Loss of Glial Glutamate and Aspartate Transporter
(Excitatory Amino Acid Transporter 1) Causes
Locomotor Hyperactivity and Exaggerated Responses
to Psychotomimetics: Rescue by Haloperidol and
Metabotropic Glutamate 2/3 Agonist
Rose-Marie Karlsson, Kohichi Tanaka, Markus Heilig, and Andrew Holmes
Background: Recent data suggest that excessive glutamatergic signaling in the prefrontal cortex may contribute to the pathophysiology
of schizophrenia and that promoting presynaptic glutamate modulation via group II metabotropic glutamate 2/3 (mGlu2/3) receptor
activation can exert antipsychotic efficacy. The glial glutamate and aspartate transporter (GLAST) (excitatory amino acid transporter 1
[EAAT1]) regulates extracellular glutamate levels via uptake into glia, but the consequences of GLAST dysfunction for schizophrenia are
Methods: We examined GLAST knockout mice (KO) for behaviors thought to model positive symptoms in schizophrenia (locomotor
hyperactivity to novelty, exaggerated locomotor response to N-methyl-D-aspartate receptor [NMDAR] antagonism) and the ability of
haloperidol and the mGlu2/3 agonist LY379268 to normalize novelty-induced hyperactivity.
Results: Glial glutamate and aspartate transporter KO consistently showed locomotor hyperactivity to a novel but not familiar environ-
ment, relative to wild-type (WT) mice. The locomotor hyperactivity-inducing effects of the NMDAR antagonist MK-801 was exaggerated in
GLAST KO relative to WT. Treatment with haloperidol or LY379268 normalized novelty-induced locomotor hyperactivity in GLAST KO.
be a pathophysiological risk factor for the disease. Our findings provide novel support for the hypothesis that glutamate dysregulation
contributes to the pathophysiology of schizophrenia and for the antipsychotic potential of mGlu2/3 agonists.
Key Words: GLAST, glutamate, mGlu2/3, NMDA, schizophrenia
in brain regions functionally compromised in schizophrenia,
such as the hippocampus and prefrontal cortex (PFC) (3).
Furthermore, treatment with N-methyl-D-aspartate receptor
(NMDAR) antagonists such as ketamine and phencyclidine (PCP)
mimic the symptoms of schizophrenia in healthy subjects and
provoke relapse in schizophrenic patients (4,5). In rodents,
NMDAR antagonists produce a range of schizophrenia-related
behavioral abnormalities including locomotor hyperactivity and
cognitive dysfunction (6,7). While the psychotomimetic effects of
NMDAR antagonists has fostered the notion of a hypoglutama-
tergic state in schizophrenia, recent data suggest that these
effects stem from PFC glutamate excess caused by disinhibition
of NMDAR-containing gamma-aminobutyric acid (GABA)ergic
lutamatergic dysfunction is increasingly implicated in the
pathophysiology of schizophrenia (1,2). Patients with
schizophrenia exhibit alterations in glutamate receptors
interneurons (8–13). In support of this hypothesis is preliminary
evidence of antipsychotic efficacy of group II metabotropic
glutamate 2/3 (mGlu2/3) agonists (14) that negatively modulate
glutamate in PFC (15).
These developments raise the possibility that molecular ab-
normalities leading to PFC glutamate excess could be risk factors
for schizophrenia. Glial glutamate and aspartate transporter
(GLAST) (excitatory amino acid transporter 1 [EAAT1]), glial
glutamate transporter-1 (GLT-1) (excitatory amino acid trans-
porter 2 [EAAT2]), excitatory amino-acid carrier 1 (EAAC1)
(excitatory amino acid transporter 3 [EAAT3]), and excitatory
amino acid transporter 4 (EAAT4) belong to a family of sodium-
dependent glutamate transporters that tightly regulate extracel-
lular concentrations (16,17). Glial glutamate and aspartate trans-
porter is mainly expressed in astrocytes within the rodent
cerebellum but is also found, albeit in lesser concentrations, in
hippocampus, cerebral cortex, and thalamus (16,17). Antisense
knockdown of GLAST has been shown to cause increased levels
of extracellular glutamate concentrations in hippocampus and
striatum and increased vulnerability to glutamatergic toxicity
(18). Intriguingly, genetic mutation of human GLAST (SLC1A3)
was recently linked to schizophrenia (19). However, the conse-
quences of disrupted GLAST function for schizophrenia-related
behavioral phenotypes have not been studied, in part due to a
lack of GLAST-specific pharmacological tools.
