Neurobiology of Disease 9, 234–243 (2002)

doi:10.1006/nbdi.2001.0457, available online at http://www.idealibrary.com on

Ca2� Influx through AMPA or Kainate ReceptorsAlone Is Sufficient to Initiate Excitotoxicityin Cultured Oligodendrocytes

Elena Alberdi, Marıa Victoria Sanchez-Gomez, Aida Marino,and Carlos Matute1

Departamento de Neurociencias, Universidad del Paıs Vasco, E-48940 Leioa, Vizcaya, Spain

Received February 15, 2001; revised October 18, 2001; accepted for publication November 6, 2001

Oligodendrocytes are vulnerable to excitotoxic insults mediated by AMPA receptors and by low andhigh affinity kainate receptors, a feature that is dependent on Ca2� influx. In the current study, we haveanalyzed the intracellular concentration of calcium [Ca2�]i as well as the entry routes of this cation,upon activation of these receptors. Selective activation of either receptor type resulted in a substantialincrease (up to fivefold) of [Ca2�]i, an effect which was totally abolished by the non-NMDA receptorantagonist CNQX or by removing Ca2� from the culture medium. Blockade of voltage-gated Ca2�

channels with La3� or nifedipine, reduced the amplitude of the Ca2� current triggered by AMPAreceptor activation by �65%, but not that initiated by low and high affinity kainate receptors. Incontrast, KB-R7943, an inhibitor of the plasma membrane Na�–Ca2� exchanger, solely attenuated therise in [Ca2�]i by �25% due to activation of low affinity kainate receptors. However, oligodendroglialdeath by glutamate receptor overactivation was largely unaffected in the presence of La3� or KB-R7943. These findings indicate that Ca2� influx via AMPA and kainate receptors alone is sufficientto initiate cell death in oligodendrocytes, which does not require the entry of calcium via otherroutes such as voltage-activated calcium channels or the plasma membrane Na�–Ca2� exchanger.

INTRODUCTION

Oligodendrocytes, the myelinating cells of CNS ax-ons, express functional AMPA and kainate receptors(Verkhratsky and Steinhauser, 2000). The physiologi-cal significance of these receptors on oligodendrocytesstill remains obscure. However, it is well establishedthat prolonged activation of these receptors can leadto oligodendroglial death both in vitro and in vivo(Yoshioka et al., 1996; Matute et al., 1997; Matute, 1998;McDonald et al., 1998). These findings suggested thattraumatic injury, ischemia-related conditions, and de-myelinating disorders may cause damage to oligoden-drocytes by excitotoxic mechanisms (see for review,Matute et al., 2001). Indeed, glutamate spillover fromaffected cells and axons in traumatic lesions of the

1 To whom correspondence and reprint requests should be ad-

dressed. Fax: 34-944 649 266. E-mail: onpmaalc@lg.ehu.es.

234

CNS can kill oligodendrocytes (Li et al., 1999; Rosen-berg et al., 1999). In addition, an AMPA/kainate re-ceptor antagonist reduced ischemia-induced whitematter injury in the spinal cord and improved loco-motor function (Kanellopoulos et al., 2000). Moreover,this antagonist ameliorated the neurological deficits ofrats in which experimental autoimmune encephalo-myelitis had been previously induced (Pitt et al., 2000).Therefore, understanding the mechanisms which trig-ger oligodendroglial death upon AMPA and kainatereceptor activation, may contribute to the develop-ment of therapeutical agents with the potential toalleviate neurological conditions involving oligoden-drocyte loss.

Glutamate receptor-mediated excitotoxicity in oli-godendrocytes, as well as in neurons, is initiated by anexcessive influx of Na� and Ca2� through the activatedreceptor (McDonald et al., 1998; Sanchez-Gomez and

�

© 2002 Elsevier Science (USA)

Matute, 1999). Subsequently, neuronal Na influx can

0969-9961/02 $35.00© 2002 Elsevier Science (USA)

All rights reserved.

in turn trigger a secondary increase in the intracellularCa2� concentration through voltage-gated Ca2� chan-nels and reverse operation of the Na� � Ca2� ex-changer (Choi, 1995; Lee et al., 1999). Cellular Ca2�

overload is a key trigger in causing neuronal and glialdeath (Choi, 1995; Verkhratsky et al., 1998). Thus, ex-citotoxicity in cells of the oligodendroglial lineage isprevented when Ca2� is absent in the culture medium(Matute et al., 1997; Sanchez-Gomez and Matute, 1999;Fern and Moller, 2000). Furthermore, Ca2�-free perfus-ate virtually abolished glutamate-induced injury inisolated spinal cord white matter (Li and Stys, 2000).However, the contribution of the various possibleroutes of Ca2� entry to Ca2� overload after glutamatereceptor activation in oligodendrocytes is still unclear,as is the possibility that the selective blockade of theseroutes may improve the viability of these cells afterexcitotoxic insults. Here, we have measured [Ca2�]i

levels after selective AMPA or kainate receptor acti-vation, following blockade of the individual routes ofCa2� entry in oligodendrocytes. We observed that thecontribution of each of the receptor subtypes to cal-cium influx is similar, and that the route of entrydepends on the receptor type involved. Interestingly,our results indicate that Ca2� influx through the recep-tor channel alone is sufficient to trigger cell death.

