A Preliminary Pharmacogenetic Investigation of Adverse Events From
Topiramate in Heavy Drinkers
Lara A. Ray
Brown University and University of California,
Robert Miranda Jr.
Brown University and University of Georgia
Providence Veterans Affairs Medical Center
and Brown University
Jennifer W. Tidey
Damaris J. Rohsenow
Providence Veterans Affairs Medical Center
and Brown University
Robert W. Swift and Peter M. Monti
Providence Veterans Affairs Medical Center
and Brown University
Topiramate, an anticonvulsant medication, is an efficacious treatment for alcohol dependence.
To date, little is known about genetic moderators of side effects from topiramate. The
objective of this study was to examine 3 single nucleotide polymorphisms (SNPs) of the
glutamate receptor GluR5 gene (GRIK1) as predictors of topiramate-induced side effects in
the context of a laboratory study of topiramate. Heavy drinkers (n ? 51, 19 women and 32
men), 75% of whom met criteria for an alcohol use disorder, completed a 5-week dose
escalation schedule to a target dose of either 200 or 300 mg or matched placebo. The
combined medication groups were compared with placebo-treated individuals for side effects
at target dose. Analyses revealed that an SNP in intron 9 of the GRIK1 gene (rs2832407) was
associated with the severity of topiramate-induced side effects and with serum levels of
topiramate. Genes underlying glutamatergic neurotransmission, such as the GRIK1 gene, may
help predict heterogeneity in topiramate-induced side effects. Future studies in larger samples
are needed to more fully establish these preliminary findings.
Keywords: alcohol, topiramate, pharmacogenetics, side effects, glutamate
The field of pharmacogenetics seeks to uncover genetic
causes of heterogeneity of pharmacotherapy treatment ef-
fects, including the susceptibility to side effects. Increased
knowledge about genetic moderators of medication re-
sponse can be applied to treatment decision making, includ-
ing dosing algorithms based on genetically determined dif-
ferences in drug metabolism (Kirchheiner et al., 2004;
O’Brien, 2008). Both pharmacokinetic and pharmacody-
namic genes are thought to influence medication response.
Genes that influence pharmacokinetic processes are those
involved in the absorption, distribution, metabolism, and
excretion of drugs in the body. In contrast, genes that
influence pharmacodynamic response are those directly re-
sponsible for the structure and/or functioning of the mole-
cules or receptors targeted by the medication. These genes
may also influence signaling or metabolic pathways in-
volved in the manifestation of a disease, such that the
mechanisms that affect the risk for a disorder may also
underlie clinical response to its pharmacological treatment.
Although pharmacokinetic genes make up the largest group
Lara A. Ray, Center for Alcohol and Addiction Studies, Brown
University, and Department of Psychology, University of Califor-
nia, Los Angeles; Robert Miranda Jr., Jennifer Tidey, and Chad
Gwaltney, Department of Psychology, University of California,
Los Angeles; James MacKillop, Center for Alcohol and Addiction
Studies, Brown University, and Department of Psychology, Uni-
versity of Georgia; John McGeary, Damaris J. Rohsenow, Robert
Swift, and Peter M. Monti, Providence Veterans Affairs Medical
Center, Providence, Rhode Island, and Center for Alcohol and
Addiction Studies, Brown University.
This research was supported in part by Career Development Award
1K23 AA014966, National Institute on Alcohol Abuse and Alcoholism
Grants 5R01 AA7850-17 and T32 AA007459, a Research Career De-
of Veterans Affairs Senior Research Career Scientist Award.
We thank Amy Christian and John-Paul Massaro for their excellent
technical assistance and Chinatsu McGeary for genotyping assistance.
Ray, Department of Psychology, University of California, Box 951563,
Los Angeles, CA 90095-1563. E-mail: firstname.lastname@example.org
Experimental and Clinical Psychopharmacology
2009, Vol. 17, No. 2, 122–129
© 2009 American Psychological Association
1064-1297/09/$12.00 DOI: 10.1037/a0015700
of known pharmacogenetic factors, some have argued that
pharmacodynamic genes have stronger effects on medica-
tion response (Goldstein, Need, Singh, & Sisodiya, 2007).
Topiramate (TOP) is an anticonvulsant medication re-
cently identified as efficacious for the treatment of alco-
hol dependence (AD). Individuals with AD, randomized
to TOP in escalating doses up to 300 mg/day, reported
reductions in drinking across several measures of alcohol
consumption, with large-magnitude effect sizes, as com-
pared with those randomized to placebo (PLAC; Johnson
et al., 2003, 2007). Results of two open-label trials found
similarly positive results (Raguraman, Priyadharshini, &
Chandrasekaran, 2005; Rubio et al., 2004). Glutamate
antagonist medications, including TOP, have also re-
ceived empirical support for the treatment of alcohol
withdrawal symptoms (Krupitsky et al., 2007).