Here, we tested a GLAST knockout (KO) model (20) for pheno-
types considered relevant to the positive symptoms of schizophre-
nia (locomotor hyperactivity to novelty, hypersensitivity to NMDAR
antagonists) (6,7). We then tested whether the novelty-induced
hyperlocomotor phenotype was reversible with a typical antipsy-
chotic (haloperidol) and an mGlu2/3 agonist (LY379268).
From the Laboratory for Clinical and Translational Studies (R-MK, MH), Na-
tional Institute on Alcoholism and Alcohol Abuse, National Institutes of
Health, Bethesda; Laboratory of Molecular Neuroscience (KT), School of
Biomedical Science and Medical Research Institute, Tokyo Medical and
Dental University Bunkyo-ku, Tokyo, Japan; and Section on Behavioral
Science and Genetics (AH), Laboratory for Integrative Neuroscience,
National Institute on Alcoholism and Alcohol Abuse, National Institutes
of Health, Rockville, Maryland.
holism, National Institutes of Health, 10 Center Drive, Building 10 Room
1-5330, Bethesda, MD 20892; E-mail: firstname.lastname@example.org.
Received December 17, 2007; revised April 8, 2008; accepted May 6, 2008.
BIOL PSYCHIATRY 2008;64:810–814
Published by Society of Biological Psychiatry
Methods and Materials
Glial glutamate and aspartate transporter KO were generated
as previously described (20) (Supplement 1). Experimental pro-
cedures were approved by the National Institute on Alcohol
Abuse and Alcoholism (NIAAA) Animal Care and Use Committee
and followed the National Institutes of Health (NIH) guidelines,
“Using Animals in Intramural Research.”
Locomotor Activity in Novel and Familiar Environment
Locomotor responses were tested in a novel square arena
(40 ? 40 ? 35 cm, 55 lux) constructed of white Plexiglas as
previously described (21). Distance traveled and time spent in
the center (20 ? 20 cm) were measured over 60 min via
Ethovision (Noldus Information Technology Inc., Leesburg, Vir-
Locomotor activity in a familiar environment was assessed
over 24 hours (after 48 hours acclimation) in individual home
cages under normal vivarium conditions via the photocell-based
Opto M3 activity monitor (Columbus Instruments, Columbus,
Locomotor Response to MK-801
To circumvent basal hyperactivity to novelty in GLAST KO
confounding assessment of the hyperactivity-inducing effects of
MK-801, mice tested for the response to novel stimulus (above)
were re-exposed to the open field and acclimated for a further 60
min. Mice were then injected with .2 mg/kg MK-801 and returned
to the open field for 60 min (further details in Supplement 1).
Locomotor Effects of Haloperidol and mGlu2/3 Agonist
Open field naïve mice were injected intraperitoneal (IP) with
vehicle or .3 mg/kg haloperidol and tested in the open field for
30 min as above.
Before testing the effects of LY379268 in GLAST KO, we first
established doses sufficient to normalize phencyclidine-induced
locomotor activity in nonmutant male C57BL/6J mice. Mice were
injected with saline or .3, 1.0, or 3.0 mg/kg LY379268 and 30 min
later, were placed in the open field. After 60 min of acclimation,
mice were injected with 5.0 mg/kg phencyclidine and tested in
the open field for 60 min as above.
On the basis of the C57BL/6J dose-response data (Supplement
2), open field naïve GLAST KO were injected with vehicle or 1.0
mg/kg LY379268 and 30 min later tested in the open field for 30
min as above.