MATERIAL AND METHODS

Optic Nerve Cultures

Oligodendrocyte cultures derived from optic nervesof 12-day-old Sprague–Dawley rats were obtained aspreviously described (Barres et al., 1992). The opticnerves were minced and enzymatically treated with1.25 mg/ml collagenase, 1.25 mg/ml trypsin, and 0.04mg/ml DNase (Sigma Chemical, St. Louis, MO) inCa2�- and Mg2�-free Hank’s balanced salt solution.The enzymatic reaction was stopped by adding 10%fetal bovine serum in Dulbecco’s modified Eagle’s me-dium and cells were resuspended in the same freshmedium and subsequently homogenized. The result-ing cell suspension was gently passed through needlesof different gauges, filtered through a nylon mesh (40�m), and seeded into 24-well plates bearing 12-mmdiameter coverslips coated with poly-d-lysine (10 �g/ml) at a density of 5–7 � 104 cells/well. Cells weremaintained at 37°C and 5% CO2 in a chemically de-fined medium (Barres et al., 1992) and fresh mediumwas added every day. After 3 days in vitro, cells wereidentified by nuclear staining with Hoechst 33258 (5

�g/ml; Sigma, U.S.A.) and stained with one of thefollowing antibodies: mouse IgM anti O4 (10 �g/ml;Boehringer-Manheim, Germany), mouse IgG3 anti-ga-lactocerebroside (GalC; 3 �g/ml; Boehringer-Man-heim, Germany), mouse IgG1 anti-myelin basic pro-tein (MBP; 0.8 �g/ml; Boehringer-Manheim, Germa-ny), polyclonal anti-glial fibrillary acidic protein(GFAP; 1:200, DAKO A/S, Copenhagen), and mono-clonal IgM A2B5 antibody (1:1000). Antibodies werevisualized by immunofluorescence using a rhodam-ine-conjugated or a fluorescein-conjugated secondaryantibody (Sigma Chemical, St. Louis, MO). Fluores-cein Bandeirae (Griffonia) Simplicifolia Lectin I (IB4;10 �g/ml, Vector, U.S.A.) was used to reveal the pres-ence of microglial cells. Cultures were composed of atleast 98% of O4/GalC� cells, the remaining cells beingGFAP� or nonidentified cells. No A2B5

� and microglialcells were detected in these cultures (Table 1). Cul-tures were used for experiments after 2–4 days invitro.

Measurement of [Ca2�]i

The [Ca2�]i was determined according to themethod of Grynkiewicz et al. (1985). Oligodendrocyteswere incubated with Fura-2 AM (Molecular Probes,Inc., Eugene, OR) at 5 �M in culture medium for 1 h at37°C. Cells were washed in Hank’s balance salt solu-tion containing 20 mM Hepes, pH 7.4, and 2 mMCaCl2 (incubation buffer) for 10 min at room temper-ature. Coverslips were placed in the cuvette of a flu-orimeter (SLM Aminco Bowman 2) containing 2 ml ofconstantly stirred incubation buffer at 37°C. The exci-tation wavelength was switched every second from340 to 380 nm (slit width 4 nm) and the light emittedat 510 nm (slit width 16 nm) was recorded. The pho-tomultiplier voltage was maintained within the rangeof 600–700 V. At the beginning of each assay, the

TABLE 1

Cellular Composition of Cultures Derived from Optic Nerve ofP12 Rats

Percentage of total labeled cells

A2B5� O4� GalC� MBP� GFAP� IB4�

3 DIV 0 98.7 � 0.3 98.6 � 0.4 84.0 � 2.0 1.0 � 0.2 0

Note. Cell number was determined by counting the percentage ofimmunolabeled cells among cells stained with Hoechst 33258 in 10fields. Total number of cells counted exceeded 4000. The resultsshown are means � SEM of at least three separate experiments.

235Excitotoxicity and Ca2� Influx in Oligodendrocytes

© 2002 Elsevier Science (USA)All rights reserved.

signal observed after excitation at 340 nm was arbi-trarily set as 50% of the maximal scale by adjustingthis voltage. At the end of the assay, in situ calibrationwas performed with the successive addition of 10 mMionomycin, 2 M Tris/50 mM EGTA, pH 8.5, and 100mM MnCl2. The [Ca2�]i concentration was estimatedby the 340/380 ratio method, using a Kd value of 224nM. Data were exported in ASCII format and ana-lyzed with Excel and Prism software.

Toxicity Assays

Cell toxicity and viability assays were carried out aspreviously described (Sanchez-Gomez and Matute,1999). After 2 days in culture, oligodendrocytes (5 �103 per coverslip) were exposed to 100 �M cyclothia-zide (CTZ) or 100 �M GYKI53655 and 20 �M con-canavalin A (con A) for 10 min prior to incubationwith AMPA or kainate, respectively. Agonists wereapplied for 15 min and then cells were incubated for24 h in fresh medium. LaCl3 (100 �M), nifedipine (10�M), and KB-R7943 (10 �M), the latter kindly sup-plied by Dr. I. Nakagawa (Kanebo Ltd., Japan), wereadded to the cells in conjunction with GYKI53655 orCTZ. Oligodendrocyte viability was assessed 24 hlater using fluorescein diacetate (60 �g/ml) as de-scribed earlier (Jones and Senft, 1985). Viable cellswere examined with a fluorescence microscope andcell counts in the entire coverslip were taken using a20� objective. All assays were carried out in duplicateand the values provided here are the average of atleast 3 independent experiments.

Data Analysis

All data are expressed as means � SEM (n), wheren represents the number of cultures assayed. Statisticaldifferences were calculated by ANOVA with theFisher PDLS test for multiple comparisons and statis-tical significance was determined at P � 0.05.

