Eur Arch Psychiatry Clin Neurosci (2010) 260 (Suppl 2):S121–S126 DOI 10.1007/s00406-010-0152-8

REVIEW

Is there a personalized medicine for mood disorders? Lucie Bartova • Andreas Berger • Lukas Pezawas

Received: 2 August 2010 / Accepted: 31 August 2010 / Published online: 19 October 2010 Ó Springer-Verlag 2010

Abstract Major Depressive Disorder (MDD) and antidepressant therapy response are largely based on behavioral criteria, which are known to correlate at best modestly with biological measures. Therefore, it is not surprising that the search for peripheral biological markers (biomarkers) being assessed in distant biological systems such as body fluids has not yet resulted in clinically convincing measures for MDD diagnostics or treatment evaluation. Imaging genetics studies, however, have been successful in the search for intermediate imaging phenotypes of MDD and treatment response that are directly related to the neurobiological underpinnings of MDD, but are not suitable for a broad clinical use today. Hence, we argue that intermediate phenotypes derived from imaging genetics studies should be utilized as substitutes of behaviorally assessed psychiatric diagnoses or therapy response in the search for easily accessible peripheral biomarkers. This article will further cover the current state of peripheral and neural biomarker research. Keywords Antidepressants  Biomarkers  Drug response  Imaging genetics  Intermediate phenotypes  Major depressive disorder

Introduction The diagnosis of Major Depressive Disorder (MDD) is based on psychopathology and clinical information such as

L. Bartova  A. Berger  L. Pezawas (&) Division of Biological Psychiatry, Department of Psychiatry and Psychotherapy, Medical University of Vienna, Waehringer Guertel 18–20, 1090 Vienna, Austria e-mail: [email protected]

disease course and global functioning, which differs significantly from diagnostic systems in other medical fields, where biological parameters are being taken into account. Nonetheless, psychiatric diagnoses as currently operationalized in DSM-IV [1] and ICD-10 [2] have promoted the standardization and comparability of psychiatric syndromes and hence reflect major steps forward compared to the situation until the mid-nineties. While objective diagnostic measures are still lacking today, antidepressants are among the most prescribed medical drugs. The development of more efficient antidepressants, however, is hampered by several disadvantages of antidepressant drugs, such as a delayed onset, a variety of undesirable side effects, as well as still unsatisfying response rates [3]. Moreover, the assessment of antidepressant response rates is currently lacking objective parameters similar to the situation in psychiatric diagnostics and hence is likely to fail to detect differences in between antidepressant compounds [4]. Although neither the psychiatric diagnosis of MDD nor the assessment of antidepressant drug response is currently based upon objective biological markers, biological psychiatric research has made large strides forward during the last decades. Genetic research, which is motivated by a substantial heritability of MDD [5], has attempted to associate numerous genetic variants to behavioral depression phenotypes [6] and antidepressant drug response [3]. Furthermore, neuroimaging studies have highlighted that these variants impact on brain regions of emotion processing being implicated in depression and antidepressant drug response [3]. Such findings support the idea that the assessment of depression and antidepressant drug response could rely entirely on neural markers that are directly linked to the underlying pathology [3]. Moreover, given the high complexity and interconnectivity of systems biology, it seems further feasible to undertake an agnostic search for

123

S122

Eur Arch Psychiatry Clin Neurosci (2010) 260 (Suppl 2):S121–S126

peripheral markers that correlate with such neural markers. Over the last couple of years, a variety of different biological marker (biomarker) candidates have been proposed, ranging from neural to peripheral measures, which will be within the scope of this article.

