Effect of Tryptophan Hydroxylase-2 Gene Variants on Suicide Risk in Major Depression

Effect of Tryptophan Hydroxylase-2 Gene Variants on Suicide Risk in Major Depression Catalina Lopez de Lara, Jelena Brezo, Guy Rouleau, Alain Lesage, Mimi Dumont, Martin Alda, Chawki Benkelfat, and Gustavo Turecki Background: Suicide and depressive disorders are strongly associated, yet not all depressed patients commit suicide. Genetic factors may partly explain this difference. We investigated whether variation at the tryptophan hydroxylase-2 (TPH2) gene and its 5’ upstream region may predispose to suicide in major depressive disorder (MDD) and whether this predisposition is mediated by impulsive-aggressive behaviors (IABs). Methods: We genotyped 14 single nucleotide polymorphisms (SNPs) in 259 depressed subjects, 114 of which committed suicide while depressed. Phenotypic assessments were carried out by means of proxy-based interviews. Single-marker and haplotype association analyses were conducted. Differences in behavioral and personality traits according to genotypic variation were investigated, as well as genetic and clinical predictors of suicide. Results: We found two upstream and two intronic SNPs associated with suicide. No direct effect of these variants was observed on IABs. However, a slight association with reward dependence scores was found. Controlling for suicide risk factors, two SNPs (rs4448731 and rs4641527) significantly predicted suicide, along with cluster B personality disorders and family history of suicide. Conclusions: The TPH2 gene and its 5’ upstream region variants may be involved in the predisposition to suicide in MDD; however, our findings do not support the role of IABs as mediators. Key Words: Aggressive-impulsive behavior, genetic predisposition, major depressive disorder, single nucleotide polymorphisms, suicide, tryptophan hydroxylase-2

S

tudies suggest a strong association between suicide and mental illness, particularly with depressive disorders (Cavanagh et al. 2003). The lifetime risk of suicide in patients with major depressive disorder (MDD) has recently been revised to 3.4% (Blair-West et al. 1999). This figure remains of considerable importance, especially if we take into account the estimated high lifetime prevalence of MDD in the United States alone (16.2%) (Kessler et al. 2003). Thus far, it is still unclear why some depressed patients do commit suicide, whereas others with the same psychiatric condition do not. Several lines of evidence suggest that biological and genetic factors are responsible for at least part of the variability in suicide risk (Mann 1998; Turecki 2001). Findings from neurobiological studies have related suicide, MDD, and impulsive-aggressive behaviors (IABs) to various aspects of altered brain serotonergic neurotransmission (Coccaro 1989; Mann et al. 2001), including brain serotonin synthesis (Leyton et al. 2001, 2006; Rosa-Neto et al. 2004). These findings point to serotonergic genes, including the tryptophan hydroxylase-2 (TPH2) gene, as promising candidates in suicide studies. The TPH2 gene is located on chromosome 12q21.1, spans a region of approximately 93.5 kilobase

From the McGill Group for Suicide Studies (CLdL, JB, MD, GT), Douglas Hospital Research Center; Study of Brain Diseases, Notre Dame Hospital; (GR), McGill University; Department of Psychiatry (AL), Centre de Recherche Fernand-Seguin, Hôpital L-H Lafontaine, Université de Montréal; and Department of Psychiatry (MA, CB), Allan Memorial Institute, McGill University, Montreal, Canada. Address reprint requests to Gustavo Turecki, M.D., Ph.D., Director, McGill Group for Suicide Studies, Douglas Hospital Research Center, McGill University, 6875 LaSalle Blvd, Montreal, QC H4H 1R3, Canada; E-mail: [email protected]. Received April 28, 2006; revised July 27, 2006; accepted September 3, 2006.

0006-3223/07/$32.00 doi:10.1016/j.biopsych.2006.09.008

(kb), and is composed of 11 exons. For many years, researchers believed that there was only one form of tryptophan hydroxylase (TPH) responsible for the synthesis of serotonin in the brain and peripheral tissues. It was not until recently that a second isoform, designated as TPH2, was identified in mice lacking the previously known TPH (now called TPH1). Moreover, TPH2 was found to be expressed exclusively in the brain of both mice and humans (Walther and Bader 2003; Zill et al. 2004b), with a predominant expression in the raphe nuclei, where the vast majority of serotonergic cell bodies are located (Patel et al. 2004; Walther and Bader 2003; Zill et al. 2007). A recent report suggests that TPH2 gene expression in the dorsal raphe nuclei of depressed suicides is upregulated, which may represent a compensatory mechanism in response to reduced serotonergic activity (Bach-Mizrachi et al. 2006). It is noteworthy that some TPH2 upstream regulatory variants have been shown to affect the responsiveness of certain brain regions involved in emotional regulation (Brown et al. 2005; Canli et al. 2005). While some studies have supported a possible role of TPH2 gene variants in MDD (Zhang et al. 2005; Zhou et al. 2005; Zill et al. 2004a), few studies have done so exploring the TPH2 gene in suicidal behavior, with the majority of them investigating suicide attempters (De Luca et al. 2004, 2005; Zhou et al. 2005) rather than suicide completers (Zill et al. 2004c), which are believed to be the most extreme phenotype. Furthermore, none of these studies have examined the role of TPH2 genetic variants while taking into account other well-established risk factors increasing susceptibility to suicide, such as IABs. More importantly, most of these studies have not controlled for existing psychopathology, making it difficult to disentangle the effects that may be attributable to suicide as opposed to those related to the underlying psychiatric condition. In this study, we investigated whether genetic variants distributed across the TPH2 gene and its 5’ upstream region may predispose to suicide in the context of MDD, and if so, whether this predisposition is mediated by IABs. BIOL PSYCHIATRY 2007;62:72– 80 © 2007 Society of Biological Psychiatry

