TPH and suicidal behavior: a study in suicide completers

Article metrics


An association between the gene that codes for tryptophan hydroxylase (TPH)—the rate-limiting enzyme in the synthesis of serotonin—and suicidal behavior has been investigated with some detail in samples of living subjects who attempted suicide. In this study, we investigated TPH and suicide completion, the most severe form of suicidal behavior. A relatively large sample of suicide completers (n = 101) was genotyped at three TPH loci (two polymorphisms in the promoter region, A-6526G and G-5806T, and one in intron 7, A218C) and compared to psychiatrically normal living controls (n = 129). Although no significant differences were found between groups for genetic variation at single loci, haplotype analysis revealed that one haplotype (-6526G -5806T 218C) was significantly more frequent among suicide cases than in normal controls (χ2 = 11.30, df = 2, P = 0.0008; OR = 2.0 CI: 1.30–3.6). Further analyses suggested that this haplotype is particularly more frequent among subjects who committed suicide using violent methods. Similar results were observed in recent haplotype analyses in suicide attempters, which found that the equivalent of haplotype -6526G -5806T 218C was more frequent in impulsive attempters (Rotondo et al, Mol Psychiatry 1999; 4: 360–368). Our results replicate in suicide completers previous data observed in suicide attempters. These and other results continue to point to the substantial role that the gene that codes for TPH may play in the neurobiology of suicidal behavior


Suicidal behavior—which is often regarded as comprising suicide completion, suicide attempt and suicidal ideation—is an important public-health problem. Worldwide, suicide ranks among the top ten causes of death in young males and consequently, it is one of the leading causes of potential years of life lost.1 Although less imposing, suicide attempts and suicidal ideation are quite frequently observed and also inflict a heavy burden on society. The etiology of suicidal behavior is certainly complex, with different factors likely playing an important role. Among these are genetic factors, as suggested by several genetic-epidemiologic studies.234567 Trying to identify specific genes that may increase the predisposition of subjects to attempt suicide, considerable attention has been paid to the possible role of the gene that codes for tryptophan hydroxylase (TPH). TPH is the enzyme that catalyzes the oxygenation of tryptophan to 5-hydroxytryptophan, which is then decarboxylated to serotonin. In light of the important body of evidence consistently suggesting that subjects with suicidal behavior have reduced serotonergic neurotransmission,8 TPH is certainly an interesting candidate gene.

Following a 1994 report9 of an association between variation at TPH and suicide attempts in a group of violent offenders, several studies were published trying to replicate this finding. Although additional reports are consistent with the presence of this association,101112 there remains some controversy because of studies which failed to confirm these findings. All previously published studies, both supportive and otherwise of an association between suicidal behavior and TPH, have investigated subjects who attempted suicide. In this study we are reporting the investigation of three polymorphisms located on the TPH gene in a relatively large sample of subjects who have completed suicide, the more extreme and less ambiguous form of suicidal behavior.

Subjects were identified as part of an ongoing collaboration with the Coroner's Office of the Montreal Central Morgue, where suicide cases are consecutively recruited and a blood or tissue sample is collected. In this study, 101 suicide completers and 129 normal controls were included. The mean ages were 32.20 ± 9.38 and 34 ± 10.56 years for suicide cases and controls, respectively. Although diagnostic information was available for a proportion (46%) of the suicide subjects, 6-month and lifetime prevalence rates of psychopathology are comparable with our previous studies in a different sample from the same population,13 as well as with data from the literature.4567891011121314151617 Approximately half of the subjects (45.7%) had a major depressive episode within 6 months prior to death and a third of the subjects (30.4%) met criteria for major depressive disorder at least once in their lifetime. Of the remaining major psychiatric disorders, as expected, life-prevalence rates were important for substance abuse or dependence (32.6%), as well as for antisocial (15.2%) and borderline (15.2%) personality disorders, or for these two conditions combined (13%).