RESULTS

AMPA Receptor Activation Causes Ca2� Influxinto Oligodendrocytes

In resting oligodendrocytes, [Ca2�]i averaged 132 �10 nM (n � 5; Fig. 1). This level increased slightly afterthe cells were exposed to AMPA (Fig. 1A), and wasabruptly raised up to 2.5-fold when CTZ, an allostericmodulator that selectively suppresses desensitizationof AMPA receptors (Patneau et al., 1993), was added tothe incubation buffer (Fig. 1A). Both AMPA and CTZwere used at 100 �M, a concentration at which AMPAreceptor activation is near saturation (Holmann andHeinemann, 1994). Activation of AMPA receptorswith the endogenous agonist glutamate (100 �M)caused a [Ca2�]i increase which was similar to thatobserved with AMPA (Fig. 1B). Addition of increasingconcentrations of 6-cyano-7-nitroquinoxaline-2,3-di-one (CNQX), an AMPA/kainate receptor antagonist,progressively reduced the [Ca2�]i to reach basal levelsat 150 �M of CNQX (Fig. 1). In contrast to thesefindings, [Ca2�]i remained unaltered after AMPA re-ceptor activation in the absence of Ca2� in the incuba-

FIG. 1. Changes in [Ca2�]i caused by AMPA receptor activation in oligodendrocytes. Cells were exposed to (A) AMPA or (B) glutamate,together with cyclothiazide (CTZ) (all at 100 �M) in incubation buffer with 2 mM Ca2� or with 0.1 M EGTA and 0 mM Ca2�. Application ofagonists together with CTZ led to an increase in the cytosolic [Ca2�]i, an effect that was diminished by increasing, cumulative concentrationsof CNQX, until reaching basal levels at 150 �M of the antagonist. (A) These responses were absolutely dependent on the presence ofextracellular calcium. Data are representative of three experiments from at least three different cultures.

236 Alberdi et al.

© 2002 Elsevier Science (USA)All rights reserved.

tion buffer (Fig. 1A) or in the presence of 30 �MCNQX (Fig. 2A).

Together, these results indicate that the increase of[Ca2�]i upon AMPA receptor activation is initiallycaused by Ca2� influx through the plasma membraneand not by Ca2� released from intracellular stores.However, the participation of Ca2�-induced Ca2� re-lease from internal stores in contributing to the overall[Ca2�]i increases cannot be excluded.

Routes of AMPA Receptor-Induced CalciumInflux

We subsequently studied the routes of Ca2� entryinto oligodendrocytes as a consequence of AMPA re-ceptor activation. AMPA receptors are permeable toNa� and Ca2� (Hollmann and Heinemann, 1994).Membrane depolarization upon receptor activationcan in turn trigger a secondary increase in the intra-cellular Ca2� concentration through voltage-gatedCa2� channels and by reverse operation of the Na� �Ca2� exchanger (Choi, 1995; Lee et al., 1999). Accord-ingly, we examined the effect of the selective blockadeof voltage-gated Ca2� channels or the Na� � Ca2�

exchanger, on the increase in [Ca2�]i caused by theactivation of AMPA receptors.

Activation of AMPA receptors in the presence ofLaCl3 (100 �M) or nifedipine (10 �M), conditions thatblock voltage-gated calcium channels or L-type Ca2�

channel respectively in oligodendrocytes (data notshown), resulted in a substantial inhibition of the in-

crease of [Ca2�]i compared to that observed in theabsence of these substances (Fig. 2A). Both blockerswere similarly effective in reducing by about 65% the[Ca2�]i observed after activation of AMPA receptors(Figs. 2A and 6). Thus, a substantial fraction of thetotal Ca2� entry elicited by AMPA occurs via L-typeCa2� channels.

The Na�–Ca2� exchanger can function in reversemode as a consequence of transient alterations in theplasma membrane Na� gradient subsequent to depo-larization, including that caused by glutamate recep-tor activation. Thus, reversal of the Na�–Ca2� ex-changer can contribute to substantial [Ca2�]i increasesafter depolarization of astrocytes (Goldman et al.,1994; Takuma et al., 1994; Holgado and Beauge, 1995),oligodendrocyte progenitors (Liu et al., 1997) and neu-rons (Hoyt et al., 1998). In order to evaluate the par-ticipation of this exchanger in the observed [Ca2�]i

increases, we used KB-R7943, an inhibitor of the re-verse mode of the Na�–Ca2� exchanger (Ladilov et al.,1999). As reported previously (Ladilov et al., 1999),KB-R7943 (10 �M) efficiently blocked Ca2� influx trig-gered by substitution of Na� by choline� (data notshown), a feature that indicates that in oligodendro-cytes this substance also prevents the reverse opera-tion of the Na�–Ca2� exchanger. The presence of KB-R7943 during stimulation of AMPA receptors withAMPA and CTZ (both at 100 �M) attenuated the[Ca2�]i increase in by about 22–25%, a reduction thatwas statistically significant (P � 0.05) during the 90 ssubsequent to the application of the agonist (Fig. 2B)

FIG. 2. AMPA receptor-induced calcium influx occurs via the receptor channel, voltage-gated Ca2� channels and reverse operation of theNa�–Ca2� exchanger. (A) Effects of LaCl3 (100 �M), nifedipine (10 �M), and CNQX (30 �M) on [Ca2�]i levels due to application of AMPAtogether with CTZ (both at 100 �M) are shown. Blockade of voltage-dependent Ca2� channels with LaCl3 and nifedipine reduced the cytosolicCa2� concentration to less than half of the peak value. (B) KB-R7943 (10 �M) reduced by about 25% the [Ca2�]i during the initial response, butthe peak value was recovered thereafter. Data represent the mean of three experiments obtained for each condition from two or three differentcultures. Data are normalized with respect to [Ca2�]i in resting cells.

