deleterious effects of A3 receptor

activation in a more integral preparation, we analyzed the

effect of adenosine receptor-mediated injury to intact optic

nerves, a preparation which has been previously used to test

glutamate receptor-mediated toxicity (Domercq et al., 2010;

Mato et al., 2013; S�anchez-G�omez et al., 2003). Freshly iso-

lated P12 rat optic nerves were exposed to adenosine (100

lM) or 2-CI-IB-MECA (30 lM) in oxygen-saturated artifi-

cial CSF for 1 h, and cell damage was monitored by meas-

uring LDH release after 2 h of terminating the stimuli.

Incubation of optic nerves with adenosine induced a

marked damage that was absent in the presence of caffeine (1

mM; Fig. 7A). In addition, selective activation of A3 receptor

triggered nerve injury that was abolished by co-treatment

with MRS 1220 (10 lM; Fig. 7A).

Next, we evaluated optic nerve damage by measuring

effector caspase-3 levels as caspase-dependent apoptosis is a

major mechanism of adenosine toxicity in oligodendrocytes.

Thus, Western blot analysis of optic nerves exposed to adeno-

sine or CI-IB-MECA using an antibody that recognizes

procaspase-3 showed a reduction in the levels of this mole-

cule. These effects were also abolished when caffeine or MRS

1220 was added to the medium during receptor stimulation

(Fig. 7B,F).

To test whether oligodendrocytes present in the optic

nerves were damaged as consequence of adenosine receptor

activation, Western blot analysis was performed with anti-

CNPase antibody to label mature and immature oligodendro-

cytes. We found a significant decrease in CNPase levels in the

optic nerves incubated with adenosine (100 lM) or selective

agonist of A3 receptor 2-CI-IB-MECA (30 lM) which was

prevented in the presence of caffeine (1 mM) or MRS 1220

(10 lM; Fig.7D,F). To further investigate myelin damage, we

also evaluated myelin basic protein (MBP) levels. Unexpect-

edly, the results showed that treatment with adenosine did

not modify MBP levels at the time points analyzed. However,

selective activation of A3 receptors caused a significant loss of

MBP, which was not observed after co-incubation with antag-

onist MRS 1220 (Fig. 7E,F). Parallel analysis of GFAP levels

revealed no changes in this astrocytic protein (Fig. 7C,F).

Finally, we examined the effects of adenosine and 2-CI-

IB-MECA, for 1 h, in the isolated intact optic nerves from

double transgenic mice (hGFAP-EGFP/PLP-DsRed) which

allows for simultaneous identification of somata and processes

of oligodendrocytes and astrocytes. High-resolution confocal z

sections of whole-mounted nerves showed that adenosine or

the selective activation of A3 receptors caused a marked loss

of oligodendrocytes and their processes, which was prevented

in the presence of caffeine (1 mM) or MRS 1220 (10 lM),

respectively. Conversely, astrocytes in the optic nerves were

not affected after their incubation with agonists of adenosine

receptors (Fig. 7G–I).

These results in whole optic nerves ex vivo further assess

the observations described above showing that oligodendro-

cytes are vulnerable to sustained activation of A3 receptors by

recruiting apoptotic pathways.

A3 Receptor Blockade Prevents OligodendrocyteDamage and Myelin Loss after Optic NerveIschemiaBecause oligodendrocytes (Goldenberg-Cohen et al., 2005;

Shibata et al., 2000) and astrocytes (Bjorklund et al., 2008;

Cao et al., 2010; Lee et al., 2009) are vulnerable to ischemia,

Gonz�alez-Fern�andez et al.: Adenosine Role in Oligodendrocyte Death

Month 2013 9

FIGURE 6: Adenosine A3 receptor induces Bax and Puma activation and its translocation into mitochondria. Histograms (A, B) and immu-noblots (C) showing the time course of the increase of Bax and Puma protein levels after exposure of oligodendrocytes to 2-CI-IB-MECA(30 lM, 15 min). Values in A and B were normalized to the actin levels and expressed as the percentage of controls (100%). Bars repre-sent the mean 6 SEM (n � 3). *P < 0.05, **P < 0.01, ***P < 0.001, compared to control; ANOVA test. (D) Bax translocation from thecytosol into the mitochondrial membrane after treatment with 2-CI-IB-MECA (30 lM) is abolished in the presence of the A3 receptorantagonist (MRS 1220; 10 lM). Scale bar: 20 lm. (E and F) Histograms and representative blots showing that activation of Bax andPuma by 2-CI-IB-MECA is prevented by MRS 1220. Bars represent the mean 6 SEM (n � 3). **P < 0.05; compared to control; #P < 0.05;compared to cultures treated with agonist; ANOVA test. [Color figure can be viewed in the online issue, which is available atwileyonlinelibrary.com.]

