Association of BDNF Gene Polymorphism with Bipolar Disorders in Han Chinese Population
Y. Fang, MD, PhD, Division of Mood Disorders, Department of Psychiatry, Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine, 600 South Wan Ping Road, Shanghai 200030, China. E-mail: email@example.com
Recent data suggest that brain-derived neurotrophic factor (BDNF) plays an essential role in neuronal plasticity and etiology of bipolar disorders (BPD). However, results from different studies have been inconsistent. In present study, 342 patients who met DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, 4th Edition) criteria for bipolar disorders type I (BPD-I) or type II (BPD-II) and 386 matched health controls were enrolled, and TaqMan® SNP Genotyping Assays (Applied Biosystems, Foster City, CA, USA) were applied to detect the functional polymorphism rs6265 (Val66Met) of BDNF gene. Treatment response to lithium and valproate was retrospectively determined. The association between Val66Met polymorphism and BPD, treatment response to mood stabilizers, was estimated. The genotype and allele distribution of Val66Met polymorphism between BPD patients and control subjects showed significant difference (genotype: χ2 = 6.18, df = 2, P = 0.046; allele: χ2 = 5.01, df = 1, P = 0.025) with Met allele as risk factor for disease susceptibility (OR = 0.79, 95%CI as 0.64–0.97). The post hoc analysis interestingly showed that Met allele had opposite effect on the treatment response for BPD-I and BPD-II separately. For BPD-I patients, the response score in Val/Val group was significantly lower than that in Met allele carriers (t = −2.27, df = 144, P = 0.025); for BPD-II patients, the response score in Val/Val group was significantly higher than that in Met allele carriers (t = 2.33, df = 26, P = 0.028). Although these results should be interpreted with caution because of the limited sample for Val/Val genotype in BPD-II patients (N = 5), these findings strengthen the hypothesis that BDNF pathway gets involved in the etiology and pharmacology of BPD and suggest the differences between BPD-I and BPD-II.
Recently, there are increasing evidences that supported the relationship between bipolar disorders (BPD) and neural plasticity (Einat & Manji 2006; Zarate et al. 2006). The brain-derived neurotrophic factor (BDNF) is widely expressed in the adult brain and plays a critical role in promoting and modifying growth, differentiation and survival of neurons (Xu et al. 2010). The effect of BDNF on neuronal plasticity is essential for memory and learning (Kaplan & Miller 2000; Minichiello et al. 1999), which has motivated investigators to study the role of BDNF in the etiology of BPD.
Among studied polymorphisms of BDNF, the functional polymorphism rs6265 (Val66Met) was mostly supported to be related to BPD susceptibility and lithium response. However, the relationship between rs6265 and BPD is still unclear (Fan & Sklar 2008; Kanazawa et al. 2007), including in the Han Chinese population (Hong et al. 2003; Liu et al. 2007; Tang et al. 2008; Xu et al. 2010; Ye et al. 2009). Recently, a cohort study of multiplex bipolar families added to the body of evidence associating Val66Met polymorphism with BPD (Sears et al. 2011). Similarly, an association of Val66Met polymorphism with a degree of prophylactic lithium response was found in a Poland population (Dmitrzak-Weglarz et al. 2008; Rybakowski et al. 2005, 2007), which was not replicated in Japanese patients with BPD (Masui et al. 2006).
Therefore, the aim of the present study is to investigate potential association between BDNF gene Val66Met functional polymorphism (rs6265) and susceptibility to BPD, treatment response to mood stabilizers in patients with BPD in a Han Chinese population.
Materials and methods
The Ethical Committee of Shanghai Mental Health Center reviewed and approved the study protocol. Written informed consent was obtained from each subject before any study-related procedures were performed. Males and females from 16 to 65 years of age, who met DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, 4th Edition) criteria for bipolar disorders type I or type II (BPD-I or BPD-II) (N = 342; BPD-I, 286 and BPD-II, 56; 152 male and 190 female; mean age 36.9 ± 14.2 years; mean age of first episode 26.0 ± 9.8 years), and unrelated, age- and gender-matched healthy controls (N = 386; 174 male and 212 female; mean age = 35.2 years; SD ±10.6), were included for the study. All subjects were Han Chinese in origin.
