Chen Yu YE，Ye Qing XU，Hua HU，Chun Bo LI，Zhi Guang LIN，Shen Xun SHI，Li Wei WANG*

Cross-sectional study of the association between serum concentration of brain-derived neurotrophic factor and bipolar disorder

Chen Yu YE1，2，Ye Qing XU3，Hua HU1，Chun Bo LI2，Zhi Guang LIN3，Shen Xun SHI1，Li Wei WANG1*

Background：Recent studies have reached different conclusions about the relationship of brain-derived neurotrophic factor(BDNF)and the symptoms of bipolar disorder.

Objective：Examine the correlation between the serum concentration of BDNF and bipolar disorder and consider whether or not family history of bipolar disorder mediates this relationship.

Methods：Serum BDNF concentrations were measured in 228 patients with bipolar disorder and 153 normal controls.Patients with bipolar disorder were assessed using the Young Mania Rating Scale(YMRS)and the Hamilton Depression Rating Scale(HAMD-17)：85 with YMRS≥20 were considered manic，14 with HAMD≥17 were considered depressed，and 129 with YMRS＜20 and HAMD＜17 were considered euthymic.

Results：The mean(SD)concentration of BDNF in bipolar patients was significantly lower than that in the control group［18.75(8.98)ng/ml vs.23.72(5.60)ng/ml;t=6.09，P＜0.001］，and this difference was present persisted in all three phases of bipolar illness(manic，depressive，euthymic).Moreover，serum BDNF was significantly higher in bipolar patients in the manic phase of illness than in those in the euthymic phase of illness.After controlling for all available variables in a multivariate regression model，the only factor significantly related to higher serum BDNF levels in bipolar patients was the YMRS score(standardized regression coefficien=0.17，P=0.011). Conclusion：The serum concentration of BDNF in bipolar patients is lower than that in normal controls and is positively related to the severity of manic symptoms.It is not related to family history of bipolar disorder.

Brain-derived neurotrophic factor;Bipolar disorder;Serum concentration

1 Introduction

First discovered and purified in 1982，brain-derived neurotrophic factor(BDNF)is the most abundant and widely distributed neurotrophin in the mammalian brain.It promotes the growth，development，differentiation，repair and regeneration of neurons;and it boosts signal conduction through the brain's synapses［1］.Studies in rats have found that the quantity of BDNF in the central nervous systemiscorrelatedwiththatintheblood stream［2，3］and studies in children report that changes in BDNF concentrations in the peripheral blood are consistent with changes in the cerebrospinal fluid［3］.Thus，it is reasonable to conclude that the serum concentration of BDNF is a proxy for brain levels of this neuropeptide.

Bipolar disorder is a serious familial disorder with a reported lifetime prevalence as high as 5.5%-7.8%［4］.The relationship between BDNF concentrations and the etiology and course of bipolar disorder has recently received increased attention.However，studies examining this relationship have conflicting results.Most studies find lower concentrations of BDNF in the peripheral blood of bipolar patients compared to that in normal controls regardless of the phase of the illness(manic，depressive or euthymic)［5］，but some studies report a negative correlation between the seriousness of manic symptoms and BDNF concentrations in peripheral blood［6］.These studies had relatively small samples—the largest previous study had a total sample of 85 individuals［7］—and the normal range for serum BDNF in humans is quite large(1-40ng/ ml)，so the contradictory results may be due tosampling biases or lack of power to identify significant differences.The current study assesses the relationship between BDNF and bipolar disorder in a larger sample of patients and considers the question of whether or not family history of bipolar disorder mediates the relationship.

2 Methods

2.1 Subjects

The bipolar patients included in this study were a convenience sample composed of inpatients and outpatients who attended the Shanghai Mental Health Center or the Pudong Mental Health Center from January to December 2008.Included subjects were individuals 16-65 years of age who met the diagnostic criteria for Bipolar Disorder as described in the Diagnostic and Statistical Manual of Mental Disorders，4th edition(DSM-IV)［8］.Exclusion criteria included：1)serious illness involving the heart，lungs，liver，kidney，or neurological system;2)substance abuse;3)prior self-harm，significant suicidal ideation or suicide attempt;and 4)currently pregnant or breast-feeding.A healthy volunteer sample was identified among individuals who received routine physical examinations at Fudan University School of Medicine.Exclusion criteria for the control sample included：1)any past or current diagnosis of a mental disorder meeting diagnostic criteria described in DSM-IV;2)current mental health dysfunction or substance abuse(judged in a brief clinical assessment by a psychiatrist);3)serious illness involving the heart，lungs，liver，kidney，or neurological system，and 4)currently pregnant or breast feeding.As shown in Figure 1，a total of 228 bipolar patients and 153 normal controls were included in the study.All patients and control subjects provided written informed consent.

