J Psychiatry Neurosci 2008;33(5)
Objective: Brain-derived neurotrophic factor (BDNF) is a key factor in neuroplasticity and has been implicated in the affective disorders;
studies have demonstrated elevated BDNF in patients taking lithium and other mood stabilizers. The objective of our study was to ana-
lyze BDNF in lithium-responsive patients with bipolar disorder (BD) to further understand the role of BDNF in the pathophysiology of BD.
Methods: Using enzyme-linked immunosorbent assay, we measured transformed B lymphocytes for BDNF protein. Results: BDNF
levels were 36% lower in lymphoblasts from patients with BD (n = 12), compared with matched control participants (n = 13), and 55%
lower when compared with their unaffected relatives (n = 14). Lithium significantly decreased BDNF levels in patients with BD and
healthy control participants, although BDNF levels remained lower (33%) in the BD group posttreatment. Conclusion: Decreased BDNF
may constitute part of the pathophysiologic process of BD in a lithium-responsive subgroup of individuals with this disease. A compen-
satory mechanism protecting the genetically predisposed unaffected relatives from phenotypic expression of BD is suggested.
Objectif : Le facteur neurotrophique dérivé du cerveau (FNDC) joue un rôle clé dans la neuroplasticité et est mis en cause dans des
troubles de l’affectivité. Des études ont démontré des concentrations élevées de FNDC chez les patients qui prenaient du lithium et
d’autres thymorégulateurs. Nous voulions analyser la concentration de FNDC chez les patients atteints d’un trouble bipolaire (TB) répon-
dant au lithium afin de comprendre davantage le rôle du FNDC dans la pathophysiologie du TB. Méthodes : Nous avons mesuré par
dosage immunoenzymatique les lymphocytes B transformés pour déterminer les protéines dans le FNDC. Résultats : Les concentrations
de FNDC étaient plus faibles de 36 % dans les lymphoblastes provenant de patients atteints d’un TB (n = 12) que dans ceux des partici-
pants témoins jumelés (n = 13), et 55 % moins élevés que chez les membres de leur parenté non atteints (n = 14). Le lithium a réduit
considérablement les concentrations de FNDC chez les patients atteints de TB et les participants témoins en bonne santé, même si
elles sont demeurées plus faibles (33 %) chez les patients atteints de TB après le traitement. Conclusion : La diminution des concentra-
tions de FNDC peut constituer un élément du processus pathophysiologique du TB chez un sous-groupe d’individus atteints de la ma-
ladie qui répondent au lithium. On pense qu’un mécanisme compensatoire protège contre l’expression phénotypique du TB les patients
apparentés non touchés qui sont génétiquement prédisposés.
BDNF protein levels are decreased in transformed
lymphoblasts from lithium-responsive
patients with bipolar disorder
Michael Tseng, MD, PhD; Martin Alda, MD; Li Xu, MSc; Xiujun Sun, MSc;
Jun-Feng Wang, PhD; Paul Grof, MD, PhD; Gustavo Turecki, MD, PhD;
Guy Rouleau, MD, PhD; L. Trevor Young, MD, PhD
Tseng, Xu, Wang, Grof, Young — Centre for Addiction and Mental Health, and Department of Psychiatry, University of
Toronto, Toronto, Ont.; Alda — Department of Psychiatry, Dalhousie University, Halifax, NS; Turecki — Department of Psych-
iatry, McGill University; Rouleau — Department of Medicine, Université de Montréal, Montréal, Que.; Young — Department of
Psychiatry, University of British Columbia, Vancouver, BC
Correspondence to: Dr. L.T. Young, Department of Psychiatry, University of British Columbia, 2255 Wesbrook Mall, Vancouver BC
V6T 2A1; fax 604 822-0399; Trevor.Young@ubc.ca
J Psychiatry Neurosci 2008;33(5):449–53.
Medical subject headings: bipolar disorder; brain-derived neurotrophic factor; enzyme-linked immunosorbent assay; lithium.
