¹Department of Psychiatry, Carémeau Hospital, Nîmes, France

²Department of Psychological Medicine and Psychiatry, Lapeyronie Hospital, INSERM E99-30, Montpellier, France

³Department of Psychiatry, Henri Mondor and Albert Chenevier Hospitals, Créteil, France

⁴University Department of Psychiatry, La Colombière Hospital, INSERM E99-30, Montpellier, France

⁵Division of Neuropsychiatry, Department of Psychiatry, Geneva University Hospital, Geneva, Switzerland

Correspondence to: P Courtet, Department of Psychological Medicine and Psychiatry, Lapeyronie Hospital, 34295 Montpellier Cedex 5, France. E-mail: courtet@aol.com

^aThese two authors contributed equally to this work and they are listed in alphabetical order.

Tryptophan hydroxylase (TPH) is the rate-limiting enzyme of serotonin synthesis. In this case-control study, we investigated whether the TPH gene was a susceptibility factor for suicidal behavior. Seven polymorphisms spanning the entire gene were studied in a case-control study including 231 individuals who had attempted suicide and 281 controls. Significant associations were found between variants in introns 7, 8 and 9 (² = 11.2, df = 1, P< 0.0008 for the allele distribution; these loci are in complete linkage disequilibrium) and in the 3' noncoding region (² = 30.94, P = 0.0014) and suicide attempt. The association was strongest for subjects who had attempted suicide by violent means and who had a history of major depression. No significant association was observed between suicide attempts and polymorphisms in the promoter, intron 1 and intron 3. The results presented here, and those of previous studies, suggest that a genetic variant of the 3' part of the TPH gene may be a susceptibility factor for a phenotype combining suicidal behavior, mood disorder and impulsive aggression. Molecular Psychiatry (2001) 6, 268-273.

tryptophan hydroxylase; genetic polymorphisms; suicidal behavior; association study

Introduction

A genetic contribution to suicidal behavior has been demonstrated in family, twin and adoption studies, as reviewed in two recent papers.^1,2 There is some evidence for a specific genetic component of susceptibility, which may be independent of or additive to the genetic transmission of the psychiatric disorders strongly related to suicidal behavior (bipolar disorder, schizophrenia, alcoholism).^2,3 Several recent studies have investigated the involvement of the tryptophan hydroxylase (TPH) gene in genetic susceptibility to suicidal behavior.^4,5,6,7,8 The principal reason for testing for an association between the TPH gene and suicidal behavior is that TPH is the first, and possibly rate-limiting, enzyme in serotonin biosynthesis.⁹ It may therefore play a key role in regulating the availability, turnover and function of serotonin. Abnormal serotonergic function has frequently been reported in individuals who commit or attempt suicide and is one of the most replicated findings in modern biological psychiatry.¹⁰ In a recent review of the neurobiology of suicide, Mann¹⁰ concluded that indicators of serotonin function are more consistent with an association between low serotonergic activity and suicidal behavior than between low serotonergic activity and primary psychiatric diagnosis. Subjects who attempted suicide, especially those using violent methods, had lower levels of cerebrospinal fluid (CSF) 5-hydroxyindoleacetic acid (5-HIAA), the breakdown product of serotonin, than did psychiatric controls. Those making subsequent suicidal acts also had lower levels of CSF 5-HIAA.¹¹ Thus, one mechanism by which genetic factors may affect the risk of suicide is by reducing serotonergic function.¹⁰

In 1992, in a case-control study, we found no difference in allele frequencies at the TPH locus (Ava II/C2-38) between suicide attempters and controls.⁴ However, subsequent studies have concluded that there is an association between suicidal behavior and TPH gene polymorphism.^5,6,8,12 These recent studies used two genetic markers located in intron 7 of this gene. The first polymorphism used (A779C) was originally identified by single strand conformational analysis¹³ and the two alleles, TPH 779A and TPH 779C, were referred to as U and L, respectively. The second polymorphism (A218C) is in complete or strong linkage disequilibrium with A779C in various European populations.¹³ These studies are therefore comparable in terms of the markers used. However, the populations studied were very different and the results conflicting.