237Excitotoxicity and Ca2� Influx in Oligodendrocytes

© 2002 Elsevier Science (USA)All rights reserved.

but the peak [Ca2�]i remained unaltered thereafter(Figs. 2B and 6).

Selective Kainate Receptor Activation alsoCauses Ca2� Influx into Oligodendrocytes

Kainate activates both AMPA and kainate receptors(Lerma, 1997). In order to dissect the effects mediatedby kainate receptors alone, we used kainate in thepresence of GYKI53655 (100 �M), an AMPA receptorantagonist (Paternain et al., 1995), to selectively acti-vate kainate receptors. In a previous recent study, weobserved that kainate receptor-elicited toxicity in oli-godendrocytes is mediated by two distinct receptortypes, each with a different affinity for the agonist; onewith a broad peak at around 3 �M kainate and asecond with an EC50 � 488 �M (Sanchez-Gomez andMatute, 1999). Thus, we used 3 �M and 1 mM kainateto activate at saturating conditions high and low-af-finity kainate receptors, respectively, in conjunctionwith concanavalin A (con A) at 20 �M, a lectin thatprevents kainate receptor desensitization (Huettner,1990). Kainate at both concentrations caused an in-crease in [Ca2�]i which was diminished by increasingconcentrations of CNQX and returned to resting levelsat 150 �M of the antagonist (Fig. 3). Increases in [Ca2�]i

were prevented when calcium was removed from theextracellular medium (Fig 3). [Ca2�]i in the presence of3 �M and 1 mM kainate increased by up to two- andthreefold, respectively (n � 5), with respect to theresting level.

Routes of Calcium Influx by High and Low-Affinity Kainate Receptor Activation

We also employed LaCl3 to analyze the contributionof Ca2� influx through voltage-gated Ca2� channels, tothe overall [Ca2�]i observed upon the activation of lowand high affinity kainate receptors. The increase of[Ca2�]i elicited by the activation of these receptors wasnot inhibited by the blockade of voltage-gated Ca2�

channels (Fig. 4). In contrast, blockade of the plasmamembrane Na�–Ca2� exchanger with KB-R7943caused a significant inhibition (about 28%) of the[Ca2�]i reached by the activation of low, but not high,affinity kainate receptors (Figs. 5 and 6).

Together these results indicate that the kainate re-ceptor-induced [Ca2�]i increase is mainly due to Ca2�

influx through the receptor channel, although theNa�–Ca2� exchanger does contribute to the entry ofcalcium following activation of kainate receptors withhigh concentrations of kainate.

Excitotoxicity in Oligodendrocytes Is Mediated byCalcium Influx via AMPA or Kainate Receptors

We have previously observed that the toxicity me-diated by AMPA and kainate receptors in oligoden-drocytes is almost completely abolished when Ca2� isremoved from the culture medium (Matute et al., 1997;Sanchez-Gomez and Matute, 1999). For this reason, wenext studied if cell death mediated by AMPA or kai-nate receptors was reduced by blocking voltage-gated

FIG. 3. Changes in [Ca2�]i by selective activation of high or low affinity kainate receptors. (A) Kainate at 3 �M and (B) at 1 mM in the presenceof GYKI53655 (100 �M) and con A (20 �M) caused an increase in [Ca2�]i which returned to basal levels by increasing, cumulative concentrationsof CNQX (30–150 �M). Increases in [Ca2�]i were abolished when calcium was removed from extracellular medium (0 mM Ca2� � 0.1 M EGTA).Data are representative of three experiments from at least three different cultures.

238 Alberdi et al.

© 2002 Elsevier Science (USA)All rights reserved.

Ca2� channels or the plasma membrane Na�–Ca2� ex-changer.

Acute exposure of oligodendrocyte cultures toAMPA (applied with CTZ, both at 100 �M) and kai-nate (at 3 �M or 1 mM, in the presence of GYKI53655at 100 �M and con A at 20 �M), in a medium contain-ing 2 mM CaCl2, caused cell death (Fig. 7), which wasprevented in Ca2�-free medium or by preincubationwith CNQX (data not shown). These findings are con-sistent with those described previously (Sanchez-Gomez and Matute, 1999). In addition, activation ofAMPA receptors in the presence of LaCl3 (100 �M),but not KB-R7943 (10 �M), significantly reduced (byabout 15%, P � 0.05) the toxicity levels observed in theabsence of this blocker (Fig. 7). In contrast, neither

LaCl3 (100 �M) nor KB-R7943 (10 �M) altered signif-icantly the proportion of oligodendrocytes which werekilled by activation of kainate receptors (Fig. 7).

Overall, these results indicate that Ca2� influxthrough AMPA or kainate receptor-channel com-plexes is the major determinant of excitotoxicity inoligodendrocytes and that the fraction of Ca2� influxvia the receptor-pore triggers most of the ensuing celldeath.

DISCUSSION

The results illustrated here indicate that in culturedoligodendrocytes derived from the perinatal rat optic

FIG. 4. Kainate receptor-induced Ca2� influx does not occur through voltage-dependent Ca2� channels. [Ca2�]i increase by kainate receptoractivation was analyzed with kainate (3 �M and 1 mM) together with GYKI53655 (100 �M) and con A (20 �M) in the absence or presence ofLaCl3 (100 �M). [Ca2�]i after activation of (A) high affinity or (B) low affinity kainate receptors was not diminished by this blocking agent. Dataare representative of three experiments from at least three different cultures, and normalized with respect to [Ca2�]i obtained in resting cellsin each experiment.