pharmacological target of selective serotonin-reuptake inhibitors (SSRIs), which are the first-line treatment of MDD [3]. A variable number of tandem repeats (VNTR) polymorphism in the 50 promoter region (5-HTTLPR) of SLC6A4 that exhibits variable 5-HTT expression occurs in two allelic variants, the lower-expressing short (S) allele as well as a long (L) allele [9]. The S allele has been associated with anxiety-related temperamental traits and constitutes a potential risk factor for MDD in the presence of environmental adversity [10]. However, this view has been questioned [11], and a recent study suggests that contradictory results are depending on assessment methods being used [12]. Two-thirds of subjects of European ancestry are carriers of the common S allele; however, allele frequencies vary significantly with ethnicity [3] and interact in a complex way with cultural styles [13]. Pharmacogenetic studies demonstrated that S allele carriers show lower response rates in MDD under antidepressant drug treatment including SSRIs in comparison with the L/L genotype [3, 14], which has been related to a favorable side effect profile for antidepressants, and is thought to be a potential predictor of treatment tolerance [3]. However, it has become obvious that the effect size of 5-HTTLPR on treatment response is at best modest and estimated to be about 2–3%, which is smaller than initially expected [3], and hence of limited clinical value per se. Moreover, a recent meta-analysis has even found no significant effect of 5-HTTLPR on antidepressant treatment response at all [15]. The importance of the serotonergic system in depression is further highlighted by pharmacogenetic studies exploring associations between drug response and C(-1019)G HTR1A (rs6295), a functional single nucleotide polymorphism (SNP) within the promoter region of the serotonin receptor 1A gene (HTR1A). Approximately 50% of European ancestry and 21% of Asians are carriers of the G allele, which has been linked to MDD in association studies [3]. Pharmacogenetic studies investigating HTR1A effect on antidepressant drug response, however, have reported inconsistent results [3], but a recent meta-analysis suggests a possible effect of HTR1A on treatment efficacy [14]. Reversible and irreversible inhibitors of monoamine oxidase A (MAOA), a key enzyme in the degradation of monoamines, are highly effective antidepressants and are alternatives for SSRI treatment. Pharmacogenetic studies investigating effects of MAOA genotype on antidepressant treatment outcome have demonstrated sparse results so far [3]. Catechol-O-methyltransferase (COMT) is predominantly involved in the degradation of dopamine in the human cortex, which is lacking dopamine transporters. Dopaminergic neurotransmission has been linked to depression, since anhedonia is a key symptom of MDD and

Peripheral biomarkers A systems biology view of MDD and antidepressant response evokes the possibility that neural alterations of this disabling disorder and its response to treatment might be reflected in peripheral biological systems such as blood, which might provide clinically useful biomarkers derived from genomic, proteomic, metabolomic as well as neuroendocrinologic parameters. Typically, a biomarker is defined as a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention [7]. Cytochrome P450 isoenzymes, ABC transporters Genes encoding proteins involved into pharmacokinetics of antidepressants act on drug absorption, distribution, metabolism, and excretion. Most evidence is available for genetic variants affecting the cytochrome P450 (CYP) isoenzyme system, which impacts on antidepressant metabolism and thereby antidepressant plasma levels [8]. Similarly important appear functional polymorphisms of the multidrug-resistance gene ABCB1 encoding for p-glycoprotein (P-gp), an energy dependent efflux pump, which protects the brain from extrinsic substances and might therefore affect the blood–brain passage of antidepressants as well as drug response [8]. Both mechanisms have clearly demonstrated their clinical importance for a personalized drug treatment; however, since they are unrelated to the underlying etiology of depression, they are likely of limited value as potential biomarkers for diagnostics or clinical course of depression. Genetic variability impacting on pharmacodynamic drug targets on the other hand is typically related to both neurobiology of depression and treatment response [3]. Over the last couple of years, numerous studies have been conducted investigating various genetic variants and their potential role in drug response and neurobiology of depression, specifically variants of the monoamine and neurotrophin system [3]. Candidate genes involved in pharmacodynamic variability The serotonin transporter (5-HTT) is expressed by the human serotonin transporter gene (SLC6A4) and is the

123

Eur Arch Psychiatry Clin Neurosci (2010) 260 (Suppl 2):S121–S126

S123

euphoria can be induced by dopaminergic drugs [6]. This extracellularly located enzyme is encoded by the COMT gene, which exhibits several polymorphic sites such as Val158Met COMT (rs4680), a functional coding SNP occurring as a higher active Val allele and the significantly lower active Met allele [3]. Pharmacogenetic studies investigating Val158Met COMT effects on antidepressant treatment response however, have reported mixed results so far [3]. Decreased serum levels of brain-derived neurotrophic factor (BDNF) have been associated with depressive symptoms [16] (Table 1), nevertheless, a simple dependence of MDD on BDNF levels appears not to be the case [17]. BDNF, which is implicated in long-term potentiation and memory encoding, is expressed by the BDNF gene and contains several allelic variants including the Val66Met BDNF (rs6265), a functional coding SNP, which has received most attention [3]. While reported results of association studies with MDD or anxious temperament are controversial, and pharmacogenetic association studies have reported negative results [3], recent meta-analyses reported superior response rates for Met allele carriers with respect to antidepressant treatment [3, 14].

subserve as a biological signature of MDD. A recent study applying a classifier approach on blood gene expression profiles of a set of unmedicated MDD patients and control subjects reported such a biological signature of genes, all of which have not been related to depression neurobiology yet [18] (Table 1). This intriguing study claims to provide the first clinical useful diagnostic blood test for an endophenotype of MDD, which needs to be independently replicated before it gets adopted by clinical guidelines.