We have previously carried out a case-control study of the A218C polymorphism in a sample of 231 suicide attempters and 161 ethnically matched controls.¹⁴ We followed up this initial study by investigating six new TPH variants spanning the entire gene,^15,16 and an additional sample of 120 controls.

Subjects and methods

Patients and controls

Two hundred and thirty one patients consecutively admitted to our psychiatric departments after a suicide attempt were included in this study. All subjects gave informed consent. A suicide attempt was defined as intentional self-harm that was not self-multilatory in nature¹⁰ and required medical evaluation and treatment in an emergency or intensive care unit. All patients were interviewed using a DSM-III-R checklist, by two experienced psychiatrists (MA and PC), and lifetime diagnoses were made according to DSM-III-R criteria. For this study, the axis I diagnoses known to be associated with a high risk of suicidal behavior were investigated: major mood disorders (MMD), schizophrenia, alcohol and substance dependence. Suicide attempts were classified as violent or nonviolent according to the criteria proposed by Asberg et al¹⁷ and adopted by Träksman et al.¹⁸ Hanging attempts, the use of firearms or knives and jumping from heights were considered to be violent attempts; drug overdoses were considered to be nonviolent suicide attempts. For all patients, we investigated whether there was a past history of violent suicide attempts. Patients with a history of violent suicide attempts were classified as violent suicide attempters, regardless of the nature of the current suicide attempt.

The 281 control subjects were blood donors, interviewed by trained psychiatrists with a French version of the Diagnostic Interview for Genetic Studies (DIGS).¹⁹ Subjects with a personal or familial (first degree) history of suicide attempts or the above cited diagnoses were excluded from the controls. In order to minimize their morbid risk, only controls older than 35 years were included. To minimize population heterogeneity, we ensured that both the patients and controls were entirely of West European Caucasian origin for at least two generations.

Laboratory methods

Fifteen-milliliter blood samples were collected into EDTA and genomic DNA was extracted from white blood cells (DNA extraction kit from Amersham-Pharmacia Biotech, Dübendorf, Switzerland). Polymorphisms in the promoter region (T1606C), intron 1b (T3792A), intron 7 (A218C), intron 8 (T465C), intron 9 (C160T), and in the 3' region, 5657 bp distal to exon 11, (CT)_m(CA)_n(CT)_p, were analyzed as described elsewhere.¹⁶

We recently identified a new imperfect [G(A)_n]_m polymorphic repetition in intron 3. We amplified a 210-220 bp fragment with the primers: TPHI4GAF (TGACAGAAC AAGACCCTGTC) and TPHI4GAR (GCTCAGGGATTGTGAATTATAG). PCR was performed with one of the primers labeled with gamma-³³P adenosine 5'-triphosphate (NEN Life Science Project, Zaventem, Belgium), using HotStar Taq DNA polymerase (Qiagen, Basel, Switzerland). PCR products (5 l) were subjected to electrophoresis in a denaturing acrylamide gel in 0.5 ´ TBE. After 2 h and 30 min of electrophoresis at 55 W, the gel was dried under vacuum and placed against Kodak (Kodak-Integra Biosciences, Wallisellen, Switzerland) BIOMAX-MR film overnight.

Statistical analysis

Allele and genotype distributions for patients and controls were compared using the chi-square (²) test (two-tailed). As some alleles of the two TPH microsatellites (intron 3 and 3' region) are rare, all ² probability distributions for the alleles of this marker were estimated using the program CLUMP, version 1.1 (5000 simulations).²⁰ Odds ratios (OR) with 95% confidence intervals were calculated using Woolf's method. As the phenotype is a discrete variable, we assessed the correlation between the severity of the phenotype and the vulnerability allele (or genotype) frequency using the Armitage linearity tendency test.²¹ As the polymorphisms in introns 7, 8 and 9 have been shown to be in complete linkage disequilibrium,¹⁶ the corresponding haplotypes (A-T-C and C-C-T) were considered to correspond to only two alleles (1 and 2 respectively). We therefore applied no correction for the analysis of these three polymorphisms.