FIG. 5. Effects of the Na�–Ca2� exchanger blocker KB-R7943 on kainate receptor-induced [Ca2�]i increases. KB-R7943 (10 �M) slightlydiminished [Ca2�]i levels after (B) low affinity kainate receptor activation, but did not modify [Ca2�]i levels after (A) high affinity kainatereceptor activation. Data represent the mean of three experiments obtained for each condition from two different cultures and are normalizedwith respect to [Ca2�]i obtained in resting cells in each experiment.

239Excitotoxicity and Ca2� Influx in Oligodendrocytes

© 2002 Elsevier Science (USA)All rights reserved.

nerve, activation of AMPA or kainate receptors causea two- to fivefold increase in the [Ca2�]i that is exclu-sively initiated by Ca2� influx through the plasmamembrane. In addition, we have characterized thecontribution of the major Ca2� routes to the overallincrease in [Ca2�]i triggered by activation of the recep-tors. Finally, we observed that AMPA and kainatereceptor-mediated cell death is principally initiated byCa2� influx through the receptor channel.

AMPA and Kainate Receptor Activation CausesCa2� Influx through Different Routes

AMPA and kainate receptors are classical ionotropicreceptors forming channels permeable to cations(Hollmann and Heinemann, 1994; Dingledine et al.,1999), though typical metabotropic functions for thesetwo types of receptors have also been reported (Wanget al., 1997; Rodrıguez-Moreno and Lerma, 1998). Inour assays, [Ca2�]i increases upon activation of eitherreceptor type were absolutely dependent on the pres-ence of Ca2� in the extracellular medium. Therefore, itis unlike that classical metabotropic receptors coupledto the PLC signaling pathway contribute to the overall[Ca2�]i increase.

Membrane depolarization subsequent to glutamatereceptor activation can lead to the opening of voltage-gated Ca2� channels (Mayer and Miller, 1990) andcause the reverse operation of the Na� � Ca2� ex-changer (Kiedrowski et al., 1994). Therefore, we usedblockers of these two routes of Ca2� entry into oligo-

dendrocytes, to evaluate their contribution to the[Ca2�]i increase observed as a consequence of AMPAand kainate receptor activation. The major types ofvoltage-gated Ca2� channels present in oligodendro-glial cells in vitro are the T (low voltage) and L (highvoltage) types (Barres et al., 1990; Kirisuchk et al.,1995). Blockade of these channels with La3� and dihy-dropyridines (Triggle and Janis, 1987) reduced byabout 65% the Ca2� influx elicited by the activation ofAMPA receptors. In contrast, La3� had no influence onthe [Ca2�]i increase triggered by kainate receptors.These findings are similar to those reported in cul-tured cerebellar granule cells, in which inhibition ofvoltage-gated Ca2� channels with Cd2� reduced byabout 62% the AMPA-stimulated Ca2� influx (Hackand Balazs, 1995). Our results are also reminiscent ofthose reported in embryonic motoneurons in whichvoltage-gated Ca2� channels are the predominant en-try route for Ca2� upon glutamate receptor activation(Metzger et al., 2000).

Inhibition of the reverse mode of the Na�–Ca2�

exchanger by KB-R7943 caused a slight attenuation ofthe initial, but not the later phase of the [Ca2�]i in-crease elicited by AMPA receptor activation. Consis-tent with these results, in cultured rat forebrain neu-

FIG. 6. Summary of the effects of AMPA and kainate receptor-induced Ca2� influx in oligodendrocytes, by blockers of voltage-gated Ca2� channels and of the plasma membrane Na�–Ca2� ex-changer. The percentages of Ca2� influx by different drugs arerepresented. The 100% value corresponds to [Ca2�]i induced byagonists (AMPA 100 �M, kainate 3 �M and 1 mM). Values wereobtained at 2 min poststimulation, from three independent experi-ments for each category. Bars represent the average value � SEM.

FIG. 7. Excitotoxicity mediated by kainate and AMPA receptors inthe presence of LaCl3 (100 �M) or by KB-R7943 (10 �M). All valueswere referred to control, nontreated cultures. Data from each groupwere compared by ANOVA, with excitotoxic levels obtained afteractivation of the receptors alone. In all cases, the observed differ-ences in cell death were not statistically significant, except for thegroup indicated with a star (P � 0.05). GYKI53655 and CTZ wereapplied at 100 �M and con A at 20 �M. Bars represent the means �SEM (n � 3).

240 Alberdi et al.

© 2002 Elsevier Science (USA)All rights reserved.

rons, KB-R7943 (10 �M) diminished the [Ca2�]i in-crease immediately after glutamate receptoractivation, but the peak of the [Ca2�]i did not changethereafter (Hoyt et al., 1998). Thus, it appears that thereverse mode of the Na�–Ca2� exchanger contributesto the [Ca2�]i increase soon after receptor activation,but it becomes inactivated later in the case of AMPAreceptor stimulation, possibly as a consequence of theaccumulation of Ca2� within the cytoplasm and a re-duction of the gradient of this cation.

In addition, KB-R7943 reduced by about 25% the[Ca2�]i increase initiated by low affinity kainate recep-tors, an effect that in contrast to AMPA receptors, wassustained during agonist exposure. Thus, the reversemode of the Na�–Ca2� exchanger contributes to[Ca2�]i increase upon AMPA and kainate receptor ac-tivation, not only in neurons, but also in oligodendro-cytes. This exchanger can thus provide an additionalroute for the regulation of cytosolic Ca2� during nor-mal glutamate neurotransmission (Yu and Choi, 1997;Hoyt et al., 1998). However, our results indicate thatthe contribution of the Na� � Ca2� exchanger to[Ca2�]i increase does not enhance excitotoxicity.

