A Brain-Derived Neurotrophic Factor Haplotype Is
Associated with Therapeutic Response in
Eva Real, Mònica Gratacòs, Virginia Soria, Geòrgia Escaramís, Pino Alonso, Cinto Segalàs, Mònica Bayés,
Rafael de Cid, José M. Menchón, and Xavier Estivill
Background: Several clinical and genetic studies have focused on the role of brain-derived neurotrophic factor (BDNF) in the pathophys-
iology of various mental disorders. Recent lines of evidence regarding the network hypothesis of treatment outcome point towards the
serotonergic system in the pathophysiology and treatment of OCD, and upregulation of BDNF has been observed with various classes of
antidepressants, including selective serotonin reuptake inhibitors (SSRI). Thus, we hypothesized that the BDNF gene might also be
associated with treatment outcome in OCD.
Methods: We performed a single-marker and haplotype association study of eight tag single nucleotide polymorphisms in the BDNF
genomic region and related this to pharmacologic response in a sample of 131 OCD patients.
be associated with treatment outcome in MD (rs908867 and rs1491850).
though these preliminary findings await replication in a follow-up sample.
Key Words: Association, BDNF, haplotype, obsessive-compulsive
disorder, TagSNP, treatment outcome
its phenotypic expression. The concordance rate of OCD symptoms
in monozygotic twins is higher than in dizygotic twins (1), and
family studies have estimated that the prevalence of the disease
among first-degree relatives of affected probands is higher than
among first-degree relatives of healthy control subjects (2,3). More-
over, familial aggregation is higher among relatives of early onset
probands, indicating a stronger genetic component in some homo-
geneous subgroups of OCD patients (4).
Brain-derived neurotrophic factor (BDNF) belongs to the
family of neurotrophins that provides trophic support to cholin-
ergic, dopaminergic, and 5-hydroxytryptamine-containing neu-
rons (5). BDNF is widely distributed in the central nervous
system, including hippocampus, neocortex, amygdala, cerebel-
lum, and hypothalamus, all of which are key regions in the
regulation of mood and behavior (6). Because of its role in
lthough obsessive-compulsive disorder (OCD) is a clini-
cally heterogeneous disease, genetic epidemiologic stud-
ies have consistently shown that genetic liability underlies
neuronal survival and differentiation, synaptic efficiency, neuro-
nal plasticity, and neurotransmitter function, BDNF has become
one of the most attractive candidate genes in the study of the
genetic susceptibility to OCD and other psychiatric disorders.
Although several case-control and family-based association
studies have studied genetic variability in the BDNF gene in OCD
patients, the findings have been inconsistent. In particular, Hall
and colleagues (7) reported a significant association between
childhood-onset OCD and all BDNF gene markers tested, includ-
ing the Met allele of the functional Val66Met polymorphism. Two
subsequent studies replicated this association, although with the
opposite allele of this variant (the Val allele) (8,9), whereas other
authors found no evidence of an association (10,11). Finally, the
BDNF Met66 allele has also been reported to have a protective
role in females with OCD (12).
Despite the efficacy of clomipramine and selective serotonin
reuptake inhibitors (SSRIs) (13,14), a great variability in treatment
response has been reported in OCD. This could be related to
a range of factors, including interindividual differences in drug
response, which may be partially due to genetic factors. Phar-
macogenetic studies focusing on the efficacy of antidepressants
have traditionally been centered on genes that encode for
metabolizing enzymes, drug targets or genes involved in drug
action (15). A complementary hypothesis to the chemical view of
antidepressant action is that these drugs may exert major effects
on signaling pathways that regulate neuronal connectivity and
cell survival (16). Accordingly, pharmacologic (or physical)
treatments may restore information processing within particular
neural networks in the brain (17). One of the proposed mecha-
nisms involves activation of the neurotrophin signaling pathway,
which would in turn induce plastic changes in neuronal connec-
Research observations in rodents and humans support the
idea that antidepressant treatments, including SSRIs, increase the
expression and signaling of BDNF (19,20), and it has been shown
that BDNF protein levels increase after antidepressant treatment
From the OCD Clinical and Research Unit (ER, PA, CS, JMM), Department of
(VS, PA, JMM) and CIBER en Epidemiología y Salud Pública (CIBERESP)
(MG, GE, XE), Instituto de Salud Carlos III, Department of Psychiatry,
Bellvitge University Hospital; Genes and Disease Program and CeGen
Barcelona Genotyping Node (MG, GE, MB, RdC, XE), Center for Genomic
Regulation; Department of Clinical Sciences (PA, JMM), Bellvitge Cam-
pus, University of Barcelona; Department of Experimental and Health
Sciences (XE), Pompeu Fabra University; and Department of Psychiatry
(CS), Clinical Psychobiology, University of Barcelona, Barcelona, Spain.
