Glutamate Release Induced by Activation ofGlycine and GABA Transporters in SpinalCord is Enhanced in a Mouse Model of

Amyotrophic Lateral SclerosisLuca Raiteri 1, Simona Zappettini 1, Sara Stigliani 1, Silvio Paluzzi 1,

Maurizio Raiteri 1,2, Giambattista Bonanno 1,2,*1Department of Experimental Medicine, Pharmacology and Toxicology Section, University of Genoa, 16148 Genoa,

Italy2Center of Excellence for Biomedical Research, University of Genoa, 16132 Genoa, Italy

Received 12 January 2005; accepted 26 January 2005

Available online 10 May 2005

Abstract

Amyotrophic lateral sclerosis is a progressive and fatal neurodegenerative disease, involving both upper and lower

motor neurons, the cause of which is obscure, although glutamate (GLU)-induced excitotoxicity has been suggested to

play a major role. We studied the release of [3H]D-aspartate ([3H]D-ASP) and endogenous glutamate evoked by glycine

(GLY) or GABA from spinal cord synaptosomes in mice expressing a mutant form of human SOD1 with a Gly93Ala

substitution ([SOD1-G93A(+)]), a transgenic model of amyotrophic lateral sclerosis, in mice expressing the non-mutated

form of human SOD1 [SOD1(+)], and in non-transgenic littermates [SOD1(�)/G93A(�)]. In parallel experiments, we

also studied the release of [3H]GABA evoked by GLYand that of [3H]GLY evoked by GABA. Mutant mice were killed at

advanced phase of pathology or during the pre-symptomatic period. In SOD1(�)/G93A(�) or SOD1(+) mice GLYevoked

[3H]D-ASPand [3H]GABA release, while GABA caused [3H]D-ASP, but not [3H]GLY, release. The GLY-evoked release of

[3H]D-ASP, but not that of [3H]GABA, and the GABA-evoked [3H]D-ASP release, but not that of [3H]GLY, were more

pronounced in SOD1-G93A(+) than in SOD1(+) or SOD1(�)/G93A(�) mice. Furthermore, the excessive potentiation of

[3H]D-ASP by GLYor GABA was already present in asymptomatic 30–40 day-old SOD1-G93A(+) mice. The releases of

endogenous glutamate and GABA also were enhanced by GLYand the GLY-evoked release of endogenous glutamate, but

not of endogenous GABA, was higher in SOD1-G93A(+) than in control animals. Potentiation of the spontaneous amino

acid release is likely to be mediated by activation of a GLY or a GABA transporter, since the effect of GLY was

counteracted by the GLY transporter blocker glycyldodecylamide but not by the GLY receptor antagonists strychnine and

5,7-dichlorokynurenate while the effect of GABA was diminished by the GABA transporter blocker SKF89976-A but not

by the GABA receptor antagonists SR9531 and CGP52432. It is concluded that the glutamate release machinery seems

excessively functional in SOD1-G93A(+) animals.

# 2005 Elsevier Inc. All rights reserved.

Keywords: Amyotrophic lateral sclerosis; SOD1-G93A(+) mice; Glutamate release; Glycine

heterotransporters; GABA heterotransporters; Release mechanisms

NeuroToxicology 26 (2005) 883–892

* Corresponding author. Tel.: +39 010 3532651;

fax: +39 010 3993360.

E-mail address: bonanno@pharmatox.unige.it (G. Bonanno).

0161-813X/$ – see front matter # 2005 Elsevier Inc. All rights reserv

doi:10.1016/j.neuro.2005.01.015

INTRODUCTION

Amyotrophic lateral sclerosis (ALS) is a chronic

neuromuscular disorder clinically characterized by

muscle wasting, weakness and spasticity reflecting a

ed.

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892884

progressive degeneration of upper and lower motor

neurons (Brown, 1995). The prevalence of the disease

is three to five patients per 100,000 individuals, making

it the most frequent paralytic disease in adults (Przed-

borski et al., 2003). To date only a few approved

treatments prolong survival in ALS patients of some

extent (Louvel et al., 1997). Development of more

effective neuroprotective therapies is impeded by lack

of understanding of the mechanisms of neuronal death

in ALS and how the disease propagates.

ALS occurs both in sporadic and familial forms

(Horton et al., 1976). Approximately, 10% of cases

are inherited, usually as an autosomal dominant trait

(Mulder et al., 1986). In about 25%of familiar cases, the

disease is caused by mutation in the gene encoding

cytosolic copper–zinc superoxide dismutase (SOD1)

(Rosen et al., 1993). Consistently, transgenic mice

over-expressing familialALS-associatedSOD1mutants

(G93A,G37R, andG85R) develop a rapidly progressive

neuromuscular disease resembling humanALS (Gurney

et al., 1994; Brown, 1995;Wong et al., 1995). Since both

sporadic and familial ALS patients show very similar

pathological traits, there is a general consensus that

clarification of the mechanisms underlying neurotoxi-

city in familialALS should help the understandingof the

pathogenic determinants in sporadic ALS.

There is increasing evidence that glutamate-

mediated excitotoxicity is implicated in ALS (see,

for reviews: Leigh and Meldrum, 1996; Morrison

and Morrison, 1999; Shaw and Eggett, 2000; Cluskey

and Ramsden, 2001). An abnormality in the function of

the glial glutamate transporter GLT-1 (EAAT-2), which

is largely responsible for removing extracellular glu-

tamate, has been proposed as one major cause of

excessive glutamate receptor activation in both spora-

dic and familiar ALS (Rothstein et al., 1995). On the

other hand, alterations in the glial glutamate transporter

may not alone account for the excitotoxic degeneration

(see Morrison and Morrison, 1999). Other reasons for

the gains of glutamatergic function present in mutant

forms of SOD1, including increase of glutamate

release, should therefore be considered.

Transporters for different transmitters often are coex-

pressed on the same neuron, particularly on the same

terminal button; that is, transporters for the reuptake of

the transmitter just released into the synapse (homo-

transporters) and transporters able to recognize and

capture transmitters/modulators coming from neigh-

bouring structures (heterotransporters) often coexist

on the same membrane. Interestingly, activation of

heterotransporters can elicit release of the transmitter

previously taken up through the coexisting homotran-

sporters or of the endogenous counterpart (Bonanno and

Raiteri, 1994; Raiteri et al., 2002; Sepkuty et al., 2002).