In cultured optic nerve oligodendrocytes, the mainAMPA receptor subunits expressed are GluR3 andGluR4 (Matute et al., 1997). The lack of GluR2 expres-sion in cultures derived from the optic nerve and alsoin those obtained from the cerebral cortex (Patneau etal., 1994; Puchalski et al., 1994), strongly suggests thatnative AMPA receptors in oligodendrocytes are per-meable to Ca2�, as described in studies with recombi-nant receptors (Hollmann and Heinemann, 1994;Dingledine et al., 1999). Consistent with these findings,we observed that a substantial fraction of the [Ca2�]i

increase triggered by AMPA receptors is due to Ca2�

influx through the oligodendrocyte membrane via thereceptor channel, as observed upon receptor activa-tion in the presence of blockers of alternative Ca2�

entry routes.Native kainate receptors in oligodendrocytes are

formed by most of the kainate receptor subunits char-acterized to date (Patneau et al., 1994; Garcıa-Barcinaand Matute, 1996; Matute et al., 1997). Among them,GluR6 appears to be a major constituent of the recep-tors present in these glial cells (Sanchez-Gomez andMatute, 1999). Moreover, the unedited form of GluR6predominates over the edited version of this subunitin oligodendrocytes (Pascual de Zulueta and Matute,1999). This finding, together with the results of recom-bination studies which have indicated that uneditedGluR6 in homomeric configuration is more permeableto calcium (Hollmann and Heinemann, 1994; Dingle-

dine et al., 1999), suggests that kainate receptors ex-pressed in cells of the oligodendroglial lineage have arelatively high Ca2� permeability. In corroborationwith these results, we found that native kainate recep-tors in oligodendrocytes are Ca2� permeable sinceblockade of the alternative routes does not abolish[Ca2�]i increase.

Ca2� Influx through AMPA and Kainate ReceptorsAlone Is Sufficient to Initiate OligodendroglialExcitotoxicity

Cell death triggered by glutamate receptor agonistsin neurons and oligodendrocytes is mediated by cyto-toxic levels of intracellular Ca2� (Lee et al., 1999). Inneurons, [Ca2�]i rise elicited by AMPA and kainateinvolves the activation of voltage-gated Ca2� channelsand the operation of the Na� � Ca2� exchanger inreverse mode (Hack and Balazs, 1995; Metzger et al.,2000; Canzoniero et al., 1996). These results indicatethat the Ca2� influx through voltage-dependent Ca2�

channels can also contribute to the Ca2� overload trig-gering neuronal death. In support of this idea, Ca2�

channel blockers substantially attenuate the excitotox-icity produced by long-term exposure to AMPA andkainate in cortical cell cultures (Weiss et al., 1990) andby deprivation of oxygen and glucose in culturedhippocampal neurons (Kimura et al., 1998). In contrast,blockade of the Na�–Ca2� exchanger in cultured ratforebrain neurons after prolonged glutamate exposureis not neuroprotective in vitro (Hoyt et al., 1998).

Our results indicate that in oligodendrocytes, theroutes contributing to [Ca2�]i increase after AMPAreceptor activation are similar to those described inneurons. Thus, in oligodendrocytes blockade of volt-age-gated Ca2� channels during AMPA receptor acti-vation has a protective, albeit weak effect, whereasinhibition of reverse operation of the Na�–Ca2� ex-changer is ineffective. The lack of a strong protectioneffect of Ca2� channel blockers is consistent with theresults obtained in the oligodendroglial cell line CG4(Yoshioka et al., 2000) and in immature oligodendro-cytes (Fern and Moller, 1999) in which damage in-duced by oxygen and glucose deprivation is not pre-vented by these blockers. In addition, selective activa-tion of kainate receptors causes oligodendroglialdeath, which is not attenuated by the blockade of Ca2�

channels or of the Na�–Ca2� exchanger. Overall, theseresults are similar to those observed in neurons indi-cating that rapid Ca2� influx through Ca2� permeableAMPA/kainate channels may result in mitochondrial

241Excitotoxicity and Ca2� Influx in Oligodendrocytes

© 2002 Elsevier Science (USA)All rights reserved.

Ca2� overload and the production of oxygen radicalswhich lead to neuronal death (Carriedo et al., 1998).

Taken together, the present study reveals the routescontributing to [Ca2�]i increase in oligodendrocytesafter activation of AMPA and kainate receptors. Inaddition, we provide evidence that Ca2� entry throughglutamate receptor channels alone is sufficient to trig-ger excitotoxicity in oligodendroglia.

ACKNOWLEDGMENTS

The authors thank Dr. J. Fogarty for reviewing the manuscript.This work was supported by grants from the Ministerio de Educa-cion y Cultura (PM97/0045), the Gobierno Vasco (PI/1999/132),and the University del Paıs Vasco. E.A. held a fellowship from theGobierno Vasco and M.V.S-G. from the Ministerio de Educacion yCultura.

REFERENCES

Barres, B. A., Chun, L. L. Y., & Corey, D. P. (1990) Ion channels invertebrate glia. Annu. Rev. Neurosci. 13, 441–474.

Barres, B. A., Hart, I. K., Coles, H. S. R., Burne, J. F., Voyvodic, J. T.,Richardson, W. D., & Raff, M. C. (1992) Cell death and control ofcell survival in the oligodendrocyte lineage. Cell 70, 31–46.