FIGURE 7: Activation of A3 receptor causes apoptotic death, oligodendrocyte damage, and myelin loss in isolated rat optic nervesand in double transgenic mice (hGFAP-EGFP/PLP-DsRed). Isolated optic nerves were treated for 1 h with adenosine (100 lM) or2-CI-IB-MECA (30 lM) alone or in the presence of caffeine (1 mM) or MRS 1220 (10 lM) respectively. (A) Cell viability was meas-ured by the LDH release assay at 1 and 2 h poststimulus. Adenosine and A3 agonist, 2-CI-IB-MECA, was toxic to the optic nerve,an effect attenuated in the presence of antagonists. Bars represent the mean 6 SEM (n � 3). *P < 0.05; **P < 0.01; ***P <0.001, compared to control; #P < 0.05, ##P < 0.01 compared to optic nerves treated with adenosine or 2-CI-IB-MECA; ANOVA.(B–F) Western blot analysis showing an A3 receptor-mediated reduction of total caspase-3 levels suggesting its conversion toactive caspase-3, as well as a decrease in CNPase and MBP, markers of oligodendrocytes and myelin. However, levels of theastrocyte marker GFAP were not modified by the stimuli. Bars represent the mean 6 SEM (n � 3). *P < 0.05; **P < 0.01, com-pared to control; #P < 0.05; ##P < 0.01, compared to cultures treated with agonist alone; non paired Student’s t test. (G–I) Pho-tographs and histograms illustrating representative fields of z-stacks of optic nerves from hGFAP-EGFP/PLP-DsRed transgenic micestimulated with adenosine (100 lM) or 2-CI-IB-MECA (30 lM). Note a loss of oligodendrocytes (red; Fig. 7G,H), which is pre-vented in the presence of caffeine (1 mM) or MRS 1220 (10 lM); whereas the astrocyte population (green; Fig. 7G,I) is unaltered.Scale bar: 100 lm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Gonz�alez-Fern�andez et al.: Adenosine Role in Oligodendrocyte Death

Month 2013 11

we hypothesized that activation of A3 receptors may mediate

oligodendroglial postischemic injury. To examine that possi-

bility, we induced ischemia (1 h, 37�C) by replacing external

O2 with N2 and simulated energy deprivation by blocking

glycolysis with iodoacetate in a medium where glucose was

replaced with sucrose. After ischemia, nerves returned to nor-

mal oxygenated artificial CSF (reperfusion) for up to 3 h.

LDH release assays showed that the presence of caffeine (1

mM) and MRS 1220 (10 lM) during ischemia and reperfu-

sion attenuated optic nerve damage (Fig. 8A). In addition,

immunoblotting experiments to evaluate postischemic injury

revealed that these A3 receptor antagonists also prevented the

activation of caspase-3, an indication of reduced overall cell

damage to the optic nerve (Fig. 8B,F), and protected oligo-

dendrocytes and myelin as assessed by evaluating CNPase and

MBP levels (Fig. 8D,E,F). In contrast, GFAP levels were not

significantly affected by ischemia (Fig. 8C,F).

Because adenosine A2a receptors may mediate oligoden-

drocyte damage after ischemia, we induced OGD in optic

nerves in the presence of the A2a antagonist. The results indi-

cated that blocking A2a receptors did not attenuate OGD-

induced damage to the optic nerve as analyzed by LDH

release and expression of total caspase-3, CNPase, and MBP

(Supporting Information, Fig. S1). These results show that

A2a does not contribute to the observed oligodendroglial

injury. Moreover, GFAP levels were unaffected by OGD in

absence or presence of antagonist SHC 58261 (Supporting

Information, Fig. S1).

In addition, we evaluated CAPs to test whether adeno-

sine receptor blockade could reverse the functional loss

induced by ischemia to optic nerves isolated from 12- and

30-day-old rats (Fig. 8G–I). In control, vehicle-treated nerves,

CAPs virtually disappeared in optic nerves electrically stimu-

lated after 1 h of ischemia, and partially recovered 1 h later

during reperfusion. Notably, caffeine and MRS 1220

improved the recovery of axon function after the ischemic

insult at the two ages analyzed (Fig. 8G,H).

Finally, to evaluate changes in number of oligodendro-

cytes, myelinating processes, and astrocytes, we examined

OGD-induced damage in the isolated intact optic nerve from

double transgenic mice hGFAP-EGFP/PLP-DsRed (Fig. 9). In

accordance with the results observed by immunoblotting in

rat optic nerves, we observed a marked and selective loss of

oligodendrocytes after OGD which was prevented in the pres-

ence of caffeine or MRS 1220. Expression levels of GFAP

and astrocytic morphology were not modified by these ische-

mic conditions (Fig. 9A–C).