Heterotransporters have also been found to exist on

neuronal cell bodies/dendrites; for example, the excita-

tory amino acid transporter 4 (EAAT4) is present on

GABAergic Purkinje cells in the cerebellum (Yamada

et al., 1996; Furuta et al., 1997; Nagao et al., 1997).

Studying the heterotransporter-mediated modula-

tion of glutamate release in the spinal cord of a

transgenic mouse model of familial ALS carrying a

human SOD1 cDNA with the G93A-associated muta-

tion (Gurney et al., 1994), we here found a potential

novel mechanism by which the extracellular glutamate

concentration can be increased.

METHODS

Animals

B6SJL-TgN SOD1-G93A(+)1Gur mice expressing

high copy number of mutant human SOD1 with a

Gly93Ala substitution [SOD1-G93A(+)] and B6SJL-

TgN (SOD1)2Gur mice expressing wild-type human

SOD1 [SOD1(+)] were obtained from Jackson Labora-

tories (Bar Harbor, ME, USA) and bred at the animal

facility of the Pharmacology and Toxicology Section,

Department of Experimental Medicine in Genoa.

Transgenic mice were originally produced by micro-

injection of the transgene in fertilized eggs obtained

from hybrid mice: the G1 line was used for G93A and

the N1029 line for SOD1 (Gurney et al., 1994). The

transgene number is different in the two lines, with G1

havingmore than double the number of gene copies. As

estimated by Gurney et al. (1994), G1 mice have a gene

copy (per diploid gene) = 18, while N1029 mice have a

gene copy (per diploid gene) = 7.2; however, it was

reported that the N1029 line expressed comparable or

even greater amounts of total brain human SOD1 (Dal

Canto and Gurney, 1994; Gurney et al., 1994). Selec-

tive breeding maintained each transgene in the hemi-

zigous state on an F1 hybrid C57Bl6 � SJL genetic

background (Gurney et al., 1994). Nontransgenic lit-

termates of SOD1-G93A(+) and SOD1(+) mice were

used as controls [SOD1(�)/G93A(�)]. All transgenic

mice were identified analyzing extracts from tail tips

(homogenized in phosphate-buffer saline, freeze/

thawed twice and centrifuged at 23,000 � g for

15 min at + 4 8C) by staining for SOD1 after Laemm-

li’s polyacrylamide gel electrophoresis (10% resolving

and 4% stacking). In the SOD1-G93A(+) strain, the

first signs of the disease appeared between 120 and 140

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892 885

days of life. Progression of the disease was then rapid

so that after 7–9 days, the animals were killed because

of their ingestion disability, due to paralysis of the

posterior limbs. Animals were housed at constant

temperature (22 � 1 8C) and relative humidity (50%)

under a regular dark-light schedule (light on 7 a.m. to 7

p.m.). Food and water were freely available. All

experiments were carried out in accordance with the

European Community Council Directive of 24 Novem-

ber 1986 (86/609/EEC). All efforts were made to

minimize animal suffering and to use only the number

of animals necessary to produce reliable results.

Preparation of Synaptosomes

Animals were sacrificed and the spinal cord rapidly

removed after exposure of the spinal column. Synapto-

somes were prepared essentially as described pre-

viously (Raiteri et al., 1984). Briefly, the tissue was

homogenized in 40 volumes of 0.32 M sucrose, buf-

fered at pH 7.4 with phosphate, using a glass-teflon

tissue grinder (clearance 0.25 mm, 12 up and down

strokes in about 1 min). The homogenate was centri-

fuged (5 min, 1000 � g) to remove nuclei and debris,

and synaptosomes were isolated from the supernatant

by centrifugation at 12,000 � g for 20 min. The synap-

tosomal pellet was then resuspended in a physiological

medium having the following composition (mM):

NaCl, 125; KCl, 3; MgSO4, 1.2; CaCl2, 1.2; NaH2PO4,

1; NaHCO3, 22; glucose, 10 (aeration with 95%O2 and

5% CO2); pH 7.2–7.4. All the above procedures were

performed at 0–4 8C. Protein was measured according

to Bradford (1976) using BSA as a standard.

Experiments of Release

Synaptosomes were incubated at 37 8C for 15 min in

the presence of 0.02 mM [3H]GABA and 50 mM of the

GABA transaminase inhibitor aminooxyacetic acid, to

avoid [3H]GABA metabolism, or of 0.02 mM [3H]D-

ASP, a non-metabolized compound often used as a

marker to mimic glutamate in release studies, or of

0.3 mM [3H]glycine. Aliquots (about 50 mg of protein)

of the synaptosomal suspension were distributed on

microporous filters placed at the bottom of a set of

parallel superfusion chambers maintained at 37 8C(Raiteri and Raiteri, 2000). Superfusion was then

started with standard medium (containing aminooxya-

cetic acid in experiments of [3H]GABA release) at a

rate of 0.5 ml/min and continued for 48 min. After

33 min of superfusion to equilibrate the system, five 3-

min fractions were collected. Synaptosomes were

exposed to glycine (1–1000 mM) or GABA (0.1–

1000 mM) at the end of the second fraction collec-

ted (t = 39 min). Strychnine, 5,7-dichlorokynurenate

(5,7-DCK), glycyldodecylamide (GDA), N[3-(40-fluorophenyl)-3-(40-phenylphenoxy)propyl]sarcosine(NFPS), amoxapine, 2-(30-carbethoxy-20-propenyl)-3-amino-6-paramethoxy-phenyl-pyridazinium bro-

mide (SR95531), [3-[[(3,4-dichlorophenyl)methyl]a-

mino]propyl](diethoxymethyl)phosphinic acid (CGP

52432), and N-(4,4-diphenyl-3-butenyl)nipecotic acid

(SKF89976A) were introduced at t = 30 min. Radio-

activity was measured in each fraction collected and in

the superfused filters. In some experiments, fractions

were analyzed for their endogenous glutamate and

GABA content.

