Recent Molecular Genetic Studies and Methodological Issues in Major Depression Research

Overview Taiwanese Journal of Psychiatry (Taipei) Vol. 27 No. 1 2013 • 11 •

Shih-Jen Tsai, M.D.1,2*

Major depressive disorder (MDD) is a genetically complex mental disorder involving numerous susceptible genes whose expression may be modulated by environmental factors. In the past decade, case-control association approaches have been the mainstream in MDD genetic studies. But whole genome gene-expression microarray and genome-wide association studies have begun to emerge in recent years. Genetic association studies have suggested several genes related to MDD, although the results to date are inconclusive and no single genetic variation has been identifi ed to increase the risk of depression substantially. Recent approaches based on gene-environment interaction have demonstrated more consistent fi ndings. Therefore, the usefulness from the study results of case-control studies is limited by present knowledge of MDD pathophysiology. Genome-wide association studies are not constrained by our limited knowledge. Although the results of several recent genome-wide association studies did not reach the desired level of statistical signifi cance, these studies do support interesting candidate genes and genomic regions for further study. In short, the fi eld of genetic studies in general has been disappointing because to fi nd common gene variants of large effect in MDD pathogenesis has been unsuccessful. However, the fi eld of psychiatric genetics is rapidly growing, and several new technological advances (e.g. whole- genome sequencing) will be soon available for large-scale studies. These approaches provide exciting new avenues to identify new candidate genes for MDD. A better understanding of the MDD-related genes may potentially lead to developing effective prevention and treatment of this disease.

Key words: genetics, association study, genome-wide association study, epigenetics (Taiwanese Journal of Psychiatry [Taipei] 2013; 27: 11-21)

1 Department of Psychiatry, Taipei Veterans General Hospital, Taipei, Taiwan 2 Division of Psychiatry, School of Medicine, National Yang-Ming University, Taipei, Taiwan Received: January 4, 2013; revised: February 26, 2013; accepted: February 26, 2013 *Corresponding author. Department of Psychiatry, Taipei Veterans General Hospital, No. 201, Section 2, Shih-Pai Road, Taipei 11217, Taiwan E-mail: Shih-Jen Tsai <[email protected]> • 12 • Genetic Studies in Depression

pathogenesis for statistical association. The clas- Introduction sic design of a genetic association analysis is a case-control study in which the frequency of a Major depressive disorder (MDD) is a het- possible risk allele of a candidate gene in unrelat- erogeneous, highly prevalent, and moderately ed individuals affected by the same disease (case heritable mental disorder. Estimated heritability group) is compared to the frequency observed in of MDD based on monozygotic vs. dizygotic twin healthy individuals of the same ethnic group (con- concordance differences exhibits a modest herita- trol group). If the risk allele/genotype analyzed is ble contribution of 37% [1]. Adoption studies also found more frequently in the case group than in have suggested both an important genetic impact controls, an association exists between the genetic of parental depression, and a signifi cant environ- variant and the disease. This means that the pres- mental impact of maternal depression in mediat- ence of this risk allele increases the risk of, or sus- ing depression among adopted adolescents [2]. ceptibility to, the disease. But researchers need to Although genetic factors play a role in MDD, interpret positive association fi nding with great identifying specifi c genes involved has proved caution since many reports have problems with challenging. In the past decade, molecular-genetic the study design and/or statistical analysis. In ad- technologies have made great advances in ge- dition, population stratifi cation, insuffi cient sam- nome-wide searches for disease-causing genes ple power, genotyping problems, and phenotype with the linkage disequilibrium (LD) approach heterogeneity could cause false positive or nega- and genome-wide studies of gene expression and tive fi ndings in genetic association studies. Given chromatin modifi cations that refl ect the epigenetic these limitations, candidate gene association re- response of the genome to environmental expo- search is still the most popular method to investi- sure. The technological revolution has shifted gate genetic factors in MDD due to its simple MDD genetic study from case-control association technique. Methodological aspects of the candi- studies (focused on hypothesis-driven candidates) date gene association studies have progressed to to genome-wide searches (employed an unbiased use increasingly larger case control samples or exploration using linked polymorphic markers meta-analytic summaries to increase the sample that span the whole genome in search of previ- size. ously unimplicated loci). Here, the author intends Using key words “depressi*” AND ((genetic to provide an overview of recent genetic studies of association) OR polymorphism*); fi lters activat- MDD with different approaches to advance ed: humans” in a PubMed search, I found about knowledge of the genetic bases of MDD, and to three hundreds of MDD genetic association stud- explore methodological issues in MDD genetic ies. Among them, 36 reports are from Taiwan studies. (Table 1) [3-38]. In 2008, Lopez-Leon et al. re- viewed MDD association studies reported in 183 Candidate Gene papers that studied 393 polymorphisms in 102 Association Studies genes [39]. They found 22 of these polymorphisms were investigated by three or more studies Candidate gene association studies test ge- allowing a meta-analysis, and found that signifi - netic polymorphisms in genes involved in disease cant evidence for association for APOE (apolipo- Tsai SJ • 13 •