TOP is a promising pharmacotherapy for AD, although
TOP’s side effect profile may limit its use in clinical prac-
tice given that approximately 20% of patients drop out of
clinical trials for alcoholism because of TOP-induced side
effects (Johnson et al., 2007). Therefore, it would be useful
to identify individuals who are more likely to report greater
severity of side effects of this medication. In a recent large
clinical trial, the most commonly reported side effects
among individuals with AD were paresthesia (51%), taste
perversion (23%), anorexia (20%), and difficulty with con-
centration (15%; Johnson et al., 2007). These results are
similar to those of previous reports (Johnson et al., 2003;
Johnson, Swift, Addolorato, Ciraulo, & Myrick, 2005).
Likewise, the literature on TOP for the treatment of epi-
lepsy, at higher doses than those used for alcoholism treat-
ment, has suggested that it causes frequent cognitive side
effects both in clinical trials and in practice (Fritz et al.,
2005; Thompson, Baxendale, Duncan, & Sander, 2000). A
number of variables have been examined as predictors of
cognitive adverse events to TOP (i.e., difficulty with atten-
tion, concentration, and memory) among patients treated for
seizure disorders, including individual differences (e.g.,
gender, age at start of treatment, and diagnosis) and clinical
(e.g., titration, maximum dose, and maintenance dose) vari-
ables, but none have successfully predicted treatment re-
sponse (Goldstein et al., 2007). Alternatively, genetic dif-
ferences may contribute to the heterogeneity of adverse
events from TOP.
TOP has multiple neuropharmacological mechanisms
of action (Johnson, 2005). These actions include the
facilitation of gamma aminobutyric acid (GABA) neuro-
transmission and the inhibition of glutamatergic recep-
tors. Acute dopamine release across the corticomesolimbic
axis plays a critical role in determining the motiva-
tional significance of alcohol (Hyman & Malenka, 2001)
and alcohol reward (Dodd, Beckmann, Davidson, &
Wilce, 2000; Weiss, Lorang, Bloom, & Koob, 1993;
Weiss & Porrino, 2002). Because corticomesolimbic do-
pamine release is under tonic inhibitory control via
GABAergic neurons and excitatory control via glutama-
tergic neurons, TOP’s concurrent GABAergic agonism
and glutamatergic antagonism have been hypothesized to
inhibit acute motivation for alcohol (i.e., craving). More
important, TOP antagonizes the ability of kainate to
activate the kainate/alpha-amino-3-hydroxy-5-methyl-4-
isoxazolepropionic acid subtype of glutamate receptors,
with no apparent activity on the N-methyl-D-aspartic acid
subtype (Johnson, 2005). Glutamatergic neurotransmis-
sion, in turn, is thought to play an important role in
various alcohol-related behaviors such as alcohol with-
drawal symptoms (Littleton, 1998) and responses to al-
cohol (Lipsky & Goldman, 2003). In addition to its
effects on glutamatergic transmission, TOP potentiates
inhibitory GABAAreceptor-mediated input, which rep-
resents another important mechanism underlying TOP’s
effects (Johnson, 2005; Johnson & Ait-Daoud, 2000).
On the basis of TOP’s putative neurobiological mech-
anisms of action, particularly its antagonist effects on
glutamatergic neurotransmission, genes coding for gluta-
mate receptors represent plausible candidate genes be-
cause they harbor allelic variants that may underlie the
pharmacodynamics of TOP. This investigation focuses
on the kainate-selective glutamate receptor GluR5 gene
(GRIK1), located on chromosome 21q, as a candidate
pharmacogenetic predictor of TOP-induced side effects,
given that animal (Gryder & Rogawski, 2003) and in
vitro (Kaminski, Banerjee, & Rogawski, 2004) studies
have found that receptors containing the GluR5 subunit
selectively bind TOP. In this study, we selected three
single nucleotide polymorphisms (SNPs) in the 3? half of
the GRIK1 gene (rs2832387, rs2186305, and rs2832407),
given that these were the only SNPs significantly asso-
ciated with AD in a recent study evaluating seven GRIK1
SNPs (Kranzler, Covault, Herman, Burian, & Arias,
2007). Additionally, the last two SNPs are located in
exons 9 and 17, respectively, for which there is evidence
of alternative splicing (Barbon & Barlati, 2000), although
their functional significance remains unclear.
Our objective in this study was to examine three SNPs of
GRIK1 as predictors of TOP-induced side effects in the
context of a recent laboratory study (Miranda et al., 2008).
A community sample of non–treatment-seeking heavy
drinkers, 75% of whom met criteria for an alcohol use
disorder, completed a 5-week titration period to two target
doses of TOP (200 mg or 300 mg) or matched PLAC.
Participants were stabilized on the target medication dose or
given an equivalent schedule of PLAC before completing a
laboratory protocol consisting of cue exposure followed by
alcohol administration (not reported here). This study fo-
cuses on genetic predictors of severity of side effects re-
ported on stabilization at the target dose of TOP versus
PLAC. A secondary aim of this study was to test GRIK1
SNPs for their association with serum levels of TOP at the
target dose. Consistent with the pharmacogenetics frame-
work, we hypothesized that genetic factors of putative in-
fluence on the targets of a given pharmacotherapy—in this
case, glutamate receptors targeted by TOP—may ultimately
predict responses to this pharmacotherapy. We also report
associations between the candidate polymorphisms and
ADVERSE EVENTS FROM TOPIRAMATE IN HEAVY DRINKERS
Inclusion criteria were the following: consuming 18 or
more alcoholic drinks per week (14 or more for women)
during the 90 days before enrollment, not currently seeking
treatment for alcoholism, and age between 21 and 65 years.