Gene expression profiling

Vitamins and trace metals

Blood cells can be viewed as biosensors, which adapt gene expression profiles depending on both individual genetic makeup and environmental factors, including surrounding body fluids and all effector molecules therein [18]. Hence, it seems plausible to detect gene expression profiles that

Dietary folic acid is essential in the development of the nervous system. Studies investigating the methioninehomocysteine-folate-B12 cycle suggest that folate deficiency is related to chronic MDD [19] and therapy resistance [20] (Table 1). Hence, folate therapy should be

Table 1 Peripheral and neurophysiological biomarker candidates of MDD

MDD major depressive disorder

Hypothalamic-pituitary-adrenocortical axis Abnormal hormonal response of the hypothalamic-pituitaryadrenocortical (HPA) axis is a frequent finding in depressive patients, which exhibit elevated plasma cortisol, corticotropin-releasing hormone (CRH), and vasopressin levels, all of which are involved in stress adaptation [8]. An increased stress hormone response following combined dexamethasone-CRH stimulation has been found in MDD patients, and a normalization of elevated stress hormone secretion is thought to be predictive for a sufficient antidepressant drug response, and has therefore been suggested as biomarker for the assessment of treatment outcome [8] (Table 1).

Biomarker candidate

Diagnostics

Antidepressant treatment

Combined gene expression profile

Increased risk [18]

–

Increased stress hormone response following dexamethasone/CRHa stimulation

Increased risk [8]

Normalization of increased stress hormone response following dexamethasone/CRH stimulation is associated with favorable response [8]

Decreased serum BDNFb levels

Increased risk [16]

Normalization of decreased serum BDNF levels is associated with favorable response [16]

Low serum folate levels

Increased risk [19]

Poor response [20, 21]

Increased serum homocysteine levels

Increased risk [22, 23]

–

Peripheral

Increased serum Vitamin B12 levels

–

Favorable response [19]

Low serum zinc levels

Increased risk [19, 24]

Normalization of low serum zinc levels is associated with favorable response [24]

–

Favorable response [8, 25]

a

CRH Corticotrophin-releasing hormone

b

Neurophysiological

c

Decreases in prefrontal quantitative EEGc cordance

BDNF Brain-derived neurotrophic factor EEG Electroencephalography

123

S124

Eur Arch Psychiatry Clin Neurosci (2010) 260 (Suppl 2):S121–S126

clinically considered [19, 21] and substituted along with antidepressant treatment in dosages sufficient to normalize increased homocysteine plasma levels [20]. Consistently, metabolically linked increased homocysteine levels should further be considered as biomarker for MDD [22, 23] (Table 1). Moreover, increased vitamin B12 levels have shown to be related to better treatment outcome in MDD [19] (Table 1). It is noteworthy that low levels of zinc have also been observed in treatment resistant depression [19] and have been shown to normalize during imipramine treatment in drug responders [24] (Table 1).

index biology is a favorable research approach [27]. While the phenotype in context of MDD is the clinical syndrome, the unknown underlying trait is called endophenotype, which is a putative entity in between an upstream genetic determinant and a downstream apparent phenotype and thus may give a hint about pathological mechanisms of the disease. Moreover, such an endophenotype has to be heritable, linked to genetic determinants, and associated with the disease [28]. Since the term ‘endophenotype’ implies that the corresponding mechanisms are hidden (and yet unknown), it has been suggested to substitute this term by using the term ‘intermediate phenotype’, which emphasizes the point that neural processes accessible to neuroimaging can characterize psychiatric syndromes in a more biological meaningful way and are hence closer related to the underlying genetic risk [26].

