This study investigated whether polymorphisms in the glycogen synthase kinase 3B (GSK-3B) gene are associated with tardive dyskinesia (TD) in schizophrenia patients. The researchers genotyped 8 single nucleotide polymorphisms in the GSK-3B gene in 215 schizophrenia patients, 169 of whom were European Caucasian. They found that minor alleles of all variants were associated with lower risk of TD and lower abnormal involuntary movement severity scores. Three variants (rs6805251, rs6438552, and rs9878473) showed a trend toward association with TD in European Caucasian patients. When age was included as a covariate, all variants were associated with abnormal involuntary movement severity scores at p

1. Association study of the GSK-3B gene with tardive
dyskinesia in European Caucasians
Renan P. Souza a,⁎, Gary Remington a,b
, Nabilah I. Chowdhury a
,
Matthew K. Lau a
, Aristotle N. Voineskos a
, Jeffrey A. Lieberman c
,
Herbert Y. Meltzer d
, James L. Kennedy a,b
a
Neurogenetics Section, Centre for Addiction and Mental Health, Toronto, Ontario, Canada
b
Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada
c
New York State Psychiatric Institute, Columbia University Medical Centre, New York City, NY, USA
d
Psychiatric Hospital, Vanderbilt University, Nashville, TN, USA
Received 28 March 2010; received in revised form 26 April 2010; accepted 2 May 2010
KEYWORDS
AIMS;
Glycogen synthase kinase;
Movement disorders;
Extrapyramidal symptoms;
Pharmacogenetics;
Genetics
Abstract
There is solid evidence of a genetic predisposition to tardive dyskinesia (TD) although the
pathophysiological mechanisms of TD are still unclear. Nevertheless, the dopamine overactivity
hypothesis of the TD etiology receives support from both pharmacological and physiological evidence.
Dopaminergic signaling modulates the glycogen synthase kinase 3B (GSK-3B), a kinase that may play a
critical role in the pathogenesis of neurodegenerative diseases. GSK-3B is an essential element of the
apoptotic signaling cascade induced by oxidative stress, which may be involved in TD pathogenesis.
We investigated whether GSK-3B polymorphisms (rs11919783, rs6805251, rs7624540, rs6438552,
rs4072520, rs9878473, rs6779828 and rs3755557) selected using tagging method were associated with
TD manifestation and abnormal involuntary movement severity. We evaluated 215 schizophrenia
subjects from whom 169 were European Caucasians. All eight evaluated variants had their minor
allele carriers consistently showing lower risk to TD and lower Abnormal Involuntary Movement Scale.
The rs6805251, rs6438552 and rs9878473 variants showed a trend for association with TD in European
Caucasian subjects (permuted p=0.07). Furthermore, all tested markers showed p≤0.0007 after we
incorporated age as covariate in the analysis of the abnormal involuntary movement severity. Our
results suggest that GSK-3B polymorphism may play a role in the genetic vulnerability to TD
manifestation in schizophrenia subjects with European Caucasian background further implicating
polymorphisms in the dopamine D2-like receptor signaling in this context. These findings should be
read with caution particularly before independent replication.
© 2010 Elsevier B.V. and ENCP. All rights reserved.
⁎ Corresponding author. Centre for Addiction and Mental Health, 250 College Street, Toronto, Canada M5T 1R8. Tel.: +1 416 5358501; fax +1 416
9794666.
E-mail addresses: renanrps@yahoo.com.br (R.P. Souza), james_kennedy@camh.net (J.L. Kennedy).
0924-977X/$ - see front matter © 2010 Elsevier B.V. and ENCP. All rights reserved.
doi:10.1016/j.euroneuro.2010.05.002
www.elsevier.com/locate/euroneuro
European Neuropsychopharmacology (2010) 20, 688–694

2. 1. Introduction
Tardive dyskinesia (TD), affecting approximately 20% of
individuals treated longer term with typical antipsychotics,
is a potentially irreversible movement disorder with a
chronic fluctuating course (Schooler and Kane, 1982). A
recent study suggests this same prevalence in subjects
treated for more than four years (Park et al., 2009).
Clinically, abnormal movements mainly involve the tongue,
face, and jaw, with choreoathetoid movements of the
extremities or trunk also observed. There is considerable
variability between individuals in degree of susceptibility to
these adverse effects, and it is widely assumed that a
substantial portion is genetically determined. The observa-
tion that TD is more likely to occur in patients with familial
history of TD provides further evidence for the involvement
of genetic factors. Thus far, most of the pharmacogenetic
studies involving TD have focused on candidate genes
selected based on pharmacokinetics or pharmacodynamics
(Thelma et al., 2008).