2.2 Measurement of serum BDNF

A coagulation tube was used to draw 5 ml blood from the antecubital vein of all enrolled patients and control subjects.A refrigerated centrifuge (large-capacity refrigerated centrifuge DLL-7.2 purchased from Taicang Huamei Biochemistry Instruments Factory)was used for the centrifugation of the blood samples at 3200 rpm for 20 minutes within 10 minutes after drawing.Then 1 ml of serum was removed，placed into 3 tubes，and stored in a-80℃refrigerator for subsequent determination of BDNF.The serum was defrosted and all the reagents were adjusted to 25℃before application.Enzyme-linked immunosorbent assay(ELISA)(Human BDNF ELISA kits purchased from the Transhold Corporation)was used to determine the serum concentration of BDNF using the Multiskan MK3microplatereader(ThermoCorporation，USA).Each sample was kept in a pair of tubes and the mean value of the measurements was used as the final concentration of BDNF.In 31 cases the results of the samples in the two tubes differed greatly，indicating that the results may be invalid，so a second analysis was conducted and the mean of this second set of values was used.

2.3 Evaluation of bipolar patients

After recruitment all identified patients were re-evaluated by a research psychiatrist using the Chinese version of the SCID［9］to confirm the diag-nosis and to collect additional demographic，clinical and family history data.The Chinese version of the Hamilton Depression Rating Scale(HAMD-17)［10］was used to assess depressive symptom.Patients with a HAMD score＞17 were classified as currently experiencing a depressive episode.The Chinese version of this instrument has acceptable reliability and validity［10］.The Chinese version of the Young Mania Rating Scale［11］(YMRS)was used to assess manic symptoms.Patients with YMRS score＞20 were classified as currently experiencing a manic episode.The internal consistency of the items on the Chinese version of the YMRS is excellent(alpha=0.968);but there is no data available on its test-retest and inter-rater reliability.Those patients with YMRS score≥20 and HAMD score≥17 were classified as experiencing a mixed bipolar episode;those with YMRS＜20 and HAMD＜17 were classified as currently euthymic.In this study patients with‘early onset'bipolar disorder were those who had an age of onset under 18 years of age［12］.

2.4 Statistical analysis

SPSS 16.0 software was employed for statistical analysis.For continuous normally distributed variables，t-tests were used.For multiple comparisons between subgroups，F(xiàn)-test was used to compare the groups and，if significantly different，leastsquares difference(LSD)post-hoc tests were used to make subsequent pairwise comparisons.Age，duration of illness and the HDRS and YMRS were divided into equal-sized tertiles for the univariate analysis.Multivariate linear regression models were used to determine the relative importance of various factors on the BDNF serum concentration.All the statistical tests were two-tailed tests and statistical significance was defined at the 0.05 level.

3 Results

3.1 Subjects

The characteristics of the 228 enrolled bipolar patients were as follows：107(46.9%)were male; their mean(SD)age was 38.3(14.0)years(range，16-65);129(56.6%)were outpatients and 99 (43.4%)were inpatients;200(87.7%)had bipolar I disorder and 28(12.3%)had bipolar II disorder;the mean age of onset was 26.8(10.8)years (range，7-62);49(21.5%)met criteria for early onset bipolar disorder;median(IQR)duration of illness was 6.9(0.9-17)years(range，2 months-41.7 years);85(37.3%)were currently in a manic episode，14(6.1%)were in a depressive episode，none were experiencing a mixed episode，and 129(56.6%)were euthymic;and 60(26.3%) had a positive family history of bipolar disorder.All of the patients were taking psychiatric medication at the time of the evaluation：205(93.4%)were taking mood stabilizers，170(74.6%)were taking antipsychotic medications，and 27(11.8%)were taking antidepressants.

The 153 healthy control subjects included 71 (46.4%)males and their mean(SD)age was 38.2(13.8)years(range 16-65).There were no significant differences in the sex ratio(χ2=0.01，

P=0.920)or age(t=0.07，P=0.947)between the two groups.