BPD-I or BPD-II patients were recruited from consecutive admissions to the Division of Mood Disorders at Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine from November 2006 to October 2010. Clinical diagnosis was made by two psychiatrists with consistency (one attending psychiatrist, the other chief psychiatrist independently interviewed again. Participants were only recruited of those patients with the same diagnosis in axis I disorders), and then confirmed and subtype-classified with Extensive Clinical Interview and Structured Clinical Interview for DSM-IV Axis I Disorders, Patient Version (SCID-P) (First et al. 1997) by a research psychiatrist. The Extensive Clinical Interview contains items to assess demographics, mental status, suicidal severity and other variables of interest. Controls were recruited from students and staff members at Shanghai Mental Health Center and interviewed by a specialized psychiatrist with SCID-P. Subjects with any psychiatric disorder and chronic physical disease were excluded.
Treatment response assessment
All patients who had been treated first with lithium or valproate were reviewed by one trained research assistant using a scale described previously (Grof et al. 2002). The scale was developed specifically for retrospective evaluation of long-term treatment response in research subjects with BPD. The scale quantifies the degree of improvement in the course of treatment (subscale A) expressed as a composite measure of change in frequency and severity of mood symptoms from ‘no change or worsening’ (score = 0) to ‘complete response, no recurrences under adequate treatment, no residual symptoms and full functional recovery’ (score = 10). The A score is weighted by five factors (subscale B), including B1 ‘number of episodes before the treatment', B2 ‘frequency of episodes before the treatment', B3 ‘duration of the treatment', B4 ‘compliance during period(s) of stability’ and B5 ‘use of additional medication during the period of stability’, score = 0–2 for each factor, which assist in determining the probability that the observed improvement is a result of the treatment rather than a spontaneous improvement or an effect of additional medication. The total score is obtained by subtracting the B score from the A score. An A score of 7 or higher was defined as full responder. An A score of 5 or higher was considered as responder, which corresponded approximately to 50% or better improvement in the course of treatment (Garnham et al. 2007). In this study, the score of subscale A was used to assess retrospectively the treatment response to lithium or valproate in patients with BPD.
Genomic DNA was extracted from peripheral blood according to standard laboratory procedures (Blood genomic DNA extraction kit, TIANGEN, Beijing, China). Genotyping was performed with TaqMan genotyping assay on an ABI PRISM 7900 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Primers and probes were purchased from Applied Biosysems. All genotypes were independently confirmed by two individuals; any discrepancies or ambiguous genotypes were resolved by repeated genotyping. Ten percent of samples were randomly chosen to genotype in duplicate, and no genotype error was found.
Demographic and clinical characteristics between patients with BPD and controls were compared with chi-square (χ2) test and t-test. The χ2 test for goodness of fit was used to check for Hardy–Weinberg equilibrium in genotype distributions in patients and controls. For the case–control genetic comparisons, differences in the genotype and allele distributions between patients and controls were examined using Pearson's χ2 test. The response scores among different genotypes were compared with analysis of variance (anova). The t-test was used for post hoc analysis to compare response scores between two groups. All statistical analyses were carried out by using the SPSS V. 17.0 software program (SPSS, Inc., Chicago, IL, USA). Criterion for statistical significance was set at α = 0.05 and two tailed.
Association analysis of BDNF polymorphisms with BPD
Comparisons of genotype and allele frequencies for Val66Met polymorphism between BPD patients and control subjects are presented in Table 1. The genotype distributions did not deviate significantly from the Hardy–Weinberg equilibrium in either patient or control group (BPD: χ2 = 0.86, df = 1, P > 0.05; BPD-I: χ2 = 0.44, df = 1, P > 0.05; BPD-II: χ2 = 0.68, df = 1, P > 0.05; control: χ2 = 0.36, df = 1, P > 0.05).
Table 1. Comparison of genotypes and alleles in BDNF polymorphisms between BPD patients and controls
|BPD-II||56||11(19.6%)||31(55.4%)||14(25.0%)||3.94||0.140||53(47.3%)||59(52.7%)||3.32||0.069||0.69 (0.47, 1.03)|
|Control||386||126(32.6%)||184(47.7%)||76(19.7%)|| || ||436(56.5%)||336(43.5%)|| || || |
The genotype and allele distribution of Val66Met polymorphism between BPD patients and control subjects showed significant difference (genotype: χ2 = 6.18, df = 2, P = 0.046; allele: χ2 = 5.01, df = 1, P = 0.025) with Met allele as risk factor for disease susceptibility (OR = 0.79, 95%CI as 0.64–0.97). When patients were divided into BPD-I and BPD-II, the genotype distribution between patients and controls did not show significant difference (BPD-I: χ2 = 4.34, df = 2, P = 0.114; BPD-II: χ2 = 3.94, df = 2, P = 0.140); however, a trend could be observed for allele (BPD-I: χ2 = 3.59, df = 1, P = 0.058; BPD-II: χ2 = 3.32, df = 1, P = 0.069).
