Methods  The TPH intron 7 A218C polymorphism was determined using a polymerase chain reaction–based method in DNA samples from 152 patients with bipolar disorder and 94 healthy control subjects.

Results  There was a significant association between TPH genotypes and manic-depressive illness. Among patients with bipolar disorder, no association was found between TPH alleles and suicidal behavior.

Conclusions  This result suggests the involvement of the TPH gene in susceptibility to manic-depressive illness. This preliminary result requires confirmation in further groups of patients and controls.

Low 5-HIAA levels in cerebrospinal fluid, indicating reduced serotonergic function, also are associated with impulsive behavior.^8-9 In addition, among impulsive subjects, those with the lowest 5-HIAA concentrations have a history of suicide attempts.¹⁰

Although genetic factors have long been implicated in the etiopathogenesis of bipolar disorders¹¹ and to a lesser extent in that of suicidal behavior,^12-14 the mode of inheritance of manic-depressive illness and suicidal behavior is unclear.

The involvement of genetic factors in manic-depressive illness mostly has been investigated using linkage studies. Several chromosomal regions have been implicated, including 6pter-p24,¹⁵ 13q13,¹⁵ 15q11-qter,¹⁵ 11p15,¹⁶ Xp27,¹⁷ Xq24-q27,¹⁸ the pericentromeric region of chromosome 18,¹⁹ and 21q22.3.²⁰ However, subsequent studies have failed to confirm some of these positive results, possibly because of genetic heterogeneity and complex modes of inheritance of major affective disorders. Thus, nonparametric strategies, which do not require any knowledge of the genetic parameters underlying the disease, may be more appropriate for identifying genes involved in complex diseases. Recent association studies between candidate genes and manic-depressive illness suggest the implication of several genes of minor effect.^21-24 Although stratification bias can cause false-positive results, case-control studies using polymorphic markers close to or within candidate genes may be an appropriate method for detecting genetic susceptibility factors. Thus, despite this bias, case-control analyses of complex diseases (eg, essential hypertension and Alzheimer disease) have proved to be useful.

Candidate genes implicated in 5-HT metabolism already have been tested for association with manic-depressive illness and suicidal behavior. Positive association between the monoamine oxidase A gene and manic-depressive illness have been reported.²⁴ This followed the report of a point mutation in the monoamine oxidase A gene in affected males, belonging to a single pedigree, who showed borderline mental retardation, and violent behavior.²⁵ Association between the 5-HT transporter gene and manic-depressive illness or suicide attempts remain controversial, because positive²⁶ and negative results have been obtained.^27-28 No association was found between 5-HT_1A receptor gene and bipolar disorder.²⁹

The rate-limiting enzyme of 5-HT synthesis—tryptophan hydroxylase (TPH)—is a candidate gene for bipolar disorder and suicidal behavior. Tryptophan hydroxylase catalyzes the oxygenation of tryptophan to 5-hydroxytryptophan, which is then decarboxylated to form 5-HT. The human gene for TPH has been cloned,³⁰ and mapped on the short arm of chromosome 11 (11p14-p15.3).³¹ Two biallelic polymorphisms have been identified in intron 7: A218C and A779C, disclosed by restriction fragment length polymorphism analysis and by single-strand conformational polymorphism analysis, respectively.³² These 2 polymorphisms are in complete linkage disequilibrium in West European Caucasian controls.³² Two other polymorphisms were disclosed by AvaII and HinfI restriction enzymes, using the human TPH copy DNA probe C2-38, but have not been accurately mapped.^{30, 33}

In a group of Finnish violent offenders meeting criteria for alcohol abuse, Nielsen et al³⁴ reported a significant difference in the conformational TPH polymorphism in intron 7 between subjects with and without a history of suicide attempts. In contrast, Abbar et al,³³ using TPH C2-38/AvaII polymorphism, did not find any significant difference between patients who attempted suicide and normal control subjects. No association was found between patients with bipolar disorder and controls with TPH/AvaII in a preliminary case-control study.³⁵

To further explore the involvement of the TPH gene, we studied the TPH A218C polymorphism in a new sample of patients with bipolar disorder, some with and some without a history of suicide attempts, and in controls.