Endogenous Glutamate and GABADetermination

Endogenous glutamate and GABA were measured

by high performance liquid chromatography analysis

following pre-column derivatization with o-phthalal-

dehyde and separation on a C18 reverse-phase chro-

matographic column (10 � 4.6 mm, 3 mm; at 30 8C;Chrompack, Middleburg, The Netherlands) coupled

with fluorometric detection (excitation wavelength

350 nm; emission wavelength 450 nm; Raiteri et al.,

2000). Buffers and the gradient program were as

follows: solvent A, 0.1 M sodium acetate (pH 5.8)/

methanol, 80:20; solvent B, 0.1 M sodium acetate (pH

5.8)/methanol, 20:80; solvent C, sodium acetate (pH

6.0)/methanol, 80:20; gradient program, 100% C for

4 min from the initiation of the program; 90% A and

10% B in 1 min; 42% A and 58% B in 14 min; 100% B

in 1 min; isocratic flow 2 min; 100% C in 3 min; flow

rate 0.9 ml/min. Homoserine was used as an internal

standard.

Calculations

Tritium released in each fraction collected was

calculated as fractional rate �100. The endogenous

glutamate and GABA released in each fraction was

expressed as pmol/mg of synaptosomal protein. Drug

effects were evaluated by performing the ratio between

the efflux in the fourth fraction collected (in which the

maximum effect of glycine or GABA was generally

reached) and that of the second fraction. This ratio was

compared to the corresponding ratio obtained under

control conditions. Appropriate controls were always

run in parallel. The concentration–response curves

shown in Figs. 1, 2 and 4 were fitted to the experimental

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892886

Fig. 1. Concentration-dependent effects of glycine on the release of [3H]D-

ASP or [3H]GABA (panel A) and of GABA on the release of [3H]D-ASP or

[3H]glycine (panel B) from superfused spinal cord synaptosomes prepared

from nontransgenic SOD1(�)/G93A(�) mice. Experiments were per-

formed in >130 days aged animals. Synaptosomes were labelled with

the radioactive tracers, layered on microporous filters and superfused as

described in Methods. Synaptosomes were exposed to glycine or GABA at

the end of the second fraction collected. Fractions were collected and

counted for radioactivity. Results are expressed as percent potentiation

respect to basal efflux. Data are means � S.E. of 3–7 experiments in

triplicate (three superfusion chambers for each experimental condition).

Fig. 2. Concentration-dependent effects of glycine on the release of [3H]D-

ASP (panel A) or [3H]GABA (panel B) in spinal cord synaptosomes

prepared from SOD1(+) and SOD1-G93A(+) transgenic mice. SOD1-

G93A(+) mice were killed when they showed ingestion disability signs

due to paralysis of the posterior limbs (age >130 days); comparably aged

SOD1(+) animals were used. Synaptosomes were labelled with the radio-

active tracers, layered on microporous filters and superfused as described in

Methods. Synaptosomes were exposed to glycine at the end of the second

fraction collected. Fractions were collected and counted for radioactivity.

Results are expressed as per cent potentiation respect to basal efflux. Data

are means � S.E. of 4–7 experiments in triplicate. *P < 0.05; **P < 0.01 vs.

the effect of glycine in SOD1(+) mice (one-way ANOVA followed by post-

hoc Newman–Keuls test).

data using the following four parameters logistic equa-

tion, provided by the software Sigma Plot version 8.0

(Jandel Corporation, UK):

y ¼ aþ ðb� aÞ½1þ ð10c=10xÞd�

( )

where a is the minimum and b the maximum value of

the data; c the EC50; d the slope of the curve.

The concentration–response curves were statisti-

cally analysed by one-way ANOVA test followed by

post-hoc Newman–Keuls test. The other data were

analyzed by the two-tailed Student’s t-test.

Materials

[3H]D-Aspartate (specific activity: 16.3 Ci/mmol),

[3H]GABA (specific activity: 100.0 Ci/mmol) and

[3H]glycine (specific activity:10.0 Ci/mmol) were pur-

chased from Amersham (Buckinghamshire, UK). Gly-

cine, GABA, aminooxyacetic acid, and strychnine

were obtained from Sigma Chemical Co. (St. Louis,

MO, USA); 5,7-dichlorokynurenic acid was from

Tocris Cookson (Bristol, UK); amoxapine was from

RBI (Sigma Aldrich, Milan, Italy); SKF89976A was

from Smith Kline & French (Welwyn, UK); SR 95531

was a gift of SANOFI (Bruxelles, Belgium); CGP

52432 was a gift from Dr. W. Froestl (Novartis, Basel,

Switzerland); glycyldodecylamide was a gift from Dr

A. Lajtha (Orangeburg, NY, USA) and NFPS (ALX-

5407) was a gift from NPS Allelix Corp (Mississauga,

Ontario, Canada) and Janssen Research Foundation

(Beerse, Belgium).

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892 887

Fig. 3. Concentration-dependent effects of GABA on the release of [3H]D-

ASP (panel A) and [3H]glycine (panel B) in spinal cord synaptosomes

prepared from SOD1(+) and SOD1-G93A(+) transgenic mice. SOD1-

G93A(+) mice were killed when they showed ingestion disability signs

due to paralysis of the posterior limbs (age >130 days); comparably aged

SOD1(+) animals were used. Synaptosomes were labelled with the radio-

active tracers, layered on microporous filters and superfused as described in

theMethods section. Synaptosomes were exposed to GABA at the end of the

second fraction collected. Fractions were collected and counted for radio-

activity. Results are expressed as per cent potentiation respect to basal

efflux. Data are means � S.E. of 4–7 experiments in triplicate. *P < 0.05 vs.

the effect of GABA in SOD1(+) mice (one-way ANOVA followed by post-

hoc Newman–Keuls test).

RESULTS

Fig. 1 (panel A) illustrates the effect of GLY, on the

basal release of [3H]D-ASP or [3H]GABA previously

taken up into spinal cord synaptosomes from nontrans-

genic SOD1(�)/G93A(�) mice aged >130 days. GLY

elicited concentration-dependent release of both

[3H]amino acids. The maximal effects amounted to

175% ([3H]D-ASP) and 200% ([3H]GABA) over basal.