Exclusion criteria were clinically significant medical con-
traindications; use of medications that could affect mood,
drinking, or TOP’s effects; current or recent alcohol treat-
ment or treatment seeking; allergy to TOP; weight of less
than 110 lb or more than 250 lb; living with another par-
ticipant in the study; consuming 60 or more drinks per week
(53 or more for women); not using reliable birth control, if
female; positive urine pregnancy screen; and Beck Depres-
sion Inventory–II score of 14 or more.
Of 78 individuals enrolled in the study, 14 withdrew, 6
for discomfort from side effects, 2 for misgivings about
participation in a medication trial, and 6 for miscellaneous
personal reasons. Of those participants who withdrew be-
cause of side effects, 4 were in the 300-mg condition, 1 was
in the 200-mg condition, and 1 was in the PLAC condition.
No medication-related serious adverse events occurred.
Three additional participants in the 300-mg group who
completed the study were determined to have zero TOP
levels by serum assay so we did not use their data in the
A total of 61 participants completed the trial. A subset of
participants (n ? 51; 19 women) provided consent for DNA
collection and represent the sample for this preliminary
investigation of TOP pharmacogenetics (n ? 32 in TOP,
n ? 19 in PLAC). The mean age of the study sample
was 43.1 years (SD ? 12.7), 90% were Caucasian, and the
mean level of education was 13.8 years (SD ? 3.0). At
baseline, participants reported drinking an average of 4.8
days per week and consuming an average of 6.3 drinks
(SD ? 2.9) per drinking day. Diagnostically, 75.5% of
participants met current criteria for an alcohol use disorder
at enrollment, 22.5% for alcohol abuse and 53% for AD.
All procedures were approved by the Brown University
Institutional Review Board. A telephone screen and a face-
to-face screening session assessed inclusion and exclusion
criteria. A medical examination by a physician determined
medical and medication-related exclusions. On enrollment
in the study, participants were provided with their 1st week
of medication and the Medication Event Monitoring System
(MEMS) and completed baseline assessments. Participants
attended weekly sessions during which they were assessed
for alcohol use, craving, and side effects and were given
their next week’s medication supply. All side effects were
communicated to the study physician on a weekly basis. The
physician and all staff were unaware of medication condi-
tions. This study focuses on the assessment completed at the
target dose of medication.
Medication and Compliance
Participants underwent a 32-day medication titration pe-
riod that is described in detail elsewhere (Miranda et al.,
2008). The target dose was maintained for up to 7 days, with
a modal stabilization period of 4 days, during which partic-
ipants completed a laboratory session that included alcohol
administration and cue reactivity. Doses were compounded
by an independent pharmacy, and all capsules appeared
identical. Compliance was assessed through two methods.
First, MEMS was used, which employs an electronic medica-
tion bottle cap to record and store the date and time at which
participants open the bottle. The MEMS also prompted the
participants to take the medication in the morning and
evening. Second, plasma levels of TOP were quantified via
blood assay. Participants provided blood samples at the
target dose (i.e., immediately before the laboratory session),
which were analyzed by an independent laboratory. Test
method consisted of a fluorescence polarization immunoas-
say, which has a high degree of sensitivity (2.0 mcg/ml).
Only samples from individuals randomized to TOP were
analyzed. Of 32 participants on TOP, 27 provided blood
samples; however, samples from 9 participants hemolyzed
in transit and were nonviable for analysis, leaving a total
of 18 valid samples (11 in the 200-mg condition and 7 in the
Alcohol use disorders were assessed through diagnostic
interviews conducted by a trained PhD-level psychologist
using the Structured Clinical Interview for DSM–IV (Patient
Version; First, Spitzer, Gibbon, & Williams, 1995). Alcohol
use before and during the study was assessed using the
Timeline Followback (Sobell, Maisto, Sobell, & Cooper,
1979), with heavy drinking days defined as six and four
standard drinks in a day for men and women, respectively
(Flannery et al., 2002). Side effect severity was the primary
dependent variable in this study and consisted of a contin-
uous measure defined as the mean ratings of the 19 side
effects at target dose. Side effects were assessed weekly in
a structured interview in which participants were asked to
rate the severity of 38 specific potential side effects using
the categories “none,” “mild,” “moderate,” and “severe.”
Participants were also asked open-ended questions to deter-
mine the incidence of any nonanticipated side effects, which
were rated using the same severity ratings. Data gathered
during Days 32–37 (target dose in TOP conditions) were
used. The following 19 side effects at the target medication
dose (or PLAC) represent the focus of the analyses: dizzi-
ness, decrease in appetite, changes in vision, difficulty
sleeping, weight loss, fatigue, difficulty with coordination
or balance, difficulty with concentration or attention, tin-
gling in finger or toes (paresthesia), word-finding difficul-
ties, memory difficulties, tremor, constipation or diarrhea,
restlessness, nervousness or anxiety, irritability, depression,
confusion, and changes in sexual function.