Address correspondence to José Manuel Menchón, Ph.D., Obsessive-Com-
pulsive Disorder Clinical Research Unit, Department of Psychiatry, Bell-
vitge University Hospital, C/Feixa Llarga s/n., 08907 L’Hospitalet de Llo-
bregat, Barcelona, Spain; E-mail: firstname.lastname@example.org.
Received Jan 3, 2009; revised May 5, 2009; accepted May 5, 2009.
BIOL PSYCHIATRY 2009;66:674–680
© 2009 Society of Biological Psychiatry
(21–24). Polymorphic variants in BDNF have also been repeat-
edly related to antidepressant response in mood disorders (25–
28). It is therefore reasonable to suggest that BDNF, involved in
disease susceptibility, might also be related to and influence at
least a small part of the phenotypic variance in the therapeutic
effect of antidepressants observed in OCD patients. The aim of
this study was to investigate whether BDNF variants were
associated with the pharmacologic response to serotonin re-
uptake inhibitors (SRIs) in OCD. To this end we performed a
single-marker and haplotype association study of eight tag single
nucleotide polymorphisms (TagSNPs) located in the BDNF
genomic region in a sample of 131 OCD patients, using antide-
pressant response as a specific subphenotype.
Materials and Methods
Subjects were recruited through the OCD Clinic of Bellvitge
University Hospital (Barcelona, Spain) between 2003 and 2005.
One hundred thirty-one consecutive Spanish Caucasian outpa-
tients with OCD (65 males and 66 females) were initially included
in the study. All probands met DSM-IV (29) criteria for OCD and
had OCD symptoms for at least 1 year. Consensus diagnosis was
reached using the Structured Clinical Interview for DSM-IV Axis
I Disorders—Clinician Version (SCID-CV) (30) by two psychia-
trists with extensive clinical experience in OCD and who inter-
viewed the patients separately. Exclusion criteria were a past or
present history of psychoactive substance abuse, a history of
psychotic disorders, age under 18 or over 65 years, mental
retardation, and severe organic or neurological pathology, ex-
cept tic disorder. Comorbidity with other DSM-IV Axis I disorders
was not considered an exclusion criterion, provided that OCD
was the primary reason for seeking medical assistance. Demo-
graphic data and medical history were documented. To be
included in the study the patients had to have received standard-
ized treatment with an SRI during a period of at least 12 weeks.
Patients taking antipsychotic drugs were excluded from the
study. Eight (4 males and 4 females) of the 131 patients initially
recruited were lost to follow-up (three preferred to switch to
cognitive-behavioral therapy (CBT) alone, one abandoned treat-
ment because of the intention to become pregnant, one devel-
oped a psychotic disorder during the study, and three dropped
out). Demographic and clinical variables of lost patients were
equivalent to those corresponding to the final sample. After a
complete description of the study, written informed consent was
obtained from every patient. The study protocol was approved
by the Hospital Universitari de Bellvitge Ethics Committee (CEIC
Ciutat Sanitària i Universitària de Bellvitge).
Clinical assessment included information on age at OCD
onset and other clinical variables such as a history of personal or
familial Tourette’s disorder or another tic disorder, comorbidity,
personality disorder, familiar history of OCD, or depression. A
clinician-administered version of the Yale-Brown Obsessive
Compulsive Scale (Y-BOCS) (31) and the 21-item Hamilton
Depression Rating Scale (HDRS) (32) were used to assess the
severity of obsessive-compulsive and depressive symptoms re-
spectively. The clinician-administered version of the Y-BOCS
Symptom Checklist (31) was used to ascertain scores on five
previously identified symptom dimensions: symmetry/ordering,
hoarding,contamination/cleaning, aggression/checking, and sexual/
religious obsessions, all classified as “absent” or “present” (33).
The presence of family psychiatric history in first-degree relatives
was established according to the family history research diagnos-
tic criteria (FH-RDC) (34). Direct diagnostic information on
first-degree relatives was obtained whenever possible.