Data were analyzed using Statistica (Statsoft, Tulsa, Okla-
homa). Effects of genotype, sex, drug treatment, and time were
analyzed using analysis of variance (ANOVA) (repeated mea-
sures for time) followed by Tukey post hoc tests.
Locomotor Activity in Novel and Familiar Environment
In the novel open field, there was a significant main effect of
genotype [F(2,45) ? 10.58, p ? .01] and time [F(11,495) ? 2.00,
p ? .05] but no genotype ? time interaction for distance traveled
(Figure 1A). Post hoc analysis revealed that KO were significantly
more active than wild-type (WT) during novel open field exposure.
We also noted a significant main effect of sex [F(1,45) ? 4.70, p ?
.05] and significant genotype ? sex interaction [F(2,45) ? 3.39, p
? .05] for distance traveled and therefore further analyzed
locomotor activity separately within each sex. This confirmed a
robust and significant effect of genotype in both male [F(2,22) ?
6.70, p ? .01] and female mice [F(2,23) ? 7.18, p ? .01], with a
more pronounced difference between female KO and WT than
between male KO and WT underlying the interaction effect in the
global analysis. Time spent in the center of open field, a putative
measure of anxiety-like behavior, was unaffected by genotype or
sex (Figure 1B). There was no significant effect of genotype, sex,
or genotype ? sex interaction for locomotor activity in the
familiar home cage (Figure 1C). Two mice were removed from
the home cage test due to scores higher than two standard
deviations (SD) above mean (SD ? 209695).
Locomotor Response to MK-801
Analyzing predrug time block, there was a main effect of time
[F(11,242) ? 19.85, p ? .01], reflecting some residual habituation
in all genotype groups, but no main effect of genotype or
genotype ? time interaction. The lack of genotype effect,
GLAST KO. (A) GLAST KO showed elevated locomotor activity than WT
during exposure to a novel open field (n ? 14–20). (B) Time spent in the
center of the open field did not differ between genotypes. (C) Home cage
locomotor activity was no different between genotypes (n ? 13–16). **p ?
HET, heterozygous; KO, knockout; WT, wild-type.
R.-M. Karlsson et al.
BIOL PSYCHIATRY 2008;64:810–814 811
together with the lower overall activity levels compared with the
novel open field, shows that hyperactivity in GLAST KOs is
selectively found under conditions of novelty.
The response to MK-801 challenge was subsequently evalu-
ated using mixed-model ANOVA for postinjection time points
and controlling for individual differences in baseline as covari-
ates using the last 15 min of predrug baseline. There was a
significant effect of genotype [F(2,21) ? 4.21, p ? .05], time
[F(5,105) ? 2.57, p ? .05], and a genotype ? time interaction
[F(22,231) ? 1.87, p ? .05] for total distance traveled. Post hoc
analysis showed that MK-801 increased activity in KO to a greater
extent than WT (Figure 2). There was also significant effect of sex
[F(2,21) ? 6.88, p ? .05] and a sex ? time interaction [F(11,231) ?
1.99, p ? .05] due to female mice responding to MK-801
challenge more profoundly than male mice. Sex did not signifi-
cantly influence the genotypic differences in responsivity to
MK-801, as shown by a lack of genotype ? sex or genotype ?
sex ? time interactions.
Locomotor Effects of Haloperidol in GLAST KO Mice
For the haloperidol rescue test, there was a significant main
effect of genotype [F(2,62) ? 5.00, p ? .01] and drug [F(1,62) ?
30.93, p ? .01] and a significant genotype ? drug interaction
[F(2,62) ? 5.97, p ? .01] for total distance traveled. Post hoc tests
showed that KO were more active than WT following vehicle
treatment but not following haloperidol treatment (Figure 3A).
Haloperidol significantly reduced locomotor activity in heterozy-
gous (HET) and KO but not WT. There was an overall effect of
sex [F(1,62) ? 4.29, p ?. 05] due to higher scores in female mice
than male mice (post hoc test: p ?. 01) but sex did not interact
with either drug or genotype.