RAY ET AL.
DNA was collected following published procedures (Free-
man et al., 1997; Walker et al., 1999). Participants swabbed
their cheeks with three cotton swabs, followed by a rinse of
the mouth with tap water. Genomic DNA was isolated from
buccal cells using published procedures (Lench, Stanier, &
Williamson, 1988; Spitz et al., 1996). An ABI 7300 instru-
ment was used to conduct 5?-nuclease (TaqMan) assays of
the three GRIK1 candidate SNPs using assays commercially
available from Applied Biosystems (Foster City, CA). This
method relies on allele-specific hybridization of oligonucle-
otide probes (Livak, 1999). For quality assurance purposes,
20% of the samples were genotyped twice. Allele frequen-
cies and tests for Hardy-Weinberg Equilibrium are reported
in Table 1.
All data were initially examined for distribution normal-
ity and outliers. We did not use multiple-test error correc-
tion because this was a preliminary exploratory study. Be-
cause of the small sample size for the analyses of genetic
moderators, we combined the two TOP groups (n ? 32) and
compared them with the PLAC group (n ? 19) to increase
the statistical power to detect group differences. A series of
t tests and chi-square tests compared the two groups for
demographics, drinking history, and number of side effects
at target dose. Cronbach’s alpha was used to test the internal
consistency of the 19 potential side effects. A series of
2 ? 2 between-groups analyses of variance (ANOVAs)
were used to test the study hypotheses regarding the main
effects of medication and Medication ? GRIK1 Genotype
interactions on mean severity ratings across side effects at
target dose. Significant interactions were followed by
planned comparisons between genes within the TOP con-
ditions. Each candidate SNP of the GRIK1 gene was tested
separately because we were not interested in the linear
combination of variables (Dar, Serlin, & Omer, 1994). A
2 ? 2 between-groups ANOVA was also used to address
the secondary aim regarding a main effect of genotype and
interaction effects with dose (200 mg vs. 300 mg) on TOP
serum levels. Associations between the candidate SNPs and
drinking at target dose were also tested using analyses of
covariance, which controlled for baseline drinking mea-
sures. Linkage disequilibrium plots for individuals of Eu-
ropean ancestry were generated from Hapmap data using
the software program Haploview version 4.1 (Barrett, Fry,
Maller, & Daly, 2005).
All dependent variables were found to be adequately
normally distributed, without outliers. The sample was pre-
dominantly made up of individuals of European ancestry
(i.e., 90%). GRIK1 genotype groups were compared with
regard to ethnicity, and results suggest that they did not
differ on ethnicity (chi-square; p ? .30); therefore, it is
highly unlikely that population stratification confounded the
analyses presented herein. Univariate ANOVAs revealed no
differences between the medication and genotype groups in
terms of demographic and drinking variables (p ? .05). In
addition, GRIK1 genotype ratios did not differ significantly
between medication conditions.
Linkage disequilibrium among the three candidate SNPs
are presented in Figure 1A and that among the three SNPs
and the ones studied by Kranzler et al. (2007) are presented
in Figure 1B. Pairwise SNP |D?| values (? 100) are pre-
sented with darkened blocks (i.e., high D? values) indicating
SNPs with limited recombination. In other words, SNPs
with high D? values indicate that markers are good surro-
gates for each other, likely to be transmitted together and to
capture similar genetic variance. As can be seen in Figure
1A, SNPs 1 and 3 in this study had the highest D? value, as
estimated from Hapmap data for pedigrees of European
ancestry. High D? values may indicate close topographic
location on the chromosome and therefore redundancy in
the information capture. Kranzler et al. (2007) built haplo-
type blocks for the GRIK1 SNPs, and in that study SNPs 1
and 3 were in the same haplotype block, whereas SNP 2 was
in Block 2, suggesting greatest overlap between SNPs 1
Allele and Genotype Frequencies for the Three Candidate Single Nucleotide Polymorphisms (SNPs) of the Glutamate
Receptor GluR5 Gene by Medication Condition and Hardy-Weinberg Equilibrium Test
SNP rs# location
(ns; n ? 32)
(ns; n ? 19)
?2(1) ? 0.07, ns
SNP 1: rs2186305
SNP 2: rs2832407
SNP 3: rs2832387
13 kb 3?