As we also analyzed four other polymorphisms and two clinical entities (MMD and violent suicide attempt), differences were considered to be statistically significant if P < 0.006 (Bonferroni corrections for eight tests). Linkage disequilibrium and haplotypes frequencies were estimated using the EH program.²²

Results

Samples

The demographic and clinical characteristics of the patients are presented in Table 1. Just over half (54.5%) of those who attempted suicide were women. One hundred and seventy-four patients (75.32%) attempted suicide by drug overdose and 57 (24.7%) attempted suicide by violent methods. Violent attempts were mainly men (70.2%) whereas nonviolent attempters were mostly women (62.6%). The proportion of patients using violent methods differed significantly between men and women (38.1% vs 13.5%, ² = 18.65, P < 2 ´ 10^-5). The number of attempts and the age at first attempt were not different between men and women, and between violent and non violent SA. The control group consisted of 170 men and 111 women, with a mean age at interview of 43.07 ± 8.67 years.

Analysis of the polymorphisms in introns 7, 8 and 9 for all patients and controls

As previously reported,¹⁶ polymorphisms in introns 7, 8 and 9 were found to be in complete linkage disequilibrium in both the patients and controls. Despite this known linkage disequilibrium, we typed all the subjects for these three polymorphisms, to eliminate most of the processing errors that might affect the results. TPH genotype distributions (Table 2) were in Hardy-Weinberg equilibrium in both patients (² = 1.48; df = 2, NS) and controls (² = 1.99; df = 2, NS).

The 1 allele (see 'Statistical analysis') frequency was significantly higher in patients who attempted suicide (44.6%) than in controls (34.3%) (² = 11.2, df = 1, P < 0.0008) (Table 2). The distribution of TPH genotypes also differed significantly between controls and suicide attempters (² = 11.65; df = 2; P = 0.003). Using bearers of the 2 allele as the reference group, the OR for the 1 allele, and the 11 genotype were 1.54 (95% CI 1.19-1.98) and 2.49 (95% CI 1.43-4.33) respectively (Figure 1).

Analysis of the other four polymorphisms for all patients and controls

We analyzed four other polymorphisms in an attempt to confirm these results and to refine the localization of the potential functional variant associated with suicidal behavior. Genotype distributions were in Hardy-Weinberg equilibrium in both patients and controls for all four polymorphisms (Table 2).

Allele and genotype frequencies at the promoter region polymorphic site (T1606C) did not differ between patients and controls (P = 0.94 for alleles and P = 0.87 for genotypes).

The frequencies of the intron 1b 3792A allele and the intron 3 '217' allele were higher in patients than in controls. However, these differences were not statistically significant (P = 0.06 for intron 1b and P = 0.08 for intron 3). Similarly, the frequencies of the 3792A/3792A and 217/217 genotypes were higher in patients than in controls, although not significantly so (P = 0.18, and P = 0.07 respectively).

A significant difference in the allelic distribution of the TPH 3' region (CT)_m(CA)_n(CT)_p microsatellite was observed between cases and controls (² = 30.94, P = 0.0014). This difference was due mainly to the higher frequency of the '194' allele in patients than in controls (OR = 1.61; 95% CI 1.24-2.09) (Figure 1). The OR for the 194/194 genotypic association was even higher (2.00; 95% CI 1.18-3.41).

Linkage disequilibrium (D') was similar in the present control sample to that previously reported.¹⁶ In those who attempted suicide, the allele 1 (introns 7, 8 and 9; see the section 'Statistical analysis') was also in incomplete linkage disequilibrium with the promoter allele 1606C (D' = 0.84), intron 1b allele A (D' = 0.95), intron 3 allele 217 (D' = 0.66) and the 194-bp microsatellite allele in the 3' region (D' = 0.87). Therefore, the allele 1 was preferentially associated with the C-A-217-194 haplotype at the promoter, intron 1b, intron 3 and 3' region loci. The estimated frequencies of the C-A-217-1-194 haplotype were 0.23 in controls and 0.31 in suicide attempters respectively. Only two haplotypes were less common in suicide attempters than in controls: C-T-non217-2-non194 (0.09 vs 0.14 in controls) and T-A-non217-2-non194 (0.18 vs 0.21 in controls).