3.2 BDNF serum concentrations

The BDNF serum concentrations of the different groups of subjects is shown in Table 1.The mean value in the bipolar group was significantly lower than in the control group(t=6.09，P＜0.001).When comparing three subtypes of bipolar disorder(currently manic，depressive or euthymic) with the control group there was an overall statistical difference between the groups(F=15.45，P＜0.001)and post-hoc tests showed that BDNF in each of the three bipolar groups was significantly lower than in the control group.The post-hoc analysis also showed that manic phase bipolar subjects have significantly higher BDNF levels than euthymic phase bipolar patients.The absolute difference between manic and depressive phase bipolar patients was slightly larger than that between manic and euthymic phase patients but the difference was not statistically significant，probably due to the much smaller number of subjects in the depressive phase group(n=14).The BDNF serum levels in euthymic phase and depressive phase bipolar disorder patients were essentially identical.

3.3 Relationship of BDNF serum concentration to characteristics of bipolar patients

Univariate analysis of the relationship of demographic and clinical variables in the bipolar group to BDNF serum concentrations are shown in Table 2.The only variable that was significantly related to BDNF serum levels in the univariate analysis was the HDRS score;the BDNF level in individuals with mild depressive symptoms was higher than in those with no depressive symptoms or with severe depressive symptoms(i.e.，the relationship was not linear).The relationship of the severity of depression to BDNF levels was different in the three clinical states：in the manic phase and the euthymic phase it was significantly positively correlated(rs=0.23，p=0.037;rs=0.21，p=0.023，respectively)but in the depressive group it was negatively correlated(rs=-0.24，p=0.290)but the number of subjects in the depressed group was too small for this fairly strong correlation to reach statistical significance.

Females had higher levels of serum BDNF than males and the BDNF level showed a stepwise increase in the three tertiles of YMRS scores，but these were only significant at trend level(i.e.，0.05＜p＜0.10).The BDNF level was not related to age，age of onset，duration of illness，family history，or type of bipolar disorder(i.e.，bipolar I or bipolar II).

Multivariate linear regression analysis was conducted to identify factors that have an independent association with BDNF in the 228 bipolar subjects. As shown in Table 3，after adjustment for the various factors in the model，serum BDNF concentration was significantly positively correlated with the score on the YMRS(i.e.，higher scores were associated with higher concentrations of BDNF)and there was a trend(p=0.051)for females to have higher BDNF levels，but none of the other factors in the model were significantly related to BDNF levels.The coefficient of determination for the full model was 0.059 and the adjusted coefficient of determination was 0.029.

4 Discussion

4.1 Main findings

Someresearchershypothesizethatmanic symptoms are the result of increased BDNF activity［13］，while others do not believe BDNF plays a causal role［14，15］.Knockout studies that alter the BDNF gene in rats to decrease BDNF production do not show any of the anticipated depressive symptoms［16］.And Eisch's group［17］found that injecting BDNF into the hippocampus of rats had an antidepressant effect，while injection of BDNF into the nucleus accumbans/ventral tegmental area resulted in increased depressive behavior.Integrating these findings，Groves and colleagues［18］conclude that BDNF acts differently on neuroplasticity in different brain regions，and that bipolar symptoms might be the final manifestation of alterations in BDNF in different brain regions.

To help clarify some of the discrepant results from small studies about the association of BDNF and bipolar disorder the current study considered this relationship in a large unselected sample of outpatients and inpatients with bipolar illness.We found that the mean serum BDNF concentrations in each phase of bipolar disorder(manic，depressive，and euthymic)were all lower than that in a healthy normal control group，a finding that is consistent with thefindingsofMachado-Vieraandcolleagues［6］.However，unlike previous studies［5］，we also found that the subgroup experiencing mania had a significantly higher BDNF score than the euthymic subgroup of bipolar patients and that more severe manic symptoms was associated with higher levels of BDNF.It is possible that increased BDNF is a marker for increased disease activity，but it is also possible that current or prior treatment with antipsychotic medication—a factor that we were unable to assess in detail in our study—induces higher serum BDNF in bipolar manic patients.A study by Gama and colleagues［19］provides evidence to support the second hypothesis：they found that BDNF plasma concentrations in patients with schizophrenia receiving long-term treatment with antipsychotic medications was significantly higher than in normal controls and than in bipolar patients during the euthymic phase of illness.

We also found an interesting relationship between the severity of depressive symptoms and BDNF serum levels in bipolar patients，a relationship that appeared different in bipolar subjects who we depressed than in bipolar subjects who were euthymic or manic.We have not found reports of this relationship in the literature so it may be an artifact，but it certainly deserves further investigation.