Although we did not find evidence of astrocyte damage

after OGD in the optic nerve, it is possible that astrocytes

contribute to OGD-triggered oligodendrocyte death by acti-

vating A3 receptors. To evaluate this possibility, we induced

ischemia in astrocyte cultures and, besides analyzing their vul-

nerability by LDH release assays, challenged oligodendrocytes

with astrocyte-conditioned media during ischemia (Support-

ing Information, Fig. S2). The data indicated that cultured

astrocytes were only slightly vulnerable to ischemia and that

FIGURE 7: (Continued)

12 Volume 00, No. 00

FIGURE 8: Blockade of A3 receptors attenuates ischemic damage to isolated optic nerves. (A) Ischemic damage at 1–3 h to isolated opticnerves, quantified by LDH release, was reduced by caffeine (1 mM) and MRS 1220 (10 lM). Bars represent the mean 6 SEM (n � 3). *P< 0.05; **P < 0.01; compared to vehicle (OGD). (B–F) Histograms of immunoblot quantification and representative blots of optic nervesafter ischemic insult shows that A3 receptor blockers prevent ischemia-induced activation of caspase-3, oligodendrocyte damage, andmyelin loss, as assessed by CNPase and MBP markers after 1 h of ischemia followed by 3 h of reperfusion. In contrast, the astrocytemarker GFAP is unaltered. Bars represent the mean 6 SEM (n � 3). *P < 0.05; compared to control; #P < 0.05; ##P < 0.01, compared tonerves treated with vehicle (OGD); nonpaired Student’s t test. (G,H) Bar graphs summarizing the extent of CAP recovery at 1 h afterischemia followed by 1 h of reperfusion (OGD) in optic nerves from 12-days-old rats (G) or 30-days-old rats (H) treated with vehicle or inthe presence of 1 mM caffeine or 10 lM MRS 1220 (*P < 0.05; **P < 0.01; ***P < 0.005). (I) Representative traces of CAPs before, dur-ing, and after the ischemic insult, in presence of MRS 1220 antagonist.

MRS 1220 prevented this effect (Supporting Information, Fig.

S2A,C,E). In addition, astrocytes under these conditions

released to the medium factors which caused low levels of oli-

godendrocyte death which was abolished by the A3 receptor

antagonist MRS1220 (Supporting Information, Fig. S2B,D,F).

Together, these results indicate that ischemia induces the

activation of apoptotic pathways, oligodendrocyte damage,

and functional loss in optic nerves which are ameliorated by

blocking adenosine A3 receptors.

DISCUSSION

The present study provides evidence that sustained activation

of A3 adenosine receptors induces oxidative stress and mito-

chondrial membrane depolarization in oligodendrocytes,

which ultimately leads to cell demise in dissociated cultures

and in the optic nerve ex vivo. In addition, we show that A3

receptor blockade attenuates ischemic injury to optic nerve

oligodendrocytes and myelin and favors functional recovery.

Although adenosine receptors have previously been

detected in brain, few studies have analyzed the adenosinergic

system in oligodendroglia. Preceding studies only showed the

expression of A1 and A2a receptors in oligodendrocytes (Mel-

ani et al., 2009; Othman et al., 2003). In the current study,

we confirm those results and provide evidence that in addi-

tion these cells also express A2b and A3 receptors, as well as

adenosine transporters, ENT1 and ENT2. In turn, key

enzymes for adenosine metabolism, ADA and ADK, are also

expressed by oligodendrocytes. Thus, it appears that adeno-

sine signaling in these cells is more complex than previously

thought.

Activation of adenosine receptors can be protective or

deleterious after brain insults and this dual effect depends on

the cell type and receptor subtype involved, as well as on the

time of activation and associated pathology conditions (Bjor-

klund et al., 2008; Melani et al., 2009; Pugliese et al., 2009;

Tsutsui et al., 2004; Wei et al., 2013). In the current study,

we observed that acute exposure to adenosine was toxic to

optic nerve oligodendrocytes mainly via activation of A3

adenosine receptors. Previous studies have shown that adeno-

sine inhibits oligodendrocyte progenitor proliferation, and

FIGURE 9: Ischemia-induced oligodendrocyte cell damage in isolated optic nerves from hGFAP-EGFP/PLP-DsRed transgenic mice. (A)Representative images of optic nerves from transgenic mice incubated in normal aCSF or in OGD medium for 1 h and reperfusion (3 h)in absence or presence of caffeine or MRS 1220. (B,C) Quantification of z-stacks of optic nerves from hGFAP-EGFP/PLP-DsRed trans-genic mice under different conditions. The significant oligodendrocyte loss (red; B) induced by OGD was reverted in the presence of caf-feine (1 mM) or MRS 1220 (10 lM). Astrocytes appearance (green; C) was unaltered by OGD. Scale bar: 100 lm. [Color figure can beviewed in the online issue, which is available at wileyonlinelibrary.com.]

14 Volume 00, No. 00

promotes their migration, differentiation, and myelination

(Stevens et al., 2002). Moreover, oligodendrocyte progenitors

express A1 receptors which stimulate their migration, but do

not affect cell viability, proliferation, or differentiation (Oth-

man et al., 2003). Likewise, mature brain oligodendrocytes

have A1 receptors and their activation does not alter cell via-

bility (Othman et al., 2003). Consistent with those findings,

we observed that selective A1 receptor activation did not

induce toxicity in cultured oligodendrocytes obtained from

the rat optic nerve. On the other hand, it has been described

that A2a receptors mediate deleterious effects to striatal oligo-

dendrocytes as their blockers reduce JNK activation in these

cells and prevent myelin disorganization after ischemia (Mel-

ani et al., 2009). In agreement with these observations, we

also have detected that selective activation of A2a receptors in

rat optic nerve oligodendrocytes causes moderate cell death,

which is reverted by its specific antagonist (Fig. 2D and data

not shown). Further analysis will be needed to determine the

mechanisms underlying this adverse effect.