Neural biomarkers Neuroimaging methods such as Magnetic Resonance Imaging (MRI) have been rapidly evolving over the last decade and provide the opportunity to indirectly assess brain function with high spatial resolution. More mature methods such as electroencephalography (EEG) have recently refined their methodology and now offer interesting new quantitative measures suggesting EEG recordings as a biomarker for antidepressant drug response. Analogous to functional MRI (fMRI) studies, which reported associations between alterations in neural activity in the lateral prefrontal cortex (PFC) and treatment response, EEG tomography studies reported decreases in quantitative prefrontal cordance that are detectable after treatment initiation and have been related to clinical outcome in treatment trials for MDD [8, 25] (Table 1). While EEG offers high temporal resolution, it suffers from poor spatial resolution. Here, functional and anatomical MRI provides superior insights, and numerous publications have demonstrated neurobiological alterations in limbic structures of the brain such as the amygdala, hippocampus, and anterior cingulate cortex in MDD [6]. Imaging genetics Genetic effects of MDD-related candidate genes are not simply expressed at a behavioral level, but are mediated by molecular and cellular mechanisms, which constitute a formidable challenge for scientists dealing with highly variable behavioral phenotypes of depression caused by complex underlying genetic mechanisms such as pleiotropy or variable penetrance. The need for reductionism has led to the development of research strategies such as imaging genetics [26] that applies anatomical or functional imaging technologies as surrogate phenotypes to evaluate genetic variation [6]. A vast number of studies have adopted such a strategy [26] and demonstrated that the study of underlying quantitative traits that more directly

123

Amygdala The almond-shaped amygdala is located in the medial temporal lobe and comprises several nuclei with specific functions, the distinction of which is beyond current MRI resolution. The primary function of the highly interconnected amygdala is fear conditioning and response [6]. Numerous fMRI studies in MDD patients have reported amygdala hyperreactivity (Table 2), which is to a large extent congruent with current mood [29], but seems to be also partly mood-state independent [30]. In healthy subjects as well as in MDD patients, the S allele of the 5-HTTLPR polymorphism has repeatedly been associated with increased activation of the amygdala, while studies regarding C(-1019)G HTR1A are still controversially discussed [6] (Table 2). Furthermore, the Met allele of Val158Met COMT has been associated with increased amygdala activation and volume (Table 2), with a minority of studies yielding opposite or lacking effects [6]. Moreover, the BDNF Met allele has also been reported to alleviate the effects of the 5-HTTLPR S allele on amygdala volume, reflecting genetic interactions (epistasis) [31]. With regard to antidepressive drug therapy, it has been shown that acute SSRI treatment results in increased amygdala activation, while chronic SSRI administration is associated with an attenuated amygdala response [3] (Table 2). Anterior cingulate cortex The belt-shaped cingulate cortex is wrapped around the corpus callosum and comprises four distinct functional regions based on pharmacological receptor distributions [6]. The anterior cingulate cortex (ACC), which is highly interconnected with the amygdala, is believed to be important for the processing of emotion and a primary region of interest in mood disorders. Specifically, the

Eur Arch Psychiatry Clin Neurosci (2010) 260 (Suppl 2):S121–S126 Table 2 Intermediate phenotypes of MDD

S125

Amygdala

sACCa

Hippocampus

MDD

Increased activation

Increased activation

Decreased volume

Acute antidepressant administration

Increased activation

Increased activation

Increased activation

Chronic antidepressant administration

Decreased activation

Decreased activation

Decreased activation

S allele of 5-HTTLPR

Increased activation, decreased volume

Increased activation, decreased volume

Met allele of Val158Met COMT (rs4680)

Increased activation, increased volume

Decreased activation related to Val allele

Phenotype

Genotype Findings are limited to studies with sufficient evidence of alterations in brain function and morphology [3, 6] MDD major depressive disorder a

sACC subgenual anterior cingulate cortex

Met allele of Val66Met BDNF (rs6265)