4.2 Limitations

There are several issues that need to be considered when interpreting these results.1)This is a cross-sectional study so we were unable to assess the causal direction of the associations identified.It is equally possible that low BDNF influences the onset of bipolar disorder or is an epiphenomenon of bipolar disorder.2)We were unable to obtain detailed treatment histories on these subjects so medication status is one important potential confounder that we were unable to adjust for.3)The small number of patients in the bipolar depression group(n=14)，which was partly due to their higher rate of refusing to participate［40 of the 54 bipolar depressive patients refused while none of the bipolar manic patients refused］，made it difficult to separately analyze this group because of potential Type II errors.Epidemiological studies estimate thatat any point in time about one-third of bipolar patients should be in the depressive phase of illness［20］so our enrolled sample was not proportionally representative of the different phases of illness.

4.3 Implications

In the relatively large sample employed in this study it was possible to confirm previous findings about lower BDNF serum levels in bipolar patients versus normal controls and to demonstrate that this difference is present during all three phases of bipolar illness(manic，depressive and euthymic).Unlike previous studies we found that BDNF levels were higher in the manic phase of bipolar disorder than in the euthymic phase of illness and that after adjustment for all available potential confounders the more severe the manic symptoms(as assessed by YMRS)the higher the level of serum BDNF.The unexpected finding of a non-linear relationship of BDNF levels with the severity of depressive symptoms in these bipolar patients(as assessed by HDRS)deserves further investigation.Another topic worthy of further assessment is the change in BDNF as bipolar patients convert from one clinical state to another(i.e.，manic，euthymic，depressed).The causal direction of these associations remains unclear so further longitudinal studies that include detailed information on medication usage and larger numbers of subjects with bipolar depression will be needed to clarify these relationships and to determine their relevance to the etiology，course and treatment of this disabling condition.

Funding

The study was sponsored by a grant from The Stanley Medical Research Institute(O2I-008).

1. Lindsay RM，Wiegand SJ，Altar CA，DiStefano PS.Neurotrophic factors：from molecule to man.Trends Neurosci，1994，17(5)：182-190.

2. Karege F，Bondolfi G，Gervasoni N，Schwald M，Aubry JM，Bertschy G.Low brainderived neurotrophic factor(BDNF) levels in serum of depressed patients probably results from lowered platelet BDNF release unrelated to platelet reactivity. Bio Psychia，2005，57(9)：1068-1072.

3. PanW，BanksWA，F(xiàn)asoldMB，Bluth J，KastinAJ.Transport of brain-derivedneurotrophic factor across the blood-brain barrier.Neuropharmacol，1998，37(12)：1553-1561.

4. Weissman MM，Bland RC，Canino GJ，F(xiàn)aravalli C，Greemwald S，Hwu HG，et al.Gross national epidemiology of major depression and bipolar disorder.JAMA，1996，276(4)：293-299. 5. Martiadis V，Monteleone P，Iodice V，Maj M.Peripheral brain derived neurotrophic factor(BDNF)in patients with unipolar depression or with bipolar I and II disorders.Eur Psychiatry，2007，22(Suppl 1)：255-257.

6. Machado-Vieira R，Dietrich MO，Leke RF，Cereser VH，Zanatto V，Kapczinski F，et al.Decreased plasma rain derived neurotrophic factor Levels in unmedicated bipolar patients during manic episode.Biol Psychiatry，2007，61(2)：142-144.

7. Monteleone P，Serritella C，Martiadis V，Maj M.Decreased levels of serum brain-derived neurotrophic factor in both depressed and euthymic patients with unipolar depression and in euthymic patients with bipolar I and II disorders.Bipolar Disord，2008，10(1)：95-100.

8. American Psychiatric Association.Diagnostic and Statistical Manual of Mental Disorders，F(xiàn)ourth Edition，Text Revision. Washington，DC：American Psychiatric Association，2000.

9. First MB，Gibbon M，Spitzer R，Williams JBW.Structured Clinical Interview for DSM-IV-TR Axis I Disorders-Research Version(SCID-I for DSM-IV-TR，F(xiàn)ebruary 2001 Revision).New York：Biometrics Research Department，New York State Psychiatric Institute.

10. Zhang MY.Rating Scales in Psychiatry，2nd Ed.Changsha：Hunan Science and Technology Press，1998：121-126.(in Chinese)

11. Young RC，Biggs JT，Ziegler VE，Meyer DA.A rating scale for mania：reliability，validity，and sensitivity.Br J Psychiatry，1978，133：429-435.

12. Perlis RH，Miyahara S，Marangell LB，Wisniewski SR，Ostacher M，DelBello MP et al.Long-term implications of early onset in bipolar disorder：data from the first 1000 participants in the systematic treatment enhancement program for bipolar disorder(STEP-BD).Biol Psychiatry，2004，55(9)：875-881.

