Positive association between the brain-derived neurotrophic factor (BDNF) gene and bipolar disorder in the Han Chinese population^†

INTRODUCTION

Bipolar disorder (BPD) is a common debilitating mental illness characterized by recurrent depressive and manic episodes, which imposes significant social, familial, and individual burdens. Genetic factors contribute to the etiology of this disorder as evidenced by twin, adoption, and family studies [Craddock and Jones, 1999], but specific genes that contribute to the illness remain to be clearly identified, which suggests that BPD is a complicated syndrome and that many genes may be involved [Hyman, 1999]. The brain-derived neurotrophic factor (BDNF), the gene encoded on human chromosome 11p13, is a member of the superfamily of the neurotrophin, which plays a critical role in promoting and modifying growth, differentiation, and survival of neurons in the central nervous system (CNS). As the most abundant of the neurotrophins in the brain, BDNF is important for guiding the neurons of CNS during their development and maintaining their survival in adulthood [Thoenen, 1995]. The function and distribution of BDNF in the CNS raise the possibility that this neurotrophin is relevant to BPD. A number of animal and clinical studies have reported the potential contribution of BDNF to the pathophysiology of the disorder and have shown that BDNF protein is lower in patients than in control subjects [Knable et al., 2004]. It is also reported that the plasma levels of BDNF are temporarily lower in BPD patients during manic episodes [Machado-Vieira et al., 2007]. When pharmacological antidepressants have been administered to patients, serum BDNF levels have increased [Shimizu et al., 2003; Aydemir et al., 2005].

Several North American and European studies have reported a significant association between Val66Met and BPD [Neves-Pereira et al., 2002; Sklar et al., 2002; Geller et al., 2004; Lohoff et al., 2005]. All have shown over-transmission of the common valine (Val) allele. However, other studies have produced negative results. Four of these involve Asian populations [Hong et al., 2003b; Nakata et al., 2003; Kunugi et al., 2004; Tang et al., 2008] and others are European based [Oswald et al., 2004; Skibinska et al., 2004; Strauss et al., 2004; Neves-Pereira et al., 2005; Schumacher et al., 2005; Green et al., 2006].

In order to clarify the possible effect of the BDNF Val66Met polymorphism in BPD, we performed an association study comparing the allele/genotype frequency of the polymorphism in the Han Chinese population.

MATERIALS AND METHODS

RESULTS

The allele and genotype distribution of Val66Met polymorphism are presented in Table I. The distribution of genotypes in healthy control subjects (χ² = 1.955, df = 1, P = 0.162) was consistent with Hardy–Weinberg equilibrium but it was not the case in patients (χ² = 7.561, df = 1, P = 0.00598). The allele frequency of patients did not significantly differ from those of control subjects (χ² = 0.636, df = 1, P = 0.425), and with regard to subtypes of bipolar II patients, we observed a slight significant difference (χ² = 4.968, df = 1, P = 0.0258 after Bonferroni correction, P = 0.0516). The genotype distribution between patients and controls showed significant difference (χ² = 9.4666, df = 2, P = 0.00884), when patients were divided into bipolar I disorders (χ² = 7.882, df = 2, P = 0.0391) and bipolar II disorders (χ² = 8.022, df = 2, P = 0.0364).

The distribution of the homozygous Met/Met genotype was much higher in patients than in control subjects. The odds ratio for patients with Met/Met was 1.44 (95% CI = 1.070–1.950; P = 0.016), for those with bipolar I disorder 1.335 (95% CI = 0.973–1.832; P = 0.146) and for those with bipolar II disorder 2.07 (95% CI = 1.242–3.454; P = 0.011), respectively.

DISCUSSION

In this study, we reported the association between Met allele of the functional Val66Met polymorphism and BPD. Several lines of investigations have also shown that the Met allele is most associated with other mental diseases such as anxiety disorders [Jiang et al., 2005] depression [Hwang et al., 2006] and suicidal behavior [Iga et al., 2007; Kim et al., 2008]. Recent meta-analysis reveal that the Met allele increases risk for anxious morbidity [Frustaci et al., 2008]. However, in previous studies the Val allele was shown to be associated with increased risk for BPD [Neves-Pereira et al., 2002; Sklar et al., 2002; Lohoff et al., 2005; Vincze et al., 2008] while some studies have indicated a lack of association between Val66Met polymorphism and BPD [Hong et al., 2003a; Nakata et al., 2003; Kunugi et al., 2004; Tang et al., 2008], which was contrary to our results in Han population.

There are several possible reasons for the discrepancy in findings: the Val66Met allele frequency between individuals of Caucasian descent (f(Val) = 77–83%) and individuals of Asian descent (f(Val) = 52–59%) is significantly different. This difference in polymorphism frequency, based on ethnic background, is difficult to control for and should be more careful to interpret the data gathered from such association studies. Another possible factor contributing to the disparity might be limited statistical power in the Asian population studies. In contrast to other studies on Asian populations, our study is the first to report that the Val66Met polymorphism is associated with BPD, and this conclusion is strengthened by fact that our study has the largest sample number among the case-control studies on Asian populations. However, larger sample sizes, in thousands, are applausive for studies of genes with small effects.

In addition, we performed the analysis by subtype of the disorder, and still found the positive association in the two subtypes (bipolar I and bipolar II). In bipolar II disorder, we detected that both the allele and the genotype showed significant difference between patients and control subjects, which implies that the BDNF gene may play an important role in bipolar II. Furthermore, to investigate the odds ratio of Val66Met polymorphism in the study, we employed the recessive model and the results suggested that the high frequency of Met/Met genotype might be a risk factor for BPD. Recent functional studies exhibited that the Met allele was associated with poorer episodic memory, abnormal hippocampal activation, and lower hippocampal n-acetyl aspartate (NAA) [Egan et al., 2003]. In vivo studies have also shown that the Met allele affects prefrontal energy metabolism in BPD [Frey et al., 2007] and may cause temporary gray matter reductions in BPD patients [McIntosh et al., 2007]. Mice model have indicated that homozygous mice with Met/Met genotype have a significant decrease in biologically relevant BDNF release, an increase in anxiety-like behavior [Chen et al., 2006]. All these studies are consistent with the result we report that the Met allele increases risk for BPD.

BPD is a spectrum disorder and is likely to involve multiple genes, each contributing only a small fraction to the overall risk. Clinical variables such as age of onset, subtype (bipolar I or II), psychotic features, or family history add to the difficulty of detecting contributing risk factors, thus making the interpretation of results is problematical. Stricter definitions patient selection, further division of patients into sub-groups and larger sample sizes are required to yield more robust results.

Considering all patients and control subjects are from Anhui Province, and our observed allele frequencies in control subjects are consistent with other Han Chinese population [Tang et al., 2008], population stratification is unlikely to occupy a critical role in our study population. Moreover, we performed the strict criterion for admission in this study, which implicated that more reliable results might yield. In addition, we also validated the allele-specific PCR results by direct sequencing assay, thus, the observation of Hardy–Weinberg disequilibrium in patients was inclined to arise from real disease association.

In conclusion, our case-control study provides the evidence that the Met allele rather than the Val allele may play a role in the susceptibility to BPD in the Han Chinese population. Given the important role of BDNF in the neural system, further studies should be carried out to elucidate the involvement of the BDNF gene in the pathophysiology of BPD.