C. Lopez de Lara et al.

Methods and Materials Sample The sample investigated consisted of 259 depressed subjects, including 114 who committed suicide during an episode of MDD (98 male subjects, 16 female subjects) and 145 depressed control individuals (124 male control individuals, 21 female control individuals). Ninety-eight percent of the suicide completers and control individuals in our sample were of French-Canadian origin. Suicide cases included in this study were 18 years-of-age and older, whose cause of death by suicide was determined by the Quebec coroner’s office. Proxy-based interviews were carried out for clinical assessments in both groups (see below). All cases met DSM-IV diagnostic criteria for MDD or depression not otherwise specified. We included suicide victims with the latter diagnosis because of the high probability that they had suffered from MDD, but due to the low sensitivity of the psychological autopsy procedure, especially for symptoms displayed immediately before death, they were not identified as such (Ernst et al. 2004). Participants in our control group were living individuals, also aged 18 years and older, suffering from MDD according to the DSM-IV criteria and whose condition was severe enough to require follow-up in a specialized psychiatric outpatient clinic. Suicide cases and control individuals who met criteria for bipolar or any psychotic disorder were excluded from the study to increase the likelihood that the depressive episode and not another condition might have been the trigger of the suicidal act. All cases and controls were recruited from 2000 to 2004. Prior to inclusion in the study, all individuals and family members of suicide completers provided written informed consent. This study was approved by the McGill University Institutional Review Board of the Faculty of Medicine. Suicide Method We classified suicide methods as violent or nonviolent according to the classification used in previous studies (Denning et al. 2000; Dumais et al. 2005). Drug overdose, carbon monoxide poisoning, and drowning were considered nonviolent. All other methods were classified as violent. The methods used to commit suicide were hanging and strangulation (56.1%), carbon monoxide poisoning (14.0%), drug overdose (9.6%), shooting (8.8%), jumping from a height or in front of a metro car (4.4%), drowning (4.4%), penetrating lesions (1.8%), and others (.9%). These figures are similar to the distribution of suicide methods in the Quebec general population (St-Laurent and Bouchard 2004). Relation of Sample to Other Reports Part of this sample has been previously described in studies investigating clinical risk factors for suicide (Dumais et al. 2005). However, only one other gene has been investigated in a subsample derived from the sample studied in this report (Lopez de Lara et al. 2006). Clinical Assessment Psychiatric illnesses and personality disorders were assessed by means of the psychological autopsy procedure, which has been well validated in earlier studies (Kelly and Mann 1996; Schneider et al. 2004). Briefly, a family member best acquainted with the deceased was identified and subsequently interviewed using a series of structured instruments adapted for use with a proxy. To ensure comparability between groups, the same proxy-based assessment procedure was also performed for con-

BIOL PSYCHIATRY 2007;62:72– 80 73 trol subjects. In this study, families of the suicide victims were first approached at the Montreal morgue and contacted again for interview after approximately 4 months. Axis I and Axis II psychopathologies were determined using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) (First et al. 2001; Spitzer et al. 1992) and the Structured Clinical Interview for DSM-IV Personality Disorders (SCID-II) (First et al. 1995), respectively. Based on the information retrieved, coroner’s notes, and medical records, a case report was written and reviewed by a panel of clinicians who reached a consensus on DSM-IV Axis I and Axis II diagnoses. By means of structured questionnaires, information on specific demographic and clinical variables, such as family history of suicide and history of physical or sexual abuse, was also obtained. Assessment of Personality Traits Information on related behavioral and personality traits was obtained with informant versions of the following instruments: Barratt Impulsiveness Scale (BIS) (Patton et al. 1995); BussDurkee Hostility Inventory (BDHI) (Buss and Durkee 1957); Brown-Goodwin Assessment of Lifetime History of Aggression (BGHA) (Brown and Goodwin 1986); and the Temperament and Character Inventory (TCI) (Cloninger et al. 1994). Estimates of internal consistency (␣) for most of these instruments were: BGHA, .88; BIS, .89; and TCI, .73 to .88 for the four temperament and three character dimensions. These estimates were overall satisfactory and similar to those reported in the literature (Brent et al. 1994). Interrater Reliability and Validity of Proxy-Based Personality Assessments More detailed information on the reliability and validity of the proxy-based assessments used in this study is provided elsewhere (Dumais et al. 2005). Estimated kappa coefficients for interrater reliability have proven to be excellent for the following specific diagnoses: MDD, .96; alcohol abuse/dependence, .98; drug abuse/dependence, 1.0; and cluster B personality disorders (cluster B PD), 1.0. No significant differences were found between the information provided by two different informants on the same subject (p values ranged from .25 to .94). Moreover, in the depressed control group, no significant differences were found between the information gathered from the informant and that obtained from the individual (p values ranged from .67 to .98). These results support the validity of proxy-based personality assessments in studies on suicidal behavior. Genetic Variants We investigated 14 single nucleotide polymorphisms (SNPs) selected from the publicly available SNP database (dbSNP) of the National Center for Biotechnology Information (NCBI) and from the literature. Selected SNPs were distributed across the TPH2 gene and within the 5’ upstream region. These SNPs were chosen based on their location, average estimated heterozygosity, and validation status reported in dbSNP. We also included the intronic SNP found to be associated with MDD and suicide in separate studies (rs1386494) (Zill et al. 2004a, 2004c), as well as the functional SNP described by Zhang et al. (2005). Of the 14 SNPs investigated, 9 were intronic (introns 1, 4, 5 [3], 6, 7, and 8 [2]), 2 were exonic (exons 8 and 11—nonsynonymous), and 3 were in the 5’ upstream region of the gene (3.5 kb, 2.5 kb, and 1.3 kb from the transcription start site, respectively) (Figure 1). www.sobp.org/journal

74 BIOL PSYCHIATRY 2007;62:72– 80

Figure 1. Schematic representation of investigated SNPs distributed across the TPH2 gene and its 5’ upstream region. Interrupted lines represent the upstream and downstream region of the TPH2 gene. Exons 1 to 11 are represented as vertical rectangles. Arrows represent the approximate location of the SNPs investigated in this study. SNP, single nucleotide polymorphism; TPH2, tryptophan hydroxylase-2.

All SNPs are named according to the identification number given by dbSNP, except for the SNP in exon 11, identified as G1463A. Genotyping Genomic DNA was extracted from blood or frozen brain tissue according to standard procedures (Sambrook et al. 1989). Genotyping was carried out using the commercially available SNaPshot method (Applied Biosystems, Foster City, California). Initial amplification of the target genomic regions was carried out as previously described (Lopez de Lara et al. 2006). Specificity of the designed primers (Alpha DNA, Montreal, Canada) to TPH2 sequences was verified by a Basic Local Alignment Search Tool (BLAST) search using Ensembl (a joint project between EMBL— the European Bioinformatics Institute [EBI] and the Wellcome Trust Sanger Institute [WTSI] [Hinxton, Cambridge, United Kingdom]). The generated data were analyzed using the GeneMapper software v.3.7 (Applied Biosystems). The genotype completion rates for all markers ranged from 90% to 99%, except for the SNP G1463A (88%). Of the 14 SNPs genotyped, we found 2 nonpolymorphic SNPs located in exon 8 (rs2887147) and exon 11 (G1463A). Hence, these SNPs were excluded from further analyses. All genotype frequencies in the control group were in conformance with Hardy-Weinberg equilibrium (HWE). Data Analysis Univariate comparisons were carried out by two-tailed Student t tests and Fisher exact tests for continuous and categorical variables, respectively. Deviations from Hardy-Weinberg equilibrium were examined using the exact test implemented in the program Haploview (Broad Institute, Cambridge, Massachusetts) version 3.2 (Barrett et al. 2005). Allelic and genotypic frequency distributions were compared between groups by chi-square tests for independence. Haplotype associations as well as linkage disequilibrium (LD) between pairs of SNPs were determined by using Haploview v.3.2. Block boundaries were defined by using a minimum average D’ value of .75 in each block. The Bonferroni method for multiple testing corrections was applied to genetic data. Hierarchical logistic regression analyses were performed to evaluate the unique contribution of identified risk factors in the prediction of completed suicide. Possible interactions between genetic variants and other risk factors were also investigated using logistic regression analyses. To further examine whether IABs may explain the relationship between TPH2 genetic variants and suicide completion, a series of regression analyses to test mediation effects were performed following the method described by Baron and Kenny (1986) www.sobp.org/journal

C. Lopez de Lara et al. (Frazier et al. 2004). Accordingly, there is evidence for mediation when the following four conditions have been met: 1) there is a significant relationship between the predictor (genetic variants) and the outcome (suicide); 2) the predictor is significantly related to the mediator (measures of IABs); 3) the mediator is significantly related to the outcome; and 4) the strength of the relationship between the predictor and the outcome is significantly decreased when the mediator is included in the analysis. Correlations between variables were calculated by Pearson’s r. The SPSS package (version 11.0 for Windows; SPSS, Chicago, Illinois) was used for statistical analyses. Our alpha risk was set at 5%.