13. Tsai SJ.Is mania caused by overactivity of central brain-derived neurotrophic factor？Med Hypotheses，2004，62(1)：19-22.

14. Russo-Neustadt A，Beard RC，Cotman CW.Exercise，antidepressant medications，and enhanced brain derived neurotrophic factor expression.Neuropsychopharmacology，1999，21 (5)：679-682.

15. Terao T.Mania is probably associated with hypoactivity of central brain-derived neurotrophic factor.Med Hypotheses，2008，70(6)：1232.

16. Castrén E，V?ikar V，Rantam?ki T.Role of neurotrophic factors in depression.Curr Opin Pharmacol，2007，7(1)：18-21.

17. Eisch AJ，Bola?os CA，de Wit J，Simonak RD，Pudiak CM，Barrot M，et al.Brain-derived neurotrophic factor in the ventral midbrain-nucleus accumbens pathway：a role in depression. Biol Psychiatry，2003，54(10)：994-1005.

18. Groves JO.Is it time to reassess the BDNF hypothesis of depression？Mol Psychiatry，2007，12(12)：1079-1088.

19. Gama CS，Andreazza AC，Kunz M，Berk M，Belmonte-de-Abreu PS，Kapczinski F.Serum levels of brain-derived neurotrophic factor in patients with schizophrenia and bipolar disorder. Neurosci Lett，2007，420(1)：45-48.

20. Kessler RC，McGonagle KA，Zhao S，Nelson CB，Hughes M，Eshleman S，et al.Life-time and 12-month prevalence of DSMIII-R psychiatry disorders in the United States.Results from the National Comorbidity Survey.Arch Gen Psychiatry，1994，51 (1)：8-19.

腦源性神經(jīng)營養(yǎng)因子血清濃度與雙相障礙關系橫斷面研究

葉塵宇1，2許燁勍3胡 華1李春波3林治光3施慎遜1王立偉1

項目基金：美國stanley醫(yī)學研究院基金（O2I-008)

作者單位：1復旦大學附屬華山醫(yī)院精神醫(yī)學科200040；2復旦大學附屬中山醫(yī)院醫(yī)學心理科200032；3上海交通大學醫(yī)學院附屬精神衛(wèi)生中心200030。通信作者：王立偉，電子信箱lwwang163@163.com

背景近年來發(fā)現(xiàn)，腦源性神經(jīng)營養(yǎng)因子（brain-derived neurotrophic factor，BDNF）的血清濃度與雙相障礙癥狀關系的研究結果不一致。

目的檢驗BDNF血清濃度與雙相障礙的關系，并討論雙相障礙家族史在兩者關系中的作用。

方法檢測了228例雙相障礙患者和153名健康對照者的BDNF血清濃度，采用楊氏躁狂量表和漢密爾頓抑郁量表（17項）評估患者的躁狂或抑郁癥狀，將楊氏躁狂量表評分≥20分定義為躁狂發(fā)作，共計85例；將漢密爾頓抑郁量表評分≥17分定義為抑郁發(fā)作，共計14例；將楊氏躁狂量表評分＜20分并且漢密爾頓抑郁量表評分＜17分定義為緩解期，共計129例。

結果 患者組平均（標準差）BDNF血清濃度低于健康對照組［18.75（8.98）ng/ml比23.72（5.60）ng/ml，t=6.09，P＜0.001］，且各個亞組（躁狂組、抑郁組和緩解期組）與健康對照組BDNF血清濃度的差異均有統(tǒng)計學意義。躁狂發(fā)作期與緩解期之間的BDNF血清濃度差異有統(tǒng)計學意義，其余各亞組間的差異無統(tǒng)計學意義。在多元線性回歸模型中控制各個因素后，發(fā)現(xiàn)僅有楊氏躁狂量表評分與BDNF血清濃度呈正相關（標準回歸系數(shù)=0.17，P=0.011）。

結論雙相障礙患者的BDNF血清濃度低于健康對照組，BDNF血清濃度與躁狂癥狀存在正相關，與是否存在家族史并不相關。

腦源性神經(jīng)營養(yǎng)因子 雙相障礙 血清濃度

date：2010-10-27;accepted date：2011-03-09)

10.3969/j.issn.1002-0829.2011.04.003

1Department of Psychiatry，Huashan Hospital，F(xiàn)udan University，Shanghai，200040;

2Department of Clinical Psychology，Zhongshan Hospital，F(xiàn)udan University，Shanghai，200032;

3Shanghai Mental Health Center，Shanghai Jiao Tong University School of Medicine，Shanghai，200030，China

*Correspondence：lwwang163@163.com