Canzoniero, L. M., Sensi, S. L., Turetsky, D. M., Finley, M. F., Choi,D. W., & Huettner, J. E. (1996) Glutamate receptor-mediatedcalcium entry in neurons derived from P19 embryonal carcinomacells. J. Neurosci. Res. 45, 226–236.

Carriedo, S. G., Yin, H. Z., Sensi, S. L., & Weiss, J. H. (1998) RapidCa2� entry through Ca2�-permeable AMPA/Kainate channelstriggers marked intracellular Ca2� rises and consequent oxygenradical production. J. Neurosci. 18, 7727–7738.

Choi, D. W. (1995) Calcium: Still center-stage in hypoxic-ischemicneuronal death. Trends Neurosci. 18, 58–60.

Dingledine, R., Borges, K., Bowie, D., & Traynelis, S. F. (1999) Theglutamate receptor ion channels. Pharmacol. Rev. 51, 7–61.

Fern, R., & Moller, T. (2000) Rapid ischemic cell death in immatureoligodendrocytes: A fatal glutamate release feedback loop. J. Neu-rosci. 20, 34–42.

Garcıa-Barcina, J., & Matute, C. (1996) Expression of kainate-selec-tive glutamate receptor subunits in glial cells of the adult bovinewhite matter. Eur. J. Neurosci. 8, 2379–2387.

Goldman, W. F., Yarowsky, P. J., Juhaszova, M., Krueger, B. K., &Blaustein, M. P. (1994) Sodium/calcium exchange in rat corticalastrocytes. J. Neurosci. 14, 5834–5843.

Grynkiewikcz, G., Poenie, M., & Tsien, R. Y. (1985) A new genera-tion of Ca2� indicators with greatly improved fluorescence prop-erties. J. Biol. Chem. 260, 3440–3450.

Hack, N., & Balazs, R. (1995) Properties of AMPA receptors ex-pressed in rat cerebellar granule cell cultures: Ca2� influx studies.J. Neurochem. 65, 1077–1084.

Holgado, A., & Beauge, L. (1995) The Na�/Ca2� exchange system inrat glial cells in culture: Activation by external monovalent cat-ions. Glia 14, 77–86.

Hollmann, M., & Heinemann, S. F. (1994) Cloned glutamate recep-tors. Annu. Rev. Neurosci. 17, 31–108.

Hoyt, K. R., Arden, S. R., Aizenman, E., & Reynolds, I. J. (1998)Reverse Na�/Ca2� exchange contributes to glutamate-inducedintracellular Ca2� concentration increases in cultured rat forebrainneurons. Mol. Pharmacol. 53, 742–749.

Huettner, J. E. (1990) Glutamate receptor channels in rat DRG neu-rons-activation by kainate and quisqualate and blockade of de-sensitization by Con-A. Neuron 5, 255–266.

Jones, K. H., & Senft, J. A. (1985) An improved method to determinecell viability by simultaneous staining with fluorescein diacetate-propidium iodide. J. Histochem. Cytochem. 33, 77–79.

Kanellopoulos, G. K., Xu, X. M., Hsu, C. Y., Lu, X., Sundt, T. M., &Kouchoukos, N. T. (2000) White matter injury in spinal cordischemia: Protection by AMPA/kainate glutamate receptor antag-onism. Stroke 31, 1945–1952.

Kiedrowski, L., Brooker, G., Costa, E., & Wroblewski, J. T. (1994)Glutamate impairs neuronal calcium extrusion while reducingsodium gradient. Neuron 12, 295–300.

Kimura, M., Sawada, K., Miyagawa, T., Kuwada, M., Katayama, K.,& Nishizawa, Y. (1998) Role of glutamate receptors and voltage-dependent calcium and sodium channels in the extracellular glu-tamate/aspartate accumulation and subsequent neuronal injuryinduced by oxygen/glucose deprivation in cultured hippocampalneurons. J. Pharmacol. Exp. Ther. 285, 178–185.

Kirischuk, S., Scherer, J., Moller, T., Verkhratsky, A., & Kettenmann,H. (1995) Subcellular heterogeneity of voltage-gated Ca2� chan-nels in cells of the oligodendrocyte lineage. Glia 13, 1–12.

Ladilov, Y., Haffner, S., Balser-Schafer, Maxeiner, H., & Piper, H. M.(1999) Cardioprotective effects of KB-R7943: A novel inhibitor ofthe reverse mode of Na�/Ca2� exchanger. Am. J. Physiol. 276,H1868–H1876.

Lee, J-M., Zipfel, G. J., & Choi, D. W. (1999) The changing landscapeof ischaemic brain injury mechanisms. Nature 399(Suppl.)A7–A14.

Lerma, J. (1997) Kainate reveals its targets. Neuron 19, 1155–1158.Li, S., Mealing, G. A. R., Morley, P., & Stys, P. K. (1999) Novel injury

mechanism in anoxia and trauma of spinal cord white matter:Glutamate release via reverse Na�-dependent glutamate trans-port. J. Neurosci. 19, RC16.

Li, S., & Stys, P. K. (2000) Mechanisms of ionotropic glutamatereceptor-mediated excitotoxicity in isolated spinal cord whitematter. J. Neurosci. 20, 1190–1198.

Liu, H., Molina-Holgado, E., & Almazan, G. (1997) Glutamate-stimulated production of inositol phosphates is mediated by Ca2�

influx in oligodendrocyte progenitors. Eur. J. Pharmacol. 338, 277–287.