Results Suicide cases and control individuals did not significantly differ on any of the demographic and clinical variables analyzed, except for family history of suicide, which was more frequently reported in suicide cases (Table 1). Our analyses showed that depressed suicides had higher comorbidity with current (last 6 months) (32.5% vs. 17.5%, p ⫽ .013) and lifetime substance use disorders (43.9% vs. 29.1%, p ⫽ .034) than control depressed patients. No significant differences were found with regard to anxiety disorders (Table 1). Higher prevalence of cluster B personality disorders was also noted in depressed suicide cases, relative to control subjects (23.9% vs. 5.9%, p ⬍ .001) (Table 1). Moreover, cases in this study exhibited higher levels of lifetime aggression than control individuals, as determined by the BGHA (11.2 [mean] ⫾ 11.7 [SD] vs. 7.9 ⫾ 9.1, p ⫽ .048) and the BDHI (33.3 ⫾ 13.4 vs. 29.8 ⫾ 11.6, p ⫽ .068). However, the latter did not reach statistical significance. Mean scores on measures of impulsivity using total BIS scores did not differ between groups. Additionally, depressed suicides showed slightly lower, yet significant, scores on the TCI reward dependence dimension in comparison with depressed control subjects (14.2 ⫾ 3.9 vs. 15.4 ⫾ 4.5, p ⫽ .048) (Table 1). Analyses of the distribution of allelic and genotypic frequencies identified four SNPs to be significantly associated with depressed suicide cases. Two were located in the TPH2 5’ upstream region (at 3.5 kb and 2.5 kb from the transcription start site) and two in introns 1 and 8 (Table 2). For the upstream region SNPs, we observed a significantly higher frequency of the T allele of SNP rs4448731 (p ⫽ .003) and the G allele of SNP rs6582071 (p ⫽ .004) in the depressed suicide group. As for the intronic SNPs, we found more frequently the G allele of SNP rs4641527 (intron 1) (p ⬍ .001) and the C allele of SNP rs1386497 (intron 8) (p ⬍ .001) in depressed suicides than depressed control subjects. After multiple testing corrections, these four SNPs remained statistically significant (Bonferroni corrected pvalue ⫽ .004 for 12 loci tested). The distribution of genotype frequencies between groups at these four loci was also significantly different (Table 2). We observed a higher frequency of homozygotes for the T allele of SNP rs4448731 in depressed suicides relative to depressed control subjects (p ⫽ .002). In addition, homozygotes for the G allele of SNPs rs6582071 and rs4641527 were found more frequently in the depressed suicide group (p ⫽ .021 and p ⫽ .001, respectively). In contrast, homozygotes, as well as heterozygotes for the C allele of SNP rs1386497, were observed more often in depressed suicides than depressed control subjects (p ⬍ .001). Furthermore, the indicated differences in genotype frequencies for these SNPs were significantly greater when genotypes were grouped in two categories based on their reported association with suicide (Table 2). After

BIOL PSYCHIATRY 2007;62:72– 80 75

C. Lopez de Lara et al. Table 1. Demographic and Clinical Characteristics of the Study Population

Variable Name Age (Mean ⫾ SD) Gender Male Female White % Catholic Completed College Employed Married (or Common Law) Parent of at Least One Child Living Alone Family History of Suicide History of Physical or Sexual Abuse Last 6 Months Prevalence Substance Use Disorders (Alcohol/Drugs) Anxiety Disorders Lifetime Prevalence Substance Use Disorders (Alcohol/Drugs) Anxiety Disorders Cluster B Personality Disorder Personality Trait Measures BGHA BDHI BIS TCI Novelty Seeking Harm Avoidance Reward Dependence Persistence Self-Directedness Cooperativeness Self-Transcendence

Depressed Suicides

Depressed Control Subjects

Mean ⫾ SD or %

Mean ⫾ SD or %

41.6 ⫾ 14.6

41.4 ⫾ 10.2

86.0 14.0 98.1 95.3 22.5 55.9 42.9 61.9 38.9 16.5 28.4

85.5 14.5 98.3 89.0 27.6 59.8 34.7 50.0 41.1 5.6 20.7

32.5 21.1

t-Test Statistic/ Fisher Exact Test .122

df

p-Value

195

.903

— — — — — — — — — — —

— — — — — — — — — — —

1.000

17.5 12.6

— —

— —

.013a .107

43.9 22.8 23.9

29.1 14.6 5.9

— — —

— — —

.034a .164 ⬍.001b

11.2 ⫾ 11.7 33.3 ⫾ 13.4 66.5 ⫾ 14.1

7.9 ⫾ 9.1 29.8 ⫾ 11.6 66.0 ⫾ 12.8

20.9 ⫾ 7.3 17.8 ⫾ 7.9 14.2 ⫾ 3.9 6.8 ⫾ 2.8 25.0 ⫾ 8.0 25.9 ⫾ 6.9 10.7 ⫾ 6.1

19.8 ⫾ 6.7 19.7 ⫾ 7.2 15.4 ⫾ 4.5 6.5 ⫾ 2.6 25.6 ⫾ 7.7 26.6 ⫾ 5.4 11.5 ⫾ 5.7

1.000 .129 .437 .585 .220 .081 .775 .021a .283

1.990 1.836 .245

158.7 177.1 193.0

.048a .068 .806

1.124 ⫺1.687 ⫺1.994 .917 ⫺.507 ⫺.723 ⫺.901

186.0 186.0 180.7 186.0 186.0 186.0 186.0

.262 .093 .048a .360 .613 .471 .369

BGHA, Brown-Goodwin Assessment of Lifetime History of Aggression; BDHI, Buss-Durkee Hostility Inventory; BIS, Barratt Impulsiveness Scale; TCI, Temperament and Character Inventory. a Statistically significant at the .05 level. b Statistically significant at the .001 level.