The dopamine overactivity hypothesis of TD has reason-
able support from pharmacological and physiological evi-
dence, and continues to be the most plausible explanation
for TD pathophysiology. This hypothesis proposes that the
nigrostriatal dopamine system develops increased sensitivity
to dopamine because of chronic dopamine receptor blockade
induced by antipsychotic drugs. In this context, the glycogen
synthase kinase 3B (GSK-3B) may be an important component
because the Akt/GSK-3 signaling cascade has been strongly
associated with the actions of dopamine and antipsychotics
(Beaulieu et al., 2009). Further, GSK-3B has shown a critical
role in the pathogenesis of neurodegenerative diseases and it
is an essential element of the apoptotic signaling cascade
induced by oxidative stress (Jope and Roh, 2006), which may
be involved in the pathogenesis of TD.
Increased GSK-3 activity has been reported to result in
neuronal apoptosis, and GSK-3 inhibitors have been shown to
exert anti-apoptotic and neuroprotective effects in several
cell and mouse models (Beurel and Jope, 2006). Previous
reports have documented GSK-3B-mediated neuroprotection
in diseases associated with a neurodegenerative process and
movement disorder symptoms. Blockage of GSK-3B activa-
tion almost completely prevents the loss of dopaminergic
neurons after 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
(MPTP) administration in rats (Wang et al., 2007) and GSK-3B
inhibitors prevent cellular polyglutamine toxicity caused by
the Huntington's disease mutation (Carmichael et al., 2002).
Moreover, genetic association studies have also implicated
GSK-3B in the susceptibility to develop neurodegenerative
diseases (Kwok et al., 2005). Polymorphisms in the GSK-3B
gene (e.g. rs6438552) alter transcription, splicing and its
interaction with Tau proteins, modifying the vulnerability of
Parkinson's disease (Kwok et al., 2005).
GSK-3B may also modulate synaptic plasticity like Akt in
the brain. However, GSK-3B is constitutively active in resting
cells, in contrast to Akt, requiring phosphorylation by kinases
such as Akt to inactivate it. Studies have consistently shown
increased GSK-3B phosphorylation after (1) clozapine treat-
ment in SH-SY5Y cells (Kang et al., 2004), (2) systemic
haloperidol and clozapine administration in rat frontal
cortex although each drug affected upstream regulators
differently (Roh et al., 2007) and (3) systemic risperidone
administration in rat cortex, hippocampus, striatum, and
cerebellum in a dose-dependent manner. Taken together,
these results indicate that both typical and atypical
antipsychotics induce GSK-3B inactivation through its phos-
phorylation in vitro and in animal models.
In a previous report, our research group reported a strong
association of gene–gene interaction between dopamine D2
receptor (DRD2) and Akt (AKT1) genes acting as a vulnera-
bility component in TD (Zai et al., 2008). We have also shown
associations in the DRD2, DRD3 and DRD4 genes with TD,
indicating that polymorphisms located in genes associated
with the D2-like receptor signaling may increase TD risk (Zai
et al., 2008; Zai et al., 2009a; Zai et al., 2009b). To
investigate further whether this pathway is associated with
TD, we evaluated eight variants in the GSK-3B gene in 215
schizophrenia subjects assessed for TD manifestation. We
also analyzed whether these polymorphisms in the GSK-3B
gene could predict the severity of abnormal involuntary
movements in these subjects because of previous reports
linking GSK-3B with other movement disorders.
2. Experimental procedures
2.1. Sample description
This research protocol was approved by the local Research Ethics
Committee, and subjects signed informed consent in accordance
with the Declaration of Helsinki. Two hundred and fifteen subjects
were recruited from the Centre for Addiction and Mental Health
(n=109, 65.1% males, mean age 41.6±9.9), Case Western Reserve
(n=65; 72.3% males, mean age 32.6±9.4) and the Hillside Hospital
(n=41; 58.5% males, mean age 34.8±7.2). Diagnosis was established
using the Structured Clinical Interview for Diagnostic and Statistical
Manual of Mental Disorders version III-R or IV (SCID). Subjects with
schizophrenia or schizoaffective depressive type were enrolled in
the study while subjects with history of major neurological
disorders, head injury with significant loss of consciousness or
history of substance abuse were excluded from the study. All
enrolled subjects had undergone at least one year of treatment with
typical antipsychotics. This sample has been evaluated in previous
studies (Souza et al., 2010a; Souza et al., 2010b).
Seventy eight percent (n=169) of the probands reported having
at least three out of four grandparents with European Caucasian
background (Centre for Addiction and Mental Health n=86; 62.9%
males, mean age 42.7±9.8; Case Western Reserve n=42; 73.8%
males, mean age 31.9±8.4 and Hillside Hospital n=41; 60.9% males,
mean age 34.8±7.2). Although ethnicity was not determined using
ancestry informative markers, modeling the possible effect of
population stratification suggests an effect that is likely to be
small in most situations, especially when the O.R. is higher than 1.20
or lower than 0.83 (Little et al., 2009).