Matute, C., Sanchez-Gomez, M. V., Martınez- Millan, L., & Miledi,R. (1997) Glutamate receptor-mediated toxicity in optic nerveoligodendrocytes. Proc. Natl. Acad. Sci. USA 94, 8830–8835.

Matute, C. (1998) Properties of acute and chronic kainate excitotoxicdamage to the optic nerve. Proc. Natl. Acad. Sci. USA 95, 10229–1034.

Matute, C., Alberdi, E., Domercq, M., Perez-Cerda, F., Perez-Sa-martın, A., & Sanchez-Gomez, M. V. (2001) The link betweenexcitotoxic oligodendroglial death and demyelinating diseases.Trends Neurosci. 24, 224–230.

Mayer, M. L., & Miller, R. J. (1990) Excitatory amino acid receptors,second messengers and regulation of intracellular Ca2� in mam-malian neurons. Trends Pharmacol. Sci. 11, 254–260.

McDonald, J. W., Althomsons, S. P., Hyrc, K. L., Choi, D. W., &Goldberg, M. P. (1998) Oligodendrocytes from forebrain are

242 Alberdi et al.

© 2002 Elsevier Science (USA)All rights reserved.

highly vulnerable to AMPA/kainate receptor-mediated excito-toxicity. Nat. Med. 4, 291–297.

Metzger, F., Kulik, A., Sendtner, M., & Ballanyi, K. (2000) Contri-bution of Ca2� permeable AMPA/KA receptors to glutamate-induced Ca2� rise in embryonic lumbar motoneurons in situ.J. Neurophysiol. 83, 50–59.

Pascual de Zulueta, M., & Matute, C. (1999) Reduced editing of lowaffinity kainate receptor subunits in optic nerve glial cells. Mol.Brain Res. 73, 104–109.

Paternain, A. V., Morales, M., & Lerma, J. (1995) Selective antago-nism of AMPA-receptors unmasks kainate receptor-mediated re-sponses in hippocampal neurons. Neuron 14, 185–189.

Patneau, D. K., Vickliky, L., Jr., & Mayer, M. L. (1993) Hippocampalneurons exhibit cyclothiazide-sensitive rapidly desensitizing re-sponses to kainate. J. Neurosci. 13, 3496–3509.

Patneau, D. K., Wright, P. W., Winters, C., Mayer, M. L., & Gallo, V.(1994) Glial cells of the oligodendrocyte lineage express bothkainate- and AMPA-preferring subtypes of glutamate receptors.Neuron 12, 357–371.

Pitt, D., Werner, P., & Raine, C. S. (2000) Glutamate excitotoxicity ina model of multiple sclerosis. Nat. Med. 6, 67–70.

Puchalski, R. B., Louis, J. C., Brose, N., Traynelis, S. F., Egebjerg, J.,Kukekov, V., Wenthold, R. J., Rogers, S. W., Lin, F., Moran, T.,Morrison, J. H., & Heinemann, S. F. (1994) Selective RNA editingand subunit assembly of native glutamate receptors. Neuron 13,131–147.

Rodrıguez-Moreno, A., & Lerma, J. (1998) Kainate receptor modu-lation of GABA release involves a metabotropic function. Neuron20, 1211–1218.

Rosenberg, L. J., Teng, Y. D., & Wrathall, J. R. (1999) 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(f)quinoxaline reduces glial loss and

acute white matter pathology after experimental spinal cord con-tusion. J. Neurosci. 19, 464–475.

Sanchez-Gomez, M. V., & Matute, C. (1999) AMPA and kainatereceptors each mediate excitotoxicity in oligodendrocyte cultures.Neurobiol. Dis. 6, 475–485.

Takuma, K., Matsuda, T., Hashimoto, H., Asano, S., & Baba, A.(1994) Cultured rat astocytes posses Na�/Ca2� exchanger. Glia 12,335–342.

Triggle, D. J., & Janis, R. A. (1987) Calcium channel ligands. Annu.Rev. Pharmacol. Toxicol. 27, 34–69.

Verkhratsky, A., & Steinhauser, C. (2000) Ion channels in glial cells.Brain Res. Rev. 32, 380–412.

Verkhratsky, A., Orkand, R. K., & Kettenmann, H. (1998) Glialcalcium: Homeostasis and signaling function. Physiol. Rev. 78,99–141.

Wang, Y., Small, D. L., Stanimirovic, D. B., Morley, P., & Durkin,J. P. (1997) AMPA receptor-mediated regulation of a Gi-protein incortical neurons. Nature 389, 502–504.

Weiss, J. H., Hartley D. M., Koh, J., & Choi, D. W. (1990) The calciumchannel blocker nifedipine attenuates slow excitatory amino acidneurotoxicity. Science 247, 1474–1477.

Yoshioka, A., Bacskai, B., & Pleasure, D. (1996) Pathophysiology ofoligodendroglial excitotoxicity. J. Neurosci. Res. 46, 427–437.

Yoshioka, A., Yamaya, Y., Saiki, S., Kanemoto, M., Hirose, G., Bees-ley, J., & Pleasure, D. (2000) Non-N-methyl-d-aspartate glutamatereceptors mediate oxygen-glucose deprivation-induced oligoden-droglial injury. Brain. Res. 854, 207–215.

Yu, S. P., & Choi, D. W. (1997) Na�-Ca2� exchange currents incortical neurons: Concomitant forward and reverse operation andeffect of glutamate. Eur. J. Neurosci. 9, 1273–1281.

243Excitotoxicity and Ca2� Influx in Oligodendrocytes

© 2002 Elsevier Science (USA)All rights reserved.