Violent suicide attempt and depression

When we adjusted for the violence of the attempt, we found that the more severe the phenotype (controls, nonviolent suicide attempt, violent suicide attempt, in ascending order of violence), the higher was the frequency of the 1 allele (Table 2, Armitage linearity tendency test: 12.36, df = 1, P < 0.001). A similar significant relationship was found for the 11 genotype (Table 2, Armitage linearity tendency test: 12.66, df = 1, P < 0.001), whereas the frequency of the 22 genotype was similar in those attempting suicide by violent and nonviolent means. The ORs of violent suicide attempt associated with one 1 allele and two 1 alleles were 1.85 (95% CI 1.23-2.77) and 3.69 (95% CI 1.64-8.29) (Figure 1) respectively.

Interestingly, these differences were more pronounced when the haplotypes were considered. In those attempting suicide by violent means the estimated frequency of the C-A-217-1-194 haplotype was 0.42. Similarly, the decreased frequencies of the C-T-non217-2-non194 and the T-A-non217-2-non194 haplotypes were more pronounced: 0.03 and 0.07 respectively. All the other haplotype frequencies did not differ between the three groups.

A similar but weaker relationship was observed for MDD. Indeed the OR of suicide attempt with a history of MDD associated with one and two 1 allele(s) was slightly lower (OR = 1.68; 95% CI 1.27-2.23 and OR = 2.97; 95% CI 1.61-5.50, respectively) than those for violent suicide attempt. This risk was even lower in nonviolent suicide attempters with a history of MDD (risk for 11: 2.58; 95% CI 1.27-5.21).

The highest ORs were those for violent suicide attempters with a history of MDD. For example, the risk for the 11 genotype was 3.93 (95% CI 1.61-9.63).

None of the tests performed for any of the other diagnoses considered gave a statistically significant difference for TPH allele and genotype frequencies (data not shown). Similarly, no significant difference in TPH allele and genotype frequencies was found for gender and for other characteristics of the suicide attempt: age at the first attempt (above or below 20 years of age), number of attempts in the patient's lifetime at the time of the interview.

Comment

In a preliminary study, we reported an association between the TPH intron 7 218A allele and suicide attempt.¹⁴ Here, we confirm this result and show that this association is even stronger for violent suicide attempt. We also analyzed other polymorphisms, spanning the TPH gene from the promoter region to about 5 kb after the last exon, and found that the putative susceptibility variant was located 3' to intron 3, a region which encodes for the catalytic part of the enzyme. This part of the TPH gene was found to be in complete linkage disequilibrium, whereas incomplete linkage disequilibrium was reported between the polymorphisms in the promoter region and introns 7, 8 and 9.^15,16

In candidate gene case-control studies, differences in allele frequencies between cases and controls may arise as an artifact of population admixture. Although we did not strictly match the geographic origins of the controls and patients in this study, we think that such a bias is unlikely. Indeed the A218C allele and genotype frequencies in our control group were similar to those previously reported for similar west European Caucasian populations (manuscript in preparation), suggesting that these populations are homogeneous in terms of A218C allele frequencies. Therefore the highest intron 7 218A allele frequency recorded here suggests that either the A allele is actually a susceptibility factor for suicidal behavior or, more likely, is in linkage disequilibrium with the actual susceptibility variant, which remains to be identified.