Submitted Oct. 26, 2007; Revised Dec. 7, 2007; Accepted Dec. 31, 2007
© 2008 Canadian Medical Association
Bipolar disorder (BD) is a highly heterogeneous disorder, with
neuropathologic findings that are often variable and inconsis-
tent.1Lithium is considered first-line treatment for BD,2,3and
an increasing body of observations support the notion that
lithium responsiveness in bipolar patients is a characteristic
that defines a more homogeneous subpopulation of people
with BD. The criteria for lithium responsiveness are quite strin-
gent and are described elsewhere.4,5There are 4 advantages to
considering this population as an entity in its own right: 1)
lithium responders have distinct clinical features, including
episodic course and absence of rapid cycling; 2) lithium re-
sponders have distinct family histories (more specific for mood
disorders) and strong genetic loading; 3) lithium responders
differ from responders to other mood stabilizers, with lower
rates of comorbidity and fewer atypical features; 4) increasing
knowledge of lithium’s mechanisms of action can aid in the
distinction of lithium-related endophenotypes.6Examining the
unaffected relatives of a bipolar population may aid in the
study of endophenotypes in BD because, although they pos-
sess a genetic complement similar to that of their relatives with
BD and are therefore at risk for developing the illness, they do
not express the disease phenotype.
Neurotrophic factors play a role in modulating cell survival
and promoting neuronal plasticity. Therefore, deficient or al-
tered expression of these factors may contribute to, or be a con-
sequence of, the pathophysiology of mental illness. In particu-
lar, brain-derived neurotrophic factor (BDNF) has been
implicated in the affective disorders. Numerous studies have
investigated the val66met BDNF single nucleotide polymorph-
ism in relation to BD, with mixed results.7,8Although studies of
major depressive disorder have shown an increase in BDNF in
the postmortem hippocampus in response to antidepressant
therapy,9findings in regard to untreated BD have been mixed.
In lymphoblasts from BD patients, BDNF messenger ribo-
nucleic acid (mRNA) expression did not differ in comparison
with control participants.10Recently, serum studies have
shown a decrease in BDNF protein in BD patients with mania
and depression but not in euthymic BD patients.11The pur-
pose of this study was to measure basal and lithium-induced
BDNF protein levels in cultured B lymphoblasts from lithium-
responsive BD patients and nonpsychiatric, nonneurologic
control participants and to examine basal BDNF protein levels
in the unaffected family members of BD patients.
We studied a group of 12 lithium-responsive BD I patients,
14 of their unaffected family members and 13 nonpsychiatric,
nonneurologic control participants. Table 1 summarizes the
demographic data of this study sample. All participants were
interviewed with the Schedule for Affective Disorders and
Schizophrenia—Lifetime version12and then diagnosed ac-
cording to Research Diagnostic Criteria.13All patients also
met the criteria of the Diagnostic and Statistical Manual of Men-
tal Disorders, fourth edition,14for BD I without comorbid diag-
noses. BD patients also had to meet the criteria for a full re-
sponse to lithium prophylaxis, which are described
elsewhere.4,5Briefly, all lithium-responsive patients had been
maintained on lithium monotherapy for a minimum of
3 years, with an average plasma concentration of 0.6 mEq/L
and no affective episodes during the course of treatment.
They were, however, known to be at high risk for recurrence
before the initiation of lithium. Healthy control participants
were married-in members of probands’ families; none had a
family history of BD or major depression among first- or
second-degree relatives. The study was approved by the Re-
search Ethics Board at the Centre for Addiction and Mental
Health in Toronto, Ontario, and the Capital District Health
Authority in Halifax, Nova Scotia, and we obtained written
informed consent regarding participation in the study from
Cell isolation, culture and drug treatment
Lymphoblastoid cell lines were generated from blood sam-
ples, as described previously.15Human Epstein–Barr
virus–transformed B lymphocytes were cultured in Iscove’s
Modified Dulbecco’s Medium (Invitrogen) containing
15% fetal calf serum with 100 units/mL penicillin,
100 μg/mL streptomycin, 2 μg/mL amphotericin B and
2 μg/mL sodium desoxycholate in 5% CO2/95% air at 37°C
in a humidified incubator. Cells from BD and control groups
were treated with 1 mM lithium chloride (LiCl) or vehicle
(dH2O) every other day for 7 days, respectively.