Reversal of PCP-Induced Hyperlocomotion by the mGlu2/3
Agonist LY379268 in C57BL/6J Mice
There was an overall significant effect of treatment [F(3,32) ?
11.08, p ? .01], time [F(3,32) ? 27.88, p ? .01], and a time ? drug
interaction for total distance traveled [F(69,736) ? 3.76, p ? .01].
Analyzing pre-PCP, there was a significant effect of treatment
[F(3,32) ? 4.58, p ? .01] and time [F(11,352) ? 32.94, p ? .01] but
no treatment ? time interaction. Post hoc tests showed that prior to
the PCP challenge, 3.0 but not .3 or 1.0 mg/kg LY379268 signifi-
cantly decreased activity relative to vehicle (Supplement 2).
The response to PCP challenge was evaluated using mixed-
model ANOVA for postinjection time points and controlling for
individual differences in baseline as covariates using the last 15
min of predrug baseline. There was a significant effect of
treatment [F(3,31) ? 11.64, p ? .01] but no time or treatment ?
time interaction effect. Post hoc test showed that 1.0 and 3.0
mg/kg LY379268 prevented PCP-induced hyperactivity relative
to vehicle (Supplement 2).
Locomotor Effects of LY379268 in GLAST KO Mice
There was a significant effect of genotype [F(2,68) ? 20.29,
p ? .01], drug [F(1,68) ? 29.33, p ? .01], and genotype ? drug
interaction [F(2,68) ? 12.31, p ? .01] for total distance traveled.
Post hoc tests showed that KO were more active than WT
following vehicle treatment but not following LY379268 treat-
ment (Figure 3B). LY379268 significantly reduced locomotor
activity in KO but not HET or WT. Because we observed a
significant effect of sex [F(1,68) ? 12.77, p ? .01] and significant
genotype ? sex interaction [F(2,68) ? 13.27, p ? .01], we also
analyzed genotype and LY379268 effects separately within each
sex and found a rescue of GLAST KO hyperlocomotion in both
male and female mice. Thus, in male mice, there was a significant
effect of genotype [F(2,32) ? 4.76, p ? .05], LY379268 [F(1,32) ?
6.18, p ? .05], and their interaction [F(2,32) ? 4.63, p ? .05]. Post
hoc analysis showed that KO were more active than WT after
vehicle but not LY379268 treatment. Similarly, in female mice,
Figure 3. Locomotor hyperactivity in GLAST KO is rescued by haloperidol
and the mGlu2/3 receptor agonist LY379268. (A) Treatment with .3 mg/kg
haloperidol normalized novel open field locomotor hyperactivity in GLAST
KO (n ? 10–14/genotype/treatment). (B) Treatment with 1 mg/kg
(n ? 10–15/genotype/treatment). *p ? .05, **p ? .01 versus WT/same
treatment; #p ? .05, ##p ? .01 versus vehicle/same genotype. Data are
mean ? SEM. GLAST, glial glutamate and aspartate transporter; HET, het-
erozygous; KO, knockout; mGlu2/3, metabotropic glutamate 2/3; WT, wild-
Figure 2. Increased sensitivity to the locomotor hyperactivity-inducing ef-
fects of MK-801 in GLAST KO. Treatment with .2 mg/kg MK-801 produced a
significantly greater increase in open field locomotor activity in GLAST KO
812 BIOL PSYCHIATRY 2008;64:810–814
R.-M. Karlsson et al.
there was a significant effect of genotype [F(2,36) ? 18.51, p ?