T ? 0.75
C ? 0.25
C ? 0.62
A ? 0.38
G ? 0.70
A ? 0.30
TT ? 0.55
TC/CC ? 0.45
CC ? 0.37
CA/AA ? 0.63
GG ? 0.53
GA/AA ? 0.47
TT ? 19
TC/CC ? 13
CC ? 11
CA/AA ? 21
GG ? 17
GA/AA ? 15
TT ? 10
TC/CC ? 9
CC ? 8
CA/AA ? 11
GG ? 10
GA/AA ? 9
?2(1) ? 0.08, ns
?2(1) ? 2.34, ns
aNonsignificant chi-square results indicate that the observed allele frequencies were in conformity with Hardy-Weinberg Equilibrium
ADVERSE EVENTS FROM TOPIRAMATE IN HEAVY DRINKERS
based on Hapmap samples of individuals of European ancestry for the three genotypes evaluated in
this study (A) and for all seven single nucleotide polymorphisms (SNPs) evaluated by Kranzler et
al. (2007) (B). Pairwise SNP |D?| values (? 100) of linkage are shown. Darkened blocks indicate
SNP pairs without evidence of extensive recombination.
Linkage disequilibrium plot from Haploview 4.1 for participants of European ancestry,
RAY ET AL.
Compliance and Side Effects
All participants met the medication compliance criterion
of 80%, with medication taken on a mean of 96.5% of days
(range ? 82%–100%) according to MEMS data. The most
frequently endorsed side effects in this sample were pares-
thesia (33.3%), difficulty with concentration or attention
(19.6%), constipation or diarrhea (17.6%), nervousness or
anxiety (15.7%), memory difficulties (13.7%), decreased
appetite (11.8%), and word-finding difficulties (11.8%). In
the TOP condition, participants reported a mean of 2.53
(SD ? 2.96) side effects, as compared with 0.84
(SD ? 1.21) in the PLAC condition, t(49) ? ?2.85, p ?
.01. Side effects were generally in the mild to moderate
range, such that 83% of the reported side effects were rated
as mild, 16% as moderate, and 1% as severe. The internal
consistency (Cronbach’s alpha) of the severity ratings for
the 19 side effects was .77.
Medication and Genetic Effects on Severity
of Side Effects
Although there was no significant main effect of medi-
cation (TOP vs. PLAC), F(1, 50) ? 2.57, p ? .12, or
GRIK1 genotype (AC/AA vs. CC), F(1, 50) ? 0.78, p ?
.38, on mean side effect severity, there was a significant
GRIK1 SNP 2 (rs2832407; see Table 1) Genotype ? Med-
ication interaction. Carriers of the A allele reported greater
severity of side effects on TOP versus PLAC, as compared
with individuals who were homozygous for the C allele,
F(1, 50) ? 4.02, p ? .05; see Figure 2. Planned compari-
sons revealed that the two genotype groups did not differ on
severity of side effects on PLAC, t(18) ? ?1.21, p ? .26.
However, in the TOP condition, carriers of the A allele
reported significantly higher side effect severity as com-
pared with individuals who were homozygous for the C
allele, t(31) ? 2.56, p ? .05. GRIK1 SNPs 1 and 3 did not
have significant main effects or interaction effects with
medication conditions on severity ratings of side effects.
Medication and Genetic Effects on TOP
As a manipulation check, serum TOP levels were com-
pared between the two active-dose groups, and the 300-mg
group had higher serum TOP levels compared with the
200-mg group (200 mg, n ? 11, M ? 4.25 mcg/ml; 300 mg,
n ? 7, M ? 7.72 mcg/ml), F(1, 17) ? 29.13, p ? .0001.
Additionally, there was a significant main effect of GRIK1
SNP 2 genotype (rs2832407) such that carriers of the A
allele had higher serum TOP levels as compared with indi-
viduals homozygous for the C allele (AA/AC, n ? 13,
M ? 6.68 mcg/ml; CC, n ? 5, M ? 5.28 mcg/ml), F(1,
17) ? 4.72, p ? .05. The interaction effect between TOP
dose (i.e., 200 mg/day vs. 300 mg/day) and GRIK1 geno-
type was not significant. GRIK1 SNPs 1 and 3 did not show
significant effects on serum TOP levels. The relationship
between serum TOP level and severity of side effects was
not significant (r ? .06, p ? .10, with less than 4% shared
Medication and Genetic Effects on Drinking
At target dose (Week 5), there was a main effect of
GRIK1 SNP 2 (rs2832407) genotype on percentage of
heavy drinking days (%HDD), F(1, 46) ? 5.46, p ? .05,
after controlling for baseline %HDD. Carriers of the A
allele reported a higher mean %HDD, as compared with
individuals who were homozygous for the C allele (40.4%
vs. 22.2%). There was no significant main effect of medi-
cation (TOP, M ? 25.2, vs. PLAC, M ? 37.4), or a
Genotype ? Medication interaction, on %HDD at target
dose. There was no significant effect of GRIK1 SNPs,
medication, or Genotype ? Medication interaction on av-
erage drinks per drinking day at target dose. Last, there was
no evidence of an association between severity of side
effects and drinking behavior at target dose in the TOP
condition, measured by %HDD (r ? ?.15, p ? .10) and
average number of drinks per drinking day (r ? .09,
p ? .10).