All probands were treated according to the clinical require-
ments of their obsessive-compulsive symptoms, as described in
the most important reviews and guidelines (35). After the first
assessment all patients received pharmacologic treatment taking
into account their response to previous therapies (see Table 1),
and full doses remained unchanged for a period of at least 12
weeks if the patient showed good tolerance. The pharmacologic
therapies employed were fluoxetine 60–80 mg/day (n ? 39,
31.7%), fluvoxamine 200–300 mg/day (n ? 31, 25.2%), and
clomipramine 225 mg/day–300 mg/day (n ? 53, 43.1%). Benzo-
diazepines were allowed during the first 2 weeks of the pharma-
cologic trial, up to doses equivalent to 20 mg/day of diazepam.
The Y-BOCS and the HDRS were used to assess progress as
regards the severity of obsessive-compulsive symptoms and
mood state at Weeks 2, 4, 8, and 12.
Standardized operational criteria described elsewhere (36)
were used to assess treatment response in OCD. The “responders
Table 1. Sociodemographic and Clinical Characteristics of OCD Patients
Included in the Study
CharacteristicsOCD Patients (n ? 123)
Age, Years, Mean ? SD (Range)
Male/Female, n (%)
Age at Onset of OCD, Mean ? SD (Range)
Baseline Y-BOCS Score, Mean ? SD (Range)
Baseline HDRS, Score, Mean ? SD (Range)
Current Comorbid Diagnosis, n (%)
Any comorbid diagnosis
Major depressive disorder, n (%)
Personality disorders, n (%)
Any personality disorder
Tic disorder or GT history
Eating disorders history
Family Psychiatric History, n (%)
Any psychiatric diagnosis
Tic disorder or GT
SRI Trials Completed, n (%)
Symptom Dimensions Present, n (%)
33.3 ? 10.5 (17–60)
61/62 (49.6%, 50.4%)
19.5 ? 8.3 (6–45)
28.2 ? 5.1 (16–40)
14.1 ? 2.5 (8–20)
14.1 ? 2.8 (8–20)
14.2 ? 5.4 (4–29)
OCD, obsessive-compulsive disorder; GT, Gilles de la Tourette’s syn-
drome; HDRS, Hamilton Depression Rating Scale; PD, personality disorder;
Y-BOCS, Yale-Brown Obsessive Compulsive Scale; SRI, serotonin reuptake
E. Real et al.
BIOL PSYCHIATRY 2009;66:674–680 675
to treatment” group (Rp) was defined as ?35% improvement in
baseline Y-BOCS (?35% postbaseline Y-BOCS reduction after
acute pharmacologic treatment at Week 12). Patients who did not
achieve this degree of improvement were considered as “nonre-
sponders” (non-Rp). We also measured depressive response
after Week 12. This was defined as a reduction in the HDRS score
from baseline of at least 50% (37).
From the HapMap project data set, we used genotypes from
public release 16 (Phase I data freeze, DbSNP b124), correspond-
ing to the 60 individuals from 30 CEPH trios of European descent
(http://www.hapmap.org). From the gene location, we extended
the search for 5–10 kilobases (kb) upstream and downstream of
the BDNF gene. Only single nucleotide polymorphisms (SNPs)
with a unique mapping location on the National Center for
Biotechnology Information (NCBI) B34 assembly and a minor
allele frequency (MAF) greater than 10% were considered for
further analysis. In the BDNF region (RefSeq: NM_170731; chro-
mosome 11: 27629852–27717072, NCBI B34 assembly), covering
87.2 kb, genotypes for 27 SNPs were available. Twenty-one of
these had an MAF ?10% (average spaced 4.4 kb).
Bins of common SNPs in strong linkage disequilibrium (LD),
as defined by an r2? .85, were identified within both data sets
by use of HapMap-LDSelect-Processor. This uses the “LD Select”
method (38) to process HapMap genotype dump format data
corresponding to the locus region. Eight TagSNPs were selected
to cover all bins in the case of BDNF, including a nonsynony-
mous variant in the coding exon (rs6265, corresponding to the
functional Val66Met amino-acid change).
SNPlex Design and Quality Control of Genotypes
Genotyping was performed at the Center for Genomic Regu-
lation (CRG) in Barcelona, linked to the “Centro Nacional de
Genotipado” (CeGen) in Spain (http://www.cegen.org). SNPs
were genotyped using the SNPlex platform (Applied Biosystems,
Foster City, California) according to manufacturer instructions
and analyzed on an Applied Biosystems 3730/3730xl DNA
Analyzer. Allele-calling was done by clustering analysis using
Genemapper software (Genemapper version 4.0). The mean
genotype call rate was 99.5%. Genotyping quality was controlled
in two ways. For quality control purposes, we used six dupli-
cated samples of two HapMap reference trios (family numbers
CEPH131 and CEPH132) that were incorporated in the genotyp-
ing process. Both genotype concordance and correct Mendelian
inheritance were verified: genotype concordance was tested
using SNPator, a Web-based tool for genotyping management
and SNP analysis developed by CEGEN, and Mendelian inheri-
tance was tested using Haploview software version 4.1.