Table 1 Reports of major depression candidate gene association studies from Taiwan Reference Polymorphism Publication Result Authors number /gene year 3 T102C /HTR2A No association Tsai SJ et al. 1999 4 A218C /TPH1 Association Tsai SJ et al. 1999 5 (AAT)n repeat /CNR1 No association Tsai SJ et al. 2001 6 C825T /GNB3 No association Lin CN et al. 2001 7 2 bp deletion /CHRNA7 Modest association Lai IC et al. 2001 8 C267T /HTR6 No association Wu WH et al. 2001 9 I/D polymorphism /ACE No association Hong CJ et al. 2002 10 5-HTTVNTR /SLC6A4 No association Yen FC et al. 2003 11 Val66Met /BDNF No association Hong CJ et al. 2003 12 2 SNPs /ESR1 No association Tsai SJ et al. 2003 13 C-511T /IL1B No association Yu YW et al. 2003 14 C276T /NOS1 No association Yu YW et al. 2003 15 Val66Met /BDNF No association Tsai SJ el al. 2003 16 Val66Met /BDNF No association Hong CJ et al. 2003 17 A-161T /HTR1B No association Tsai SJ et al. 2004 18 uVNTR /MAOA No association Yu YW et al. 2005 19 C-634G /IL6 No association Hong CJ et al. 2005 20 C1976T /ADORA2A No association Tsai SJ et al. 2006 21 Val66Met /BDNF Met66 allele was associated with geriatric MDD Hwang JP et al. 2006 22 epsilon /APOE A lower frequency of epsilon2 allele in MDD Fan PL et al. 2006 23 epsilon /APOE No association Hwang JP et al. 2006 24 T-182C, G1287A / No association Chang CC et al. 2007 SLC6A2 25 5 SNPs /NGFR No association Gau YT et al. 2008 26 4 SNPs /GSK3B No association Tsai SJ et al. 2008 27 6 SNPs /SERPINE1 rs2227684-G and rs7242-T alleles were more Tsai SJ et al. 2008 frequent in MDD 28 3 SNPs /S100A10 No association Tzang RF et al. 2008 29 uVNTR, rs1137070/ Association in female MDD Huang SY et al. 2009 MAOA 30 5 SNPs /TPH2 rs17110747-G homozygote was more frequent in Tsai SJ et al. 2009 MDD 31 6 SNPs /TPH2 2755A allele was associated with female MDD Lin YM et al. 2009 32 C-511T /IL1B No association Hwang JP et al. 2009 33 7 SNPs /VEGFA No association Tsai SJ et al. 2009 34 5 SNPs /KCNK2 rs6686529 was associated with MDD Liou YJ et al. 2009 35 Val66Met /BDNF & 4 BDNF and NTRK2 interacted in MDD suceptibility Lin E et al. 2009 SNPs /NTRK2 36 6 SNPs /PAWR rs8176874 was associated with MDD Liou YJ et al. 2011 37 17 SNPs /SLC6A3 Modest association Huang CC et al. 2011 38 uVNTR /MAOA MAOA 4R allele was more frequent in male MDD Lung FW et al. 2011 SNP, single-nucleotide polymorphism; Bp, base pairs. uVNTR, upstream variable-number tandem repeat; I/D, insertion- deletion. • 14 • Genetic Studies in Depression