subgenual ACC (sACC), the rostral subregion of the ACC, exhibits prominent metabolic as well as anatomical alterations in patients [6], similarly to functional studies of transiently induced sadness in healthy subjects [32]. Imaging genetics studies in healthy subjects have reported decreased sACC volume and increased sACC activation in 5-HTTLPR S allele carriers [6] (Table 2). With regard to Val158Met COMT, decreased activation has been associated with the Val allele [6] (Table 2). While short-term SSRI treatment in healthy subjects leads to an increased sACC activation, effective long-term antidepressant treatment as well as deep brain stimulation results in decreased sACC activation [3] (Table 2). It is noteworthy that the ACC features the highest density of 5-HTT of the human cortex [6]. However, 5-HTT is not only expressed in neurons, but also in peripheral cells such as blood platelets, and is highly correlated with synaptosomal 5-HT uptake [33], which is inhibited by SSRI treatment. Preliminary findings have revealed that sACC activation can be predicted by 5HT platelet uptake rates in healthy and unmedicated subjects [34], which suggests that platelet 5-HT uptake might be an eligible biomarker candidate for depression and treatment response. Hippocampus The seahorse-shaped hippocampus located in the medial temporal lobe receives inputs from sensory, association and prefrontal cortical areas, projects to neocortical areas, and has been demonstrated to be of essential importance in the formation of memories as well as stress adaptation [6]. With regard to MDD, the most frequent finding is hippocampal volume loss [6] (Table 2). Since it has also been shown that hippocampal volume loss is induced within stress paradigms [6], the question arises, whether hippocampal volume loss can qualify as a trait marker for

Decreased volume

MDD or simply reflects accumulated stress occurring during untreated episodes of depression [30]. While imaging genetics studies investigating Val66Met BDNF support the concept of hippocampal volume loss being a trait marker of depression, since the defective Met allele has shown to decrease hippocampal volume [6] (Table 2), other studies suggest a neuroprotective effect of antidepressant treatment, because hippocampal volume loss has been found to correlate with the duration of untreated depression [35].

Conclusion While peripheral biomarker candidates raise the possibility to detect the biological signature of a brain disease such as MDD in distant biological systems, neural biomarker candidates like neurophysiologic or imaging markers are attempting to index biological alterations that are directly related to the disease. However, neither approach has been widely incorporated in clinical routine with the exception of vitamin B12 measurements and cytochrome P450 isoenzyme diagnostics. Several reasons may account for the lack of clinical utilization in MDD diagnostics and treatment: (1) Lacking specificity or sensitivity for MDD or treatment outcome, which applies basically to all currently proposed markers. (2) Economic limitations going along with the assessment of expensive and time-consuming analytical procedures of neural imaging biomarker candidates. (3) The technical challenges inherent in imaging procedures, which are currently only used for research purposes and have not yet been developed for other environments so far. The shortfall of available objective criteria of MDD diagnostics and treatment response along with its associated high biological variability may have contributed to a low specificity and sensitivity and calls the search for a

123

S126

Eur Arch Psychiatry Clin Neurosci (2010) 260 (Suppl 2):S121–S126

single clinical useful biomarker for MDD into question. Imaging genetics studies have suggested that intermediate phenotypes derived from neuroimaging studies are superior to behaviorally defined psychiatric diagnoses since they are closer tied to the underlying neurobiology [36]. Hence, imaging intermediate phenotypes might be utilized as surrogate endpoints in the search for agnostic peripheral biomarkers, which are easy accessible and thereby more suitable for widespread clinical use.