applying the above-mentioned Bonferroni correction for 12 loci tested, the results on genotype distributions were still significant, except for the SNP rs6582071. These suicide-related TPH2 SNPs were not associated with the use of more violent methods by depressed suicides (data not shown). We determined the LD pattern and haplotype frequencies for the SNPs analyzed in this study. We identified two blocks of LD in our sample. The first block was composed of SNPs rs4448731, rs6582071, rs7963803, rs4641527, and rs2129575. D’ values ranged between .43 and 1.0, with an average D’ value in this block of .76. The second block consisted of SNPs rs1386496, rs1386494, rs1843812, rs1023990, rs1386497, and rs1487276. In this block, D’ values ranged between .55 and 1.0, with an average of .88. Haplotype frequencies were estimated and association tests performed on the mentioned blocks of LD. Our results (Table 3) for the first block initially revealed four haplotypes associated with suicide; however, after correction for multiple testing (Bonferroni corrected p value ⫽ .003 for 15 haplotypes tested), there were only two haplotypes that remained significant (TGCGG [p ⫽ .002] and TGAGG [p ⫽ .001]). Analysis of the second block of LD showed two haplotypes initially associated with depressed suicides (TGGTCG and TGGCCG). However,

these haplotypes were not significant after multiple testing corrections by the Bonferroni method. We then investigated whether the two upstream regions (rs4448731 and rs6582071) and two intronic (rs4641527 and rs1386497) SNPs associated with depressed suicides could increase the susceptibility to suicide through influencing levels of IABs, as well as temperament dimensions. To do so, we compared the mean scores of the corresponding questionnaires according to genotypic variation at these four SNPs in the total sample (Table 4). We did not find any significant differences on levels of IABs measured by BGHA, BDHI, and BIS scales. However, homozygotes for the G allele of SNP rs4641527 were related to lower levels of reward dependence compared with the other genotypes (p ⫽ .039). Finally, we investigated the unique contribution of risk factors individually identified in this study by carrying out a hierarchical logistic regression analysis, which included seven steps. Because of the high correlation between SNPs rs6582071 and rs4641527 (r ⫽ .85, p ⬍ .001), we only included in the analysis the SNP rs4641527, which showed a stronger association with suicide. Similarly, current and lifetime prevalence of substance use disorders were highly correlated (r ⫽ .63, p ⬍ .001); therefore, www.sobp.org/journal

76 BIOL PSYCHIATRY 2007;62:72– 80

C. Lopez de Lara et al.

Table 2. Allelic Frequencies for SNPs in the TPH2 Gene and Its 5’ Upstream Region Between Depressed Suicides and Depressed Control Individuals, as well as Genotypic Distributions of the Four TPH2 SNPs Associated with Suicide Marker

Location

Allele

Suicide (%)

Control (%)

Genotype

Suicide (%)

Control (%)

Chi-Square

rs4448731

Upstream

C T

76 (37.3) 128 (62.7)

147 (50.7) 143 (49.3)

C/C C/T T/T C/T, C/C A/A A/G G/G A/G, A/A

23 (22.5) 30 (29.4) 49 (48.0) 53 (52.0) 2 (1.9) 25 (23.8) 78 (74.3) 27 (25.7)

40 (27.6) 67 (46.2) 38 (26.2) 107 (73.8) 10 (7.0) 49 (34.3) 84 (58.7) 59 (41.3)

8.73a

.003g

13.00b 12.51c 8.13a

.002g ⬍.001g .004g

7.70b 6.46d 3.12a

.021 .011 .077

G/G G/T T/T G/T, T/T

85 (80.2) 16 (15.1) 5 (4.7) 21 (19.8)

84 (57.9) 50 (34.5) 11 (7.6) 61 (42.1)

12.30a

⬍.001g

14.05b 13.79e 2.26a

.001g ⬍.001g .132

2.88a

.090

.94a

.332

.00a

.995

2.98a

.084

.39a

.533

rs6582071

Upstream

rs7963803

Upstream

rs4641527

Intron 1

rs2129575

Intron 4

rs4760747

Intron 5

rs1386496

Intron 5

rs1386494

Intron 5

rs1843812

Intron 6

rs1023990

Intron 7

rs1386497

Intron 8

rs1487276

Intron 8

A G

A C G T

G T A G C T A G A G C T A C

A G

29 (13.8) 181 (86.2)

30 (14.2) 182 (85.8) 186 (87.7) 26 (12.3)

183 (81.7) 41 (18.3) 27 (13.5) 173 (86.5) 34 (16.2) 176 (83.8) 25 (12.1) 181 (87.9) 41 (18.6) 179 (81.4) 56 (29.8) 132 (70.2) 162 (71.7) 64 (28.3)

39 (19.3) 163 (80.7)

69 (24.1) 217 (75.9)

26 (9.1) 260 (90.9) 218 (75.2) 72 (24.8)

221 (76.2) 69 (23.8) 25 (8.7) 263 (91.3) 38 (13.1) 252 (86.9) 35 (12.2) 253 (87.8) 37 (13.0) 247 (87.0) 76 (27.1) 204 (72.9) 246 (85.4) 42 (14.6)

A/A A/C C/C C/C, C/A

53 (46.9) 56 (49.6) 4 (3.5) 60 (53.1)

106 (73.6) 34 (23.6) 4 (2.8) 38 (26.4)

47 (16.8) 233 (83.2)

p-Value

14.59a

⬍.001g

19.59b 19.14f .51a

⬍.001g ⬍.001g .476

SNP, single nucleotide polymorphism; TPH2, tryptophan hydroxylase-2. Chi-square for the between-groups comparison of allelic frequencies for the corresponding SNP. Chi-square for the between-groups comparison of genotypic frequencies for the corresponding SNP. c Chi-square for the between-groups comparison of the T/T and the C/T, C/C genotypes frequencies for the corresponding SNP. d Chi-square for the between-groups comparison of the G/G and the A/G, A/A genotypes frequencies for the corresponding SNP. e Chi-square for the between-groups comparison of the G/G and the G/T, T/T genotypes frequencies for the corresponding SNP. f Chi-square for the between-groups comparison of the C/C, C/A and the A/A genotypes frequencies for the corresponding SNP. g Significant after Bonferroni correction for multiple testing (corrected p-value ⫽ .004). a

b

we created a variable called history of substance use disorders that comprised both. In the first step of the hierarchical logistic regression, we controlled for possible confounders by entering age and sex. The order in which other variables were entered was the following: step 2, family history of suicide; step 3, BGHA and BDHI scores (aggression); step 4, BIS scores (impulsivity); step 5, reward dependence scores; step 6, cluster B PD and history of substance use disorders; and step 7, TPH2 genetic variants in the upstream region (rs4448731), intron 1 (rs4641527), and intron 8 (1386497). The results of this model (Table 5), which correctly predicted 78% of the suicides in MDD and explained 41% of the variance, showed the upstream and intron 1 TPH2 genetic variants, a family history of suicide, and cluster B PD as significant and independent predictors of suicide. As noted, once the effect of other risk factors was accounted for, the SNP in intron 8 was no longer significant. Measures of aggression, particularly BGHA scores, added significantly to the model (B ⫽ .05, p ⫽ .038). However, when cluster B PD and substance use disorders were added, BGHA scores were no longer significant (Table 5). Age, sex, and the measures of impulsivity (BIS scores) www.sobp.org/journal