.01], LY379268 [F(1,36) ? 26.47, p ? .01], and their interaction
[F(2,36) ? 9.77, p ? .01]. Again, post hoc tests demonstrated that
KO were more active than WT after vehicle but not after
The first major novel finding of the present study was that
GLAST KO showed a significant and robust open field locomotor
hyperactivity relative to WT. This was restricted to conditions of
novelty and not seen after prior habituation to the open field (see
MK-801-challenge experiment) or in the familiar environment of
the home cage. Together, these data demonstrate an exaggerated
GLAST KO locomotor response under conditions of sufficient
The basal locomotor hyperactivity in GLAST KO phenocopies
the effects of NMDAR antagonists in nonmutant mice (6,7). These
drugs also aggravate symptoms in schizophrenic patients and
simulate psychosis in normal subjects (4,5). Therefore, providing
further support for a schizophrenia-related abnormality in GLAST
KO, these mice exhibited an exaggerated locomotor hyperactiv-
ity response to the noncompetitive NMDAR antagonist MK-801.
A final important observation was that basal GLAST KO locomo-
tor hyperactivity was rescued by the prototypical antipsychotic
haloperidol and the mGlu2/3 agonist LY379268 recently shown
to have therapeutic efficacy in schizophrenia (14). These findings
provide a pharmacological validation that reversal of GLAST KO
induced hyperactivity may have predictive activity for clinical
efficacy in schizophrenia and therefore provide a useful tool for
screening novel antipsychotics.
Mechanistically, the ability of NMDAR antagonists to provoke
human psychosis and produce schizophrenia-related behaviors
in rodents has been linked to a loss of NMDAR-mediated
GABAergic inhibition, leading to excessive glutamate release and
neuronal hyperexcitability in PFC (8–13). Loss of GLAST is
predicted to cause glutamate excess and extrasynaptic spillover
under conditions that provoke glutamate release, including
novelty and NMDAR-mediated GABA inhibition. Although vol-
tametry or microdialysis measures of PFC extracellular glutamate
in behaving GLAST KO would provide direct evidence of this,
indirect support for this notion is given by the ability of LY379268
to rescue GLAST KO phenotype and is entirely consistent with
the notion that extrasynaptic mGlu2/3 receptors facilitate inhib-
itory control of glutamate release and mitigate the effects of
impaired clearance (22). In this context, our data offer novel
support for the hypothesis that excessive glutamate contributes
to the pathophysiology of schizophrenia and lend further cre-
dence to the antipsychotic potential of mGlu2/3 agonists (14).
A number of issues await clarification. First, the relative
contribution of GLAST to forebrain glutamate signaling remains
to be established. Glial glutamate and aspartate transporter
messenger RNA and protein expression is most heavily concen-
trated in the cerebellum rather than the forebrain regions impli-
cated in schizophrenia, such as the PFC and hippocampus, and
neurons in these areas are not surrounded by high numbers of
astrocytes where the majority of GLAST-mediated glutamate
reuptake is thought to occur (16,17). Second, although acute loss
of glutamate clearance provides a plausible mechanism for the
schizophrenia-related abnormalities in GLAST KO, this does not
exclude the potential for constitutive loss of GLAST to cause more
permanent neural changes resulting from excitotoxicity (18,20) or
developmental abnormalities. The neurodevelopment issue is par-
ticularly salient given that high transient GLAST gene promoter
activity is present during early postnatal mouse development in
both the cortex and hippocampus (23) and the posited neurode-
velopmental ontogeny of schizophrenia. Third, possible GLAST KO
abnormalities on behaviors pertinent to the negative and cognitive
symptoms of schizophrenia await further study.
In summary, present data demonstrate that GLAST KO exhibit
abnormalities on behavioral measures thought to model positive
symptoms of schizophrenia. These disturbances were normal-
ized by the prototypical antipsychotic haloperidol and the
mGlu2/3 receptor agonist LY379268. Present findings suggest
that loss of GLAST-mediated glutamate clearance could be a
pathophysiological risk factor for schizophrenia and are partic-
ularly intriguing given recent evidence linking genetic mutation
of human GLAST (SLC1A3) with schizophrenia (19). More gen-
erally, our data provide novel support for the hypothesis that
excessive glutamate neurotransmission contributes to psychosis.