Association analysis of BDNF polymorphisms with treatment response
Data were shown in Table 2. For BPD patients, the mean scores of treatment response were not different among three genotypes of Val66Met polymorphism (F = 1.43, df = 173, P = 0.242). Similarly, no significant difference was found when patients were divided into manic episode (F = 1.70, df = 105, P = 0.187) and depressive episode (F = 0.69, df = 67, P = 0.508). However, when patients were divided into BPD-I and BPD-II, the response scores showed significant difference (BPD-I: F = 3.98, df = 145, P = 0.021; BPD-II: F = 4.24, df = 27, P = 0.026). For BPD-I patients, the response score in Val/Val group was significantly lower than that in Val/Met group (t = −2.65, df = 118, P = 0.009); for BPD-II patients, the response score in Val/Val group was significantly higher than that in Val/Met group (t = 2.86, df = 15, P = 0.012).
Table 2. Comparison of response scores among different genotypes of rs6265 in BPD patients
When grouping Met allele carriers (including Val/Met and Met/Met genotypes) and comparing them with Val/Val homozygous, the post hoc analyses showed that there were no significant difference for response scores in BPD patients and patients with index episode of mania or depression (BPD: t = −1.30, df = 172, P = 0.196; manic episode: t = −0.80, df = 104, P = 0.428; depressive episode: t = −1.11, df = 66, P = 0.273). For BPD-I patients, the response score in Val/Val group was significantly lower than that in Met allele carriers (t = −2.27, df = 144, P = 0.025); for BPD-II patients, the response score in Val/Val group was significantly higher than that in Met allele carriers (t = 2.33, df = 26, P = 0.028).
The findings of present study showed that the functional polymorphism Val66Met of BDNF gene was related to BPD susceptibility. Moreover, this is the first study to explore the relationship between BPD treatment response and BDNF gene polymorphism in a Han Chinese population. Our results showed that the Val66Met polymorphism was related to treatment response to mood stabilizers in both BPD-I and BPD-II patients.
Interestingly, the finding of the association between the Val66Met polymorphism and BPD in this study was consistent with the results of a recent study in another Han Chinese population, in which Met allele was associated with increased risk for BPD (Xu et al. 2010). On the other hand, other four studies in Han Chinese population did not show an association between Val66Met polymorphism and the risk for BPD (Hong et al. 2003; Liu et al. 2007; Tang et al. 2008; Ye et al. 2009). In contrast, other studies in non-Chinese population showed that the Val allele was associated with increased risk for BPD (Lohoff et al. 2005; Neves-Pereira et al. 2002; Sklar et al. 2002; Vincze et al. 2008). These inconsistent results may be related to variation in ascertainment, phenotype definition and control selection, limited power (due to their small sample size) and possible confounding by population substructure (Fan & Sklar 2008; Petryshen et al. 2010).
In addition, when patients were divided into BPD-I and BPD-II, the differences in polymorphism between patients and controls were weak, even though a significant trend for allele, which could be interpreted with the smaller sample size after division. The inconsistent of genotypes and alleles distribution between BPD-I and BPD-II suggested that they could have different biological mechanisms. The similar proportion of recruited BPD-I and BPD-II patients could interpret the consistent findings in two studies (our study and Xu et al. 2010). Moreover, the interplay between BDNF genotype and environment (stressful life events and enhanced stress response) in the etiology of BPD also could not be ignored (Hosang et al. 2010; Vinberg et al. 2009). Thus the multiplex bipolar families could be selected for further study excluding environmental factors (Sears et al. 2011).