Subjects committed suicide as follows: hanging (56.4%), shooting (10.9%), CO poisoning (9.9%), drug overdose (8.9%), jumping (6.9%), drowning (3%), blunt or penetrating lesion (2.0%), others (1.0%). These figures are similar to available data on suicide methods for the Canadian population.18

TPH genetic variation was investigated using three different polymorphisms: A218C, a variant located on intron 7 of the TPH gene,19 and two common polymorphisms located on the promoter region, A-6526G and G-5806T.20 These three markers were in linkage disequilibrium. Total normalized disequilibrium values (D′) were 0.79, 0.84 and 0.97 respectively for A218C vs A-6526G, A218C vs G-5806T and A-6526G vs G-5806T. We could not reject the hypothesis of three-way disequilibrium between the three loci following a forward-selection testing strategy as suggested by Long et al21 and using an empirical distribution (5000 simulates) to assess significance (G = 139.46, P < 0.0002). The forward selection strategy consists of increasingly adding significant component sets of disequilibrium to the most restricted model and contrasting these models with the appropriate null hypotheses.

Table 1 shows allelic and genotypic distributions for the three loci typed. No significant differences were found between groups. The results of genotypic comparisons were as follows for each polymorphism: A218C (χ2 = 1.47, df = 2, P = 0.48), A-6526G (χ2 = 0.28, df = 2, P = 0.86) and G-5806T (χ2 = 0.428, df = 2, P = 0.81). Similar results were observed for allelic comparisons: (χ2 = 0.01, df = 1, P = 0.91), A-6526G (χ2 = 0.08, df = 1, P = 0.77) and G-5806T (χ2 = 0.455, df = 1, P = 0.50).

Table 1 Genotype and allele distributions for the three TPH loci investigated in completed suicide cases and healthy controls

Estimated haplotype frequencies are listed in Table 2 and their distribution in cases and controls is shown in Figure 1. Overall frequency comparisons indicated a significant difference between cases and controls at the haplotype level (χ2 = 23.31, df = 6, P = 0.00069). This difference reflected mainly different distributions at haplotype -6526G -5806T 218C, which was significantly more frequent among suicide cases (χ2 = 11.30, df = 2, P = 0.0008; OR = 2.0 CI: 1.30–3.6). After correction for multiple testing (seven haplotype, three allelic and three genotype comparisons), this difference remained significant (P = 0.01). Visual inspection of Figure 1 also suggests that haplotype -6526A -5806G 218C is more frequent among controls. However, this difference (χ2 = 5.22, df = 2, P = 0.02; OR = 0.59 CI: 0.37–0.95) did not remain significant after correcting for multiple testing.

Table 2 Estimated haplotype frequencies for completed suicide cases and healthy controls
Figure 1

Haplotype distributions in suicide completers and normal controls. All eight possible haplotypes are on the x axis and frequency (estimated percentage) on the y axis. ▪ Haplotype -6526G -5806T 218C was significantly more frequent among suicide cases (χ2 = 11.30, df = 2, P = 0.0008); O Haplotype -6526A -5806G 218C was more common among controls (χ2 = 5.22, df = 2, P = 0.02). After correction for multiple testing only the difference observed for haplotype -6526G -5806T 218C remains significant.

To investigate possible differences between violent and non-violent suicide cases, we compared haplotype distributions between these two groups. Interestingly, haplotype -6526G -5806T 218C was present in 38.17% of violent completers, but only in 25% of the non-violent cases, and thus, resembled the figures found for controls (23.58%). However, these results should be considered with caution as the non-violent sample of suicide completers consisted of only nine subjects.

Although the precise relationship between the different forms of suicidal behaviors is not yet known, some authors believe that they represent different degrees of severity of a common underlying process.22 Thus, it is possible to speculate that by focusing on suicide completers we have access to a more severe phenotype. In this sense, the study of genetic variation at TPH in suicide completers is of particular interest given the previous results obtained among suicide attempters.

In this study, we have been unable to reproduce among suicide completers previous findings from TPH studies in suicide attempters that suggested a positive association between genetic variation at intron 7 loci (A779C, A218C) and suicidal behavior.9101112 Moreover, our results do not suggest that polymorphisms located in the promoter region of the TPH gene (A-6526G and G-5806T) are associated with suicide. Nevertheless, our results are consistent with an association between genetic variation at TPH and suicide, as haplotype analysis suggested that haplotype distributions are different between groups. More specifically, the -6526G -5806T 218C haplotype was significantly more frequent among suicide cases than in normal controls. Are the negative association results observed with each single locus and the positive results observed in the haplotype analysis consistent with each other? With the exception of situations where the single marker tested is the predisposing allele itself, in principle, haplotypes provide increased statistical power to detect the presence of a disease-predisposing locus.23 Thus, there seems to be no conflict between the results observed in this study, providing that none of the single polymorphisms tested is the ‘disease’ locus itself.