Allelic variation in the GRIK1 (rs2832407) gene was
associated with the severity of TOP-induced side effects in
a community sample of non–treatment-seeking heavy drink-
ers. The same SNP (rs2832407) was associated with differ-
ences in serum levels of TOP among participants in the TOP
condition. Specifically, carriers of the A allele of SNP 2 of
the GRIK1 gene reported greater severity of side effects
when receiving TOP as compared with those homozygous
for the C allele. Carriers of the A allele of SNP 2 of the
GRIK1 gene also had higher serum TOP levels and higher
percentage of heavy drinking days at target dose than did
those homozygous for the C allele. Interestingly, less than
poTc a l P
Mean Severity of Side Effects
tion on mean severity of side effects at target dose (topiramate
[TOP] vs. placebo [PLAC]) along with standard error bars. Sig-
nificant (p ? .05) genotype differences within the TOP condition
are indicated by an asterisk.
GRIK1 genotype (rs2832407) by medication interac-
ADVERSE EVENTS FROM TOPIRAMATE IN HEAVY DRINKERS
4% of the variance was shared between serum TOP levels
and severity of side effects, suggesting that mechanisms
other than drug metabolism may account for the effects of
genotype on TOP-induced side effect severity.
Thus, meaningful allelic variation in the gene responsible
for the structure of one of the receptors targeted by TOP, the
GluR5 kainate receptor, may contribute to the heterogeneity
of treatment effects for TOP, especially side effect liability.
These findings suggest that individuals with this genetic
variant of the GRIK 1 gene may require additional medica-
tion management to keep from discontinuing treatment with
TOP because of side effects, or the treating physician may
consider alternative pharmacotherapies for these individu-
als. Studies of treatment-seeking samples and larger cell
sizes are needed to more fully uncover the role of allelic
variation in the GRIK1 gene, drinking behavior, and clinical
response to TOP, including side effect liability. If sup-
ported, these preliminary findings may lead to a more per-
sonalized use of TOP for alcoholism.
The mechanisms by which this polymorphism may influ-
ence TOP-induced side effects and TOP serum levels re-
main unclear. Although there is some evidence of alterna-
tive splicing mechanisms operating at exon 9 of the GRIK1
gene (Barbon & Barlati, 2000) where SNP 2 (rs2832407) is
located, the functional significance of this polymorphism
remains unknown. On the basis of the pharmacogenetics
framework used in this article, one may speculate that
genetic factors of putative influence on the targets of a given
pharmacotherapy—in this case, glutamate receptors tar-
geted by TOP—may ultimately predict responses to this
pharmacotherapy. Specifically, animal (Gryder & Rogaw-
ski, 2003) and in vitro (Kaminski et al., 2004) studies have
suggested that inhibition of GluR5 kainate receptors repre-
sents an important mechanism underlying TOP’s effects.
Therefore, it is plausible to hypothesize that genetic varia-
tion leading to structural and/or functional variation in the
GluR5 kainate receptor, a selective target of TOP coded by
the GRIK1 gene, may in turn predict differential response to
this pharmacotherapy. Similar findings have been demon-
strated, for example, for the Asn40Asp allele of the ?-opi-
oid receptor gene, a functional polymorphism leading to
greater binding affinity for ?-endorphins (Bond et al.,
1998). This SNP has been found to predict more favorable
responses to naltrexone in the laboratory (Ray & Hutchison,
2007) and in clinical trials (Anton et al., 2008; Oslin et al.,
2003). Much work remains to be done before similar asser-
tions can be made for the pharmacogenetics of TOP. Nev-
ertheless, the rationale for this pharmacogenetic investiga-
tion is similar to that of naltrexone and largely based on the
premise that genetic variation in the GRIK1 gene, which
codes for the GluR5 kainate receptor subunit, may be highly
relevant to the pharmacogenetics of TOP because receptors
containing this subunit selectively bind TOP.
As is the case for genetic associations to polymorphisms
of unknown functional significance, linkage disequilibrium
represents an important and plausible alternative explana-
tion, whereby the signal detected at this locus in this study
may be a result of its close proximity to a functional
polymorphism. Similarly, population stratification and other
unmeasured third variables, both environmental and genetic
in nature, may account for these results. Future studies are
undoubtedly needed to establish the role of glutamatergic
genes, including GRIK1, to the pharmacogenetics of TOP,
as well as to elucidate the neurobiological and molecular
mechanisms underlying potential effects. Attention to
mechanisms of action of TOP beyond the glutamate system,
such as its effects on GABAergic neurotransmission, for
instance, is needed to fully inform pharmacogenetic inves-
tigations. Nevertheless, this study contributes preliminary
data using an empirically driven approach to the pharma-
cogenetics of TOP, a promising pharmacotherapy for alco-
holism (Johnson et al., 2007).