Enzyme-linked immunosorbent assay (ELISA)
After treatment with 1 mM LiCl or dH2O for 7 days, cultured
B lymphocytes from BD, unaffected relatives and control par-
ticipants were collected and homogenized in 100 mM Tris-
HCl pH 7.0, with 2% bovine serum albumin, 1 M NaCl,
4 mM EDTA-Na2, 2% Triton X-100, 0.1% Na-azide, aprotinin
0.2 μg/mL, pepstatin A 0.2 μg/mL, leupeptin 0.5 μg/mL and
0.01 mM phenylmethylsulphonyl fluoride. Homogenates
were centrifuged (12000 × g, 4°C, 40 min), and the resulting
supernatant was used to perform the ELISA according to the
instructions of a BDNF Sandwich ELISA Kit (Chemicon). The
BDNF concentration for each sample was determined by
interpolating the sample’s optical density at 450 nm into the
linear range of a standard curve generated with serial con-
centrations of BDNF protein.
Data and statistical analyses
We tested between-group differences in the demographic
characteristics of age and sex, using analyses of variance
Tseng et al.
Rev Psychiatr Neurosci 2008;33(5)
Table 1: Demographic data for study participants
Group; mean (and SD) [range]*
n = 13
n = 12
n = 14
Age at onset, yr
Episodes before Li
Li treatment, yr
BD = bipolar disorder; Li = lithium; NA = not applicable; SD = standard deviation.
*Unless otherwise indicated.
BDNF is decreased in bipolar disorder lymphoblasts
J Psychiatry Neurosci 2008;33(5)
(ANOVA) and χ2tests, respectively. We used repeated-
measures ANOVA to analyze paired data, including cells
with and without lithium treatment, and we examined the
remaining measures with 1-way ANOVA followed by
Tukey post hoc tests to assess the difference between
groups in BDNF expression. All results were expressed as
the mean (and standard error of the mean [SEM]). Statistic-
al significance was set at p < 0.05 for all analyses. All statis-
tical analyses were carried out with BDMP version 8.1 for
The BDNF levels were significantly lower (36%, F1,23= 6.21,
p = 0.02) in transformed lymphoblasts from lithium-
responsive BD patients (mean 18.6, SEM 2.7 optical density
[OD] units), compared with healthy control participants
(mean 27.7, SEM 2.9 OD units) (Fig. 1). Of additional interest,
BDNF levels in lymphoblasts from BD patients were 55%
lower (p = 0.003) than in cells from unaffected relatives (mean
41.4, SEM 6.8 OD units), but they did not differ significantly
between healthy control participants and unaffected relatives
of BD patients. Further, lithium treatment significantly de-
creased BDNF levels in both treatment groups (12%,
F1,23= 4.36, p = 0.05). After lithium treatment, BDNF levels
were still significantly lower (33%) in BD patients (mean 15.9,
SEM 2.6 OD units), compared with control participants
(mean 24.8, SEM 2.7 OD units). No age- or sex-related effects
were observed in BDNF expression as examined by ANOVA
and χ2tests, respectively (data not shown).
The present study reveals that basal BDNF protein levels are
decreased in cultured lymphoblasts from lithium-responsive
BD patients when compared with both their unaffected rela-
tives and with healthy control participants. There was no sig-
nificant difference in BDNF levels between unaffected rela-
tives and healthy control participants. Treatment of cultured
lymphoblasts with lithium decreased BDNF levels in all par-
ticipants, but the difference between BD patients and healthy
control participants remained.