A3 adenosine receptors play a dual role as they are

involved in both cell protective and deleterious mechanisms

(Cheong et al., 2013). Thus, in nonneuronal cells, mild acti-

vation of A3 receptors protected against cell death while toxic

effects were observed as a consequence of intense activation

(Yao et al. 1997). Our results show that A3 receptors are the

major mediator of adenosine toxicity in oligodendrocytes as

the selective agonist 2-CI-IB-MECA impaired oligodendro-

cyte viability and the selective antagonist MRS 1220 was pro-

tective of adenosine damage.

We observed that incubation of oligodendrocytes with

activation of A3 receptors-induced ROS generation and mito-

chondrial depolarization, two events that typically lead to

mitochondrial disruption and the release of soluble proapop-

totic proteins into the cytosol (Brady et al., 2004; Cao et al.,

2011). Previous studies indicated that adenosine could trigger

apoptosis by activating extrinsic (Tourneur et al., 2008) and

intrinsic pathways (Sai et al., 2006). In the present study, we

found that apoptotic cell death was initiated by adenosine

mainly through the intrinsic apoptotic pathway, with involve-

ment of mitochondria/caspase-9. The extrinsic pathway, with

activation of caspase-8, was recruited only with low concen-

trations of adenosine. In addition, selective activation of A3

receptors killed oligodendrocytes exclusively via the caspase-9/

caspase-3 pathway. These findings are consistent with data

from other studies that examined 2-Cl-IB-MECA stimulation

of A3 adenosine receptors as reported in human lung cancer

cells (Kim et al., 2008; Otsuki et al., 2012) and in human

glioma cells (Kim et al., 2012).

Apoptosis depends on the balance between Bcl-2 family

members. A recent study showed that A3 receptor activation

reduced the expression of antiapoptotic molecules, while the

expression of Bax protein was steadily increased (Aghaei

et al., 2011). In oligodendrocytes, we previously described

that Bax is basally expressed and localized in the cytosol but,

under excitotoxic insults, this proapoptotic protein is upregu-

lated and translocated into the outer mitochondrial mem-

brane (S�anchez-G�omez et al., 2011). Similarly, in the current

study, we found that A3 receptor-mediated insults elevate

active Bax levels, and that active Bax translocates into the

mitochondria whereby may associate with Puma, a “BH3

domain-only” proapoptotic protein, to execute apoptosis fol-

lowing well-established signaling paradigms (Itoh et al., 2003;

Ren et al., 2010).

Consistent with the results in dissociated oligodendrocyte

cultures, both adenosine and selective agonist of A3 receptors

caused oligodendrocyte and myelin damage in isolated rat

optic nerves, activating apoptotic pathways with involvement

of caspase-3. Surprisingly, A3 receptor stimulation induced a

significant reduction in the expression of CNPase and MBP,

whereas adenosine mainly promoted reduction in CNPase

presence, without MBP loss. This apparent contradiction

might be explained by the presence in the myelin sheath of

adenosine receptors, other than A3 subtype, which counteract

the deleterious effects of A3 receptor overactivation by adeno-

sine under the conditions assayed in this study. In turn,

CNPase1 premyelinating oligodendrocytes are abundant in

young optic nerves and therefore CNPase may be a better

marker than MBP for monitoring oligodendrocyte damage.

Our data in vitro and in isolated rat optic nerves ex vivosuggest that enhanced activation of adenosine receptors, and

mainly A3 receptor subtype, in oligodendrocytes may contrib-

ute to the pathophysiology of these cells. To test this hypoth-

esis, we induced ischemic insults as adenosine levels are

dramatically increased as a consequence of energy metabolism

failure and A3 receptors are known to contribute to this

pathology (Latini and Pedata, 2001; Pedata et al., 2007). In

addition, in developing CNS white matter tracts the period

of early myelination is characterized by a heightened sensitiv-

ity to ischemic conditions (Fern and M€oller, 2000; Melani

et al., 2009). Indeed, we observed that ischemic injury in iso-

lated optic nerves involves A3 adenosine receptors as blockade

with caffeine and MRS 1220 prevented tissue damage,

caspase-3 activation, as well as oligodendrocyte and myelin

injury. Consistent with these results, targeting adenosine

receptors was shown to have therapeutic potential in hypoxia

(Melani et al., 2009); however, and in contrast with our

results, the A2a receptor subtype was involved in protecting

oligodendroglial damage and attenuating neurological deficits

after cerebral ischemia. These differences may be explained by

the different experimental settings used to induce ischemia

(OGD vs. vessel occlusion) and the regions studied (optic

nerve vs. brain).