Our data are consistent with two sets of previous results. First, an association between the 218A allele and suicide attempt has been reported in bipolar patients^12,23 and in patients with major depression.⁸ In French bipolar patients, the 218A allele frequency was 0.50 in those who had not attempted suicide and 0.58 in those who had (Armitage linearity test, P<0.001, vs 0.36 in controls).^12,23 In a small sample of Chinese suicide attempters with major depression (n = 41), ORs very similar to those reported here were observed for the 218A allele and 218A/218A genotype (1.83, 95% CI 1.13-2.97 and 2.26, 95% CI 1.20-4.27).⁸ Second, two recent studies have suggested that there is a correlation between low central serotonergic activity, and the 218A allele. Jönsson et al²⁴ found that men with UU (218A/218A) and UL (218A/218C) genotypes have lower levels of CSF 5-HIAA, and more recently Manuck et al²⁵ reported that in men with a U allele, the peak prolactin response to fenfluramine was attenuated. As low central serotonergic activity is a known factor for susceptibility to violent suicide attempt,¹⁰ these findings are consistent with our own.

In our initial study,^4,26 we found no association between a TPH polymorphism (an AvaII restriction polymorphism revealed by Southern blotting using the C2-38 probe) and suicide attempt. There are two possible reasons for this difference in results. First, the case and control sample sizes were small (62 cases and 52 controls). Based on the ORs presented here, these samples were not large enough to detect an association. Second, the precise location of this polymorphic site is unknown (we did not analyze this polymorphism for the samples presented here because the Southern blot method is much more cumbersome than that used here), but it may lie to the 5' side of the gene, a region that we have found to be not associated with suicide attempt.

Mann et al⁵ reported a statistically significant higher frequency of the U allele in depressed patients who had attempted suicide (0.41) than in depressed patients who had not (0.20). It is difficult to compare this result with our own, because the number of depressed patients who had attempted suicide was very small in Mann's study (n = 29) and because we did not include depressed patients who had not attempted suicide in this study.

The first case-control study using the TPH intron 7 polymorphism was carried out by Nielsen et al.²⁷ The results of this preliminary study were confirmed in larger samples.⁶ The authors asserted that 'a functional variant in or close to the TPH gene may predispose individuals to suicidality and other behaviors thought to be influenced by serotonin'.⁶ This conclusion was based on the significant difference they found between the genotype and the allele frequencies in offenders with a history of suicide attempt and offenders without a history of suicide attempt. This difference was even more significant in impulsive offenders. However, these differences could not be attributed to offenders who attempted suicide, because the frequency of the U allele was similar to that estimated for controls (0.41 in impulsive offenders with a history of suicide attempt, 0.42 in nonimpulsive offenders and 0.43 in controls). In fact, higher U allele frequencies were observed in offenders without a history of suicide attempt: 0.48 in nonimpulsive and 0.57 in impulsive offenders. Therefore, it cannot be concluded from these results that TPH is a susceptibility gene for suicide attempt. These results therefore conflict with our own results and with those of Tsai et al.⁸

The main unresolved question is that of the exact phenotype associated with the TPH gene. Two thirds of our suicide attempters had a history of MDD. So, it is possible that the allele and genotype frequency differences reported here are due to MDD only and have nothing to do with suicide attempt. However, higher ORs were found for violent suicide attempters than for suicide attempters with a history of MDD, and this is not due to a difference in the distribution of violent suicide attempts between patients with and without a history of major mood disorder (data not shown). We therefore conclude that the difference in allele and genotype frequencies between patients and controls reported here is not due to MDD alone. This view is supported by the lower 218A allele frequency found by Tsai et al⁸ in patients with major depression who did not attempt suicide (0.37) compared to controls (0.45). Our data support the notion that TPH is mainly associated with the combination of both a history of MDD and violent suicide attempt. The recently reported association between the TPH 218A allele and individual differences in aggressive disposition²⁵ suggests that the precise phenotype related to this TPH variant probably combines three dimensions: mood disorders, suicidal behavior (mainly violent) and aggressive disposition. Clinically, this is consistent with the 'stress-diathesis' model of suicidal behavior proposed by Mann et al.²⁸

This model is also consistent with that proposed by Depue²⁹ who suggested that 'suicidal behavior and impulsive aggression may represent two different indicators of a central (behavioral and biochemical) trait, such as a low threshold of emotional facilitation'.²⁸ This hypothesis is supported by the fact that the cooccurrence of these two indicators is associated with the lowest CSF 5-HIAA values. On the other hand, a low central serotonergic activity would be associated with the frequency of depressive episodes, rather than with the depressive disorder per se, contributing to the appearance of 'neurotic depression' in some subjects with personality disorders. Thus, in some individuals, a TPH variant, may, by reducing central serotonergic activity, confer vulnerability for a spectrum of behavioral disorders including mood disorders, suicidal behavior and impulsive aggression. The risk of presenting one or more of these disorders would increase with the magnitude of serotonergic dysfunction, which may be correlated with the number of TPH 218A alleles.