protein E), DRD4 (D4 subtype of the dopamine frequency of 5-HTTLPR alleles in 94 MDD pa- receptor), GNB3 (guanine nucleotide binding pro- tients who had history of attempted suicide with tein (G protein), beta polypeptide 3), MTHFR 94 controls who were free of any psychiatric dis- (methylenetetrahydrofolate reductase), SLC6A3 order, showing no distributional difference be- (gene encoding the dopamine transporter) and tween those two groups [43]. SLC6A4 (gene encoding the serotonin transporter). Tryptophan hydroxylase-1 (TPH1) Several candidate genes deserve mentioning, TPH1 encodes tryptophan hydroxylase-1, as they have been suggested repeatedly to be im- which is the rate-limiting enzyme in serotonin plicated in MDD genetic association studies. synthesis. The TPH1 A218C (rs1800532) polymorphism has been the most extensively studied Serotonin transporter (5-HTT) TPH1 polymorphism in MDD. In 1997, Mann et Following serotonin release, the serotonin al. reported that the TPH1 A218C polymorphism transporter is the major site of serotonin reuptake is associated with suicidal behaviors in patients into the presynaptic neuron. By regulating the with major depression [44]. Our study replicated reuptake of the released serotonin, the serotonin their fi nding that this TPH1 polymorphism is as- transporter is central to fi ne-tune the serotonergic sociated with major depression and the associa- neurotransmission. The SLC6A4 gene encoding tion was most signifi cant in patients who attempt- serotonin transporter has been one of the genes ed suicide [4]. But a meta-analysis across eight most studied in MDD because serotonin trans- MDD samples (1,812 cases and 2,223 controls) porter is the major therapeutic target of selective has not shown any association for any rs1800532 serotonin reuptake inhibitors (SSRIs) and sero- genotype and MDD [45]. tonin-and-norepinephrine reuptake inhibitors (SNRIs). Among the 5-HTT genetic variants, a Brain-derived neurotrophic factor (BDNF) 44-base pairs (bp) insertion-deletion in the pro- Brain-derived neurotrophic factor is a mem- moter region (5-HTTLPR) polymorphism has ber of the neurotrophin family and plays an im- been extensively studied for association with portant role in neuronal survival and brain plastic- MDD. The 5-HTTLPR short variant suppresses ity. Thus, BDNF is considered an attractive transcriptional activity of the promoter [40]. A candidate gene for the study of healthy and dis- meta-analysis in 2010 indicated evidence of a eased brain function and behaviors. Evidence small but statistically signifi cant association be- from animal and clinical studies has implicated tween the 5-HTTLPR polymorphism and MDD decreased central BDNF activity in the pathogen- (odds ratio [OR] 1.08, 95% confi dence interval esis of MDD; therefore, activating the BDNF- [CI] 1.03-1.12) [41]. A recent study was also dependent pathway plays an important role in the found that 5-HTTLPR is functionally triallelic mechanism of antidepressant therapeutic action [46]. [42]. The presence of an A-to-G single nucleotide In 2003, we studied the association between polymorphism further subdivides the long allele the functional BDNF Val66Met polymorphism into La or Lg, which is equivalent to the S allele in and MDD outpatients [15] and inpatients [16], reduced functional activity. Using the triallelic and demonstrated negative fi ndings. The negative classifi cation, Segal et al. in 2009 compared the association between BDNF Val66Met polymor- Tsai SJ • 15 •

phism and MDD is supported by most of the fol- neglect) and genetic risk factors that are alone nei- lowing studies, suggesting that the BDNF ther necessary nor suffi cient for the disorder to Val66Met polymorphism is not a major contribut- develop [1]. Gene-environment interaction ing factor to MDD susceptibility. In 2006, we (G×E) studies suggest that by moderating the ef- tested the association between the BDNF fects of the environment, genetic variation ex- Val66Met polymorphism and geriatric MDD, and plains why some individuals are vulnerable and found that a highly signifi cant excess of the Met66 others are resistant to the effects of adversity. allele is present in the geriatric MDD patients Caspi et al. were the fi rst investigators to [21]. The depolarization-induced secretion of show that childhood maltreatment and later stress- BDNF was reduced in Met66 BDNF-transfected ful life events can predict the onset of depressive neurons compared with Val66 BDNF analogs symptoms only in genetically predisposed indi- [47]. Since subjects carrying the Met66/Met66 viduals with a short allele of the 5-HTTLPR poly- genotype may have lower BDNF activity than morphism, while the long allele carriers were Val66 carriers, the increased Met66/Met66 fre- found to be more resilient to depression after ex- quency in geriatric MDD patients is consistent posing to adverse life events [50]. In a recent me- with the hypothesis that MDD is associated with ta-analysis of 54 studies, 5-HTTLPR has been reduced BDNF activity in the brain [46]. A com- found to moderate the relationship between stress parison study with 245 elderly depressed white and depression, with the 5-HTTLPR short allele subjects and 94 elderly white normal controls rep- associated with an increased risk of developing licated our fi ndings indicating geriatric depressed depression under stress (p = 0.00002) [51]. subjects are more likely to be Met66 allele carriers Signifi cant G×E interactions on MDD suscepti- than the normal controls [48]. Verhagen et al. re- bility have also been reported for BDNF, FKBP5 cently did a meta-analysis of the BDNF Val66Met and CRHR1 genetic variants [52]. polymorphism in MDD of 14 studies involving Because MDD may be related to multiple 2,812 MDD cases and 10,843 non-depressed con- genes, several studies have investigated gene- trols [49]. It revealed that the BDNF Val66Met gene (G×G) interaction in MDD. For example, in polymorphism is not signifi cantly associated with genetic study of geriatric depression, we found MDD in the total sample; however, the gender that BDNF Val66Met polymorphism and NTRK2 stratifi ed analyses revealed signifi cant effects in (BDNF receptor gene) genetic variants interact in both the allelic and genotypic analyses in men the 2-locus, 3-locus, 4-locus, and 5-locus gene- (OR [Met66], 95% CI; 1.27 [1.10-1.47]) ; OR gene interaction models using a generalized mul- (Met66/Met66), 95% CI; 1.67 [1.19-2.36]) but not tifactor dimensionality reduction (GMDR) meth- in women. od [35]. The results suggest that the BDNF and NTRK2 genes may contribute to the risk of geriat- Gene-environment interaction and gene- ric depression in an interactive manner. gene interaction Epidemiologic studies suggest that MDD re- Linkage Studies sults from multiple interdependent environmental factors (such as childhood maltreatment, includ- Linkage studies are genetic analyses of pedi- ing psychological, physical, and sexual abuse and grees and are well-suited for studying disorders • 16 • Genetic Studies in Depression