Gonz�alez-Fern�andez et al.: Adenosine Role in Oligodendrocyte Death

Month 2013 15

Astrocytes are also susceptible to ischemic damage (Cao

et al., 2010; Gabryel et al., 2010; Lee et al., 2009). However,

we did not find evidence of astrocyte alterations after overac-

tivation of A3 receptors or in the ischemia conditions assayed

in our study. Yet, we noticed that media conditioned by cul-

tured astrocytes undergoing OGD were slightly toxic to oligo-

dendrocytes and, therefore may contribute to damage induced

by ischemia (Supporting Information, Fig. S2). These find-

ings are in line with previous data indicating that astrocytes

release adenosine during ischemia (Martin et al., 2007).

Caffeine consumption reduces the risk of neurodegener-

ative diseases like Parkinson’s (Blum et al., 2003; Schwarzs-

child et al., 2007), Alzheimer’s (Arendash et al., 2006;

Ribeiro et al., 2002), and MS (Chen et al., 2010). In addi-

tion, there are also multiple evidences about the beneficial

effect of caffeine in stroke models (Back et al., 2007; Ragab

et al., 2004; Riksen et al., 2006). Our results in the optic

nerve expand those observations to CNS axonal tracts and

define A3 adenosine receptors as a target of the neuroprotec-

tive effects of caffeine.

In summary, the results of this study indicate that oligo-

dendrocytes are endowed with the molecular machinery to

process adenosine signals, and that those signals can be detri-

mental to oligodendrocytes, identifying A3 adenosine recep-

tors as a relevant intermediary of oligodendroglial injury and

myelin loss after ischemia-reperfusion insults. In turn, our

findings show that endogenously released adenosine under

pathological conditions may induce white matter demise via

activation of A3 receptors and suggest that A3 receptor antag-

onists have therapeutic potential for treating neurological dis-

eases involving white matter.

Acknowledgment

Grant sponsors: Ministerio de Ciencia e Innovaci�on (E. G.-F.);

Grant sponsor: Ministerio de Ciencia e Innovaci�on,

CIBERNED, Ikerbasque, and Gobierno Vasco; Grant spon-

sor: PASPA-UNAM and Ikerbasque (R.O.A.)

The technical assistance of Silvia Mart�ın, Hazel G�omez,

Saioa Marcos, the staff of the animal facility of the University

of the Basque Country and technical and human support

provided by SGIker (UPV/EHU, MICINN, GV/EJ, ERDF,

and ESF) is gratefully acknowledged.

References

Aghaei M, Panjehpour M, Karami-Tehrani F, Salami S. 2011. Molecular mech-anisms of A3 adenosine receptor-induced G1 cell cycle arrest and apoptosisin androgen-dependent and independent prostate cancer cell lines: Involve-ment of intrinsic pathway. J Cancer Res Clin Oncol 137:1511–1123.

Alberdi E, S�anchez-G�omez MV, Marino A, Matute C. 2002. Ca21 influxthrough AMPA or kainate receptors alone is sufficient to initiate excitotoxicityin cultured oligodendrocytes. Neurobiol Dis 9:234–243.

Alberdi E, Wyssenbach A, Alberdi M, S�anchez-G�omez MV, Cavaliere F,Rodr�ıguez JJ, Verkhratsky A, Matute C. 2013. Ca21-dependent endoplasmicreticulum stress correlates with astrogliosis in oligomeric amyloid b-treatedastrocytes and in a model of Alzheimer’s disease. Aging Cell 12:292–302.

Arendash GW, Schleif W, Rezai-Zadeh K, Jackson EK, Zacharia LC, Cracchiolo JR,Shippy D, Tan J. 2006. Caffeine protects Alzheimer’s mice against cognitive impair-ment and reduces brain beta-amyloid production. Neuroscience 142:941–952.

Back SA, Craig A, Kayton RJ, Luo NL, Meshul CK, Allcock N, Fern R. 2007.Hypoxia-ischemia preferentially triggers glutamate depletion from oligoden-droglia and axons in perinatal cerebral white matter. J Cereb Blood FlowMetab 27:334–347.

Barres BA, Hart IK, Coles HS, Burne JF, Voyvodic JT, Richardson WD, RaffMC. 1992. Cell death and control of cell survival in the oligodendrocyte line-age. Cell 70:31–46.

Bjorklund O, Shang M, Tonazzini I, Dare E, Fredholm BB. 2008. AdenosineA1 and A3 receptors protect astrocytes from hypoxic damage. Eur J Pharma-col 596:6–13.

Blum D, Hourez R, Galas MC, Popoli P, Schiffmann SN. 2003. Adenosinereceptors and Huntington’s disease: Implications for pathogenesis and thera-peutics. Lancet Neurol 2:366–374.

Brady NR, Elmore SP, van Beek JJ, Krab K, Courtoy PJ, Hue L, WesterhoffHV. 2004. Coordinated behavior of mitochondria in both space and time: Areactive oxygen species-activated wave of mitochondrial depolarization. Bio-phys J 87:2022–2034.