This work was supported by the National Institute on Alcohol
Abuse and Alcoholism (NIAAA)-Intramural Research Program
(IRP). Tanaka was supported by The Novartis Foundation (Japan)
for the Promotion of Science, Takeda Science Foundation, The
Tokyo Biochemical Research Foundation, Research Foundation of
Opto-Science and Technology, and Grants-in-Aids for Scientific
Research on Priority Area (20022013 and 18053006) provided by
the Ministry of Education, Culture, Sports, Science, and Technology
We thank Dr. Jeffrey Rothstein at the John Hopkins University
for providing the mutant mice for breeding and Dr. Judith Davis
and Monique Melige for assistance with breeding.
The authors reported no biomedical financial interests or
potential conflicts of interest.
Supplementary material cited in this article is available
1. Coyle JT (2006): Glutamate and schizophrenia: Beyond the dopamine
2. Harrison PJ, Weinberger DR (2005): Schizophrenia genes, gene expres-
sion, and neuropathology: On the matter of their convergence. Mol
Psychiatry 10:40–68; image 45.
3. Beneyto M, Kristiansen LV, Oni-Orisan A, McCullumsmith RE, Meador-
Woodruff JH (2007): Abnormal glutamate receptor expression in the
medial temporal lobe in schizophrenia and mood disorders. Neuropsy-
5. Malhotra AK, Pinals DA, Adler CM, Elman I, Clifton A, Pickar D, et al.
(1997): Ketamine-induced exacerbation of psychotic symptoms and
6. Arguello PA, Gogos JA (2006): Modeling madness in mice: One piece at
a time. Neuron 52:179–196.
7. Powell CM, Miyakawa T (2006): Schizophrenia-relevant behavioral test-
ing in rodent models: A uniquely human disorder? Biol Psychiatry 59:
neurons. J Neurosci 27:11496–11500.
9. Jackson ME, Homayoun H, Moghaddam B (2004): NMDA receptor hypo-
function produces concomitant firing rate potentiation and burst activity
chotic drugs reverse the disruption in prefrontal cortex function pro-
duced by NMDA receptor blockade with phencyclidine. Proc Natl Acad
R.-M. Karlsson et al.
BIOL PSYCHIATRY 2008;64:810–814 813
11. Lopez-Gil X, Babot Z, Amargos-Bosch M, Sunol C, Artigas F, Adell A Download full-text
(2007): Clozapine and haloperidol differently suppress the MK-801-in-
frontal cortex of the rat. Neuropsychopharmacology 32:2087–2097.
of ketamine and N-methyl-D-aspartate on glutamate and dopamine
release in the rat prefrontal cortex: Modulation by a group II selective
group II metabotropic glutamate receptor agonist in rats. Science 281:
15. Conn PJ, Pin JP (1997): Pharmacology and functions of metabotropic
16. Torres GE, Amara SG (2007): Glutamate and monoamine transporters:
New visions of form and function. Curr Opin Neurobiol 17:304–312.
17. Tzingounis AV, Wadiche JI (2007): Glutamate transporters: Confining
runaway excitation by shaping synaptic transmission. Nat Rev Neurosci
(1996): Knockout of glutamate transporters reveals a major role for
19. Walsh T, McClellan JM, McCarthy SE, Addington AM, Pierce SB, Coo-
per GM, et al. (2008): Rare structural variants disrupt multiple genes
in neurodevelopmental pathways in schizophrenia. Science 320:
20. Watase K, Hashimoto K, Kano M, Yamada K, Watanabe M, Inoue Y, et al.
injury in GLAST mutant mice. Eur J Neurosci 10:976–988.
al. (2008): Mice lacking the AMPA GluR1 receptor exhibit striatal hyper-
tropic glutamate 2/3 receptors reverses the effects of NMDA receptor
hypofunction on prefrontal cortex unit activity in awake rats. J Neuro-
(2007): Variations in promoter activity reveal a differential expression
mature CNS. J Neurosci 27:6607–6619.
814 BIOL PSYCHIATRY 2008;64:810–814
R.-M. Karlsson et al.