Genotyping for Population Admixture
We tested potential genetic stratification in our sample by
analyzing 48 anonymous SNPs from an SNP set derived from a
panel of 52 ancestral informative markers (AIM) reported to be
polymorphic in European, Asian, and African populations (39).
Only the SNPs in Hardy-Weinberg equilibrium (HWE) and a
genotyping rate ?90% were retained for the analysis (n ? 47). In
addition, to objectively avoid any bias due to low LD between
markers, we also dropped those markers with r2? .1 between
them (n ? 10). Thus, we finally used a total of 37 out of the 48
SNPs initially genotyped for these analyses. Clusters of geneti-
cally similar individuals were identified by means of Structure
software (version 2.0) (40). An admixture model with correlated
frequencies was used, using five putative k values (1 to 5).
Analysis was performed both with and without prior population
Deviations from HWE were tested using an exact test, con-
sidering a p value of .05 as a threshold. To test the hypothesis of
association between alleles and phenotypes, we used the chi-
square test as implemented in Haploview software version 4.1
(41). For the genotype association analyses, we used multivariate
methods based on logistic regression analyses, with obsessive-
compulsive treatment outcome as the dependent variable and
analyzed under dominant, recessive, and additive inheritance
models. The allele with the highest frequency was assumed to be
the common allele and the genotype homozygous for it was
considered as the reference category. Odds ratios (OR) and 95%
confidence intervals (CIs) were calculated for each genotype
compared with the reference category. For each polymorphism,
p values were computed using the likelihood ratio test (LRT) by
comparing the model adjusted according to the following covari-
ates: sex, age at onset, HDRS score, total score on Y-BOCS,
presence of obsessive-compulsive personality disorder, and
presence of tics. For the multiple comparison correction, we
considered eight independent SNPs and three inheritance mod-
els, which gave a corrected p value of .0045. All analyses were
performed using the SNPassoc R package (42). For the multiple
comparison correction, we used a permutation procedure (n ?
1000) as implemented in Haploview software version 4.1, and we
forced the inclusion of both single-marker comparisons and
haplotype analysis (discussed later).
Linkage disequilibrium between polymorphisms and haplo-
type block structures was evaluated by Haploview software
version 4.1. Regions of strong LD were defined according to the
four gamete rules algorithm (43). We compared haplotypes
between Rp and non-Rp present at a frequency of at least 5%.
Statistical significance was assessed with a permutation proce-
dure to estimate the significance of the best result (1000 permu-
tations) taking into account both single markers and haplotypes
in blocks as implemented in Haploview software. We then used
the SNPassoc R package to obtain the OR and 95% CI interval
corresponding to the significant haplotype.
The robustness of these individual findings was investigated
by a bootstrap analysis. The bootstrap approach selects random
samples of size n (n1 cases ? n2 controls) drawn from the
original data set (44). Through the repetition of the sampling
procedure, the association between each of the polymorphisms
and treatment outcome phenotype was determined and informa-
tion was obtained on the variability and validity of the parameter
estimate. We repeated the selection procedure on 1000 bootstrap
data sets generated from the original sample and reported the
number of times that a particular association between each
polymorphism and the phenotype was statistically significant.
The significance level for each sampling repetition was deter-
mined with a permutation test (1000 permutations).
Sample Power Calculations
We computed power calculations using Quanto software (45)
and determined that, under a log-additive model and an MAF of
10%, the sample had 80% power to detect an OR of at least 2.8,
90% power to detect an OR of at least 3.2%, and 95% power to
detect an OR of at least 3.6.
676 BIOL PSYCHIATRY 2009;66:674–680
E. Real et al.
The sample consisted of 61 male and 62 female OCD patients,
with a mean age of 33.3 years (? 10.5) and a mean age at onset
of 19.5 years (? 8.3). Of these 123 patients, 37 were classified as
Rp and 86 as non-Rp. Other clinical variables are summarized in
Assessment of Population Admixture
We tested the sample for population stratification as de-
scribed in Methods and Materials. The set of anonymous un-
linked markers showed no evidence of stratification, because
allelic frequencies were similar and the best solution for each
sample was a single population.