and reward dependence did not contribute to the predictive value of the model. No interactions were found between genetic variation and IABs or cluster B personality disorders (data not shown). The hypothesized mediating role of IABs was further examined by a series of logistic and linear regression analyses. From the results obtained in the hierarchical logistic regression, we decided to test the upstream region (rs4448731) and intron 1 (rs4641527) SNPs as predictors and BGHA scores and cluster B PD as mediators. Brown-Goodwin Assessment of Lifetime History of Aggression scores were included because of its significant contribution to suicide prediction before the inclusion of cluster B PD to which it was moderately correlated (r ⫽ .55, p ⬍ .001). Results showed that TPH2 genetic variants were significantly related to suicide and thus the first condition was met (upstream region SNP, B ⫽ .93, p ⫽ .001; and intron 1 SNP, B ⫽ 1.03, p ⫽ .001, respectively). However, they did not show a significant association between genetic variants and levels of aggression measured by BGHA scores (upstream region SNP, B ⫽ ⫺.05, p ⫽ .495; and intron 1 SNP,

BIOL PSYCHIATRY 2007;62:72– 80 77

C. Lopez de Lara et al. Table 3. Haplotype Frequencies for SNPs in Block 1 and Block 2 Between Depressed Suicide Cases and Depressed Control Individuals Frequency Haplotypea Block 1 CGCGG TGCGG TACTT TAATT TGCGT TGAGG TACTG CGAGG TGCTT Block 2 TGGTAG TGGCAG CAATCA TGGTAA TGGTCG TGGCCG

Depressed Suicides

Depressed Control Subjects

Chi-Square Statistic

p-Value

.34 .36 .05 .02 .04 .05 .02 .03 .02

.50 .23 .14 .07 .01 .003 .02 .01 .01

12.55 9.87 11.54 6.25 7.34 11.39 .001 4.43 .36

⬍.001b .002b .001b .012 .007 .001b .978 .035 .548

.46 .21 .11 .03 .04 .03

.54 .26 .11 .05 .004 .004

3.209 1.462 .030 1.351 7.879 6.775

.073 .227 .863 .245 .005 .009

SNP, single nucleotide polymorphism. a Haplotypes with frequencies above 1% were analyzed. b Significant haplotypes after Bonferroni correction for multiple testing (corrected p-value ⫽ .003).

B ⫽ .01, p ⫽ .881, respectively) or cluster B PD (upstream region SNP, B ⫽ ⫺.34, p ⫽ .433; and intron 1 SNP, B ⫽ .06, p ⫽ .885, respectively). Hence, the second condition for mediation was not met, and thus the third and fourth conditions were not further tested. Thus, our results did not reveal a mediating role of levels of aggression (measured by BGHA) or cluster B PD in the relationship between TPH2 genetic variants and suicide completion in MDD.

Discussion In this study, we provide some evidence that variation at the TPH2 gene and its 5’ upstream region may predispose, at least partially, to the risk of suicide completion in MDD. However, our findings do not support the hypothesis that this predisposition might be mediated by increased levels of IABs. Moreover, TPH2 genetic variation, cluster B PD, and family history of suicide may act as independent predictors of suicide. We identified two SNPs in the TPH2 5’ upstream region (rs4448731, and rs6582071) and two SNPs in introns 1 (rs4641527) and 8 (rs1386497) implicated in suicide susceptibility. Alleles T, G, G, and C of SNPs rs4448731, rs6582071, rs4641527, and rs1386497, respectively, were overrepresented in depressed suicide completers (Table 2). Results from the genotype analyses indicated that homozygotes for the T allele of SNP rs4448731, the G allele of SNP rs4641527, as well as carriers of the C allele of SNP rs1386497, were more often found in depressed suicide cases. Further support for the implication of these TPH2 variants in suicide came from the haplotype analyses where two suicide risk haploytpes in the first block of LD (Table 3), containing the associated SNP alleles in the upstream region and intron 1, were identified. None of the SNPs tested here had been previously reported in studies of suicidal behavior, except for the SNP in intron 5 (rs1386494), which was shown to be associated with suicide completion in a German population (Zill et al. 2004c). However,

we did not find such association in the present sample. Some possible explanations follow: 1) Given the lack of information on psychiatric disorders in that study, it is plausible that the association found might have been related to comorbid psychopathology, rather than suicide itself. Supporting this view is the fact that the same group had previously reported the same intronic SNP to be linked to MDD (Zill et al. 2004a). 2) True biological differences between populations may account for the results obtained in that study and the one reported here. Indeed, the respective variants found to be associated with suicide in either of the two studies may, in fact, be in LD with the actual causal variant. Despite these differences, both studies support a role of TPH2 genetic variation in the risk of suicide completion. The lack of variability for the putative TPH2 functional SNP (G1463A) in exon 11 (Zhang et al. 2005) found in our sample agrees well with results from other studies (Delorme et al. 2006; Garriock et al. 2005). Our findings do not support a mediating effect of IABs in the relationship between TPH2 genetic variants and suicide vulnerability. We did not find any significant difference in these trait measures for the SNPs found overrepresented in suicide. Neither did the series of regression analyses carried out demonstrate a mediation role for these traits. Provided that the investigated TPH2 SNPs result in altered brain serotonin synthesis, some possible explanations are as follows: 1) A lack of power may have prevented us from detecting a possible effect of these genetic variants on IABs. 2) The magnitude of the differences observed in our groups of depressed suicides and control subjects with regard to IABs may be lower than those observed in studies suggesting a central serotonergic deficit associated to IABs (Brown et al. 1982; Linnoila et al. 1983) 3) Impulsivity and aggression cover a wide range of behaviors that may involve different biological mechanisms (Evenden 1999; Lesch and Merschdorf 2000) and, therefore, may introduce phenotypic and genetic heterogeneity. For instance, variation at a murine TPH2 gene SNP (Zhang et al. 2004) had a lowering effect on baseline extracellular serotonin brain levels; however, it did not have an impact on a specific measure of impulsive behaviors such as impulsive choice (Isles et al. 2005). Moreover, a study on mice showed an association between a functional TPH2 SNP and intermale aggression, lending support to the role of TPH2 variation in the regulation of specific behavioral traits (Kulikov et al. 2005). Hence, we might have missed a possible association of TPH2 gene variants with these traits by looking at the overall scores of the instruments used instead of focusing on more specific aspects. 4) Although unlikely, we cannot rule out that other TPH2 variants not investigated here may have an effect on IABs. Higher lifetime aggression and impulsivity have been reported in depressed patients with a history of alcoholism compared with those patients without such history (Sher et al. 2005). Here, we report more substance use disorders in depressed suicide victims. It is possible that by investigating those depressed suicides with a history of substance use disorders, we would be more likely to find an association between TPH2 gene variants and IABs. While no effect of TPH2 variants was observed on IABs, homozygotes for the G allele of SNP rs4641527 had slightly lower scores on the reward dependence temperament dimension. Reward dependence is believed to be moderately heritable and stable throughout life. Individuals with low reward dependence scores are usually cold and socially detached. They prefer to be alone and have difficulties connecting with other people (Clonwww.sobp.org/journal