2.2. Tardive dyskinesia assessment
All three clinicians (GR, HYM, and JAL) are highly experienced in TD
assessment, and consistency was further enhanced by exchange
visits across sites (Zai et al., 2009a). TD was assessed using the
Abnormal Involuntary Movement Scale (AIMS) and the Schooler and
Kane criteria (Schooler and Kane, 1982), as well as the Hillside
Simpson Dyskinesia Scale (HSDS) for patients recruited from Hillside
Hospital (JAL). The seven body area items and overall global item of
the HSDS match those of the AIMS; thus, assessment for presence of
TD was equivalent for all four sites. The majority of patients
collected at the Centre for Addiction and Mental Health were on
689
Tardive dyskinesia and GSK-3b

3. atypical antipsychotics at the time of TD assessment; as noted
though, they had all been previously treated with typical antipsy-
chotic medication for at least one year (Table 1). Subjects collected
at the Case Western Reserve and Hillside Hospital were part of a
clinical trial where all patients had received typical antipsychotics
before assessment for TD. Subjects have not received any treatment
for abnormal involuntary movements prior of the TD assessment.
2.3. Genetic analysis
The GSK-3B gene has 192 polymorphisms described in the Caucasian
(CEU) Hap Map population spanning 267 kbp. Of considerable
interest, these variants are grouped in a single linkage disequilib-
rium block comprised of 264 kbp. We selected eight tag SNPs
(Table 2) from variants with minor allele frequency higher than 15%
(n=55) due to our sample size. This tag SNP selection covers 87% of
the common genetic variants located in the GSK-3B gene (para-
meters: r2
threshold=0.8; 2- and 3-marker haplotype tagging
approach; mean maximum r2
=0.992). Tagging was performed
using Haploview 4.1.
DNA isolation was performed using high-salt method (Lahiri and
Nurnberger, 1991). Genotyping was performed at The Centre for
Applied Genomics (Toronto, ON, Canada) using Illumina platform
(Illumina, San Diego, CA, USA) (Souza et al., 2008). We regenotyped
rs4072520 using TaqMan SNP genotyping assays (Applied Biosystems,
Foster City, CA, USA), and performed allelic discrimination using
automated procedure implemented in the 7500 System Sequence
Detection Software version 1.3.1.21 (Applied Biosystems, Foster
City, CA, USA). In accordance with general guidelines for association
studies (Little et al., 2009), we used six negative controls for each
96-well plate; research staff was blind to clinical data and genetic
data was matched with automated database. The common allele for
each variant was designated as “1” while the minor allele was
identified as “2”.
2.4. Statistical analysis
Hardy–Weinberg equilibrium was accessed using a χ2
test. We
evaluated categorical data using the χ2
test and continuous data
using analysis of variance or covariance implemented on the
Statistical Package for the Social Sciences version 13.0. We
calculated pairwise linkage disequilibrium (LD) coefficient from
expectation–maximization algorithm-derived haplotype frequen-
cies through Haploview 4.1 (Barrett et al., 2005). We defined a
haplotype block as a region over which less than five percent of
pairwise comparisons among informative SNPs showed strong
evidence of historical recombination (upper confidence bound on
D′ less than 0.9). Haplotype analyses were carried out using
likelihood ratio test combined with expectation–maximization
algorithm implemented on Unphased 2.0 (Dudbridge et al., 2003)
and Haploview 4.1 (Barrett et al., 2005). We report individual
haplotype effects (less conservative result) instead of global
haplotype effects (more conservative result). We calculated power
for genotype association using the unmatched case–control design
under gene only hypothesis and log additive inheritance model
implemented on Quanto 1.24 (Gauderman, 2002) assuming popula-
tion risk of 20%. Nominal significant association threshold was
pb0.05 and Bonferroni corrected was pb0.006. Permuted p values
were obtained after 10000 permutations using Haploview 4.1.
3. Results
3.1. Tardive dyskinesia manifestation
Thirty seven percent of the total subjects presented with TD
(n=81), of which 68 were European Caucasian. This translated
to a 40.2% TD occurrence rate in the Caucasian population in
this study. We observed significantly lower TD (25.5%) in the 43
subjects not identified as European Caucasian; therefore, we
focused our association analysis on the European Caucasian
sample (n=169). Our overall sample had statistical power
ranging from 60.6% to 91.6% to detect odds ratios as low as 1.6
(n=215, 1.65 controls per case, minor allelic frequency range in
the overall sample=10.2% to 38.1%). Our Caucasian sample had
statistical power ranging from 58.3% to 90.8% to detect the
same odds ratio (n=169, 1.45 controls per case, minor allelic
frequency range in the European Caucasian sample=10.1% to
38.2%). We observed a statistical trend for a gender effect in
our TD sample (p=0.06), and age was strongly associated with a
diagnosis of TD diagnosis in our sample (p=0.002).