BDNF Single-SNP Association Analyses
We genotyped eight TagSNPs located in the BDNF genomic
region, all of them consistent with HWE (Table S1 in Supplement
1) and with a mean genotyping rate of 99.5%. Initial analysis of
the SNP allele frequency distributions for the Rp versus the
non-Rp group revealed significant differences for four SNPs
(rs1103009, rs10501087, rs6265 and rs1491850; Table 2), although
only rs1491850 remained significant after permutation correction
(permutation p value ? .002). The permutation correction was
applied considering both single markers and haplotypes in
blocks to take into account all analyses performed in the study.
We further examined the genotype frequency distributions in
a fully adjusted model, covariating by sex, age at onset, HDRS
score, total score on Y-BOCS, presence of obsessive-compulsive
personality disorder, and presence of tics. We found significant
differences between Rp and non-Rp for the same four SNPs:
rs1103009 (p ? .04, additive model), rs10501087 (p ? .01,
dominant model), rs6265 (p ? .03, dominant model), and
rs1491850 (p ? .004, additive model). After permutation correc-
tion, only rs1491850 remained statistically significant and associ-
ated with a higher probability of being a responder (OR ? 2.40;
95% CI ? 1.29–4.47; Table 3). We used a “bootstrap” strategy to
test the likelihood that results would be replicated and explored
the stability of results across numerous configurations of the
subjects. Following the 1000 bootstrap data sets generated from
the original sample, rs1491850 was selected (statistically signifi-
cant after permutation correction) with the highest frequency
(802 of 1000).
We performed an exploratory analysis comparing the geno-
type frequency distribution of rs1491850 with regard to demo-
graphic and clinical variables of the clinical sample. We did not
find statistically significant associations regarding age, sex, main
OCD clinical dimension symptoms, and previous SRI trials (data
Table 2. Results of the Allelic Association for Responders Versus Nonresponders with Eight BDNF TagSNPs
dbSNP IDAllele Non-Rp n (%)Rp n (%)
Permutation p Valuea
rs1491851 .85 .99
Rp, responders; TagSNP, tag single nucleotide polymorphisms.
aPermutation p value.
Table 3. Genotype Frequencies of Brain-Derived Neurotrophic Factor Single Nucleotide Polymorphism
rs1491850 in Responders Versus Nonresponders Among OCD Patients
Genetic ModelNon-Rp n (%)Rp n (%)OR (95% CI)Adjusted p Valuea
responder; Y-BOCS, Yale-Brown Obsessive Compulsive Scale.
aPermuted p value adjusted by sex, Hamilton Depression Rating Scale baseline score, Y-BOCS baseline score,
obsessive-compulsive personality disorder, age at onset of OCD and presence of tic disorder.
bAttempts to find the minimal model that correctly explains the data.
E. Real et al.
BIOL PSYCHIATRY 2009;66:674–680 677
not shown). We detected nominal statistically significant differ-
ences between mean baseline Y-BOCS scores (p ? .037) and
mean baseline HDRS scores (p ? .015) with the rs1491850
genotypes. However, our main analysis of OCD treatment re-
sponse is adjusted for these variables, and so our results are
controlled for the severity of OCD and depression.
BDNF Haplotype Association Analysis
Calculation of pairwise LD for the SNPs analyzed in the whole
sample produced two LD blocks within BDNF: one block with
two SNPs (1 kb; mean r2? .3) and a second block with five SNPs
(79 kb; mean r2? .2; Figure S1 in Supplement 1). Based on
haplotype block structure, we performed a haplotype-specific
association study between the Rp and non-Rp groups. A com-
posite haplotype in Block 2 was found to be significantly
associated to outcome: being a carrier of haplotype T-G-T-G-T
confers a twofold higher probability of being a non-Rp (OR ?
2.27; 95% CI ? 1.01–5.0; permutation p value ? .027) (Table 4).