78 BIOL PSYCHIATRY 2007;62:72– 80

C. Lopez de Lara et al.

Table 4. BGHA, BDHI, BIS, and TCI RD Scores for the Total Sample According to Genotypic Variation at the Four TPH2-SNPs Scale

SNPsa

Mean ⫾ SD (n)

BGHA

rs4448731- T/T C/T,C/C rs6582071- G/G A/A,A/G rs4641527- G/G G/T,T/T rs1386497- A/C,C/C A/A rs4448731- T/T C/T, CC rs6582071- G/G A/A,A/G rs4641527- G/G G/T, TT rs1386497- A/C, CC A/A rs4448731- T/T C/T, CC rs6582071- G/G A/A,A/G rs4641527- G/G G/T, TT rs1386497- A/C, CC A/A rs4448731- T/T C/T, CC rs6582071- G/G A/A,A/G rs4641527- G/G G/T, TT rs1386497- A/C, CC A/A

9.5 ⫾ 11.3 (63) 10.0 ⫾ 10.6 (96) 9.5 ⫾ 11.0 (108) 9.8 ⫾ 10.3 (52) 9.9 ⫾ 11.0 (114) 9.1 ⫾ 10.0 (50) 8.3 ⫾ 10.0 (72) 10.6 ⫾ 11.1 (95) 30.3 ⫾ 11.6 (61) 32.2 ⫾ 13.2 (112) 31.8 ⫾ 13.0 (119) 31.3 ⫾ 12.5 (54) 32.1 ⫾ 12.7 (126) 31.2 ⫾ 12.4 (50) 29.5 ⫾ 12.3 (76) 32.9 ⫾ 12.7 (103) 65.9 ⫾ 13.4 (67) 66.8 ⫾ 13.3 (118) 66.5 ⫾ 13.1 (125) 65.8 ⫾ 14.4 (61) 66.7 ⫾ 13.2 (133) 66.5 ⫾ 14.0 (56) 65.4 ⫾ 13.5 (79) 66.7 ⫾ 13.6 (114) 15.0 ⫾ 4.2 (62) 14.9 ⫾ 4.1 (116) 14.7 ⫾ 4.1 (123) 15.4 ⫾ 4.4 (56) 14.4 ⫾ 4.0 (131) 15.8 ⫾ 4.5 (51) 14.3 ⫾ 4.6 (78) 15.2 ⫾ 3.9 (108)

BDHI

BIS

TCI RD

t-Test Statistic

df

p-value

.31

157

.756

.13

158

.895

⫺.43

162

.670

1.36

165

.177

.90

171

.368

⫺.25

171

.805

⫺.47

174

.639

1.79

177

.075

.44

183

.662

⫺.35

184

.725

⫺.13

187

.896

.62

191

.533

⫺.09

176

.925

1.13

177

.258

2.08

180

.039b

1.38

184

.168

BGHA, Brown-Goodwin Assessment of Lifetime History of Aggression; BDHI, Buss-Durkee Hostility Inventory; BIS, Barratt Impulsiveness Scale; TCI, Temperament and Character Inventory; RD, Reward dependence. a SNPs previously found in association with suicide are shown. b Significant at the .05 level.

inger et al. 1994). We may hypothesize, pending replication of this finding, that the effect of TPH2 gene variants on lower reward dependence scores could contribute to suicide vulnerability by preventing some individuals from having beneficial social affiliations (Cloninger et al. 1994). Finally, we detected a strong and unique contribution of the TPH2 upstream region (rs4448731) and intron 1 (rs4641527) variants to suicide risk, while adjusting for the effect of each other and other risk factors, as well as confounders (Table 5). Having the T and G alleles of the upstream and intron 1 variants, respectively, in homozygosis translated into an approximately fivefold increase in suicide risk. Whereas the upstream SNP— located at 3.5 kb from the transcription start site—may be a causal one, the presence of LD between this SNP and other SNPs in this or other genes is also possible. Although this upstream SNP may seem to be far from the transcription start site, studies have suggested that regulatory elements necessary for correct gene expression can be found as much as 1 Mb either upstream or downstream from the transcription unit and furthermore, that some of these elements can also be found within introns (Kleinjan and van Heyningen 2005). In addition, variants in the upstream regulatory region of the TPH2 have been shown to affect the reactivity of the amygdala, hence, providing support to the involvement of TPH2 gene in disorders with emotional dysregulation (Brown et al. 2005; Canli et al. 2005). www.sobp.org/journal

Cluster B personality disorders and family history of suicide also appeared as independent risk factors for suicidal behavior in MDD, which is in accordance with previous research (Dumais et al. 2005; Roy 1993). A nearly eightfold increase in suicide risk was found in individuals with cluster B PD, while an approximately fourfold increase was related to having a family history of suicide. Although the increased risk related to a family history of suicide might be due to a genetically determined familial transmission (Turecki 2001), shared environmental factors also need to be considered (Fu et al. 2002). Limitations of this study are those typically related to association studies, such as population stratification (Lander and Schork 1994). However, the vast majority of our sample consisted of individuals of French-Canadian origin, which is an isolated and young population with a well-known founder effect (Heyer and Tremblay 1995). Therefore, the total genetic variance involved in suicide predisposition is likely reduced, and population stratification effects are less likely to occur. The use of proxy-based assessments represents another limitation. However, their validity has been well demonstrated and discussed elsewhere (Conner et al. 2001; Schneider et al. 2004). The main concern appears to be related to their sensitivity (Ernst et al. 2004); nevertheless, by using proxy-based techniques in both groups, this possible bias was controlled. Last, our sample size may have reduced power to detect small genetic effects or interac-

BIOL PSYCHIATRY 2007;62:72– 80 79

C. Lopez de Lara et al. Table 5. Hierarchical Logistic Regression Analysis for Suicide Completion Stepa 1 2 3 4 5 6 7

Variable

B

SE

Wald

p Value

OR

Age Sex Family history of suicide BGHA BDHI BIS Reward dependence Cluster B PD History of substance use disorders Upstream variantb Intron 2 variantc Intron 8 variantd