BDNF is a member of the neurotrophin family, and plays an
important role in neuroplasticity and neuroprotection. In BD
patients, serum BDNF has been reported as low during manic
and depressive episodes11and has been negatively correlated
with severity of manic episodes.16Further, the BDNF val66met
single nucleotide polymorphism has been associated with BD
with mixed results7; a recent meta-analysis found a lack of as-
sociation of this polymorphism with BD.8Our study revealed a
decrease in BDNF protein levels in the transformed lymph-
oblasts from lithium-responsive bipolar patients when com-
pared with BDNF levels in cells from their unaffected relatives
and from healthy control participants. In addition, we did not
find significant differences between the latter 2 groups. Taken
together, our data suggest that BDNF levels in lymphoblasts
may be indicative of a pathophysiologic process within this
disease entity and may also be a suitable biomarker for trait-
related factors in BD. These results also suggest that lithium
responsiveness may confer pathophysiologic changes on BD
lymphoblasts because a previous lymphoblast study did not
reveal BDNF mRNA expression differences between BD
patients and healthy control participants.10This may provide
a further rationale for future studies comparing lithium-
responsive and lithium-unresponsive BD patients.
The study of endophenotypes in BD is useful because it
permits the identification and examination of a more homo-
geneous clinical population. The use of a lithium-responsive
population is advantageous in this regard. If BDNF protein
levels are indeed an endophenotypic marker of BD, one might
predict the expression of BDNF in unaffected relatives to be in
the mid-range between that found in control participants and
that found in BD patients, according to recently modified BD
endophenotype criteria.17However, we found no significant
differences in BDNF levels between transformed lymph-
oblasts of control participants and those of unaffected rela-
tives, and paradoxically, we found higher BDNF levels in the
relatives compared with their BD counterparts. One possible
explanation for this discrepancy is a compensatory mech-
anism wherein unaffected relatives of lithium-responsive BD
patients upregulate peripheral BDNF protein levels, ultim-
ately protecting against expression of the bipolar phenotype
in a genetically predisposed cohort. According to this hypoth-
esis, one might expect BDNF levels to drop in unaffected rela-
tives in whom BD is later diagnosed. Prospectively examining
BDNF levels from transformed lymphoblasts in this subpopu-
lation may improve our understanding of the role of BDNF in
the emergence and progression of BD pathology. BDNF’s role
in neuroprotective signal transduction pathways makes it an
Fig. 1: Brain-derived neurotrophic factor (BDNF) protein levels in
transformed lymphoblasts from nonpsychiatric control subjects and
lithium-responsive bipolar disorder (BD) patients (with and without
1 mM lithium treatment for 7 days) and unaffected relatives. BDNF
protein levels were 36% lower in subjects with BD when compared
with nonpsychiatric control subjects (*p = 0.02), and this decrease
remained (33%, §p = 0.02) after lithium treatment. Lithium treat-
ment decreased BDNF levels in both BD and control populations
(†p = 0.05). BDNF protein levels were 55% lower in BD patients
when compared with their unaffected relatives (+p = 0.003). All
measures are expressed as mean (and standard deviation).
ideal candidate in such a compensatory mechanism, and ex-
ploring the role of related signal transduction factors such
as the cAMP response element-binding protein or mitogen-
activated protein kinases in unaffected relatives may inform
our understanding of such a mechanism.
Treatment with antidepressants and mood stabilizers in-
creases BDNF levels, as observed in human postmortem hip-
pocampus,9in rat hippocampus and frontal cortex18and in
several serum studies of individuals with depression.19–22In
contrast, our study demonstrates a decrease in BDNF levels
after lithium treatment of transformed lymphoblasts (Fig. 1)
in both bipolar and healthy control populations. This study is
the first to examine lymphoblast BDNF levels in response to
lithium, a response that may not be directly comparable with
the effects in brain tissue or serum. The decrease in BDNF
may be a dose-dependent response to lithium, or it may be
related to length of treatment. Dosage and duration of
lithium treatment were chosen on the basis of established
methods from previous studies.23–25However, lithium may
also have a biphasic effect.26Further studies examining a
wider range of dosages and treatment lengths may elaborate
on BDNF response to lithium in lymphoblasts.