The calculated EC50 values were 68.3 and 209.2 mM

for the GLY-evoked [3H]D-ASP and [3H]GABA

release, respectively. Fig. 1 (panel B) also shows the

effect of GABA on the basal release of [3H]D-ASP or

[3H]GLY previously taken up into spinal cord synapto-

somes from nontransgenic mice. GABA elicited con-

centration-dependent release of [3H]D-ASP (maximal

effect: 100% over basal; EC50 = 3.7 mM). In contrast,

GABA exhibited a very modest effect on [3H]GLY

release.

Fig. 2 describes the results obtained when synapto-

somes prepared from the spinal cord of SOD1(+) or

SOD1-G93A(+) symptomatic mice (aged >130 days)

were labeled with [3H]D-ASP or [3H]GABA and

exposed in superfusion to varying concentrations of

GLY. In SOD1(+) mice, GLY increased the basal

release of [3H]D-ASP (panel A) and of [3H]GABA

(panel B). These results are similar to those obtained

with nontransgenic mice (cf. with Fig. 1, panel A). The

maximal effects of GLY amounted to 150% ([3H]D-

ASP release) and 200% ([3H]GABA release); the

calculated EC50 were 96.1 and 170.0 mM, respectively.

Interestingly, the effect of GLY on [3H]D-ASP release

was significantly more pronounced in synaptosomes

prepared from the spinal cord of SOD1-G93A(+) mice

than from that of SOD1(+) animals; the concentration–

response curve indicates similar potency (EC50 = 94.3)

but higher efficacy (maximal effect 250%). Fig. 2

(panel B) shows that the [3H]GABA released by

GLY did not differ between synaptosomes from

SOD1(+) and SOD1-G93A(+) mice.

The effects of exogenously addedGABAon the basal

release of [3H]D-ASP and [3H]GLY are illustrated in

Fig. 3. Higher amounts of [3H]D-ASP were released by

GABA added to SOD1-G93A(+) synaptosomes than to

SOD1(+) nerve terminals (panel A). The EC50 values

amount to 1.97 mM and to 1.84 mM in SOD1(+) and

SOD1-G93A(+), respectively. The effects of GABA on

the release of [3H]GLY were very modest and identical

in SOD1(+) and SOD1-G93A(+) (panel B).

In order to ascertain whether the potentiation of

[3H]D-ASP release observed in nerve endings from the

spinal cord of symptomatic mice is already present in

pre-symptomatic animals, experiments were performed

with 70–90 and 30–40 days old mice. The results show

that GLY could release more [3H]D-ASP from SOD1-

G93A(+) than from SOD1(+) or SOD1(�)/G93A(�)

synaptosomes also in 70–90 days oldmice (Fig. 4, panel

A). In younger mice (30–40 days), the GLY-evoked

super-release of [3H]D-ASP observed in SOD1-

G93A(+) with respect to SOD1(+) or SOD1(�)/

G93A(�) nerve endings (Fig. 4, panel B) was even

more pronounced than at 70–90 days or at >130 days

(cf. with Fig. 2). Also the GABA-evoked [3H]D-ASP

release was more pronounced in pre-symptomatic 30–

40 days oldmice than in control animals of the same age

(Fig. 5).

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892888

Fig. 4. Concentration-dependent effects of glycine on the release of [3H]D-

ASP in spinal cord synaptosomes prepared from SOD1(�)/G93A(�),

SOD1(+) and SOD1-G93A(+) mice. Experiments were performed using

70–90 (panel A) or 30–40 (panel B) day-old mice. Synaptosomes were

labelled with the radioactive tracer layered on microporous filters and

superfused as described in Methods. Synaptosomes were exposed to glycine

at the end of the second fraction collected. Fractions were collected and

counted for radioactivity. Results are expressed as per cent potentiation

respect to basal efflux. Data are means � S.E. of 4–6 experiments in

triplicate. *P < 0.05; **P < 0.01 when compared to the effect of glycine

in SOD1(+) mice (two-tailed Student’s t-test).

Fig. 5. Concentration-dependent effects of GABA on the release of [3H]D-

ASP in spinal cord synaptosomes prepared from SOD1(�)/G93A(�),

SOD1(+) and SOD1-G93A(+) mice. Experiments were performed using

30–40 days old mice. Synaptosomes were labelled with the radioactive

tracer, layered on microporous filters and superfused as described in

Methods. Synaptosomes were exposed to GABA at the end of the second

fraction collected. Fractions were collected and counted for radioactivity.

Results are expressed as per cent potentiation respect to basal efflux. Data

are means � S.E. of 3–5 experiments in triplicate. *P < 0.05 when com-

pared to the effect of GABA in SOD1(+) mice (two-tailed Student’s t-test).

Table 1

Effects of glycine on the release of endogenous glutamate and

GABA in spinal cord synaptosomes prepared from nontransgenic

SOD1(�)-G93A(�) mice and transgenic SOD1(+) or SOD1-

G93A(+) mice

Percentage of potentiation (means � S.E.)

Glutamate GABA

SOD1(�)-G93A(�) 74.8 � 7.8 (7) 140.0 � 10.1 (5)

SOD1(+) 73.7 � 5.4 (6) 148.0 � 7.1 (5)

SOD1-G93A(+) 102.1 � 10.8 (5)* 131.0 � 6.8 (5)

Synaptosomes were layered on microporous filters and superfused as

described in Methods. Synaptosomes were exposed to glycine (100 mM)

at the end of the second fraction collected. Results are expressed as per cent

potentiation respect to the basal efflux. The basal efflux of endogenous

glutamate and GABA in the 3-min fraction collected before the addition of

glycine amounted to 278.0 � 27.9 pmol/mg and 150.1 � 11.5 pmol/mg

synaptosomal protein, respectively. Data are means � S.E. of the number

of experiments run in triplicate indicated in parentheses.* P < 0.05 vs. the effect of glycine alone (two-tailed Student’s t-test).