In this study we characterized the common genetic variation
in BDNF and tested whether inherited differences in this gene
influence the pharmacologic response to antidepressants in
a sample of 123 OCD patients. We identified a single SNP
(rs1491850) and a composite haplotype (rs1491850–rs908867)
located in the BDNF 5= upstream region that was associated with
a better therapeutic effect in our sample (measured by Y-BOCS
scores). To our knowledge, this is the first attempt to investigate
the involvement of BDNF variants in pharmacologic response in
Recently, Chen and colleagues (46) assessed anxiety-like
behavior in a knock-in mouse model in which BDNFMetwas
endogenously expressed. Interestingly, the authors found that,
compared with their wildtype control mice, homozygous (Met/
Met) and heterozygous (?/Met) mice showed similarly dimin-
ished responses to antidepressant treatment with fluoxetine. In
humans, few studies have been carried out relating antidepres-
sant response and genetic variability in the BDNF gene. One
initial study in major depressive disorder patients treated with
fluoxetine showed a trend toward improved antidepressant
response among heterozygous carriers of the amino acid change
Val66Met (rs6265), although it failed to provide evidence of a
significant association (27). Similarly, another study found a
higher frequency of Met allele carriers among responders to
citalopram (25). These findings were further replicated by an
independent study that found that heterozygous carriers of this
same polymorphism were associated with a significantly better
therapeutic effect, irrespective of which antidepressant (fluvox-
amine or milnacipram) was administered (28). Finally, a BDNF
haplotype has recently been reported to be associated with the
patient’s state of remittance after an adequate trial with a single
antidepressant in mood disorders (26).
A relationship between disorder risk genes and genes related
to treatment response has been suggested for different mental
disorders (47). The BDNF gene is one of the candidate genes
most frequently associated with both the pathogenesis of mental
disorders and the treatment outcome of these diseases. In this
regard, our results suggest that BDNF could indeed be involved
in the response of OCD to serotonergic drugs and not only in the
development of specific mental conditions. This reasoning sup-
ports the network hypothesis as an accurate explanation for the
mechanism of action of antidepressant drugs (48). The network
hypothesis, initially formulated for mood disorders, proposes
that mental diseases reflect problems in information processing
within particular neural networks and that the action of antide-
pressant drugs might result in morphological and physiological
reorganization of neuronal connections, thus improving informa-
tion processing within these networks (18). Activity-dependent
plastic processes seem to underlie these effects, and an increased
turnover in synaptic connections (without any net change in
synaptic number) might lead to significant reorganization of
neuronal connectivity (49).
The evidence for the network hypothesis of antidepressant
action is limited and mostly indirect. Several reports support the
neurotrophic-like effects of antidepressant drugs. Upregulation
of BDNF has been observed with chronic administration of
different classes of antidepressants: SSRI, norepinephrine selec-
tive reuptake inhibitors, monoamine oxidase inhibitors, and
atypical antidepressants (19,20). Nibuya and colleagues (20)
found a significant increase of BDNF mRNA in rat hippocampus
after the administration of several antidepressants, and the time
course for these changes was consistent with the time delay
required for the drugs to have a therapeutic action. Furthermore,
an increased proliferation of new neurons in rodent hippocam-
pus after chronic administration of different antidepressants has
been found (50). The time needed for the mature neurons to
develop was also correlated with the delayed onset of clinical
drug effects. It is important to note that other classes of psycho-
tropic drugs (including antipsychotics) have not been related to
an increase in BDNF expression, thus demonstrating the phar-
macologic specificity of antidepressants for this effect.
The way in which antidepressants enable the induction of
BDNF is still unknown. Recently, Tsankova et al. (51) showed
that antidepressants modified the chromatin state by increasing
histone acetylation at the BDNF promoter, leading to an activa-
tion of BDNF gene expression. Our results suggest that polymor-
phic variations in BDNF could be involved in the signaling
pathways that regulate neuroplastic restoration of the neural
circuits involved in OCD pathophysiology.
Although we ruled out population admixture, which could
lead to false-positive associations, and performed a bootstrap
analysis to investigate the robustness of the results, our findings
must be considered preliminary until further confirmation is
obtained in independent samples or in family studies. This is so
for two reasons; first, because of a small sample size; second
because patients were recruited from a specialized OCD Unit,
and so the population will have presented a higher degree of
severity than a sample from the community.
Table 4. Summary of Haplotype-Based Analysis of Treatment Outcome
Phenotype in Nonresponders (non-Rp) Versus Responders (Rp) (Block2:
rs1050187, rs6265, rs12273363, rs908867 and rs1491850)
(% non-Rp; % Rp) OR95% CI
CI, confidence interval; OR, odds ratio.
aPermutation p value considering single-markers and haplotypes in
score, total score on Yale-Brown Obsessive Compulsive Scale, presence of
obsessive-compulsive personality disorder, and presence of tics.
678 BIOL PSYCHIATRY 2009;66:674–680
E. Real et al.
In summary, our data support the notion that genetic variants
in BDNF may be related to antidepressant response in OCD.