⫺.19 ⫺.24 1.47e .01 ⫺.003 ⫺.01 ⫺.05 2.05e .86 1.66f 1.65f .33

.02 .69 .69 .03 .02 .02 .06 .90 .51 .58 .57 .53

1.01 .001 4.50 .19 .02 .08 .73 5.20 2.89 8.18 8.33 .40

.316 .973 .034 .666 .901 .772 .392 .023 .089 .004 .004 .528

.98 .98 4.35 1.01 1.00 .99 .95 7.75 2.37 5.26 5.22 1.39

95% CI .94 .25 1.12 .95 .96 .96 .85 1.33 .88 1.69 1.70 .50

1.02 3.81 16.95 1.07 1.04 1.03 1.06 45.12 6.41 16.44 16.04 3.91

n ⫽ 124. B, unstandardized logistic regression coefficient; Wald, Wald statistic; OR, odds ratio; CI, confidence interval; BGHA, Brown-Goodwin Assessment of Lifetime History of Aggression; BDHI, Buss-Durkee Hostility Inventory; BIS, Barratt Impulsiveness Scale; Cluster BPD, Cluster B personality disorders; SNP, single nucleotide polymorphism. a Step of the logistic regression analysis. All the statistics shown correspond to those of the last step in the analysis. b SNP rs4448731. c SNP rs4641527. d SNP rs1386497. e Statistically significant at the .05 level. f Statistically significant at the .01 level.

tions between genetic and clinical risk factors; however, the most extreme phenotype (suicide completion) was investigated. In summary, our results indicate that variants in the TPH2 gene and upstream region may be implicated in the predisposition to suicide completion in MDD. However, our findings do not support the hypothesis that this predisposition might be mediated by increased levels of IABs. Nevertheless, a slight effect of genetic variation on a TCI temperament dimension was detected. Finally, TPH2 genetic variation, cluster B PD, and family history of suicide increased significantly and independently the risk of suicide completion in depressed individuals. Further studies in larger and independent homogeneous samples are required to have stronger evidence of the implication of these variants in suicide, as well as to establish interactions between genetic variants and other risk factors. Future studies should control for associated psychopathology so that the effects attributable to suicide can be separated. Progress is also needed in establishing the physiological relevance of the TPH2 variants investigated. This study was supported by the Canadian Institute of Health Research (CIHR) MOP-38078. GT is a CIHR scholar. The authors declare no conflict of interest related to this work. Bach-Mizrachi H, Underwood MD, Kassir SA, Bakalian MJ, Sibille E, Tamir H, et al. (2006): Neuronal tryptophan hydroxylase mRNA expression in the human dorsal and median raphe nuclei: Major depression and suicide. Neuropsychopharmacology 31:814 – 824. Baron RM, Kenny DA (1986): The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J Pers Soc Psychol 51:1173–1182. Barrett JC, Fry B, Maller J, Daly MJ (2005): Haploview: Analysis and visualization of LD and haplotype maps. Bioinformatics 21:263–265. Blair-West GW, Cantor CH, Mellsop GW, Eyeson-Annan ML (1999): Lifetime suicide risk in major depression: Sex and age determinants. J Affect Disord 55:171–178. Brent DA, Johnson BA, Perper J, Connolly J, Bridge J, Bartle S, et al. (1994): Personality disorder, personality traits, impulsive violence, and completed suicide in adolescents. J Am Acad Child Adolesc Psychiatry 33: 1080 –1086.

Brown GL, Ebert MH, Goyer PF, Jimerson DC, Klein WJ, Bunney WE, Goodwin FK (1982): Aggression, suicide, and serotonin: Relationships to CSF amine metabolites. Am J Psychiatry 139:741–746. Brown GL, Goodwin FK (1986): Cerebrospinal fluid correlates of suicide attempts and aggression. Ann N Y Acad Sci 487:175–188. Brown SM, Peet E, Manuck SB, Williamson DE, Dahl RE, Ferrell RE, et al. (2005): A regulatory variant of the human tryptophan hydroxylase-2 gene biases amygdala reactivity. Mol Psychiatry 10:805. Buss AH, Durkee A (1957): An inventory for assessing different kinds of hostility. J Consult Psychol 21:343–349. Canli T, Congdon E, Gutknecht L, Constable RT, Lesch KP (2005): Amygdala responsiveness is modulated by tryptophan hydroxylase-2 gene variation. J Neural Transm 112:1479 –1485. Cavanagh JT, Carson AJ, Sharpe M, Lawrie SM (2003): Psychological autopsy studies of suicide: A systematic review. Psychol Med 33:395– 405. Cloninger CR, Przybeck TR, Svrakic DM, Wetzel RD (1994): The Temperament and Character Inventory (TCI): A Guide to Its Development and Use. St. Louis: Center for Psychobiology of Personality, Washington University. Coccaro EF (1989): Central serotonin and impulsive aggression. Br J Psychiatry Suppl Dec(8):52– 62. Conner KR, Conwell Y, Duberstein PR (2001): The validity of proxy-based data in suicide research: A study of patients 50 years of age and older who attempted suicide. II. Life events, social support and suicidal behavior. Acta Psychiatr Scand 104:452– 457. Delorme R, Durand CM, Betancur C, Wagner M, Ruhrmann S, Grabe HJ, et al. (2006): No human tryptophan hydroxylase-2 gene R441H mutation in a large cohort of psychiatric patients and control subjects. Biol Psychiatry 60:202–203. De Luca V, Mueller DJ, Tharmalingam S, King N, Kennedy JL (2004): Analysis of the novel TPH2 gene in bipolar disorder and suicidality. Mol Psychiatry 9:896 – 897. De Luca V, Voineskos D, Wong GW, Shinkai T, Rothe C, Strauss J, et al. (2005): Promoter polymorphism of second tryptophan hydroxylase isoform (TPH2) in schizophrenia and suicidality. Psychiatry Res 134:195–198. Denning DG, Conwell Y, King D, Cox C (2000): Method choice, intent, and gender in completed suicide. Suicide Life Threat Behav 30:282–288. Dumais A, Lesage AD, Alda M, Rouleau G, Dumont M, Chawky N, et al. (2005): Risk factors for suicide completion in major depression: A case-control study of impulsive and aggressive behaviors in men. Am J Psychiatry 162:2116 –2124. Ernst C, Lalovic A, Lesage A, Seguin M, Tousignant M, Turecki G (2004): Suicide and no axis I psychopathology. BMC Psychiatry 4:7. Evenden JL (1999): Varieties of impulsivity. Psychopharmacology (Berl) 146: 348 –361.