with a clear pattern of Mendelian inheritance and the resultant infl ation of type I error), the standard disorders driven by genes of large effect. While in the fi eld has been to set “genome-wide signifi - both conditions are less likely to operate in de- cance” at p ≤ 1.0×10-8, which is highly restrictive pression, there have been few reports of linkage and diffi cult to attain. studies in MDD. For MDD, there are now fi ve published Supportive fi ndings of many MDD linkage GWAS studies. In 2009, Sullivan et al. fi rst used a studies include regions on chromosomes 1, 3, 4, 6, sample of 1,738 MDD cases and 1,802 controls 8, 11, 12, 15, and 18 [53]. A major limitation of from the Netherlands [54]. Among the 435,291 family-based linkage studies is an inherent diffi - single-nucleotide polymorphisms (SNPs), the au- culty to ascertain sizable numbers of families with thors found the top signal to be at rs1558477 multiple affected relatives sharing restrictive phe- (trend-test p-value of 1×10-6), 12.4 kb down- notype characteristics that will allow a large stream of ADCYAP1R1 gene. Additional GWAS enough sample power to detect variants with a of MDD have followed. Disappointingly, there small effect size. Further, large chromosomal re- were no genome-wide signifi cant fi ndings in these gions shared among family members constrain the GWAS studies. Although the results of each indi- narrowing down of a linkage signal suffi ciently to vidual GWAS studies did not reach the desired identify a causative gene. level of statistical signifi cance, they do suggest interesting candidate genes that may be worth- Genome-wide Scans and while following up in future studies. It is becom- DNA Microarray ing increasingly clear that individual genetic susceptibility factors for MDD are likely to have only Technical improvements in genotyping in the minor effects, and large sample size will be neces- past decade have led to markedly increased geno- sary to identify them. typing capacity at reduced cost. Combined with Microarray technology examines the ex- the identifi cation of millions of common polymor- pression levels of thousands of gene transcripts phisms through the HapMap project, high- in postmortem brains of suicide completers. This throughput genotyping approaches and bioinfor- technology has opened new avenues for discov- matics analysis have made genome-wide ering MDD pathogenesis and may aid in identi- association study (GWAS) a feasible study design fying candidate genes for a further case-control to identify genetic loci and ultimately candidate association study. Using this technique, Sequeira genes associated with complex psychiatric disor- et al. demonstrated lower expression of the ders. This method is based on genotyping arrays spermine/sperminidine N1-acetyltransference or microarrays that allow the variability of the hu- (SSAT) gene, the rate-limiting enzyme in the ca- man genome to be traced for assessing the hypoth- tabolism of polyamines, in the brain of suicide esis of common disease-common variant without completers and further confi rmed the role of the need to conduct a hypothesis-driven study of SSAT in suicide behaviors by genetic case-con- the etiology of the disease. Limitations of this ap- trol analysis [55]. proach are the immense amount of data, costs, and issues regarding multiple testings. Because of the unprecedented amount of hypothesis testing (and Tsai SJ • 17 •

conclude this overview with six comments on Epigenetics methodological issues for future MDD genetic studies to improve future consistent fi ndings be- Epigenetics comprises mechanisms of gene tween studies. expression modifi cations that do not alter the ge- ‧ Many MDD genetic association studies to date nomic code itself. The two major epigenetic address only hypothesis-driven candidate genes mechanisms are DNA methylation and histone that mostly revolve around known mechanisms modifi cation. Epigenetic mechanisms may be in- of antidepressant drugs. Novel ground-break- volved in the reprogramming of gene expression ing hypotheses for MDD are needed to produce in response to stressful stimuli and have long-last- new sets of candidate genes. Furthermore, most ing effects within mature neurons. Studies in ani- genetic studies have devoted disproportionate mals have shown that epigenetic mechanisms focus to being on few highly reported function- with histone modifi cations or DNA methylation al polymorphisms (DNA sequence variations affect diverse pathways leading to depression-like that alter the expression and/or functioning of behaviors [56]. Human studies have revealed the the gene product), neglecting comprehensive