Does haplotype -6526G -5806T 218C increase predisposition to suicidal behavior? Though tempting, such conclusions should be carefully examined, in the light of the following considerations. First, it will be important to assess the functional role associated with this particular haplotype. The A218C polymorphism does not alter amino acid sequence and neither appears to influence gene transcription.19 Preliminary data investigating protein–nucleic acid interactions in sequences containing the different promoter variants suggest that both polymorphisms may alter protein binding.20 Thus, it is possible that different haplotypes confer functional variability. Second, this result should be confirmed in larger samples and by independent investigators. Recently, Rotondo et al20 investigated among violent offenders (with and without a history of suicide attempts) the frequency of haplotypes containing variation at A-6526G and C779A, another intron 7 polymorphism which is reportedly in complete disequilibrium with A218C. These authors found that subjects with suicidal behavior more frequently had the equivalent of the -6526G 218C haplotype. Our results are in agreement with Rotondo et al’s findings.

An interesting preliminary observation was that the frequency of the associated haplotype (-6526G -5806T 218C) seems to be increased only in suicide cases that used violent methods to complete suicide. Similarly, the frequency of the haplotype -6526A -5806G 218C in non-violent cases resembles that of the control subjects. Although these data should be regarded with caution, as only a few subjects used non-violent methods, it is consistent with reports from previous association studies between TPH and suicidal behavior that observed positive results particularly among attempters with aggressive and impulsive traits,91020 as well as with aggression and anger-related traits.24 If confirmed in larger numbers, this observation will be congruent with the notion that serotonergic alterations in suicidal behavior are more likely to be present in subjects who used violent methods.

It would have been interesting to investigate the effect of possible interactions with specific psychiatric diagnoses on the observed association. Unfortunately, the lack of psychiatric diagnoses in all subjects made such analysis impossible at this time. However, we are in the process of completing the psychological autopsies in the whole sample and should be able to investigate possible interactions with specific morbidity as well as with variation at other loci in the near future.

In summary, we carried out an association study investigating genetic variation at the TPH locus in suicide completers. Although our results do not directly confirm previous association findings with individual loci, it is consistent with previous haplotype studies in suicide attempters and supports the role of TPH in the predisposition to suicidal behavior.



Subjects for this study were 101 suicide completers and 129 normal controls. Diagnoses were made as described elsewhere.13 Briefly, a psychological autopsy was carried out with at least one family member identified as the best respondent and best consensus diagnoses were made by a panel of psychiatrists according to the DSM-IV criteria. Cases were classified according to the violence of the method used to commit suicide. Following the usual classification, subjects who died by overdose were considered non-violent, whereas all other cases were considered violent. Controls in this study were living subjects without a history of suicidal behavior and without major psychiatric diagnoses. Cases and controls were Caucasian of French-Canadian origin. All subjects included in this study (or their families in the case of the suicide cases) provided written informed consent, which was approved by the local IRB.


Genomic DNA was extracted either from blood or frozen brain tissue using standard procedures. PCR was carried out in a total volume of 12.5 μl containing 40 ng of genomic DNA; 125 ng of the specific primers; 200 μM each of dGTP, dCTP, and dTTP; 25 μM dATP; 0.5 units of Taq DNA polymerase (Bio/Can Scientific, Mississauga, Ontario, Canada); and 2.0 μl of 10 × buffer (Bio/Can Scientific) with MgCl2 included in the final concentration of 1.5 mM. Samples were processed throughout 35 cycles of denaturation at 94°C, annealing specific to each primer pair, and elongation at 72°C. PCR products were digested with Bfa I (A218C), Sau3AI (A-6526G) and MslI (G-5806T) and analyzed on a 2% agarose gel. Genotypes were read blind to the group status of the sample.