These results must be interpreted in light of the study’s
strengths and limitations. Strengths include the community
sample composed of heavy drinkers, most of whom met
criteria for an alcohol use disorder, the careful assessment
and medication compliance protocols, and the collection of
a biomarker of medication compliance via serum TOP lev-
els. Limitations include the small sample size, necessitating
the combination of the two active medication conditions
into one group, which in turn precluded the examination of
dose-dependent effects. In addition, because participants
were afforded some flexibility in scheduling the laboratory
session in Week 5, the number of possible drinking days
during Week 5 varied across participants (see Miranda et
al., 2008). As such, replication of the observed association
between genotype and %HDD will be important. In these
analyses, we did not implement corrections for multiple
comparisons because of the preliminary nature of this in-
vestigation and the small sample size. On balance, these
preliminary results suggest that genetic variation at the
GluR5 kainate receptor may be a useful predictor of TOP-
induced side effects. Future studies are needed to more fully
ascertain these results and to examine the role of these
polymorphisms as predictors of outcomes in clinical trials
Anton, R. F., Oroszi, G., O’Malley, S., Couper, D., Swift, R.,
Pettinati, H., et al. (2008). An evaluation of mu-opioid receptor
(OPRM1) as a predictor of naltrexone response in the treatment
of alcohol dependence: Results from the Combined Pharmaco-
therapies and Behavioral Interventions for Alcohol Dependence
(COMBINE) study. Archives of General Psychiatry, 65, 135–
Barbon, A., & Barlati, S. (2000). Genomic organization, proposed
alternative splicing mechanisms, and RNA editing structure of
GRIK1. Cytogenetics and Cell Genetics, 88, 236–239.
Barrett, J. C., Fry, B., Maller, J., & Daly, M. J. (2005). Haploview:
Analysis and visualization of LD and haplotype maps. Bioin-
formatics, 21, 263–265.
Bond, C., LaForge, K. S., Tian, M., Melia, D., Zhang, S., Borg, L.,
et al. (1998). Single-nucleotide polymorphism in the human mu
opioid receptor gene alters beta-endorphin binding and activity:
Possible implications for opiate addiction. Proceedings of the
National Academy of Sciences, USA, 95, 9608–9613.
Dar, R., Serlin, R. C., & Omer, H. (1994). Misuse of statistical
tests in three decades of psychotherapy research. Journal of
Consulting and Clinical Psychology, 62, 75–82.
RAY ET AL.
Dodd, P. R., Beckmann, A. M., Davidson, M. S., & Wilce, P. A. Download full-text
(2000). Glutamate-mediated transmission, alcohol, and alcohol-
ism. Neurochemistry International, 37, 509–533.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W.
(1995). Structured Clinical Interview for DSM–IV Axis I Dis-
orders—Patient ed. (SCID-I/P, version 2.0). New York: Bio-
metrics Research Department, New York State Psychiatric In-
Flannery, B. A., Allen, J. P., Pettinati, H. M., Rohsenow, D. J.,
Cisler, R. A., & Litten, R. Z. (2002). Using acquired knowledge
and new technologies in alcoholism treatment trials. Alcohol-
ism: Clinical and Experimental Research, 26, 423–429.
Freeman, B., Powell, J., Ball, D., Hill, L., Craig, I., & Plomin, R.
(1997). DNA by mail: An inexpensive and noninvasive method
for collecting DNA samples from widely dispersed populations.
Behavior Genetics, 27, 251–257.
Fritz, N., Glogau, S., Hoffmann, J., Rademacher, M., Elger, C. E.,
& Helmstaedter, C. (2005). Efficacy and cognitive side effects
of tiagabine and topiramate in patients with epilepsy. Epilepsy &
Behavior, 6, 373–381.
Goldstein, D. B., Need, A. C., Singh, R., & Sisodiya, S. M. (2007).
Potential genetic causes of heterogeneity of treatment effects.
American Journal of Medicine, 120(4, Suppl. 1), S21–S25.
Gryder, D. S., & Rogawski, M. A. (2003). Selective antagonism of
GluR5 kainate-receptor-mediated synaptic currents by topira-
mate in rat basolateral amygdala neurons. Journal of Neuro-
science, 23, 7069–7074.
Hyman, S. E., & Malenka, R. C. (2001). Addiction and the brain:
The neurobiology of compulsion and its persistence. Nature
Reviews Neuroscience, 2, 695–703.
Johnson, B. A. (2005). Recent advances in the development of
treatments for alcohol and cocaine dependence: Focus on topi-
ramate and other modulators of GABA or glutamate function.
CNS Drugs, 19, 873–896.
Johnson, B. A., & Ait-Daoud, N. (2000). Neuropharmacological
treatments for alcoholism: Scientific basis and clinical findings.
Psychopharmacology (Berlin), 149, 327–344.
Johnson, B. A., Ait-Daoud, N., Bowden, C. L., DiClemente, C. C.,
Roache, J. D., Lawson, K., et al. (2003, May 17). Oral topira-
mate for treatment of alcohol dependence: A randomised con-
trolled trial. Lancet, 361, 1677–1685.
Johnson, B. A., Rosenthal, N., Capece, J. A., Wiegand, F., Mao,
L., Beyers, K., et al. (2007, October 10). Topiramate for treating
alcohol dependence: A randomized controlled trial. JAMA, 298,
Johnson, B. A., Swift, R. M., Addolorato, G., Ciraulo, D. A., &
Myrick, H. (2005). Safety and efficacy of GABAergic medica-
tions for treating alcoholism. Alcoholism: Clinical and Experi-
mental Research, 29, 248–254.