All the above experiments were conducted by mon-

itoring release of radiolabelled neurotransmitters. To

establish if the release of [3H]D-ASP and [3H]GABA,

representing the newly taken up amino acid transmit-

ters, reflects release of endogenous glutamate and

GABA, spinal cord synaptosomes from nontransge-

nic SOD1(�)/G93A(�) and transgenic SOD1(+) or

SOD1-G93A(+) mice were superfused in the presence

of 100 mM GLY and the release of endogenous gluta-

mate and GABA was measured by HPLC. GLY

increased the basal release of both amino acids in

SOD1(�)/G93A(�), SOD1(+) and in SOD1-

G93A(+) mice. Similar to the results obtained with

[3H]D-ASP and [3H]GABA, the effects of GLY on

GABA release were comparable in the three groups

of mice, whereas the GLY-evoked release of endogen-

ous glutamate was significantly higher in SOD1-

G93A(+) animals than in SOD1(+) or in nontransgenic

mice (Table 1).

Finally we performed experiments to clarify the

mechanisms by which GLY and GABA evoked neuro-

transmitter release. Table 2 shows that, in non-trans-

genic mice, the GLY-evoked release of [3H]D-ASP was

unaffected by the GLY receptor antagonists strychnine

and 5,7-dichlorokynurenic acid (5,7-DCK), but were

prevented by the GLYT-1/GLYT-2 transporter blocker

glycyldodecylamide (Javitt and Frusciante, 1997). The

effect of GLY was also partly inhibited by 0.1 mM

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892 889

Table 2

Effects of glycine receptor antagonists and of glycine uptake

inhibitors on the glycine-evoked release of [3H]D-ASP and effects

of GABA receptor antagonists and of GABA uptake inhibitors on

the GABA-evoked release of [3H]D-ASP in spinal cord synapto-

somes prepared from nontransgenic SOD1(�)-G93A(�) mice and

transgenic SOD1(+) or SOD1-G93A(+) mice

Drugs Percentage of

potentiation

(means � S.E.)

SOD1(�)-G93A(�)

Glycine (100 mM) 91.0 � 7.4 (5)

Glycine (100 mM) + strychnine (0.1 mM) 81.6 � 12.1 (4)

Glycine (100 mM) + 5,7-DCK (1 mM) 93.1 � 11.5 (4)

Glycine (100 mM) + GDA (50 mM) 21.0 � 4.2 (5)**

Glycine (100 mM) + NFPS (0.1 mM) 37.5 � 1.7 (4) **

Glycine (100 mM) + amoxapine (50 mM) 50.5 � 8.6 (4)*

GABA (10 mM) 73.9 � 8.0 (5)

GABA(10 mM) + SR95531 64.9 � 8.2 (4)

GABA (10 mM) + CGP 52432 57.6 � 6.1 (4)

GABA (10 mM) + SKF89976A 21.5 � 4.6 (5) *

SOD1(+)

Glycine (100 mM) 81.1 � 7.4 (6)

Glycine (100 mM) + strychnine (0.1 mM) 72.1 � 11.1 (3)

Glycine (100 mM) + 5,7-DCK (1 mM) 75.5 � 12.5 (3)

Glycine (100 mM) + GDA (50 mM) 18.7 � 4.2 (3)**

Glycine (100 mM) + NFPS (0.1 mM) 28.3 � 8.5 (4)**

Glycine(100 mM) + amoxapine (50 mM) 36.5 � 5.2 (3)**

SOD1-G93A(+)

Glycine (100 mM) 115.0 � 14.1 (4)

Glycine (100 mM) + strychnine (0.1 mM) 100.3 � 14.5 (3)

Glycine (100 mM) + 5,7-DCK (1 mM) 112.0 � 12.0 (3)

Glycine (100 mM) + GDA (50 mM) 28.7 � 7.8 (3)**

Synaptosomes were labelled with the radioactive tracer, layered on micro-

porous filters and superfused as described in Methods. Synaptosomes were

exposed to glycine (100 mM) or GABA (10 mM) at the end of the second

fraction collected. Receptor antagonists and uptake inhibitors were added to

the superfusion medium 9 min before glycine or GABA. Results are

expressed as per cent potentiation respect to the basal efflux. Data are

means � S.E. of the number of experiments run in triplicate indicated in

parentheses.* P < 0.05 vs. the effect of glycine or GABA alone (two-tailed Student’s

t-test).** P < 0.01 vs. the effect of glycine or GABA alone (two-tailed Student’s

t-test).

NFPS, a potent and selective non-transportable blocker

ofGLY transporters of theGLYT-1 type (Atkinson et al.,

2001; Herdon et al., 2001) or by amoxapine, added at

50 mM, a concentration reported to be selective for the

GLYT-2 transporter (Nunez et al., 2000). Similar results

were obtained with transgenic SOD1-G93A(+) or

SOD1(+) (Table 2). To note, also the GLY-evoked

[3H]GABA release was insensitive to GLY receptor

antagonists but it was abolished by GLY transport

blockers (data not shown). Table 2 also shows that

the release of [3H]D-ASP evoked by GABAwas insen-

sitive to the GABAA receptor antagonist SR95531 or the

GABAB receptor antagonist CGP 52432, but was pre-

vented by the GABA transport blocker SKF89976A, a

selective GAT-1 inhibitor (Borden et al., 1994; Zhou

et al., 2004) in SOD1(�)/G93A(�) mice.

DISCUSSION

The aim of this work has been to study on the effects

of glycine and GABA on the release of glutamate in a

transgenic mouse model of familial ALS, using spinal

cord synaptosomes in superfusion. Synaptosomal pre-

parations contain the different nerve ending families

present in the mammalian CNS. To study the effects of

a modulatory compound on the release of a given

neurotransmitter, indirect effects have to be minimized.

This can be obtained by stratifying synaptosomes on

microporous filters (diameter 2.5 cm) in such an

amount (<100 mg protein/filter) that, based on pre-

vious estimates (Raiteri et al., 1986), the particles

constitute no more than a monolayer. Many studies

have demonstrated that, when such a synaptosomal

layer is up–down superfused, any compound released

is removed before it can activate synaptosomal targets

(transporters, presynaptic receptors and so on). These

targets remain therefore ligand-free and can be selec-

tively activated by ligands added to the superfusion

medium (Raiteri and Raiteri, 2000). Under such con-

ditions, changes in the release of a neurotransmitter A

produced by a compound B can be attributed to direct

action of B on the particles releasing A.