Cao Q, Zhang LL, Hu ZY, Wu BY. 2010. [Protective effect of mild hypothermiaon astrocytes with traumatic or ischemic injury]. Nan Fang Yi Ke Da Xue XueBao 30:61–63.

Cao X-h, Zhao S-s, Liu D-y, Wang Z, Niu L-l, Hou L-h, Wang C-l. 2011. ROS-Ca21 is associated with mitochondria permeability transition pore involved insurfactin-induced MCF-7 cells apoptosis. Chem Biol Interact 190:16–27.

Conde SV, Obeso A, Vicario I, Rigual R, Rocher A, Gonzalez C. 2006. Caf-feine inhibition of rat carotid body chemoreceptors is mediated by A2A andA2B adenosine receptors. J Neurochem 98:616–628.

Chen GQ, Chen YY, Wang XS, Wu SZ, Yang HM, Xu HQ, He JC, Wang XT,Chen JF, Zheng RY. 2010. Chronic caffeine treatment attenuates experimen-tal autoimmune encephalomyelitis induced by guinea pig spinal cord homog-enates in Wistar rats. Brain Res 1309:116–125.

Cheong SL, Federico S, Venkatesan G, Mandel AL, Shao YM, Moro S,Spalluto G, Pastorin G. 2013. The A3 adenosine receptor as multifacetedtherapeutic target: Pharmacology, medicinal chemistry, and in silicoapproaches. Med Res Rev 33:235–335.

Dar�e E, Schulte G, Karovic O, Hammarberg C, Fredholm BB. 2007. Modula-tion of glial cell functions by adenosine receptors. Physiol Behav 92:15–20.

Domercq M, Perez-Samartin A, Aparicio D, Alberdi E, Pampliega O, MatuteC. 2010. P2X7 receptors mediate ischemic damage to oligodendrocytes. Glia58:730–740.

Dunwiddie TV, Masino SA. 2001. The role and regulation of adenosine in thecentral nervous system. Annu Rev Neurosci 24:31–55.

Fatokun AA, Stone TW, Smith RA. 2008. Adenosine receptor ligands protectagainst a combination of apoptotic and necrotic cell death in cerebellar gran-ule neurons. Exp Brain Res 186:151–160.

Fern R, Moller T. 2000. Rapid ischemic cell death in immature oligodendro-cytes: a fatal glutamate release feedback loop. J Neurosci 20:34–42.

Fredholm BB, AP IJ, Jacobson KA, Klotz KN, Linden J. 2001. InternationalUnion of Pharmacology. XXV. Nomenclature and classification of adenosinereceptors. Pharmacol Rev 53:527–552.

Fredholm BB, Arslan G, Halldner L, Kull B, Schulte G, Wasserman W. 2000.Structure and function of adenosine receptors and their genes. NaunynSchmiedebergs Arch Pharmacol 362:364–374.

Gabryel B, Bielecka A, Stolecka A, Bernacki J, Langfort J. 2010. Cytosolicphospholipase A(2) inhibition is involved in the protective effect of

16 Volume 00, No. 00

nortriptyline in primary astrocyte cultures exposed to combined oxygen-glucose deprivation. Pharmacol Rep 62:814–826.

Gallenne T, Gautier F, Oliver L, Hervouet E, Noel B, Hickman JA, Geneste O,Cartron PF, Vallette FM, Manon S and others. 2009. Bax activation by theBH3-only protein Puma promotes cell dependence on antiapoptotic Bcl-2family members. J Cell Biol 185:279–290.

Galluzzi L, Blomgren K, Kroemer G. 2009. Mitochondrial membrane perme-abilization in neuronal injury. Nat Rev Neurosci 10:481–494.

Goldenberg-Cohen N, Guo Y, Margolis F, Cohen Y, Miller NR, Bernstein SL.2005. Oligodendrocyte dysfunction after induction of experimental anterioroptic nerve ischemia. Invest Ophthalmol Vis Sci 46:2716–2725.

Green DR, Kroemer G. 2004. The pathophysiology of mitochondrial celldeath. Science 305:626–629.

Hammarberg C, Fredholm BB, Schulte G. 2004. Adenosine A3 receptor-mediated regulation of p38 and extracellular-regulated kinase ERK1/2 viaphosphatidylinositol-30-kinase. Biochem Pharmacol 67:129–134.

Iijima T, Mishima T, Akagawa K, Iwao Y. 2006. Neuroprotective effect of pro-pofol on necrosis and apoptosis following oxygen-glucose deprivation--rela-tionship between mitochondrial membrane potential and mode of death.Brain Res 1099:25–32.

Itoh T, Itoh A, Pleasure D. 2003. Bcl-2-related protein family gene expressionduring oligodendroglial differentiation. J Neurochem 85:1500–1512.