Furthermore, the findings are consistent with current views
regarding the pathophysiology of OCD, which has been linked
to a disturbance of specific cortico-striato-thalamic circuits in-
volving orbitofrontal cortex and caudate nucleus. Because BDNF
plays a major role both in modulating neuronal circuits and in
antidepressant response, improvement in OCD could be related
to BDNF variants, which may modulate the functional restoration
of the circuits involved in OCD.
This study was supported in part by the Instituto de Salud,
Carlos (III) Centro de Investigación en Red de Salud Mental
(CIBERSAM), Centro de Investigación en Red en Epidemiología y
Salud Pública (CIBERESP), Psychiatry Genetics Network (GO3/
184) and Fondo de Investigaciones Sanitarias de la Seguridad
Social, FIS PI071029 and FIS PI071044. CS was funded by the
Agència de Gestió d’Ajuts Universitaris I de Recerca (AGAUR,
2005 FI00738 and Grants 010,210 from the Fundació la Marató
TV3, Barcelona. ER was funded by the Institut d’Investigació
Biomèdica de Bellvitge (IDIBELL). The authors thank all the
study participants and staff from the Department of Psychiatry of
Hospital Universitari de Bellvitge who helped enroll the study
sample. We also thank Michael Maudsley for reviewing the
The authors have no biomedical financial interests or poten-
tial conflicts of interest.
Supplementary material cited in this article is available
and obsessive disorders. In: Klein D, Rabkin J, editors. Anxiety: New
2. Nestadt G, Samuels J, Riddle M, Bienvenu OJ, III, Liang KY, LaBuda M, et
al. (2000): A family study of obsessive-compulsive disorder. Arch Gen
3. Pauls DL, Alsobrook JP 2nd, Goodman W, Rasmussen S, Leckman JF
4. do Rosario-Campos MC, Leckman JF, Curi M, Quatrano S, Katsovitch L,
5. Skaper SD (2008): The biology of neurotrophins, signaling pathways,
and functional peptide mimetics of neurotrophins and their receptors.
6. Drevets WC (2000): Neuroimaging studies of mood disorders. Biol Psy-
7. Hall D, Dhilla A, Charalambous A, Gogos JA, Karayiorgou M (2003):
Sequence variants of the brain-derived neurotrophic factor (BDNF)
gene are strongly associated with obsessive-compulsive disorder. Am J
protective haplotypes against obsessive-compulsive disorder. Biol Psy-
9. Hemmings SM, Kinnear CJ, Van der Merwe L, Lochner C, Corfield VA,
Moolman-Smook JC, et al. (2008): Investigating the role of the brain-
10. Dickel DE, Veenstra-VanderWeele J, Bivens NC, Wu X, Fischer DJ, Van
Etten-Lee M, et al. (2007): Association studies of serotonin system can-
didate genes in early-onset obsessive-compulsive disorder. Biol Psychi-
11. Wendland JR, Kruse MR, Cromer KR, Murphy DL (2007): A large case-
control study of common functional SLC6A4 and BDNF variants in ob-
(BDNF) in the phenotypic expression of obsessive-compulsive disorder
(OCD) [published online ahead of print February 13]. Am J Med Genet B
13. Greist JH, Jefferson JW, Kobak KA, Katzelnick DJ, Serlin RC (1995): Effi-
pulsive disorder. A meta-analysis. Arch Gen Psychiatry 52:53–60.
14. The Clomipramine Collaborative Study Group (1991): Clomipramine in
15. Binder EB, Holsboer F (2006): Pharmacogenomics and antidepressant
drugs. Ann Med 38:82–94.
16. Manji HK, Drevets WC, Charney DS (2001): The cellular neurobiology of
depression. Nat Med 7:541–547.
17. Duman RS, Heninger GR, Nestler EJ (1997): A molecular and cellular
theory of depression. Arch Gen Psychiatry 54:597–606.
neurotrophin receptor trkB regulates the expression of plasticity-re-
latedgenesinmousebrain. BrainResMolBrainRes 130:81–94.
19. Duman RS, Monteggia LM (2006): A neurotrophic model for stress-
related mood disorders. Biol Psychiatry 59:1116–1127.
20. Nibuya M, Morinobu S, Duman RS (1995): Regulation of BDNF and trkB
sant drug treatments. J Neurosci 15:7539–7547.
21. Hellweg R, Ziegenhorn A, Heuser I, Deuschle M (2008): Serum concen-
22. Karege F, Perret G, Bondolfi G, Schwald M, Bertschy G, Aubry JM (2002):
Decreased serum brain-derived neurotrophic factor levels in major de-
pressed patients. Psychiatry Res 109:143–148.
al. (2003): Alterations of serum levels of brain-derived neurotrophic
factor (BDNF) in depressed patients with or without antidepressants.