Due to the above considerations, the augmentation

of glutamate release provoked by GLY or by GABA

indicates that GLY or GABA directly acts on gluta-

mate-releasing terminals of SOD1(�)/G93A(�),

SOD1(+) and SOD1-G93A(+) mice. Since the GLY

effect was strychnine-insensitive, GLY does not acti-

vate strychnine-sensitive receptors, which sometimes

mediate stimulation of transmitter release (Darstein

et al., 2000). GLY does not act as glutamate co-agonist

on NMDA receptors either, because its releasing effect

was insensitive to 5,7-DCK, a selective antagonist at

the NMDA receptor glycine site. Instead, the GLY

effect was prevented by the GLY transporter blocker

glycyldodecylamide (Javitt and Frusciante, 1997) sug-

gesting that GLYuptake into glutamatergic terminals is

the primary trigger for glutamate release. We can

conclude that axon terminals able to take up [3H]D-

ASP through glutamate transporters also possess GLY

transporters, the activation of which elevates basal

glutamate release. The GLY effect has an EC50 of

about 90 mM, comparable with reported Km values

for GLYuptake (Lopez-Corcuera et al., 1998; Morrow

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892890

et al., 1998). Two GLY transporter subtypes, termed

GLYT-1 and GLYT-2, have been reported to exist (see,

for a review, Lopez-Corcuera et al., 2001) and non-

transportable selective blockers have become recently

available. The GLY-evoked glutamate release was

partly sensitive to NFPS, a potent GLYT-1 inhibitor

(Atkinson et al., 2001; Herdon et al., 2001), and partly

to amoxapine, which exhibits GLYT-2 selectivity

(Nunez et al., 2000; see Table 2), thus indicating that

GLYT-1 and GLYT-2 are present on glutamatergic

nerve terminals of mouse spinal cord. Considering that

GLYT-1 and GLYT-2 display different expression pat-

terns in the spinal cord (Adams et al., 1995; Zafra et al.,

1995a,b), the possibility that some glutamate nerve

terminals express either GLYT-1 or GLYT-2 appears

more likely than the existence of glutamatergic axon

terminals endowed with both GLYT-1 and GLYT-2.

The GABA-evoked glutamate release was insensi-

tive to GABAA and GABAB receptor antagonists, but it

was blocked by a GABA transporter inhibitor, indicat-

ing involvement of GABA transporters coexisting with

glutamate transporters in spinal cord glutamatergic

nerve terminals. The EC50 of GABA amounted to

about 3 mM, in line with Km values previously reported

in uptake studies (Levi, 1970; Debler and Lajtha,

1987). The finding that the GABA releasing effect

was blocked by the selective GAT-1 inhibitor

SKF89976-A (Borden et al., 1994; Zhou et al.,

2004) indicates that the GABA heterotransporters

involved belongs to the GAT-1 subtype.

Our results show for the first time that glutamate

release can be abnormally enhanced in an animal model

of familial ALS.As shown in Figs. 1–3, GLYandGABA

are more efficacious in releasing glutamate when added

to nerve endings from SOD1-G93A(+) than from

SOD1(+) or SOD1(�)/G93A(�) mice. In contrast,

the GLY-evoked GABA release exhibits identical pat-

tern in SOD1-G93A(+) and in SOD1(+) or SOD1(�)/

G93A(�) animals. Furthermore, GABA was almost

ineffective on the release of GLY. If these effects take

place also in vivo, they could induce unbalances

between inhibitory and excitatory transmitters.

The mechanisms underlying the increased release of

glutamate may be manifold including augmented

expression of GLY and GABA heterotransporters in

the plasmalemmal membrane of glutamatergic axon

terminals, possibly consequent to improved transporter

trafficking. Another possibility is represented by an

augmented efficiency of the glutamate translocation

processes throughout the terminal membrane following

heterotransporter activation. Accordingly, it has been

reported that GLY- or GABA-evoked glutamate release

occurs by reversal of the homotransporter or by the

opening of anion channels (Raiteri et al., 2005). Finally,

the excessive release of glutamate could be triggered by

modifications of the intracellular pathways involved in

heterotransporter-induced neurotransmitter release, pos-

sibly linked to the co-transport of Na+ occurring during

the process of heterotransporter activation. Clarification

of these mechanisms deserves further investigation

because their precise understanding is crucial to the

development of appropriate therapeutic interventions.

The observation that the release of glutamate eli-

cited by GLYor GABAwas higher in SOD1-G93A(+)

mice than in SOD1(+) animals also in experiments

with presymptomatic mice (30–40 or 70–90 days old)

gives support to the idea that the observed effects are

related to the mutation and are not an aspecific con-

sequence of the loss of motor neurons, which is not

present in 8-week-old (Bendotti et al., 2001) or 12-

week-old (Howland et al., 2002) transgenic mice.

Particularly relevant is the observation that the poten-

tiation of GLY-evoked glutamate release was most

pronounced in asymptomatic 30–40 days old SOD1-

G93A(+) mice (Fig. 4). In fact, no changes in GLT-1

(EAAT2) were observed at 60 days of age, before the

appearance of clinical symptoms (Bendotti et al.,

2001). Similarly, glutamate uptake was not decreased

between 8 and 17 weeks in transgenic mice, but was

significantly reduced only at 21 weeks (Canton et al.,

1998). Therefore, the increase of glutamate release

evoked by GLY and by GABA here observed seems

to represent the earliest sign of disease in SOD1-

G93A(+) animals. If the decrease in the glial EAAT2

transporter observed in ALS is a consequence of the

increased glutamate release remains to be established.

Interestingly, elevation of glutamate concentrations

in plasma and cerebrospinal fluid has been documented

in ALS patients (Perry et al., 1990; Rothstein et al.,

1990). Since in our superfused synaptosomes indirect

effects are prevented (Raiteri and Raiteri, 2000), it

seems reasonable to propose that SOD1 mutation

causes modifications in glutamatergic nerve endings,

particularly in the glutamate release machinery. If this

is the case, a thorough examination of the mechanisms

involved in glutamate release should clarify, at least in

part, the relations between SOD1 mutation and

enhanced release of glutamate.