PATIENTS AND METHODS

After giving informed consent, 152 patients with bipolar disorder (103 patients with bipolar I disorder and 49 patients with bipolar II disorder) and 94 healthy volunteers were included in this study. Patients and controls were all French (at least 3 grandparents born in France). Patients with bipolar disorder (84 women and 68 men) were recruited from consecutive admissions to the psychiatric unit and controls (37 women and 57 men) were blood donors from Hôpital Pitié-Salpêtrière, Paris, France, and Hôpital Corentin Celton, Issy-les-Moulineaux, France.

Patients and controls were interviewed by trained psychiatrists (F.B. and M.L.) with a French version of the Diagnostic Interview for Genetic studies.³⁶ Diagnosis with DSM-III-R,³⁷ Research Diagnostic Criteria, and history of suicide attempts were assessed with this instrument. Demographic and clinical characteristics of patients with bipolar disorder and controls are given in Table 1.

View this table: [in this window] [in a new window] Table 1. Demographic and Clinical Characteristics

To minimize morbidity among subjects in the control group, only blood donors older than 35 years were included. Controls with a personal or family history of mood disorders or suicide attempt disclosed by interview were excluded.

Suicide attempts were classified as violent or nonviolent according to the criteria proposed by Asberg et al.² The criterion for suicide attempts was death intent requiring medical treatment at a hospital. Death intent and medical seriousness of the attempt were rated according to the 6-point scale of the Diagnostic Interview of Genetic Studies. Hanging attempts, use of firearms or knives, and jumping from heights were considered as violent attempts; drug overdoses were considered as nonviolent suicide attempts.

LABORATORY METHODS

Because the 2 intron 7 TPH polymorphisms (A218C and A779C) are in complete disequilibrium,³² we tested A218C, which is more easily revealed. Twenty-milliliter edetic acid blood samples were collected, and DNA was prepared from lymphocyte pellets by sodium dodecyl sulfate lysis, proteinase K digestion, phenol and chlorophorm extraction, and ethanol precipitation and was resuspended in TRIS buffer and edetic acid. For genotyping, target sequences were amplified from 200 ng of genomic DNA using polymerase chain reaction (PCR) in a 50-µL reaction volume using a Hybaid cycler (MWG-Biotech GmbH, Ebersberg, Germany). The PCR primers were HTHSSCPA, 5'-TTC AGA TCC CTT CTA TAC CCC AGA-3'; and HTHSSCP5, 5'-GGA CAT GAC CTA AGA GTT CAT GGC A-3'; in 1.5-mmol/L magnesium chloride, 50 mmol of potassium chloride, 10 mmol of TRIS buffer, and 20 pmol of each oligonucleotide, 2 U of Taq DNA polymerase (Brunschwig, Basel, Switzerland) and 200 mmol of each dNTP per tube. The PCR program was as follows: first cycle for 3 minutes at 95°C, 30 seconds at 58°C, and 1 minute at 72°C; and 29 cycles for 1 minute at 92°C, 30 seconds at 58°C, and 1 minute at 72°C. After amplification, PCR products were digested with 8 U of BfaI (BioConcept, Allschwil, Switzerland) overnight, then electrophoresed on a 1.4% agarose gel (Brunschwig). DNA was visualized by ethidium bromide staining and UV transillumination. The uncut amplicon is 918 base pairs (bp) long. After digestion, the "A" allele gives 860-bp and 58 (not seen in the gel)-bp fragments, and the "C" allele gives 615-, 245-bp, and 58 (not seen in the gel)-bp fragments.

STATISTICAL ANALYSIS

Allelic and genotypic distributions between patients and controls were compared using the ² test. Odds ratios and 95% confidence intervals were calculated using the Woolf's method. The Armitage linearity tendency test³⁸ was used to identify any dose effect of a susceptibility allele. Differences were considered statistically significant at P=.05.

RESULTS

No association was found between the TPH gene polymorphism (A218C) and suicidal behavior (presence vs absence of lifetime suicide attempt) in the patients with bipolar disorder. Among those who attempted suicide, no association was found for violent vs nonviolent suicide subgroups.