GABAA receptor promoter [57] and the BDNF coverage of the coding and promoter regions of promoter hypermethylation [58] in the brain of the gene of interest. depressed suicide completers. Recent study using ‧ One strategy to investigate gene effects in etio- genome-wide DNA methylation scan in postmor- logically complex diseases is through decon- tem frontal cortex of 39 MDD samples and 26 structing complexity into parts that are likely to controls identifi ed that 224 candidate regions with be less etiologically heterogeneous, such as in- DNA methylation differences are greater than termediate phenotypes [60]. Heritable interme- 10% [59]. These regions are highly enriched for diate phenotypes that are disease-associated neuronal growth and development genes. The have been termed endophenotypes. Several above-mentioned fi ndings suggested that epigen- promising endophenotypes proposed for MDD etic pharmaceuticals, such as DNA methylation genetic studies include neuroticism, the dexa- inhibitors, are promising and deserves further at- methasone suppression test and the neuroimag- tention in developing depression treatment. ing [60]. ‧ We must not overlook the role of the environ- Conclusion ment factors in studying genetics. The most consistent MDD genetic fi ndings are the GxE MDD is a complex and multifactorial men- studies (e.g. 5-HTTLPR polymorphism and ad- tal disorder with important genetic and nonge- verse life events). It should become possible to netic contributory factors. Although many ge- study gene-environment interactions by con- netic association studies have investigated MDD, ducting larger-scale molecular studies on sam- at present no identifi ed gene are likely to be ples drawn from longitudinal cohorts and by found as the specifi c cause of MDD. Researchers measuring these variables in samples collected have reported many genetic polymorphisms are for genetic studies. More intensive studies to associated with MDD, but their studies cannot be measure factors like childhood trauma and life duplicated later. Here, the author would like to stress are needed. • 18 • Genetic Studies in Depression

‧ Although candidate gene case-control studies netic factors and non-genetic factors. Genes have value in MDD genetic studies, they are interact not only among each other, but also naturally limited to the study of genes that have environmental factors may interplay with SB. some contributing function to the development Using statistical methods that account for of MDD. Previous genetic studies have focused multi-factorial etiology of MDD would be im- on monoaminergic systems, particularly the se- portant in future research. For example, the rotonergic system, and recent studies have fo- GMDR method, which extends the MDR meth- cused on genes in neurotrophic systems. But od, is a non-parametric data-mining approach many genes with unknown function may con- applied to continuous and dichotomous pheno- tribute to causing MDD. Genome-wide studies types. Additionally, GMDR permits adjusting of gene expression and genetic variation are discrete and quantitative covariates, and has now possible and are not constrained by our been used in various population-based studies limited knowledge. Although these unbiased with unbalanced groups of case and control approaches are at the infancy stage, they have subjects. already provided some new avenues to identify new candidate genes for MDD genetic studies Summary and to confi rm the fi ndings in genome-wide Many genes are likely responsible for MDD studies. susceptibility. A better understanding of the genes ‧ Genome-wide association studies using indi- involved in MDD and their interaction with non- rect mapping have been successful in localizing genetic factors such age, gender, and stress, may genes associated with complex diseases when help develop more effective prevention, diagnosis common variants are the underlying cause of and management of MDD. disease etiology. A strong LD between tagSNPs and underlying causal variants makes it feasible Acknowledgements to use indirect mapping. To detect associations with rare variants, indirect LD mapping will be Fundings for this report were provided by low-powered due to weak correlations between grant NSC 101-2314-B-075-040 from the common tagSNPs and rare causal variants. But National Science Council, Taiwan and by grant for associations with rare variants, to perform VGHUST102-G1-2-1 from the Taipei Veterans direct mapping and to fi rst identify rare variants General Hospital. within a sample are necessary. Sequencing candidate genes or entire genomes are the optimal References way to identify rare variants. Compared with present clinical genetic testing, whole genome 1. Sullivan PF, Neale MC, Kendler KS: Genetic epide- sequencing greatly expands the breadth of test- miology of major depression: review and meta-anal- ing from genes associated with MDD to the ysis. Am J Psychiatry 2000; 157: 1552-62. whole genome and, potentially, all the informa- 2. Tully EC, Iacono WG, McGue M: An adoption study tion that the genome contains about MDD. of parental depression as an environmental liability for adolescent depression and childhood disruptive ‧ MDD is a complex disorder and the predisposi- disorders. Am J Psychiatry 2008; 165: 1148-54. tion towards MDD consists of numerous ge- Tsai SJ • 19 •