Acknowledgements

This research was supported by grant 32-47315.96 and grant 31-53849.98 from the Swiss National Foundation and grants from 'Assistance Publique des Hôpitaux de Paris' (Contrat de Recherche Clinique), Institut National de la Santé et de la Recherche Médicale, France (F Bellivier, poste d'accueil and Réseau INSERM No. 4M305D).

1 Abbar M, Courtet P, Malafosse A, Castelanu D. Epidemiological and molecular genetics of suicidal behavior. Encephale 1996; 4: 19-24,

2 Roy A, Rylander G, Sarchiapone M. Genetic studies of suicidal behavior. Psychiatr Clin North Am 1997; 20: 595-611, MEDLINE

3 Brent DA, Bridge J, Johnson BA, Connolly J. Suicidal behavior runs in families. A controlled family study of adolescent suicide victims. Arch Gen Psychiatry 1996; 53: 1145-1152, MEDLINE

4 Abbar M, Amadeo S, Malafosse A, Shenk L, Mallet J, Castelnau D. An association study between suicidal behavior and tryptophan hydroxylase markers. Clin Neuropharmacol 1992; 15: (Suppl 1) 229, MEDLINE

5 Mann JJ, Malone KM, Nielsen DA, Goldman D, Erdos J, Gelernter J. Possible association of a polymorphism of the tryptophan hydroxylase gene with suicidal behavior in depressed patients. Am J Psychiatry 1997; 154: 1451-1453, MEDLINE

6 Nielsen DA, Virkkunen M, Lappalainen J, Eggert M, Brown GL, Long JC et al. A tryptophan hydroxylase gene marker for suicidality and alcoholism. Arch Gen Psychiatry 1998; 55: 593-602, MEDLINE

7 Kunugi H, Ishida S, kato T, Sakai T, Tatsumi M, Hirose T et al. No evidence for an association of polymorphisms of the tryptophan cnhydroxylase gene with affective disorders or attempted suicide among Japanese patients. Am J Psychiatry 1999; 156: 774-776, MEDLINE

8 Tsai SJ, Hong CJ, Wang YC. Tryptophan hydroxylase gene polymorphism (A218C) and suicidal behaviors. NeuroReport 1999; 10: 3773-3775, MEDLINE

9 Frazer A, Hensler JG. Serotonin. In: Siegel GJ et al (eds). Basic Neurochemistry: Molecular, Cellular and Medical Aspects, 6th edn. Lippincott-Raven Publishers: Philadelphia, 1999, pp 263-292.

10 Mann JJ. The neurobiology of suicide. Nature Med 1998; 4: 25-30, MEDLINE

11 Lester D. The concentration of neurotransmitter metabolites in the cerebrospinal fluid of suicidal individuals: a meta-analysis. Pharmacopsychiatry 1995; 28: 45-50, MEDLINE

12 Bellivier F, Leboyer M, Courtet P, Buresi C, Samolyk D, Mallet J et al. Association between the tryptophan hydroxylase gene and suicidal behavior in bipolar patients. Neuropsychiat Genet 1998; 81: 512,

13 Nielsen DA, Jenkins GL, Stefanisco KM, Jefferson KK, Goldman D. Sequence, splice site and population frequency distribution analyses of the polymorphic human tryptophan hydroxylase intron 7. Mol Brain Res 1997; 45: 145-148, MEDLINE

14 Buresi C, Courtet P, Leboyer M, Feingold J, Malafosse A. Association between suicide attempt and the tryptophan hydroxylase (TPH) gene. Am J Hum Genet 1997; 61: A270,