Limitations of this study include the lack of data on
lithium-treated lymphoblasts from the unaffected relatives of
lithium-responsive BD patients. Also, the transformation
process of the lymphocytes may induce changes in BDNF ex-
pression, although in this experiment, any such effect should
have been uniform across all experimental groups. Addition-
ally, the relevance of BDNF levels in peripheral cells to BDNF
expression and function in the central nervous system of BD
patients is not entirely clear. In support of this approach,
BDNF levels in serum and cortex have been positively correl-
ated in rats,27and serum BDNF levels have recently been
associated with expression of the neuronal integrity marker
N-acetylaspartate in the human cerebral cortex.28Despite
their peripheral source, transformed lymphoblasts are easily
isolated and cultured and have the tremendous advantage of
allowing the study of putative biomarkers in their native, liv-
ing context, free of medication or hormonal influences. A
prolonged study duration might have revealed longer-term
lithium effects on BDNF levels. Finally, our findings might
have been strengthened with a larger number of study par-
ticipants, particularly in regard to increasing the power of
statistical analyses applied to discern family-of-origin effects.
In conclusion, we demonstrate a decrease in BDNF levels
in transformed lymphoblasts from lithium-responsive BD pa-
tients when compared with BDNF levels in cells from their
unaffected relatives and healthy control participants, even
when they are compared with healthy control participants
after lithium treatment. This raises the possibility that meas-
uring BDNF expression in transformed lymphoblasts could
be a useful diagnostic marker for a predisposition to develop
BD. Our results also suggest that, in unaffected relatives who
are genetically predisposed to BD, a potential BDNF-related
compensatory effect may be taking place and preventing
emergence of a bipolar phenotype. Finally, we observed a de-
crease in BDNF levels after lithium treatment, suggesting that
the mechanism of BDNF response to lithium in lymphoblasts
may be different when compared with that in cortex or
serum. A more detailed characterization of BDNF’s role in
the response to lithium therapy in both human and animal
models is warranted.
1. Alda M. The phenotypic spectra of bipolar disorder. Eur Neuro-
psychopharmacol 2004;14(Suppl 2):S94-9.
2.American Psychiatric Association. Practice guideline for the treat-
ment of patients with bipolar disorder (revision). Am J Psychiatry
3. Yatham LN, Kennedy SH, O’Donovan C, et al. Canadian Network
for Mood and Anxiety Treatments (CANMAT) guidelines for the
management of patients with bipolar disorder: consensus and con-
troversies. Bipolar Disord 2005;7(Suppl 3):5-69.
4.Grof P, Alda M, Grof E, et al. Lithium response and genetics of af-
fective disorders. J Affect Disord 1994;32:85-95.
5. Turecki G, Grof P, Cavazzoni P, et al. Evidence for a role of phos-
pholipase C-gamma1 in the pathogenesis of bipolar disorder. Mol
6.Alda M, Grof P, Rouleau GA, et al. Investigating responders to
lithium prophylaxis as a strategy for mapping susceptibility genes
for bipolar disorder. Prog Neuropsychopharmacol Biol Psychiatry
7.Post RM. Role of BDNF in bipolar and unipolar disorder: clinical
and theoretical implications. J Psychiatr Res. In press.
8.Kanazawa T, Glatt SJ, Kia-Keating B, et al. Meta-analysis reveals
no association of the Val66Met polymorphism of brain-derived
neurotrophic factor with either schizophrenia or bipolar disorder.
Psychiatr Genet 2007;17:165-70.
9.Chen B, Dowlatshahi D, MacQueen G, et al. Increased hippocam-
pal BDNF immunoreactivity in subjects treated with antidepres-
sant medication. Biol Psychiatry 2001;50:260-5.
10.Karege F, Schwald M, Papadimitriou P, et al. The cAMP-
dependent protein kinase A and brain-derived neurotrophic factor
expression in lymphoblast cells of bipolar affective disorder.
J Affect Disord 2004;79:187-92.
11.Cunha AB, Frey BN, Andreazza AC, et al. Serum brain-derived
neurotrophic factor is decreased in bipolar disorder during de-
pressive and manic episodes. Neurosci Lett 2006;398:215-9.