www.sobp.org/journal

80 BIOL PSYCHIATRY 2007;62:72– 80 First MB, Gibbon M, Spitzer RL, Williams JBW (2001): Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Patient Edition with psychotic screen. New York, Biometrics Research, New York State Psychiatric Institute. First MB, Spitzer RL, Gibbon M, Williams JBW, Lorna B (1995): Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). New York, Biometrics Research Department, New York State Psychiatric Institute. First MB, Spitzer RL, Gibbon M, Williams J (1995): The Structured Clinical Interview for DSM-III-R Personality Disorders (SCID-II). Part I: Description. J Personal Disord 9:83–91. Frazier PA, Tix AP, Barron KE (2004): Testing moderator and mediator effects in counselling psychology research. J Couns Psychol 51:115–134. Fu Q, Heath AC, Bucholz KK, Nelson EC, Glowinski AL, Goldberg J, et al. (2002): A twin study of genetic and environmental influences on suicidality in men. Psychol Med 32:11–24. Garriock HA, Allen JJ, Delgado P, Nahaz Z, Kling MA, Carpenter L, et al. (2005): Lack of association of TPH2 exon XI polymorphisms with major depression and treatment resistance. Mol Psychiatry 10:976 –977. Heyer E, Tremblay M (1995): Variability of the genetic contribution of Quebec population founders associated to some deleterious genes. Am J Hum Genet 56:970 –978. Isles AR, Hathway GJ, Humby T, de la Riva C, Kendrick KM, Wilkinson LS (2005): An mTph2 SNP gives rise to alterations in extracellular 5-HT levels, but not in performance on a delayed-reinforcement task. Eur J Neurosci 22:997–1000. Kelly TM, Mann JJ (1996): Validity of DSM-III-R diagnosis by psychological autopsy: A comparison with clinician ante-mortem diagnosis. Acta Psychiatr Scand 94:337–343. Kessler RC, Berglund P, Demler O, Jin R, Koretz D, Merikangas KR, et al. (2003): The epidemiology of major depressive disorder: Results from the National Comorbidity Survey Replication (NCS-R). JAMA 289:3095–3105. Kleinjan DA, van Heyningen V (2005): Long-range control of gene expression: Emerging mechanisms and disruption in disease. Am J Hum Genet 76:8 –32. Kulikov AV, Osipova DV, Naumenko VS, Popova NK (2005): Association between Tph2 gene polymorphism, brain tryptophan hydroxylase activity and aggressiveness in mouse strains. Genes Brain Behav 4:482– 485. Lander ES, Schork NJ (1994): Genetic dissection of complex traits. Science 265:2037–2048. Lesch KP, Merschdorf U (2000): Impulsivity, aggression, and serotonin: A molecular psychobiological perspective. Behav Sci Law 18:581– 604. Leyton M, Okazawa H, Diksic M, Paris J, Rosa P, Mzengeza S, et al. (2001): Brain regional alpha-[11C]methyl-L-tryptophan trapping in impulsive subjects with borderline personality disorder. Am J Psychiatry 158:775–782. Leyton M, Paquette V, Gravel P, Rosa-Neto P, Weston F, Diksic M, et al. (2006): Alpha-[(11)C]methyl-l-tryptophan trapping in the orbital and ventral medial prefrontal cortex of suicide attempters. Eur Neuropsychopharmacol 16:220 –223. Linnoila M, Virkkunen M, Scheinin M, Nuutila A, Rimon R, Goodwin FK (1983): Low cerebrospinal fluid 5-hydroxyindoleacetic acid concentration differentiates impulsive from nonimpulsive violent behavior. Life Sci 33: 2609 –2614. Lopez de Lara C, Dumais A, Rouleau G, Lesage A, Dumont M, Chawky N, et al. (2006): STin2 variant and family history of suicide as significant predictors of suicide completion in major depression. Biol Psychiatry 59:114 –120. Mann JJ (1998): The neurobiology of suicide. Nat Med 4:25–30.

www.sobp.org/journal

C. Lopez de Lara et al. Mann JJ, Brent DA, Arango V (2001): The neurobiology and genetics of suicide and attempted suicide: A focus on the serotonergic system. Neuropsychopharmacology 24:467– 477. Patel PD, Pontrello C, Burke S (2004): Robust and tissue-specific expression of TPH2 versus TPH1 in rat raphe and pineal gland. Biol Psychiatry 55: 428 – 433. Patton JH, Stanford MS, Barratt ES (1995): Factor structure of the Barratt impulsiveness scale. J Clin Psychol 51:768 –774. Rosa-Neto P, Diksic M, Okazawa H, Leyton M, Ghadirian N, Mzengeza S, et al. (2004): Measurement of brain regional alpha-[11C]methyl-L-tryptophan trapping as a measure of serotonin synthesis in medication-free patients with major depression. Arch Gen Psychiatry 61:556 –563. Roy A (1993): Features associated with suicide attempts in depression: A partial replication. J Affect Disord 27:35–38. Sambrook J, Fritsch E, Maniatis T (1989): Molecular Cloning. A Laboratory Manual 2. New York: Cold Spring Harbor Press. Schneider B, Maurer K, Sargk D, Heiskel H, Weber B, Frolich L, et al. (2004): Concordance of DSM-IV Axis I and II diagnoses by personal and informant’s interview. Psychiatry Res 127:121–136. Sher L, Oquendo MA, Galfalvy HC, Grunebaum MF, Burke AK, Zalsman G, et al. (2005): The relationship of aggression to suicidal behavior in depressed patients with a history of alcoholism. Addict Behav 30:1144 – 1153. Spitzer RL, Williams JB, Gibbon M, First MB (1992): The Structured Clinical Interview for DSM-III-R (SCID). I: History, rationale, and description. Arch Gen Psychiatry 49:624 – 629. St-Laurent D, Bouchard C (2004): L’épidémiologie du suicide au Québec: Que savonsnous de la situation récente? Institut national de santé publique du Québec. Turecki G (2001): Suicidal behavior: Is there a genetic predisposition? Bipolar Disord 3:335–349. Walther DJ, Bader M (2003): A unique central tryptophan hydroxylase isoform. Biochem Pharmacol 66:1673–1680. Zhang X, Beaulieu JM, Sotnikova TD, Gainetdinov RR, Caron MG (2004): Tryptophan hydroxylase-2 controls brain serotonin synthesis. Science 305:217. Zhang X, Gainetdinov RR, Beaulieu JM, Sotnikova TD, Burch LH, Williams RB, et al. (2005): Loss-of-function mutation in tryptophan hydroxylase-2 identified in unipolar major depression. Neuron 45:11–16. Zhou Z, Roy A, Lipsky R, Kuchipudi K, Zhu G, Taubman J, et al. (2005): Haplotype-based linkage of tryptophan hydroxylase 2 to suicide attempt, major depression, and cerebrospinal fluid 5-hydroxyindoleacetic acid in 4 populations. Arch Gen Psychiatry 62:1109 –1118. Zill P, Baghai TC, Zwanzger P, Schule C, Eser D, Rupprecht R, et al. (2004a): SNP and haplotype analysis of a novel tryptophan hydroxylase isoform (TPH2) gene provide evidence for association with major depression. Mol Psychiatry 9:1030 –1036. Zill P, Buttner A, Eisenmenger W, Bondy B, Ackenheil M (2004b): Regional mRNA expression of a second tryptophan hydroxylase isoform in postmortem tissue samples of two human brains. Eur Neuropsychopharmacol 14:282–284. Zill P, Buttner A, Eisenmenger W, Moller HJ, Bondy B, Ackenheil M (2004c): Single nucleotide polymorphism and haplotype analysis of a novel tryptophan hydroxylase isoform (TPH2) gene in suicide victims. Biol Psychiatry 56:581–586. Zill P, Buttner A, Eisenmenger W, Moller HJ, Ackenheil M, Bondy B (2007): Analysis of tryptophan hydroxylase I and II mRNA expression in the human brain: A postmortem study. j Psychiatr Res 41:168 –173.