We observed low frequency of homozygous subjects carrying
minor alleles for all polymorphisms. Therefore, we conducted
statistical evaluations comparing allelic instead of genotypic
frequencies. Minor alleles of all eight variants consistently
showed lower risk of TD in European Caucasian subjects (O.
R.b1.0), but only three variants reached nominal statistical
significance: rs6805251, rs6438552 and rs9878473 (p=0.023)
(Table 2). These results did not remain significant after
correction for multiple testing using either Bonferroni (thresh-
old pb0.006) or permutation (permuted p=0.07) approaches.
The rs6805251, rs6438552 and rs9878473 variants showed high
Table 1 Clinical and demographical data of the European Caucasian subjects.
n Male (%) Age (SD) TD (%) AIMS (SD) Life-time antipsychotic
treatment
Antipsychotic at
assessment
Clinical sites
Case Western Reserve 42 73.80 32.0 (8.4) 35.7 4.0 (5.6) Typicals No exposure to atypicals
Hillside Hospital 41 59.50 35.0 (7.2) 43.9 9.0 (8.9) a
Typicals No exposure to atypicals
Centre for Addiction and
Mental Health
86 60.97 43.0 (9.9) 40.7 7.0 (7.8) Typicals for at least one year Atypical agents
Overall subjects
TD present 68 55.88 41.0 (10.7) 100.0 13.0 (7.6) b
48.63% never exposed to atypicals
TD absent 101 70.29 36.0 (9.3) 0 2.0 (2.0) c
49.51% never exposed to atypicals
a
25 subjects had quantitative measures missing.
b
7 subjects had quantitative measures missing.
c
18 subjects had quantitative measures missing.
690 R.P. Souza et al.

4. LD in our sample (Fig. 1). Due to high linkage disequilibrium
across seven of the eight variants, we analyzed only
haplotypes included in the linkage disequilibrium block
(Fig. 1). Haplotype analysis revealed significant results. The
haplotype composed of the common alleles for the seven
variants included in the linkage block was associated with
increased risk of TD (1-1-1-1-1-1-1 haplotype), while
subjects carrying the haplotype comprised of the minor
allele variants for the markers rs6805251, rs6438552 and
rs9878473 showed lower risk of TD (2-1-2-1-2-1-1 haplo-
type). Both haplotype results remained significant after
permutation (Fig. 1).
3.2. Abnormal involuntary movement severity in
the Caucasian subjects
We had AIMS scores for 144 European Caucasian subjects
because 25 subjects collected at the Hillside Hospital could
not have their HSDS score converted to AIMS score. Age was
associated with TD severity in our European Caucasian
sample (p=0.0001), while gender was not (p=0.191).
Homozygous subjects carrying minor alleles for all variants
consistently showed lower AIMS scores compared to those
homozygous for the common allele (Table 3). However,
none of the variants reached statistical significance.
Haplotype results also showed no significant association.
Nevertheless, we observed strong association of all markers
with TD severity after incorporating age as a covariate in
the analysis of covariance using generalized linear model.
Table
2
Genetic
association
results
of
GSK-3B
variants
with
tardive
dyskinesia
diagnosis
in
European
Caucasian
subjects.
All
O.R.
were
calculated
using
allele
2
as
risk
allele.
We
reported
allelic
positivity
instead
of
genotypic
results
because
of
the
low
frequency
of
2/2
genotypes
in
all
evaluated
polymorphisms.
Bolded
values
represent
p
b
0.05.
Genetic
variant
Allele
Tardive
dyskinesia
absent
Tardive
dyskinesia
present
Allelic
positivity
[11x
(12
+
22)]
Allelic
frequency
difference
(1
×
2)
1
2
1/1
1/2
2/2
HWE
1/1
1/2
2/2
HWE
χ
2
p
O.R.
(C.I.)
χ
2
p
O.R.
(C.I.)
rs11919783
G
A
80
20
1
0.839
57
10
1
0.478
0.56
0.452
0.735
(0.329–1.644)
0.38
0.535
0.792
(0.378–1659)
rs6805251
C
T
30
55
16
0.267
31
32
5
0.398
4.45
0.034
0.504
(0.266–0.957)
5.12
0.023
0.591
(0.374–0.934)
rs7624540
C
A
67
32
2
0.411
47
20
1
0.485
0.14
0.705
0.880
(0.455–1.703)
0.15
0.694
0.890
(0.497–1.592)
rs6438552
A
G
30
55
16
0.267
31
32
5
0.398
4.45
0.034
0.504
(0.266–0.957)
5.12
0.023
0.591
(0.374–0.934)
rs4072520
C
A
67
32
2
0.267
47
20
1
0.485
0.14
0.705
0.880
(0.455–1.703)
0.15
0.694
0.890
(0.497–1.592)
rs9878473
T
C
30
55
16
0.267
31
32
5
0.398
4.45
0.034
0.504
(0.266–0.957)
5.12
0.023
0.591
(0.374–0.934)
rs6779828
C
T
65
34
2
0.305
48
19
1
0.563
0.71
0.398
0.752
(0.880–1.458)
0.64
0.423
0.788
(0.440–1.413)
rs3755557
T
A
70
28
3
0.920
53
14
1
0.945
1.66
0.198
0.659
(0.348–1.248)
1.53
0.216
0.639
(0.313–1.303)
Figure 1 Linkage disequilibrium (LD) across the GSK-3B variants.