12. Endicott J, Spitzer RL. A diagnostic interview: the schedule for af-
fective disorders and schizophrenia. Arch Gen Psychiatry 1978;35:
Tseng et al.
Rev Psychiatr Neurosci 2008;33(5)
Acknowledgements: This study was supported by grant #64410
from the Canadian Institutes of Health Research.
Competing interests: None declared.
Contributors: Drs. Alda, Wang, Turecki, Rouleau and Young de-
signed the study. Drs. Tseng, Alda, Xu, Sun, Grof and Rouleau col-
lected the data, which Drs. Alda, Xu and Young analyzed. Drs. Tseng
and Young wrote and reviewed the article, which Drs. Alda, Xu, Sun,
Wang, Grof, Turecki and Rouleau also reviewed. All authors gave fi-
nal approval for publication.
BDNF is decreased in bipolar disorder lymphoblasts Download full-text
J Psychiatry Neurosci 2008;33(5)
13. Spitzer RL, Endicott J, Robins E. Research diagnostic criteria:
rationale and reliability. Arch Gen Psychiatry 1978;35:773-82.
14. American Psychiatric Association. Diagnostic and statistical manual
of mental disorders. 4th ed. Washington: The Association; 1994.
15.Alda M, Keller D, Grof E, et al. Is lithium response related to
G(s)alpha levels in transformed lymphoblasts from subjects with
bipolar disorder? J Affect Disord 2001;65:117-22.
16. Machado-Vieira R, Dietrich MO, Leke R, et al. Decreased plasma
brain derived neurotrophic factor levels in unmedicated bipolar
patients during manic episode. Biol Psychiatry 2007;61:142-4.
17.Hasler G, Drevets WC, Gould TD, et al. Toward constructing an
endophenotype strategy for bipolar disorders. Biol Psychiatry 2006;
18. Einat H, et al. The role of the extracellular signal-regulated kinase
signaling pathway in mood modulation. J Neurosci 2003;23:7311-6.
19.Aydemir O, Deveci A, Taneli F. The effect of chronic antidepres-
sant treatment on serum brain-derived neurotrophic factor levels
in depressed patients: a preliminary study. Prog Neuropsycho-
pharmacol Biol Psychiatry 2005;29:261-5.
20. Gonul AS, Akdeniz F, Taneli F, et al. Effect of treatment on serum
brain-derived neurotrophic factor levels in depressed patients. Eur
Arch Psychiatry Clin Neurosci 2005;255:381-6.
21. Huang T, Lee C, Liu Y. Serum brain-derived neurotrophic factor
levels in patients with major depression: effects of antidepressants.
J Psychiatr Res 2008;42:521-5.
22. Piccinni A, Marazziti D, Catena M, et al. Plasma and serum brain-
derived neurotrophic factor (BDNF) in depressed patients during
1 year of antidepressant treatments. J Affect Disord 2008;105:279-83.
23. Young LT, Woods CM. Mood stabilizers have differential effects
on endogenous ADP ribosylation in C6 glioma cells. Eur J
24.Ozaki N, Chuang DM. Lithium increases transcription factor bind-
ing to AP-1 and cyclic AMP-responsive element in cultured neur-
ons and rat brain. J Neurochem 1997;69:2336-44.
25. Kato T, Shioiri T, Inubushi T, et al. Brain lithium concentrations
measured with lithium-7 magnetic resonance spectroscopy in pa-
tients with affective disorders: relationship to erythrocyte and
serum concentrations. Biol Psychiatry 1993;33:147-52.
26. Manji HK, Lenox RH. Long-term action of lithium: a role for tran-
scriptional and posttranscriptional factors regulated by protein
kinase C. Synapse 1994;16:11-28.
27. Karege F, Schwald M, Cisse M. Postnatal developmental profile
of brain-derived neurotrophic factor in rat brain and platelets.
Neurosci Lett 2002;328:261-4.
28.Lang UE, Hellweg R, Seifert F, et al. Correlation between serum
brain-derived neurotrophic factor level and an in vivo marker of
cortical integrity. Biol Psychiatry 2007;62:530-5.