Recently, a number of candidate genes for lithium prophylactic efficacy have been proposed, some of them being also associated with a predisposition to bipolar illness, i.e. BDNF, DRD1, COMT and CREB genes (Dmitrzak-Weglarz et al. 2008; Lee & Kim 2010; Mamdani et al. 2008; Rybakowski et al. 2009). In addition, Rybakowski et al. (2007) reported the interaction between 5-HTTLPR and BDNF polymorphism and response to lithium. However, those previous studies were mainly focused on BPD-I or did not specify the diagnostic types of BPD. In our study, the post hoc analysis interestingly showed that Met allele had opposite effect on the treatment response for BPD-I and BPD-II separately. The BPD-I patients with Met allele (Val/Met and Met/Met genotypes) had better response to lithium/valproate, which was consistent with previous study in a Poland population (Dmitrzak-Weglarz et al. 2008). On the contrary, the BPD-II patients with Met allele had poorer response to mood stabilizers. These findings suggested that BPD-I and BPD-II could have different pharmacological mechanisms.
Evidence exists for functional effects of the BDNF Val66Met polymorphism. In both healthy individuals (Egan et al. 2003; Hariri et al. 2003; Pezawas et al. 2004) and patients with BPD (Chepenik et al. 2009; Frey et al. 2007; Matsuo et al. 2009), this functional polymorphism Val66Met can affect human memory, i.e. lower activity of BDNF and poorer performance of memory with Met allele, and the anatomy and function of the hippocampus and prefrontal cortex by influencing intracellular trafficking and activity-dependent secretion of BDNF. Moreover, the genetic effect BDNF on variation of human hippocampal volume may be greater among patients with psychiatric disorders compared to healthy volunteers (Szeszko et al. 2005). These observations imply that the deleterious Met allele increases risk for BPD, which was supported by our study. However, until now, it is not possible to determine if this variant [or some variant in linkage disequilibrium (LD) with it] is causally associated with increased risk of BPD.
The limitations encountered in this study were those that are common to most human genetic research, which requires to account for various demographic and clinical variables that may influence genetic heterogeneity. There were also other several limitations in this study. First, the sample size of this study was modest. The sample powers to detect difference between genotypes based on treatment response were 0.58 for BPD-I and 0.99 for BPD-II. On the basis of this difference, a sample of at least 86 for BPD-I or 9 for BPD-II per arm is needed to detect a significant difference at α = 0.05 and 80% power. Clearly, it is feasible to conduct a large sample study to support or refute our findings. Second, this study retrospectively, not prospectively, assessed the treatment response to mood stabilizers. Third, we only reviewed those patients (174/342) for treatment response who had been treated first with lithium or valproate. The relationship between BDNF gene polymorphisms and antipsychotics response is needed to be further investigated. Moreover, we did not compare the difference between BDNF gene polymorphisms and individual mood stabilizer response such as lithium versus valproate.
Together, these initial findings strengthen the hypothesis that BDNF gene Val66Met variation is a risk factor for BPD susceptibility and treatment response to mood stabilizers. Although these results should be interpreted with caution because of the limited sample for Val/Val genotype in BPD-II patients (N = 5), the role of BDNF pathway in the etiology of different BPD subtypes is deserved to further studies. Further genetic studies from large-scale population or multiplex bipolar families are necessary to elucidate the relevance of BDNF gene variation(s) as risk factor for BPD susceptibility and clinical efficacy of mood stabilizer treatment. The pharmacogenetics of lithium response in bipolar disorder remains a field in its infancy. There is a need for large-scale, prospective studies of biologically plausible candidate gene genes. The advent of genome-wide association studies holds particular promise for future studies.
Funding for this study was provided by the National Natural Science Foundation of China (Grant# 30971047), Shanghai Jiaotong University School of Medicine Training Program for Excellent Doctoral Dissertation (Grant# YBPY2010010), Shanghai Key Medicine Specialties Program (Grant# Shanghai Health 12), Key Medicine Specialties Program of Hongkou District (Grant# Hongkou Health 91) and Hongkou District Medicine Study Program (Grant# Hongkou Health 1101-02). This study was partly supported by the National Key Clinical Disciplines at Shanghai Mental Health Center (OMA-MH, 2011-873), National High-tech R&D Program (863 Program) (2006AA02Z430, Ministry of Science and Technology of China) and the Key Disciplines Program at Shanghai Jiaotong University School of Medicine (Clinical Psychiatry, SJTUSM2008-6). The authors thank Keming Gao, Jinbo Fan and Joseph R. Calabrese from Mood Disorders Program, Case Western Reserve University School of Medicine for revising the manuscript. They also thank Martin Alda from Dalhousie University for providing an authorization of the scale ‘Retrospective Criteria of Long-Term Treatment Response in Research Subjects with Bipolar Disorders'.