Values in the LD plot represent pairwise in the R-squared. Bolded
values in table represent pb0.05. TD − = subject without tardive
dyskinesia. TD + = subjects with tardive dyskinesia.
691
Tardive dyskinesia and GSK-3b

5. These associations remain significant after permutation proce-
dure. Similar findings were found in the sample collected at the
Centre for Addiction and Mental Health.
4. Discussion
Our results suggest that polymorphisms in the GSK-3B gene
may play a minor role in the genetic vulnerability of TD and
its severity in European Caucasian subjects with schizophre-
nia. All eight evaluated variants had their minor allele
carriers consistently showing lower risk of TD and lower AIMS
scores, but just three variants showed nominal significant
association with TD. All tested markers showed p≤0.001
after we incorporated age as a covariate in analysis of the
AIMS scores. We found no gender effect in our sample but
older age was related to increased TD risk. Previous reports
have indicated that advancing age and, to a lesser extent,
female sex are the two variables most consistently associ-
ated with increased TD prevalence (Kane and Smith, 1982).
All three variants nominal associated with TD (rs6805251,
rs6438552 and rs9878473) were in very high LD in our sample
(Fig. 1) although they are located in different introns (11, 5
and 3, respectively) spanning over 90 kb. Thus, these results
should not be interpreted as three independent findings.
Two studies have reported on rs6438552 function (Inkster
et al., 2009; Kwok et al., 2005). Kwok et al. (Kwok et al.,
2005) demonstrated that rs6438552 modulates the utiliza-
tion of splice acceptor sites in downstream introns leading to
a: (1) 1.9-fold increase in the brain GSK-3B mRNA level in
subjects carrying A/A genotype compared to G/G carriers;
(2) 1.4-fold increase in transcriptional strength of the A
allele relative to the G allele in HEK293; and (3) 7.4-fold
increase of protein Tau phosphorylation in the brain of
Parkinson's subjects with A/A versus G/G genotypes (Kwok
et al., 2005). In a second study, Inkster et al. (Inkster et al.,
2009) reported less grey matter in the right and left superior
temporal gyri, as well as the right hippocampus, in carriers of
the rs6438552 A allele compared to the G allele in subjects
with major depressive disorder. Although these results were
Table 3 Genetic association results of GSK-3B variants with the Abnormal Involuntary Movement Scale (AIMS) score in European
Caucasian subjects. We reported allelic positivity [11x (12+22)] instead of genotypic results. Bolded values represent pb0.05.
Genetic
variant
Abnormal Involuntary Movement Scale (AIMS) score
1/1 genotype 1/2 genotype 2/2 genotype ANOVA ANCOVA a
n Mean SD n Mean SD n Mean SD F p F p
Case Western Reserve+Hillside Hospital
rs11919783 44 6.61 7.56 13 3.15 4.76 1 3.00 – 2.624 0.111 1.682 0.195
rs6805251 41 6.61 7.60 16 3.94 5.43 1 1.00 – 1.973 0.166 1.391 0.258
rs7624540 36 6.67 7.96 20 4.60 5.34 2 1.50 2.12 1.151 0.223 1.442 0.245
rs6438552 18 8.33 8.77 31 4.55 6.09 9 4.89 5.80 3.557 0.064 2.653 0.079
rs4072520 36 7.08 8.00 20 3.85 4.83 2 1.50 2.12 3.370 0.072 2.416 0.099
rs9878473 18 8.33 8.77 31 4.55 6.09 9 4.89 5.80 3.557 0.064 2.653 0.079
rs6779828 36 6.67 7.96 20 4.60 5.34 2 1.50 2.12 1.151 0.223 1.442 0.245
rs3755557 18 8.33 8.77 31 4.55 6.09 9 4.89 5.80 3.557 0.064 2.653 0.079
Centre for Addiction and Mental Health
rs11919783 71 6.69 7.78 14 7.36 8.84 1 1.00 – 0.012 0.914 8.440 0.00042
rs6805251 63 6.68 7.63 21 7.29 9.02 2 2.50 3.54 0.009 0.923 8.470 0.00041
rs7624540 61 6.93 8.09 24 6.46 7.61 1 1.00 – 0.136 0.713 8.628 0.00039
rs6438552 36 7.47 8.13 45 6.71 8.00 5 1.60 2.07 0.541 0.464 8.793 0.00034
rs4072520 61 6.93 8.09 24 6.46 7.61 1 1.00 – 0.136 0.713 8.628 0.00039
rs9878473 36 7.47 8.13 45 6.71 8.00 5 1.60 2.07 0.541 0.464 8.793 0.00034