Although the concentrations of GLY and GABA in

the synaptic cleft remain unknown, they could be

inferred from the Km values of the GLY and GABA

uptake processes. The EC50 for the GABA-evoked

glutamate release amounts to about 3 mM, quite com-

patible with the Km values reported for the high affinity

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892 891

uptake by different authors (Levi, 1970; Debler and

Lajtha, 1987). As to the GLY-evoked glutamate

release, the Km values obtained in uptake studies with

transiently or stably expressed GLYT-1 and GLYT-2

transporters (Smith et al., 1992; Lopez-Corcuera et al.,

1998; Morrow et al., 1998) are well in the range of the

EC50 calculated from the present concentration–

response curves. Based on the information available,

the possibility therefore exists that, in particular con-

ditions, GLY and GABA activate their respective het-

erotransporters on glutamatergic terminals causing

excessive glutamate release. The GLY- and the

GABA-evoked release of glutamate may have obvious

pathophysiological significance in presence of a defi-

ciency in the astroglial glutamate transporter EAAT2.

In this scenario, selective GLY and GABA transporter

subtype blockers could be useful not only to limit the

GLY and GABA heterotransporter-induced release of

glutamate, but also to increase the extracellular avail-

ability of the two inhibitory amino acids, thus offering

substantial benefits as future treatments.

ACKNOWLEDGEMENTS

This work was supported by grants from Italian

MIUR (COFIN 2002) and from Italian Ministry of

Health (1% Project). The authors wish to thank Mrs.

Maura Agate for her skilful secretarial assistance.

REFERENCES

Adams RH, Sato K, Shimada S, Tohyama M, Puschel AW, Betz H.

Gene structure and glial expression of the glycine transporter

GlyT1 in embryonic and adult rodents. J Neurosci

1995;15:2524–32.

Atkinson BN, Bell SC, De Vivo M, Kowalski LR, Lechner SM,

Ognyanov VI, et al. ALX 5407: a potent, selective inhibitor of

the hGlyT1 glycine transporter. Mol Pharmacol 2001;60:1414–

20.

Bendotti C, Tortarolo M, Suchak SK, Calvaresi N, Carvelli L,

Bastone A, et al. Transgenic SOD1 G93A mice develop

reduced GLT-1 in spinal cord without alterations in cerebrosp-

inal fluid glutamate levels. J Neurochem 2001;79:737–46.

Bonanno G, Raiteri M. Release-regulating presynaptic heterocar-

riers. Prog Neurobiol 1994;44:451–62.

Borden LA, Dhar TGM. Smith KF, Weinshank RL, Branchek TA,

Gluchowski C. Tiagabine, SK&F 89976-A, CT-966, and NNC-

711 are selective for the cloned GABA transporter GAT-1. Eur J

Pharmacol Mol Pharmacol Sect 1994;269:219–24.

Bradford MM. A rapid and sensitive method for the quantitation of

microgram quantities of protein utilizing the principle of protein

dye binding. Anal Biochem 1976;72:248–54.

Brown RH. Amyotrophic lateral sclerosis: recent insights from

genetic and transgenic mice. Cell 1995;80:687–92.

Canton T, Pratt J, Stutzmann J-M, Imperato A, Boireau A. Glu-

tamate uptake is decreased tardively in the spinal cord of FALS

mice. Neuroreport 1998;9:775–8.

Cluskey S, Ramsden DB. Mechanisms of neurodegeneration in

amyotrophic lateral sclerosis. J Clin Pathol Mol Pathol

2001;54:386–92.

Dal Canto MC, Gurney ME. Development of central nervous

system pathology in a murine transgenic model of human

amyotrophic lateral sclerosis. Am J Pathol 1994;145:1271–9.

Darstein M, Landwehrmeyer GB, Klig C, Becker C-M, Feuerstein

TJ. Strychnine-sensitive glycine receptors in rat caudatoputa-

men are expressed by cholinergic interneurons. Neuroscience

2000;96:33–9.

Debler EA, Lajtha A. High-affinity transport of gamma-aminobu-

tyric acid, glycine, taurine, L-aspartic acid, and L-glutamic acid

in synaptosomal (P2) tissue: a kinetic and substrate specificity. J

Neurochem 1987;48:1851–6.

Furuta A, Martin LJ, Lin CLG. Dykes-Hoberg M, Rothstein JD.

Cellular and synaptic localization of the neuronal glutamate

transporters excitatory amino acid transporters 3 and 4. Neu-

roscience 1997;81:1031–42.

GurneyME, Pu H, Chiu A, Dal CantoMC, Polchow CY, Alexander

DD, et al. Motor neuron degeneration in mice that express a

human Cu, Zn superoxide dismutase mutation. Science

1994;264:1772–5.

Herdon HJ, Godfrey FM, Brown AM, Coulton S, Evans JR, Cairns

WJ. Pharmacological assessment of the role of the glycine

transporter GLYT1 in mediating high-affinity glycine uptake

by rat cerebral cortex and cerebellum synaptosomes. Neuro-

pharmacology 2001;41:88–96.

Horton WA, Eldridge R, Brody JA. Familial motor neuron disease.

Evidence for at least three different types. Neurology

1976;26:460–5.

Howland DS, Liu J, She Y, Goad B, Maragakis NJ, Kim B, et al.

Focal loss of the glutamate transporter EAAT2 in a transgenic

rat model of SOD1mutant-mediated amyotrophic lateral sclero-

sis (ALS). Proc Natl Acad Sci 2002;99:1604–9.

Javitt DC, Frusciante M. Glycyldodecylamide, a phencyclidine

behavioral antagonist, blocks cortical glycine uptake: implica-

tions for schizophrenia and substance abuse. Psychopharmacol-

ogy 1997;129:96–8.

Leigh PN, Meldrum BS. Excitotoxicity in ALS. Neurology

1996;47:S221–7.

Levi G. Cerebral amino acid transport in vitro during development:

a kinetic analysis. Arch Biochem Bioph 1970;138:347–9.