Kaminski, R. M., Banerjee, M., & Rogawski, M. A. (2004).
Topiramate selectively protects against seizures induced by
ATPA, a GluR5 kainate receptor agonist. Neuropharmacol-
ogy, 46, 1097–1104.
Kirchheiner, J., Nickchen, K., Bauer, M., Wong, M. L., Licinio, J.,
Roots, I., et al. (2004). Pharmacogenetics of antidepressants and
antipsychotics: The contribution of allelic variations to the phe-
notype of drug response. Molecular Psychiatry, 9, 442–473.
Kranzler, H. R., Covault, J., Herman, A., Burian, L., & Arias, A.
(2007). Glutamate receptor gene variation and alcohol depen-
dence risk. Alcoholism: Clinical and Experimental Re-
search, 31, 71A.
Krupitsky, E. M., Rudenko, A. A., Burakov, A. M., Slavina, T. Y.,
Grinenko, A. A., Pittman, B., et al. (2007). Antiglutamatergic
strategies for ethanol detoxification: Comparison with placebo
and diazepam. Alcoholism: Clinical and Experimental Re-
search, 31, 604–611.
Lench, N., Stanier, P., & Williamson, R. (1988, June 18). Simple
non-invasive method to obtain DNA for gene analysis. Lan-
cet, 1, 1356–1358.
Lipsky, R. H., & Goldman, D. (2003). Genomics and variation of
ionotropic glutamate receptors. Annals of the New York Acad-
emy of Sciences, 1003, 22–35.
Littleton, J. (1998). Neurochemical mechanisms underlying alco-
hol withdrawal. Alcohol Health & Research World, 22(1),
Livak, K. J. (1999). Allelic discrimination using fluorogenic
probes and the 5? nuclease assay. Genetic Analysis, 14, 143–
Miranda, R., Jr., MacKillop, J., Monti, P. M., Rohsenow, D. J.,
Tidey, J., Gwaltney, C., et al. (2008). Effects of topiramate on
urge to drink and the subjective effects of alcohol: A prelimi-
nary laboratory study. Alcoholism: Clinical and Experimental
Research, 32, 489–497.
O’Brien, C. P. (2008). Prospects for a genomic approach to the
treatment of alcoholism. Archives of General Psychiatry, 65,
Oslin, D. W., Berrettini, W., Kranzler, H. R., Pettinati, H., Gelern-
ter, J., Volpicelli, J. R., et al. (2003). A functional polymorphism
of the mu-opioid receptor gene is associated with naltrexone
response in alcohol-dependent patients. Neuropsychopharma-
cology, 28, 1546–1552.
Raguraman, J., Priyadharshini, R. K., & Chandrasekaran, R.
(2005). Effects of topiramate in alcohol dependence. Australian
and New Zealand Journal of Psychiatry, 39, 736–737.
Ray, L. A., & Hutchison, K. E. (2007). Effects of naltrexone on
alcohol sensitivity and genetic moderators of medication re-
sponse: A double-blind placebo-controlled study. Archives of
General Psychiatry, 64, 1069–1077.
Rubio, G., Ponce, G., Jimenez-Arriero, M. A., Palomo, T., Man-
zanares, J., & Ferre, F. (2004). Effects of topiramate in the
treatment of alcohol dependence. Pharmacopsychiatry, 37, 37–
Sobell, L. C., Maisto, S. A., Sobell, M. B., & Cooper, A. M.
(1979). Reliability of alcohol abusers’ self-reports of drinking
behavior. Behaviour Research and Therapy, 17, 157–160.
Spitz, E., Moutier, R., Reed, T., Busnel, M. C., Marchaland, C.,
Roubertoux, P. L., et al. (1996). Comparative diagnoses of twin
zygosity by SSLP variant analysis, questionnaire, and dermato-
glyphic analysis. Behavior Genetics, 26(1), 55–63.
Thompson, P. J., Baxendale, S. A., Duncan, J. S., & Sander, J. W.
(2000). Effects of topiramate on cognitive function. Journal of
Neurology, Neurosurgery and Psychiatry, 69, 636–641.
Walker, A. H., Najarian, D., White, D. L., Jaffe, J. F., Kanetsky,
P. A., & Rebbeck, T. R. (1999). Collection of genomic DNA by
buccal swabs for polymerase chain reaction-based biomarker
assays. Environmental Health Perspectives, 107, 517–520.
Weiss, F., Lorang, M. T., Bloom, F. E., & Koob, G. F. (1993). Oral
alcohol self-administration stimulates dopamine release in the
rat nucleus accumbens: Genetic and motivational determinants.
Journal of Pharmacology and Experimental Therapeutics, 267,
Weiss, F., & Porrino, L. J. (2002). Behavioral neurobiology of
alcohol addiction: Recent advances and challenges. Journal of
Neuroscience, 22, 3332–3337.
Received September 24, 2008
Revision received February 26, 2009
Accepted February 27, 2009 ?
ADVERSE EVENTS FROM TOPIRAMATE IN HEAVY DRINKERS