COMMENT

The risk of bipolar disorder is increased by the presence of at least 1 copy of the TPH A allele, and the risk is higher for TPH A–homozygous subjects. In our sample, there was a dose effect for the risk associated with the A allele for bipolar disorder as assessed by the linearity tendency test (²=12.19, df=1; P<.001).³⁸

The allelic frequencies in our control population (Table 2) were similar to those reported by Nielsen et al³² in West European Caucasian control populations. Our sample of patients with bipolar disorder was recruited with the same geographical criteria as controls, ie, Caucasian and of French origin for at least 2 generations. Despite these precautions, there may have been a stratification bias, and replication is required in other samples of patients with bipolar disorder, using other nonparametric methods.

View this table: [in this window] [in a new window] Table 2. Allelic and Genotypic Distributions and Frequencies of the Tryptophan Hydroxylase Gene Marker*

It is unclear how the TPH A allele acts as a risk factor for bipolar disorder. The A218C TPH polymorphism is located in a potential GATA transcription-factor binding site and therefore might affect TPH gene expression.³² Alternatively, this polymorphism may be nonfunctionally significant and our results may indicate only linkage disequilibrium between the TPH A allele and a different mutation in the TPH gene, or another nearby gene. Although the other TPH intron 7 polymorphism (A779C SSCP) is located upstream from the 3 acceptor splice site, it has no functional significance because no aberrant splice product from the TPH gene has been detected.³²

We did not find any association between a previously studied small sample of patients with bipolar disorder (N=70) and another restriction fragment length polymorphism of the TPH gene (AvaII/C2-38).³⁵ Because the location of the AvaII polymorphic site in the TPH locus is unknown, these discrepant results may be due to large distance between the 2 polymorphic sites (ie, the AvaII and the A218C polymorphism).

Possibly, a clinical or biological trait underlying manic-depressive illness could be associated with this TPH polymorphism. We, therefore, looked for an association between TPH polymorphism and suicide attempts. However, in our sample, there was no difference for A allele frequency between patients with bipolar disorder with and without a history of suicide attempts (0.56 and 0.48, respectively). This observation does not confirm the suggestion by Abbar et al³³ that carrying the U allele (A779C polymorphism), which is in strong disequilibrium with the A allele (A218C polymorphism), protects against suicidal behavior in patients with bipolar disorder.

Furthermore, in our sample of patients who attempted suicide, no difference was found for the A allelic frequency between violent and nonviolent attempt subgroups (0.57 and 0.55, respectively). Nevertheless, this negative result could be due to lack of power because of the small sample size. Our sample of patients with bipolar disorder who have attempted suicide (14 violent and 38 nonviolent attempts) provides a power of less than 20% for detecting a difference between frequencies of 0.55 and 0.57 (unilateral test).

In summary, TPH intron 7 polymorphism seems to be associated with violent behavior without a history of suicide attempts³⁴ and with manic-depressive illness. This TPH polymorphism might be associated with a trait related to violence and bipolar disorders rather than directly to these phenotypes themselves.

This research was supported by grant CRC 932208 from Assistance Publique, Institut National de la Santé et de la Recherche Médicale, Paris, France (Dr Bellivier) and grant 32-47315-96 from Fond National de la Recherche Suisse, Geneva, Switzerland.

We thank Jean-Louis Beaumont, MD, and his collaborators, as well as nurses Mireille Barré, Marie-Christine Benhouidga, Caroline Brachet, Nadine Gaudot, Marinette Naudier, Martine Platt, and Jean-Luc Thomas for technical assistance. We thank D. A. Nielsen, G. L. Jenkins, K. M. Stefanisco, K. K. Jefferson, and D. Goldman for providing their data before publication.

Reprints: Frank Bellivier, MD, Laboratoire de recherche sur les personnalités et conduites adaptatives, CNRS URA 1957 Pavillon Clérambault, Hôpital Pitié-Salpétrière, 47 boulevard de l'Hôpital, Paris 75013, France (e-mail: fbell{at}ext.jussieu.fr).