3. Tsai SJ, Hong CJ, Hsu CC, et al.: Serotonin-2A re- genetic polymorphism with major depressive disor- ceptor polymorphism (102T/C) in mood disorders. der, associated symptomatology, and antidepressant Psychiatry Res 1999; 87: 233-7. response. Neuropsychopharmacology 2003; 28: 4. Tsai SJ, Hong CJ, Wang YC: Tryptophan hydroxy- 1182-5. lase gene polymorphism (A218C) and suicidal be- 14. Yu YW, Chen TJ, Wang YC, Liou YJ, Hong CJ, Tsai haviors. Neuroreport 1999; 10: 3773-5. SJ: Association analysis for neuronal nitric oxide 5. Tsai SJ, Wang YC, Hong CJ: Association study be- synthase gene polymorphism with major depression tween cannabinoid receptor gene (CNR1) and patho- and ﬂ uoxetine response. Neuropsychobiology 2003; genesis and psychotic symptoms of mood disorders. 47: 137-40. Am J Med Genet 2001; 105: 219-21. 15. Tsai SJ, Cheng CY, Yu YW, Chen TJ, Hong CJ: 6. Lin CN, Tsai SJ, Hong CJ: Association analysis of a Association study of a brain-derived neurotrophic- functional G protein beta3 subunit gene polymor- factor genetic polymorphism and major depressive phism (C825T) in mood disorders. Neuropsy- disorders, symptomatology, and antidepressant re- chobiology 2001; 44: 118-21. sponse. Am J Med Genet B Neuropsychiatr Genet 7. Lai IC, Hong CJ, Tsai SJ: Association study of nico- 2003; 123B: 19-22. tinic-receptor variants and major depressive disorder. 16. Hong CJ, Huo SJ, Yen FC, Tung CL, Pan GM, Tsai J Affect Disord 2001; 66: 79-82. SJ: Association study of a brain-derived neurotroph- 8. Wu WH, Huo SJ, Cheng CY, Hong CJ, Tsai SJ: ic-factor genetic polymorphism and mood disorders, Association study of the 5-HT(6) receptor polymor- age of onset and suicidal behavior. Neuropsychobio- phism (C267T) and symptomatology and antidepres- logy 2003; 48: 186-9. sant response in major depressive disorders. Neuro- 17. Tsai SJ, Hong CJ, Yu YW, Chen TJ, Wang YC, Lin psychobiology 2001; 44: 172-5. WK: Association study of serotonin 1B receptor 9. Hong CJ, Wang YC, Tsai SJ: Association study of (A-161T) genetic polymorphism and suicidal behav- angiotensin I-converting enzyme polymorphism and iors and response to ﬂ uoxetine in major depressive symptomatology and antidepressant response in ma- disorder. Neuropsychobiology 2004; 50: 235-8. jor depressive disorders. J Neural Transm 2002; 109: 18. Yu YW, Tsai SJ, Hong CJ, Chen TJ, Chen MC, Yang 1209-14. CW. Association study of a monoamine oxidase a 10. Yen FC, Hong CJ, Hou SJ, Wang JK, Tsai SJ: gene promoter polymorphism with major depressive Association study of serotonin transporter gene disorder and antidepressant response. Neuropsycho- VNTR polymorphism and mood disorders, onset age pharmacology 2005; 30: 1719-23. and suicide attempts in a Chinese sample. 19. Hong CJ, Yu YW, Chen TJ, Tsai SJ: Interleukin-6 ge- Neuropsychobiology 2003; 48: 5-9. netic polymorphism and Chinese major depression. 11. Hong CJ, Huo SJ, Yen FC, Tung CL, Pan GM, Tsai Neuropsychobiology 2005; 52: 202-5. SJ: Association study of a brain-derived neurotroph- 20. Tsai SJ, Hong CJ, Hou SJ, Yen FC: Association study ic-factor genetic polymorphism and mood disorders, of adenosine A2a receptor (1976C > T) genetic poly- age of onset and suicidal behavior. Neuropsychobio- morphism and mood disorders and age of onset. logy 2003; 48: 186-9. Psychiatr Genet 2006; 16: 185. 12. Tsai SJ, Wang YC, Hong CJ, Chiu HJ: Association 21. Hwang JP, Tsai SJ, Hong CJ, et al.: The Val66Met study of oestrogen receptor alpha gene polymor- polymorphism of the brain-derived neurotrophic- phism and suicidal behaviours in major depressive factor gene is associated with geriatric depression. disorder. Psychiatr Genet 2003; 13: 19-22. Neurobiol Aging 2006; 27: 1834-7. 13. Yu YW, Chen TJ, Hong CJ, Chen HM, Tsai SJ: 22. Fan PL, Chen CD, Kao WT, Shu BC, Lung FW: Association study of the interleukin-1 beta (C-511T) Protective effect of the apo epsilon2 allele in major • 20 • Genetic Studies in Depression