Lopez-Corcuera B, Geerlings A, Aragon C. Glycine neurotrans-

mitter transporters: an update. Mol Membr Biol 2001;18:13–20.

Lopez-Corcuera B, Martinez-Maza R, Nunez E, Roux M, Suppli-

sson S, Aragon C. Differential properties of two stably

expressed brain-specific glycine transporters. J Neurochem

1998;71:2211–9.

Louvel E, Hugon J, Double A. Therapeutic advances in amyo-

trophic lateral sclerosis. Trends Pharmacol Sci 1997;18:196–

203.

Morrison BM, Morrison JH. Amyotrophic lateral sclerosis asso-

ciated with mutations in superoxide dismutase: a putative

mechanism of degeneration. Brain Res Rev 1999;29:121–35.

Morrow JA, Collie IT, Dunbar DR,Walker GB, Shahid M, Hill DR.

Molecular cloning and functional expression of the human

L. Raiteri et al. / NeuroToxicology 26 (2005) 883–892892

glycine transporter GlyT2 and chromosomal localisation of the

gene in the human genome. FEBS Lett 1998;439:334–40.

Mulder D, Kurland L, Offord K, Beard C. Familial adult motor

neuron disease: amyotrophic lateral sclerosis. Neurology

1986;36:511–7.

Nagao S, Kwak S, Kanazawa I. EAAT4, a glutamate transporter

with properties of a chloride channel, is predominantly loca-

lized in Purkinje cell dendrites, and forms parasagittal compart-

ments in rat cerebellum. Neuroscience 1997;78:929–33.

Nunez E, Lopez-Corcuera B, Vazquez J, Gimenez C, Aragon C.

Differential effects of the tricyclic antidepressant amoxapine on

glycine uptake mediated by the recombinant GLYT1 and

GLYT2 glycine transporters. Br J Pharmacol 2000;129:200–6.

Perry TL, Krieger C, Hansen S, Eisen A. Amyotrophic lateral

sclerosis amino acid levels in plasma and cerebrospinal fluid.

Ann Neurol 1990;28:12–7.

Przedborski S, Vila M, Jackson Lewis V. Neurodegeneration: what

is it and where are we? J Clin Invest 2003;111:3–10.

Raiteri L, Raiteri M. Synaptosomes still viable after 25 years of

superfusion. Neurochem Res 2000;25:1265–74.

Raiteri L, Raiteri M, Bonanno G. Coexistence and function of

different neurotransmitter transporters in the plasma membrane

of CNS neurons. Prog Neurobiol 2002;68:287–309.

Raiteri L, Stigliani S, Siri A, Passalacqua M, Melloni E, Raiteri M,

et al. Glycine taken up through GLYT1 and GLYT2 hetero-

transporters into glutamatergic axon terminals of mouse spinal

cord elicits release of glutamate by homotransporter reversal

and through anion channels. Biochem Pharmacol 2005;69:159–

68.

Raiteri M, Bonanno G, Marchi M, Maura G. Is there a functional

linkage between neurotransmitter uptake mechanisms and pre-

synaptic receptors? J Pharmacol Exp Ther 1984;231:671–7.

Raiteri M, Marchi M, Caviglia A. Studies on a possible functional

coupling between presynaptic acetylcholinesterase and high-

affinity choline uptake in the rat brain. J Neurochem

1986;47:1696–9.

Raiteri M, Sala R, Fassio A, Rossetto O, Bonanno G. Entrapping of

impermeant probes of different size into non-permeabilized

synaptosomes as a method to study presynaptic mechanisms.

J Neurochem 2000;74:423–31.

Rosen DR, Siddique T, Patterson D, Figlewicz DA, Sapp P, Hentati

A, et al. Mutations in Cu/Zn superoxide dismutase gene are

associated with familial amyotrophic lateral sclerosis. Nature

1993;362:59–62.

Rothstein JD, Tsai G, Kuncl RW, Clawson L, Cornblath DR,

Drachman DB, et al. Abnormal excitatory amino acid meta-

bolism in amyotrophic lateral sclerosis. Ann Neurol 1990;28:

18–25.

Rothstein JD, Van Kammen M, Levey AI, Martin LJ, Kuncl RW.

Selective loss of glial glutamate transporter GLT-1 in amyo-

trophic lateral sclerosis. Ann Neurol 1995;38:73–84.

Sepkuty JP, Cohen AS, Eccles C, Rafiq A, Behar K, Ganel R, et al.

A neuronal glutamate transporter contributes to neurotransmit-

ter GABA synthesis and epilepsy. J Neurosci 2002;22:

6372–9.

Shaw PJ, Eggett CJ. Molecular factors underlying selective vulner-

ability of motor neurons to neurodegeneration in amyotrophic

lateral sclerosis. J Neurol 2000;247:17–27.

Smith KE, Borden LA, Hartig PR, Branchek T, Weinshank RL.

Cloning and expression of a glycine transporter reveal coloca-

lization with NMDA receptors. Neuron 1992;8:927–35.

Wong PC, Pardo CA, Borchelt DR, LeeMK, Copeland NG, Jenkins

NA, et al. An adverse property of a familial ALS-linked SOD1

mutation causes motor neuron disease characterized by

vacuolar degeneration of mitochondria. Neuron 1995;14:

1105–16.

Yamada K, Watanabe M, Shibata T, Tanaka K, Wada K, Inoue Y.

EAAT4 is a post-synaptic glutamate transporter at Purkinje cell

synapses. Neuroreport 1996;7:2013–7.

Zafra F, Aragon C, Olivares L, Danbolt NC, Gimenez C, Storm-

Mathisen J. Glycine transporters are differentially expressed

among CNS cells. J Neurosci 1995;15:3952–69.

Zafra F, Gomeza J, Olivares L, Aragon C, Gimenez C. Regional

distribution and developmental variation of the glycine trans-

porters GLYT1 and GLYT2 in the rat CNS. Eur J Neurosci

1995;7:1342–52.

Zhou YG, Bennett ER, Kanner BI. The aqueous accessibility in the

external half of transmembrane domain I of the GABA trans-

porter GAT-1 is modulated by its ligands. J Biol Chem

2004;279:13800–8.