rs6779828 61 6.93 8.09 24 6.46 7.61 1 1.00 – 0.136 0.713 8.628 0.00039
rs3755557 36 7.47 8.13 45 6.71 8.00 5 1.60 2.07 0.541 0.464 8.793 0.00034
Overall
rs11919783 115 6.66 7.67 27 5.33 7.35 2 2 1.4 0.98 0.32 7.712 0.0006
rs6805251 54 7.76 8.27 76 5.83 7.32 14 3.71 5 3.05 0.08 9.148 0.0001
rs7624540 97 6.84 8 44 5.61 6.67 3 1.33 1.5 1.24 0.27 8.26 0.0004
rs6438552 54 7.76 8.27 76 5.83 7.32 14 3.71 5 3.05 0.08 9.148 0.0001
rs4072520 97 6.84 8 44 5.61 6.67 3 1.33 1.5 1.24 0.27 8.26 0.0004
rs9878473 54 7.76 8.27 76 5.83 7.32 14 3.71 5 3.05 0.08 9.148 0.0001
rs6779828 97 6.99 8.01 44 5.27 6.56 3 1.33 1.5 2.16 0.14 8.778 0.0002
rs3755557 104 6.65 7.58 37 5.84 7.77 3 2 2.7 0.61 0.44 7.583 0.0007
a
Generalized linear model including age as covariate.
692 R.P. Souza et al.

6. acquired in non-schizophrenia patients, they report robust
evidence of the rs6438552 function. In summary, there is an
indication that rs6438552 A allele is associated with
increased GSK-3B expression and activity.
Our results indicated that subjects carrying the rs6438552
A allele were more likely to have TD than carriers of the G
allele. Moreover, the A allele carriers showed higher AIMS
scores and the analysis of covariance results of rs6438552
was one of the most significant. Considering that A allele
carriers present increased expression and activity of the
GSK-3B (Kwok et al., 2005), and GSK-3B inhibition leads to
neuroprotection in different models (Beurel and Jope, 2006;
Carmichael et al., 2002; Wang et al., 2007), our results
corroborate the suggestion of increased neurotoxicity in TD.
There is one report that evaluated GSK-3B variants in the
genetic susceptibility to TD in a Korean cohort; here, G/G
homozygous subjects for the rs334558 were over-repre-
sented in the non-TD group carrying the brain derived
neurotrophic factor (BDNF) rs6265 Val allele (Park et al.,
2009). We have not evaluated the rs334558 variant in the
current study but this variant shows high LD with rs6438552
in Asian and Caucasian populations (Kwok et al., 2005).
Because the flanking genes do not appear to be in LD with
GSK-3B, it has been suggested that both rs334558 and
rs6438552 fulfill a key criterion required for regulatory
variants as predicted in the common disease-common variant
hypothesis, namely that the pathogenic polymorphism is
ancient and LD around it is low (Kwok et al., 2005; Weiss and
Clark, 2002). There is no evidence of functional activity for
rs6805251 and rs9878473 variants but both variants are in
high LD with the rs6438552. Moreover, these variants can be
in LD with other exonic non-synonymous variants in this
gene.
It is important to highlight that this study has its own
limitations. First, we have tested eight variants across this
gene, which required correction for multiple testing. Dealing
with multiple testing is a controversial issue and has been
intensely debated. Considering the exploratory nature of this
study, without a priori hypotheses for most of our variants
there is no clear structure in the multiple tests (Bender and
Lange, 1999). Further, Bonferroni correction, while possibly
the most straight-forward approach, was likely too stringent
particularly since, each variant should likely not be considered
independently due to high LD observed (Nyholt, 2004). Second,
our sample is underpowered to evaluate small genetic effects
(0.625bO.R.b1.6) for some of the variants although we had
reasonable power to detect the reported nominal associations
(O.R.=0.591). Nevertheless, there is a need for replication of
the nominal significant findings in independent and larger
samples. Third, we cannot rule out differences in important
factors that may alter TD risk: smoking status, length of typical
antipsychotic exposure and individual pharmacokinetics back-
ground. Fourth, we have no consistent data regarding
comorbidities across the samples.