Kim SJ, Min HY, Chung HJ, Park EJ, Hong JY, Kang YJ, Shin DH, Jeong LS,Lee SK. 2008. Inhibition of cell proliferation through cell cycle arrest and apo-ptosis by thio-Cl-IB-MECA, a novel A3 adenosine receptor agonist, in humanlung cancer cells. Cancer Lett 264:309–315.

Kim TH, Kim YK, Woo JS. 2012. The adenosine A3 receptor agonist Cl-IB-MECA induces cell death through Ca(21)/ROS-dependent down regulationof ERK and Akt in A172 human glioma cells. Neurochem Res 37:2667–2677.

Latini S, Pedata F. 2001. Adenosine in the central nervous system: releasemechanisms and extracellular concentrations. J Neurochem 79:463-84.

Lauro C, Di Angelantonio S, Cipriani R, Sobrero F, Antonilli L, Brusadin V,Ragozzino D, Limatola C. 2008. Activity of adenosine receptors type 1 Isrequired for CX3CL1-mediated neuroprotection and neuromodulation in hip-pocampal neurons. J Immunol 180:7590–7596.

Lee WT, Hong S, Yoon SH, Kim JH, Park KA, Seong GJ, Lee JE. 2009. Neuropro-tective effects of agmatine on oxygen-glucose deprived primary-cultured astro-cytes and nuclear translocation of nuclear factor-kappa B. Brain Res 1281:64–70.

Martin ED, Fernandez M, Perea G, Pascual O, Haydon PG, Araque A, CenaV. 2007. Adenosine released by astrocytes contributes to hypoxia-inducedmodulation of synaptic transmission. Glia 55:36–45.

Mato S, S�anchez-G�omez MV, Bernal-Chico A, Matute C. 2013. Cytosolic zincaccumulation contributes to excitotoxic oligodendroglial death. Glia 61:750–764.

McCarthy KD, de Vellis J. 1980. Preparation of separate astroglial and oligo-dendroglial cell cultures from rat cerebral tissue. J Cell Biol 85:890–902.

Meester BJ, Shankley NP, Welsh NJ, Meijler FL, Black JW. 1998. Pharmaco-logical analysis of the activity of the adenosine uptake inhibitor, dipyrida-mole, on the sinoatrial and atrioventricular nodes of the guinea-pig. Br JPharmacol 124:729–741.

Melani A, Cipriani S, Vannucchi MG, Nosi D, Donati C, Bruni P, GiovanniniMG, Pedata F. 2009. Selective adenosine A2a receptor antagonism reducesJNK activation in oligodendrocytes after cerebral ischaemia. Brain 132:1480–1495.

Mills JH, Thompson LF, Mueller C, Waickman AT, Jalkanen S, Niemela J,Airas L, Bynoe MS. 2008. CD73 is required for efficient entry of lymphocytesinto the central nervous system during experimental autoimmune encephalo-myelitis. Proc Natl Acad Sci USA 105:9325–9330.

Miyazaki N, Nakatsuka T, Takeda D, Nohda K, Inoue K, Yoshida M. 2008.Adenosine modulates excitatory synaptic transmission and suppresses neuro-nal death induced by ischaemia in rat spinal motoneurones. Pflugers Arch457:441–451.

O’Meara RW, Michalski JP, Kothary R. 2011. Integrin signaling in oligoden-drocytes and its importance in CNS myelination. J Signal Transduct 2011:354091.

Othman T, Yan H, Rivkees SA. 2003. Oligodendrocytes express functional A1adenosine receptors that stimulate cellular migration. Glia 44:166–172.

Otsuki T, Kanno T, Fujita Y, Tabata C, Fukuoka K, Nakano T, Gotoh A,Nishizaki T. 2012. A3 adenosine receptor-mediated p53-dependent apopto-sis in Lu-65 human lung cancer cells. Cell Physiol Biochem 30:210–220.

Pedata F, Melani A, Pugliese AM, Coppi E, Cipriani S, Traini C. 2007. Therole of ATP and adenosine in the brain under normoxic and ischemic condi-tions. Purinergic Signal 3:299–310

Pugliese AM, Traini C, Cipriani S, Gianfriddo M, Mello T, Giovannini MG,Galli A, Pedata F. 2009. The adenosine A2A receptor antagonist ZM241385enhances neuronal survival after oxygen-glucose deprivation in rat CA1 hip-pocampal slices. Br J Pharmacol 157:818–830.

Ragab S, Lunt M, Birch A, Thomas P, Jenkinson DF. 2004. Caffeine reducescerebral blood flow in patients recovering from an ischaemic stroke. AgeAgeing 33:299–303.

Ren D, Tu HC, Kim H, Wang GX, Bean GR, Takeuchi O, Jeffers JR, ZambettiGP, Hsieh JJ, Cheng EH. 2010. BID, BIM, and PUMA are essential for activationof the BAX- and BAK-dependent cell death program. Science 330:1390–1393.

Ribeiro JA, Sebasti~ao AM, de Mendonca A. 2002. Adenosine receptors in thenervous system: Pathophysiological implications. Prog Neurobiol 68:377–392.