15 Rotondo A, Schuebel K, Bergen A, Aragon R, Virkkunen M, Linnoila M et al. Identification of four variants in the tryptophan hydroxylase promoter and association to behavior. Mol Psychiatry 1999; 4: 360-368, MEDLINE

16 Paoloni-Giacobino A, Mouthon D, Lambercy C, Vessaz M, Coutant-Zimmerli S, Rudolph W et al. Identification and analysis of new sequence variants in the human tryptophan hydroxylase gene. Mol Psychiatry 2000; 5: 49-55, MEDLINE

17 Asberg M, Träskman L, Thoren P. 5-HIAA in the cerebrospinal fluid. A biochemical suicide predictor? Arch Gen Psychiatry 1976; 33: 1193-1197,

18 Träksman L, Asberg M, Bertilsson L, Sjöstrand L. Monoamine metabolites in CSF and suicidal behavior. Arch Gen Psychiatry 1981; 38: 631-636, MEDLINE

19 Nurnberger JL, Blehar MC, Kaufman CA, York-Cooler C, Simpson SG, Harkavy-Freedman Freedman J et al. Diagnostic interview for genetic studies. Arch Gen Psychiatry 1994; 51: 849-864, MEDLINE

20 Sham PC, Curtis D. Monte Carlo tests for association between disease and alleles at highly polymorphic loci. Ann Hum Genet 1995; 59: 97-105, MEDLINE

21 Armitage P, Berry G. Statistical Methods in Medical Research. Blackwell Scientific Publications, Oxford 1987; pp 371-407,

22 Terwilliger JD, Ott J. Handbook of Human Genetic Linkage. Johns Hopkins University Press, Baltimore, 1994,

23 Bellivier F, Leboyer M, Courtet P, Buresi C, Beaufils B, Samolyk D et al. Association between the tryptophan hydroxylase gene and manic-depressive illness. Arch Gen Psychiatry 1998; 55: 33-37, MEDLINE

24 Jönsson EG, Goldman D, Spurlock G, Gustavsson JP, Nielsen DA, Linnoila M et al. Tryptophan hydroxylase and catechol-O-methyltransferase gene polymorphisms: relationships to monoamine metabolite concentrations in CSF of healthy volunteers. Eur Arch Psychiatry Clin Neurosci 1997; 247: 297-300, MEDLINE

25 Manuck SB, Flory JD, Ferell RE, Dent KM, Mann JJ, Muldoon MF. Aggression and anger-related traits associated with a polymorphism of the tryptophan hydroxylase gene. Biol Psychiatry 1999; 45: 603-614, MEDLINE

26 Abbar M, Courtet P, Amadeo S, Caer Y, Mallet J, Baldy-Moulinier M et al. Suicidal behaviors and the tryptophan hydroxylase gene. Arch Gen Psychiatry 1995; 52: 846-849, MEDLINE

27 Nielsen DA, Goldman D, Virkkunen M, Tokola R, Rawlings R, Linnoila M. Suicidality and 5-hydroxyindoleacetic acid concentration associated with a tryptophan hydroxylase polymorphism. Arch Gen Psychiatry 1994; 51: 34-38, MEDLINE

28 Mann JJ, Waternaux C, Haas GL, Malone KM. Toward a clinical model of suicidal behavior in psychiatric patients. Am J Psychiatry 1999; 156: 181-189, MEDLINE

29 Depue RA. A neurobiological framework for the structure of personality and emotion: implications for personality disorders. In: Clarckin JF, Lenzenweger MF (eds). Major Theories of Personality Disorder. The Guilfold Press: New York, 1996, pp 347-390.

Figure 1 Localization of the seven tryptophan hydroxylase gene polymorphisms studied and odds ratios for alleles in suicide attempters (SA, below) and for genotypes in violent suicide attempters (V SA, above).

Table 1 Demographic and clinical characteristics of the patients

Table 2 Population association between TPH polymorphisms and history of suicidal behavior