GSK-3B is a serine/threonine kinase that is present in most
tissues, and it is particularly abundant in the central nervous
system. GSK-3 participates in multiple signaling pathways
coupled to receptors for a variety of signaling molecules such
as insulin. Beyond its function in glucose metabolism, GSK-3
plays numerous roles in differentiation and development,
intracellular trafficking, apoptosis, and regulation of gene
transcription. Our results suggest that GSK-3B polymorphism
may play a minor role in the genetic vulnerability to TD in
individuals with schizophrenia subjects taking antipsycho-
tics, further implicating genetic variants in the dopamine D2-
like receptor signaling in this context.
Role of the funding source
Canadian Institutes of Mental Health (Grant MOP-79525; Grant GMH-
79044 and Fellowship XWY-93967 to Dr. Souza). Funding agencies
had no further role in study design; in the collection, analysis and
interpretation of data; in the writing of the report; and in the
decision to submit the paper for publication.
Contributors
Authors RPS, JAL, HYM, GR and JLK designed the project. RPS
managed the literature searches and analyses. Authors ANV, HYM,
JAL, GR and JLK collected research subjects, and author RPS
collected genotypes. RPS undertook the statistical analysis and
wrote the first draft of the manuscript. All authors contributed to
and have approved the final manuscript.
Conflict of interest
ANV and GR have nothing to declare.RPS has received a Wyeth
fellowship.JAL has served as a consultant/advisor or grantee of
Acadia, Astra Zeneca, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithK-
line, Janssen Pharmaceutica, Lundbeck, Merck, Organon, Pfizer and
Wyeth; and holds a patent from Repligen.HYM declares that he is a
consultant or grantee of Abbott, Acadia, ARYx, Astra Zeneca, Bristol-
Myers Squibb, Eli Lilly, Janssen, Memory, Minster, Organon, Pfizer,
Solvay, Wyeth, and Vanda.JLK declares that he is a consultant for
GlaxoSmithKline.
References
Barrett, J.C., Fry, B., Maller, J., Daly, M.J., 2005. Haploview:
analysis and visualization of LD and haplotype maps.
Bioinformatics 21, 263–265.
Beaulieu, J.M., Gainetdinov, R.R., Caron, M.G., 2009. Akt/GSK3
signaling in the action of psychotropic drugs. Annu. Rev.
Pharmacol. Toxicol. 49, 327–347.
Bender, R., Lange, S., 1999. Multiple test procedures other than
Bonferroni's deserve wider use. BMJ 318, 600–601.
Beurel, E., Jope, R.S., 2006. The paradoxical pro- and anti-apoptotic
actions of GSK3 in the intrinsic and extrinsic apoptosis signaling
pathways. Prog. Neurobiol. 79, 173–189.
Carmichael, J., Sugars, K.L., Bao, Y.P., Rubinsztein, D.C., 2002.
Glycogen synthase kinase-3beta inhibitors prevent cellular
polyglutamine toxicity caused by the Huntington's disease
mutation. J. Biol. Chem. 277, 33791–33798.
Dudbridge, F., 2003. Pedigree disequilibrium tests for multilocus
haplotypes. Genet. Epidemiol. 25, 115–121.
Gauderman, W.J., 2002. Sample size requirements for matched
case–control studies of gene–environment interaction. Stat.
Med. 21, 35–50.
Inkster, B., Nichols, T.E., Saemann, P.G., Auer, D.P., Holsboer, F.,
Muglia, P., Matthews, P.M., 2009. Association of GSK3beta
polymorphisms with brain structural changes in major depressive
disorder. Arch. Gen. Psychiatry 66, 721–728.
693
Tardive dyskinesia and GSK-3b

7. Jope, R.S., Roh, M.S., 2006. Glycogen synthase kinase-3 (GSK3) in
psychiatric diseases and therapeutic interventions. Curr. Drug
Targets 7, 1421–1434.
Kane, J.M., Smith, J.M., 1982. Tardive dyskinesia: prevalence and
risk factors, 1959 to 1979. Arch. Gen. Psychiatry 39, 473–481.
Kang, U.G., Seo, M.S., Roh, M.S., Kim, Y., Yoon, S.C., Kim, Y.S.,
2004. The effects of clozapine on the GSK-3-mediated signaling
pathway. FEBS Lett. 560, 115–119.
Kwok, J.B., Hallupp, M., Loy, C.T., Chan, D.K., Woo, J., Mellick, G.
D., Buchanan, D.D., Silburn, P.A., Halliday, G.M., Schofield, P.
R., 2005. GSK3B polymorphisms alter transcription and splicing in
Parkinson's disease. Ann. Neurol. 58, 829–839.
Lahiri, D.K., Nurnberger Jr., J.I., 1991. A rapid non-enzymatic
method for the preparation of HMW DNA from blood for RFLP
studies. Nucleic Acids Res. 19, 5444.