Statistical analysis

Genotype and allele frequency distributions were compared between suicide cases and controls using the χ2 test and odds ratio when applicable. Statistical analyses were carried out using the SPSS package (SPSS standard version, release 10.0.5). Haplotype frequencies were estimated using maximum likelihood based on the expectation maximization algorithm described by Long et al21 and implemented in the computer program 3LOCUS. Correction for multiple testing was carried out using the Bonferroni procedure.


  1. 1

    Diekstra RFW . The epidemiology of suicide and parasuicide Acta Psychiatr Scand 1993 371: 9–20

  2. 2

    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

  3. 3

    Tsuang M . Risk of suicide in the relatives of schizophrenics, manics, depressives, and controls J Clin Psychiatry 1983 44: 396–400

  4. 4

    Roy A, Segal NL, Sarchiapone M . Attempted suicide among living co-twins of twin suicide victims Am J Psychiatry 1995 152: 1075–1076

  5. 5

    Roy A, Segal NL, Centerwall BS, Robinette CD . Suicide in twins Arch Gen Psychiatry 1991 48: 29–32X

  6. 6

    Wender PH, Seymour SK, Rosenthal D, Schulsinger F, Ortmann J, Lunde I . Psychiatric disorders in the biological and adoptive families of adopted individuals with affective disorders Arch Gen Psychiatry 1986 43: 923–929

  7. 7

    Statham DJ, Heath AC, Madden PA et al. Suicidal behaviour: an epidemiological and genetic study Psychol Med 1998 28: 839–855

  8. 8

    Mann JJ . The neurobiology of suicide Nature Med 1998 4: 25–30

  9. 9

    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: 34–38X

  10. 10

    Nielsen DA, Virkkunen M, Lappalainen J et al. A tryptophan hydroxylase gene marker for suicidality and alcoholism Arch Gen Psychiatry 1998 55: 593–602

  11. 11

    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

  12. 12

    Tsai SJ, Hong CJ, Wang YC . Tryptophan hydroxylase gene polymorphism (A218C) and suicidal behaviors Neuroreport 1999 10: 3773–3775

  13. 13

    Lesage AD, Boyer R, Grunberg F et al. Suicide and mental disorders: a case-control study of young men Am J Psychiatry 1994 151: 1063–1068

  14. 14

    Conwell Y, Duberstein PR, Cox C, Herrmann JH, Forbes NT, Caine ED . Relationships of age and axis I diagnoses in victims of completed suicide: a psychological autopsy study Am J Psychiatry 1996 153: 1001–1008

  15. 15

    Beautrais AL, Joyce PR, Mulder RT, Fergusson DM, Deavoll BJ, Nightingale SK . Prevalence and comorbidity of mental disorders in persons making serious suicide attempts: a case-control study Am J Psychiatry 1996 153: 1009–1014

  16. 16

    Barner-Rasmussen P, Dupont A, Bille H . Suicide in psychiatric patients in Denmark, 1971–81. I. Demographic and diagnostic description Acta Psychiatr Scand 1986 73: 441–448

  17. 17

    Barraclough B, Bunch J, Nelson B, Sainsbury P . A hundred cases of suicide: clinical aspects Br J Psychiatry 1974 125: 355–373

  18. 18

    Strategie quebecoise d'action face au suicide . S'entraider pour la vie. Gouvernament du Quebec Ministere de la sante et des services sociaux 1998

  19. 19

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

  20. 20

    Rotondo A, Schuebel K, Bergen A et al. Identification of four variants in the tryptophan hydroxylase promoter and association to behavior Mol Psychiatry 1999 4: 360–368

  21. 21

    Long JC, Williams RC, Urbanek M . An E-M algorithm and testing strategy for multiple-locus haplotypes Am J Hum Genet 1995 56: 799–810

  22. 22

    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

  23. 23

    Devlin B, Risch N . A comparison of linkage disequilibrium measures for fine-scale mapping Genomics 1995 29: 311–322

  24. 24

    Manuck SB, Flory JD, Ferrell 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

Download references


This study was funded in part by the Réseau de Santé Méntale du Québec—Suicide Axis and the Canadian Medical Research Council grant 38078.

Author information

Correspondence to G Turecki.

Rights and permissions

Reprints and Permissions

About this article


  • suicide completion
  • tryptophan hydroxylase gene
  • haplotype analysis
  • suicide method

Further reading