15. Taylor MJ, Sen S, Bhagwagar Z (2010) Antidepressant response and the serotonin transporter gene-linked polymorphic region. Biol Psychiatry 16. Sen S, Duman R, Sanacora G (2008) Serum brain-derived neurotrophic factor, depression, and antidepressant medications: meta-analyses and implications. Biol Psychiatry 64(6):527–532 17. Groves JO (2007) Is it time to reassess the BDNF hypothesis of depression? Mol Psychiatry 12(12):1079–1088 18. Spijker S et al (2010) Stimulated gene expression profiles as a blood marker of major depressive disorder. Biol Psychiatry 68(2): 179–186 19. Mossner R et al (2007) Consensus paper of the WFSBP task force on biological markers: biological markers in depression. World J Biol Psychiatry 8(3):141–174 20. Coppen A, Bailey J (2000) Enhancement of the antidepressant action of fluoxetine by folic acid: a randomised, placebo controlled trial. J Affect Disord 60(2):121–130 21. Morris DW, Trivedi MH, Rush AJ (2008) Folate and unipolar depression. J Altern Complement Med 14(3):277–285 22. Almeida OP et al (2008) Homocysteine and depression in later life. Arch Gen Psychiatry 65(11):1286–1294 23. Tolmunen T et al (2004) Association between depressive symptoms and serum concentrations of homocysteine in men: a population study. Am J Clin Nutr 80(6):1574–1578 24. Siwek M et al (2010) Serum zinc level in depressed patients during zinc supplementation of imipramine treatment. J Affect Disord. doi: 10.1016/j.jad.2010.04.024 25. Bares M et al (2010) The change of prefrontal QEEG theta cordance as a predictor of response to bupropion treatment in patients who had failed to respond to previous antidepressant treatments. Eur Neuropsychopharmacol 20(7):459–466 26. Meyer-Lindenberg A, Pezawas L (2010) Imaging genetics: progressing by leaps and bounds. Neuroimage 53:801–803 27. Meyer-Lindenberg A, Weinberger DR (2006) Intermediate phenotypes and genetic mechanisms of psychiatric disorders. Nat Rev Neurosci 7(10):818–827 28. Gottesman II, Gould TD (2003) The endophenotype concept in psychiatry: etymology and strategic intentions. Am J Psychiatry 160(4):636–645 29. Wang L, La Bar KS, McCarthy G (2006) Mood alters amygdala activation to sad distractors during an attentional task. Biol Psychiatry 60(10):1139–1146 30. Savitz JB, Drevets WC (2009) Imaging phenotypes of major depressive disorder: genetic correlates. Neuroscience 164(1): 300–330 31. Pezawas L et al (2008) Evidence of biologic epistasis between BDNF and SLC6A4 and implications for depression. Mol Psychiatry 13(7):709–716 32. Mayberg HS et al (1999) Reciprocal limbic-cortical function and negative mood: converging PET findings in depression and normal sadness. Am J Psychiatry 156(5):675–682 33. Rausch JL et al (2005) Serotonin transport kinetics correlated between human platelets and brain synaptosomes. Psychopharmacology (Berl) 180(3):391–398 34. Scharinger C et al (2010) Peripheral 5-HT uptake reflects cingulate cortex activation (Abstract). Neuroimage 35. Sheline YI, Gado MH, Kraemer HC (2003) Untreated depression and hippocampal volume loss. Am J Psychiatry 160(8): 1516–1518 36. Meyer-Lindenberg A (2010) Intermediate or brainless phenotypes for psychiatric research? Psychol Med 40(7):1057–1062

Acknowledgments This article is part of the educational efforts of the Special Research Project SFB-35 (Project No. F3514-B11 and F3506-B11), funded by the Austrian Science Fund (FWF). Conflict of interest All authors declare that they have no conflict of interest.

References 1. APA (2000) Diagnostic and statistical manual of mental disorders: DSM-IV-TR. 2000. American Psychiatric Association, Washington, DC 2. WHO (2007) International statistical classification of disease and related health problems. World Health Organization 3. Rabl U et al (2010) Imaging genetics of mood disorders: implications for the understanding of variable treatment response. Expert Rev Clin Pharmacol 3(4):471–489 4. Della Pasqua O, Santen GW, Danhof M (2010) The missing link between clinical endpoints and drug targets in depression. Trends Pharmacol Sci 31(4):144–152 5. Wong ML, Licinio J (2001) Research and treatment approaches to depression. Nat Rev Neurosci 2(5):343–351 6. Scharinger C et al (2010) Imaging genetics of mood disorders. Neuroimage 53(3):810–821 7. Atkinson AJ et al (2001) Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther 69(3):89–95 8. Holsboer F (2008) How can we realize the promise of personalized antidepressant medicines? Nat Rev Neurosci 9(8):638–646 9. Lesch KP et al (1996) Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science 274(5292):1527–1531 10. Caspi A et al (2003) Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science 301(5631):386–389 11. Risch N et al (2009) Interaction between the serotonin transporter gene (5-HTTLPR), stressful life events, and risk of depression: a meta-analysis. JAMA 301(23):2462–2471 12. Uher R, McGuffin P (2010) The moderation by the serotonin transporter gene of environmental adversity in the etiology of depression: 2009 update. Mol Psychiatry 15(1):18–22 13. Chiao JY, Blizinsky KD (2010) Culture-gene coevolution of individualism-collectivism and the serotonin transporter gene. Proc Biol Sci 277(1681):529–537 14. Kato M, Serretti A (2010) Review and meta-analysis of antidepressant pharmacogenetic findings in major depressive disorder. Mol Psychiatry 15(5):473–500

123