Little, J., Higgins, J.P., Ioannidis, J.P., Moher, D., Gagnon, F., von
Elm, E., Khoury, M.J., Cohen, B., Davey-Smith, G., Grimshaw, J.,
Scheet, P., Gwinn, M., Williamson, R.E., Zou, G.Y., Hutchings,
K., Johnson, C.Y., Tait, V., Wiens, M., Golding, J., van Duijn, C.,
McLaughlin, J., Paterson, A., Wells, G., Fortier, I., Freedman,
M., Zecevic, M., King, R., Infante-Rivard, C., Stewart, A.,
Birkett, N., 2009. STrengthening the REporting of Genetic
Association Studies (STREGA)—an extension of the STROBE
statement. Genet. Epidemiol. 33, 581–598.
Nyholt, D.R., 2004. A simple correction for multiple testing for
single-nucleotide polymorphisms in linkage disequilibrium with
each other. Am. J. Hum. Genet. 74, 765–769.
Park, S.W., Lee, J.G., Kong, B.G., Lee, S.J., Lee, C.H., Kim, J.I.,
Kim, Y.H., 2009. Genetic association of BDNF val66met and
GSK-3beta-50T/C polymorphisms with tardive dyskinesia
Psychiatry Clin. Neurosci. 63, 433–439.
Roh, M.S., Seo, M.S., Kim, Y., Kim, S.H., Jeon, W.J., Ahn, Y.M., Kang,
U.G., Juhnn, Y.S., Kim, Y.S., 2007. Haloperidol and clozapine
differentially regulate signals upstream of glycogen synthase
kinase 3 in the rat frontal cortex. Exp. Mol. Med. 39, 353–360.
Schooler, N.R., Kane, J.M., 1982. Research diagnoses for tardive
dyskinesia. Arch. Gen. Psychiatry 39, 486–487.
Souza, R.P., de Luca, V., Remington, G., Lieberman, J.A., Meltzer,
H.Y., Kennedy, J.L., Wong, A.H., 2010b. Glial cell line-derived
neurotrophic factor receptor alpha 2 (GFRA2) gene is associated
with tardive dyskinesia. Psychopharmacology (Berl.) 210,
347–354.
Souza, R.P., Remington, G., Lieberman, J.A., Meltzer, H.Y., Kennedy,
J.L., Wong, A.H., 2010a. Phosphodiesterase 4B genetic variants are
not associated with antipsychotic-induced tardive dyskinesia. Int.
Clin. Psychopharmacol. doi:10.1097/YIC.0b013e32833a5ff9.
Souza, R.P., Romano-Silva, M.A., Lieberman, J.A., Meltzer, H.Y.,
Wong, A.H., Kennedy, J.L., 2008. Association study of GSK3 gene
polymorphisms with schizophrenia and clozapine response.
Psychopharmacology (Berl.) 200, 177–186.
Thelma, B., Srivastava, V., Tiwari, A.K., 2008. Genetic
underpinnings of tardive dyskinesia: passing the baton to
pharmacogenetics. Pharmacogenomics 9, 1285–1306.
Wang, W., Yang, Y., Ying, C., Li, W., Ruan, H., Zhu, X., You, Y., Han,
Y., Chen, R., Wang, Y., Li, M., 2007. Inhibition of glycogen
synthase kinase-3beta protects dopaminergic neurons from MPTP
toxicity. Neuropharmacology 52, 1678–1684.
Weiss, K.M., Clark, A.G., 2002. Linkage disequilibrium and the
mapping of complex human traits. Trends Genet. 18, 19–24.
Zai, C.C., Romano-Silva, M.A., Hwang, R., Zai, G.C., Deluca, V., Muller,
D.J., King, N., Voineskos, A.N., Meltzer, H.Y., Lieberman, J.A.,
Potkin, S.G., Remington, G., Kennedy, J.L., 2008. Genetic study of
eight AKT1genepolymorphismsandtheirinteractionwithDRD2gene
polymorphisms in tardive dyskinesia. Schizophr. Res. 106, 248–252.
Zai, C.C., Tiwari, A.K., Basile, V., De Luca, V., Muller, D.J., King, N.,
Voineskos, A.N., Remington, G., Meltzer, H.Y., Lieberman, J.A.,
Potkin, S.G., Kennedy, J.L., 2009a. Association study of tardive
dyskinesia and five DRD4 polymorphisms in schizophrenia
patients. Pharmacogenomics J. 9, 168–174.
Zai, C.C., Tiwari, A.K., DeLuca, V., Muller, D.J., Bulgin, N., Hwang, R.,
Zai, G.C., King, N., Voineskos, A.N., Meltzer, H.Y., Lieberman, J.
A., Potkin, S.G., Remington, G., Kennedy, J.L., 2009b. Genetic
study of BDNF, DRD3, and their interaction in tardive dyskinesia.
Eur. Neuropsychopharmacol. 19, 317–328.
694 R.P. Souza et al.