24. Yoshimura R, Mitoma M, Sugita A, Hori H, Okamoto T, Umene W, et al.
(2007): Effects of paroxetine or milnacipran on serum brain-derived
neurotrophic factor in depressed patients. Prog Neuro-Psychopharma-
25. Choi MJ, Kang RH, Lim SW, Oh KS, Lee MS (2006): Brain-derived neuro-
in major depressive disorder. Brain Res 1118:176–182.
26. Gratacos M, Soria V, Urretavizcaya M, Gonzalez JR, Crespo JM, Bayes M,
et al. (2008): A brain-derived neurotrophic factor (BDNF) haplotype is
associated with antidepressant treatment outcome in mood disorders.
a brain-derived neurotrophic-factor genetic polymorphism and major
depressive disorders, symptomatology, and antidepressant response.
28. Yoshida K, Higuchi H, Kamata M, Takahashi H, Inoue K, Suzuki T, et al.
(2007): The G19 6A polymorphism of the brain-derived neurotrophic
ine. J Psychopharmacol 21:650–656.
ual of Mental Disorders, 4th ed. Washington, DC: American Psychiatric
30. First M, Spitzer R, Gibbon M, Williams J (1997): Structured Clinical Inter-
view for DSM-IV Axis I Disorders-Clinician Version (SCID-CV). Washington,
DC: American Psychiatric Publishing.
31. Goodman WK, Price LH, Rasmussen SA, Mazure C, Fleischmann RL, Hill
ment, use, and reliability. Arch Gen Psychiatry 46:1006–1011.
32. Hamilton M (1960): A rating scale for depression. J Neurol Neurosurg
33. Mataix-Cols D, Rauch SL, Manzo PA, Jenike MA, Baer L (1999): Use of
factor-analyzed symptom dimensions to predict outcome with seroto-
nin reuptake inhibitors and placebo in the treatment of obsessive-
34. Andreasen NC, Endicott J, Spitzer RL, Winokur G (1977): The family
history method using diagnostic criteria. Reliability and validity. Arch
E. Real et al.
BIOL PSYCHIATRY 2009;66:674–680 679
35. Fineberg NA, Gale TM (2005): Evidence-based pharmacotherapy of Download full-text
obsessive-compulsive disorder. Int J Neuropsychopharmacology 8:
(2002): Treatment non-response in OCD: Methodological issues and
Conceptualization and rationale for consensus definitions of terms in
major depressive disorder. Remission, recovery, relapse, and recur-
Selecting a maximally informative set of single-nucleotide polymor-
39. Sanchez JJ, Phillips C, Borsting C, Balogh K, Bogus M, Fondevila M, et al.
(2006): A multiplex assay with 52 single nucleotide polymorphisms for
human identification. Electrophoresis 27:1713–1724.
40. Pritchard JK, Stephens M, Donnelly P (2000): Inference of population
structure using multilocus genotype data. Genetics 155:945–959.
41. Barrett JC, Fry B, Maller J, Daly MJ (2005): Haploview: Analysis and
visualization of LD and haplotype maps. Bioinformatics 21:263–265.
42. Gonzalez JR, Armengol L, Sole X, Guino E, Mercader JM, Estivill X, et al.
(2007): SNPassoc: An R package to perform whole genome association
studies. Bioinformatics 23:644–645.
43. Wang N, Akey JM, Zhang K, Chakraborty R, Jin L (2002): Distribution of
play of population history, recombination, and mutation. Am J Hum
44. Efron B, Tibshirani R (1993): An Introduction to the Bootstrap. New York:
Chapman & Hall.
of gene-gene interaction. Am J Epidemiol 155:478–484.
46. Chen ZY, Jing D, Bath KG, Ieraci A, Khan T, Siao CJ, et al. (2006): Genetic
47. McGough JJ (2005): Attention-deficit/hyperactivity disorder pharma-
cogenomics. Biol Psychiatry 57:1367–1373.
Neurobiology of depression. Neuron 34:13–25.
49. Hua JY, Smith SJ (2004): Neural activity and the dynamics of central
nervous system development. Nat Neurosci 7:327–332.
50. Malberg JE, Eisch AJ, Nestler EJ, Duman RS (2000): Chronic antidepres-
Sustained hippocampal chromatin regulation in a mouse model of
depression and antidepressant action. Nat Neurosci 9:519–525.
680 BIOL PSYCHIATRY 2009;66:674–680
E. Real et al.