depressive disorder in Taiwanese. Acta Psychiatr sociation to major depression and anxiety in Chinese Scand 2006; 113: 48-53. peripartum women. Arch Womens Ment Health 2009; 23. Hwang JP, Yang CH, Hong CJ, Lirng JF, Yang YM, 12: 401-8. Tsai SJ: Association of APOE genetic polymorphism 32. Hwang JP, Tsai SJ, Hong CJ, Yang CH, Hsu CD, with cognitive function and suicide history in geriat- Liou YJ: Interleukin-1 beta-511C/T genetic polymor- ric depression. Dement Geriatr Cogn Disord 2006; phism is associated with age of onset of geriatric de- 22: 334-8. pression. Neuromolecular Med 2009; 11: 322-7. 24. Chang CC, Lu RB, Chen CL, et al.: Lack of associa- 33. Tsai SJ, Hong CJ, Liou YJ, et al.: Haplotype analysis tion between the norepinephrine transporter gene and of single nucleotide polymorphisms in the vascular major depression in a Han Chinese population. J endothelial growth factor (VEGFA) gene and antide- Psychiatry Neurosci 2007; 32: 121-8. pressant treatment response in major depressive dis- 25. Gau YT, Liou YJ, Yu YW, et al.: Evidence for asso- order. Psychiatry Res 2009; 169: 113-7. ciation between genetic variants of p75 neurotrophin 34. Liou YJ, Chen TJ, Tsai SJ, Yu YW, Cheng CY, Hong receptor (p75NTR) gene and antidepressant treat- CJ: Support for the involvement of the KCNK2 gene ment response in Chinese major depressive disorder. in major depressive disorder and response to antide- Am J Med Genet B Neuropsychiatr Genet 2008; pressant treatment. Pharmacogenet Genomics 2009; 147B: 594-9. 19: 735-41. 26. Huang SY, Lin MT, Lin WW, Huang CC, Shy MJ, Lu 35. Lin E, Hong CJ, Hwang JP, et al.: Gene-gene interac- RB: Association of monoamine oxidase A (MAOA) tions of the brain-derived neurotrophic-factor and polymorphisms and clinical subgroups of major de- neurotrophic tyrosine kinase receptor 2 genes in geri- pressive disorders in the Han Chinese population. atric depression. Rejuvenation Res 2009; 12: World J Biol Psychiatry 2009; 10: 544-51. 387-93. 27. Tsai SJ, Liou YJ, Hong CJ, Yu YW, Chen TJ: 36. Liou YJ, Chen TJ, Tsai SJ, et al.: Evidence of in- Glycogen synthase kinase-3beta gene is associated volvement of the human Par-4 (PAWR) gene in ma- with antidepressant treatment response in Chinese jor depressive disorder. World J Biol Psychiatry major depressive disorder. Pharmacogenomics J 2011; 12: 288-95. 2008; 8: 384-90. 37. Huang CC, Lu RB, Shih MC, Yen CH, Huang SY: 28. Tsai SJ, Hong CJ, Liou YJ, Yu YW, Chen TJ: Association study of the dopamine transporter gene Plasminogen activator inhibitor-1 gene is associated with personality traits and major depressive disorder with major depression and antidepressant treatment in the Han Chinese population. Pharmacogenet response. Pharmacogenet Genomics 2008; 18: Genomics 2011; 21: 94-7. 869-75. 38. Lung FW, Tzeng DS, Huang MF, Lee MB: 29. Tzang RF, Hong CJ, Liou YJ, Yu YW, Chen TJ, Tsai Association of the MAOA promoter uVNTR poly- SJ: Association study of p11 gene with major depres- morphism with suicide attempts in patients with ma- sive disorder, suicidal behaviors and treatment re- jor depressive disorder. BMC Med Genet 2011; 12: sponse. Neurosci Lett 2008; 447: 92-5. 74. 30. Tsai SJ, Hong CJ, Liou YJ, et al.: Tryptophan hy- 39. Lopez-Leon S, Janssens AC, Gonzalez-Zuloeta Ladd droxylase 2 gene is associated with major depression AM, et al.: Meta-analyses of genetic studies on major and antidepressant treatment response. Prog Neuro- depressive disorder. Mol Psychiatry 2008; 13: psychopharmacol Biol Psychiatry 2009; 33: 637-41. 772-85. 31. Lin YM, Ko HC, Chang FM, Yeh TL, Sun HS: 40. Collier DA, Stober G, Li T, et al.: A novel functional Population-speciﬁ c functional variant of the TPH2 polymorphism within the promoter of the serotonin gene 2755C > A polymorphism contributes risk as- transporter gene: possible role in susceptibility to af- Tsai SJ • 21 •