Riksen NP, Zhou Z, Oyen WJ, Jaspers R, Ramakers BP, Brouwer RM,Boerman OC, Steinmetz N, Smits P, Rongen GA. 2006. Caffeine preventsprotection in two human models of ischemic preconditioning. J Am Coll Car-diol 48:700–707.

Sai K, Yang D, Yamamoto H, Fujikawa H, Yamamoto S, Nagata T, Saito M,Yamamura T, Nishizaki T. 2006. A(1) adenosine receptor signal and AMPKinvolving caspase-9/-3 activation are responsible for adenosine-induced RCR-1 astrocytoma cell death. Neurotoxicology 27:458–467.

S�anchez-G�omez MV, Alberdi E, Ibarretxe G, Torre I, Matute C. 2003. Cas-pase-dependent and caspase-independent oligodendrocyte death mediatedby AMPA and kainate receptors. J Neurosci 23:9519–9528.

S�anchez-G�omez MV, Alberdi E, Perez-Navarro E, Alberch J, Matute C. 2011.Bax and calpain mediate excitotoxic oligodendrocyte death induced by acti-vation of both AMPA and kainate receptors. J Neurosci 31:2996–3006.

S�anchez-G�omez MV, Matute C. 1999. AMPA and kainate receptors eachmediate excitotoxicity in oligodendroglial cultures. Neurobiol Dis 6:475–485.

Schwarzschild MA. 2007. Adenosine A2A antagonists as neurotherapeutics:Crossing the bridge. Prog Neurobiol 83:261–262.

Shibata M, Hisahara S, Hara H, Yamawaki T, Fukuuchi Y, Yuan J, Okano H,Miura M. 2000. Caspases determine the vulnerability of oligodendrocytes inthe ischemic brain. J Clin Invest 106:643–653.

Smith KJ, Lassmann H. 2002. The role of nitric oxide in multiple sclerosis.Lancet Neurol 1:232–241.

Steckley D, Karajgikar M, Dale LB, Fuerth B, Swan P, Drummond-Main C,Poulter MO, Ferguson SS, Strasser A, Cregan SP. 2007. Puma is a dominantregulator of oxidative stress induced Bax activation and neuronal apoptosis. JNeurosci 27:12989–12999.

Stevens B, Porta S, Haak LL, Gallo V, Fields RD. 2002. Adenosine: A neuron-glial transmitter promoting myelination in the CNS in response to actionpotentials. Neuron 36:855–868.

Tekkok SB, Ye Z, Ransom BR. 2007. Excitotoxic mechanisms of ischemic injuryin myelinated white matter. J Cereb Blood Flow Metab 27:1540–1552.

Tourneur L, Mistou S, Schmitt A, Chiocchia G. 2008. Adenosine receptorscontrol a new pathway of Fas-associated death domain protein expressionregulation by secretion. J Biol Chem 283:17929–17938.

Trincavelli ML, Melani A, Guidi S, Cuboni S, Cipriani S, Pedata F, MartiniC. 2008. Regulation of A(2A) adenosine receptor expression and function-ing following permanent focal ischemia in rat brain. J Neurochem 104:479–490.

Gonz�alez-Fern�andez et al.: Adenosine Role in Oligodendrocyte Death

Month 2013 17

Tsutsui S, Schnermann J, Noorbakhsh F, Henry S, Yong VW, Winston BW,Warren K, Power C. 2004. A1 adenosine receptor upregulation and activationattenuates neuroinflammation and demyelination in a model of multiple scle-rosis. J Neurosci 24:1521–1529.

Walker EJ, Rosenberg GA. 2010. Divergent role for MMP-2 in myelin break-down and oligodendrocyte death following transient global ischemia. J Neu-rosci Res 88:764–773.

Wei W, Du C, Lv J, Zhao G, Li Z, Wu Z, Hasko G, Xie X. 2013. Blocking A2Badenosine receptor alleviates pathogenesis of experimental autoimmuneencephalomyelitis via inhibition of IL-6 production and Th17 differentiation. JImmunol 190:138–146.

Xu Z, Park SS, Mueller RA, Bagnell RC, Patterson C, Boysen PG. 2005. Aden-osine produces nitric oxide and prevents mitochondrial oxidant damage inrat cardiomyocytes. Cardiovasc Res 65:803–812.

Yang Y, Jalal FY, Thompson JF, Walker EJ, Candelario-Jalil E, Li L, ReichardRR, Ben C, Sang QX, Cunningham LA and others. 2011. Tissue inhibitor ofmetalloproteinases-3 mediates the death of immature oligodendrocytes viaTNF-alpha/TACE in focal cerebral ischemia in mice. J Neuroinflammation 8:108.

Yao Y, Sei Y, Abbracchio MP, Jiang JL, Kim YC, Jacobson KA. 1997. Adeno-sine A3 receptor agonists protect HL-60 and U-937 cells from apoptosisinduced by A3 antagonists. Biochem Biophys Res Commun 232:317–322.

18 Volume 00, No. 00