fective disorders. Mol Psychiatry 1996; 1: 453-60. 50. Caspi A, Sugden K, Moffi tt TE, et al.: Infl uence of 41. Clarke H, Flint J, Attwood AS, Munafò MR: life stress on depression: moderation by a polymor- Association of the 5-HTTLPR genotype and unipolar phism in the 5-HTT gene. Science 2003; 301: 386-9. depression: a meta-analysis. Psychol Med 2010; 40: 51. Karg K, Burmeister M, Shedden K, Sen S: The sero- 1767-78. tonin transporter promoter variant (5-HTTLPR), 42. Hu XZ, Lipsky RH, Zhu G, et al.: Serotonin trans- stress, and depression meta-analysis revisited: evi- porter promoter gain-of-function genotypes are dence of genetic moderation. Arch Gen Psychiatry linked to obsessive-compulsive disorder. Am J Hum 2011; 68: 444-54. Genet 2006; 78: 815-26. 52. Keers R, Uher R: Gene-environment interaction in 43. Segal J, Schenkel LC, Oliveira MH, et al.: Novel al- major depression and antidepressant treatment re- lelic variants in the human serotonin transporter gene sponse. Curr Psychiatry Rep 2012;14:129-37. linked polymorphism (5-HTTLPR) among depressed 53. Levinson DF: The genetics of depression: a review. patients with suicide attempt. Neurosci Lett 2009; Biol Psychiatry 2006; 60: 84-92. 451: 79-82. 54. Sullivan PF, de Geus EJ, Willemsen G, et al.: 44. Mann JJ, Malone KM, Nielsen DA, Goldman D, Genome-wide association for major depressive dis- Erdos J, Gelernter J: Possible association of a poly- order: a possible role for the presynaptic protein pic- morphism of the tryptophan hydroxylase gene with colo. Mol Psychiatry 2009; 14:359-75. suicidal behavior in depressed patients. Am J 55. Sequeira A, Gwadry FG, Ffrench-Mullen JM, et al.; Psychiatry 1997; 154: 1451-3. Implication of SSAT by gene expression and genetic 45. Chen C, Glatt SJ, Tsuang MT: The tryptophan hy- variation in suicide and major depression. Arch Gen droxylase gene infl uences risk for bipolar disorder Psychiatry 2006; 63: 35-48. but not major depressive disorder: results of meta- 56. Schroeder M, Krebs MO, Bleich S, Frieling H: analyses. Bipolar Disord 2008; 10: 816-21. Epigenetics and depression: current challenges and 46. Duman RS, Heninger GR, Nestler EJ: A molecular new therapeutic options. Curr Opin Psychiatry 2010; and cellular theory of depression. Arch Gen 23: 588-92. Psychiatry 1997; 54: 597-606. 57. Poulter MO, Du L, Weaver IC, et al.: GABAA recep- 47. Egan MF, Kojima M, Callicott JH, et al.: The BDNF tor promoter hypermethylation in suicide brain: im- val66met polymorphism affects activity-dependent plications for the involvement of epigenetic process- secretion of BDNF and human memory and hippo- es. Biol Psychiatry 2008; 64: 645-52. campal function. Cell 2003; 112: 257-69. 58. Keller S, Sarchiapone M, Zarrilli F, et al.: Increased 48. Taylor WD, Züchner S, McQuoid DR, Steffens DC, BDNF promoter methylation in the Wernicke area of Speer MC, Krishnan KR: Allelic differences in the suicide subjects. Arch Gen Psychiatry 2010; 67: brain-derived neurotrophic factor Val66Met poly- 258-67. morphism in late-life depression. Am J Geriatr 59. Sabunciyan S, Aryee MJ, Irizarry RA, et al.: Psychiatry 2007; 15: 850-7. Genome-wide DNA methylation scan in major de- 49. Verhagen M, van der Meij A, van Deurzen PA, et al.: pressive disorder. PLoS One 2012; 7: e34451. Meta-analysis of the BDNF Val66Met polymorphism 60. Shyn SI, Hamilton SP: The genetics of major depres- in major depressive disorder: effects of gender and sion: moving beyond the monoamine hypothesis. ethnicity. Mol Psychiatry 2010; 15: 260-271. Psychiatr Clin North Am 2010; 33: 125-40.