Haplotype Analysis Reveals Tryptophan Hydroxylase (TPH) 1 Gene Variants Associated with Major Depression

ORIGINAL ARTICLES

Haplotype Analysis Reveals Tryptophan Hydroxylase (TPH) 1 Gene Variants Associated with Major Depression Rinat Gizatullin, Ghazal Zaboli, Erik G. Jönsson, Marie Åsberg, and Rosario Leopardi Background: Tryptophan hydroxylase (TPH) is the rate-limiting enzyme in the biosynthesis of serotonin (5-HT) and might be related to the pathogenesis of major depression (MD). Two isoforms are known, TPH-1 and TPH-2. Tryptophan hydroxylase-1 association with MD is still debated. Methods: A single nucleotide polymorphism (SNP) screening strategy was used to define TPH-1 haplotypes spanning over 23 kilobase (kb) of the 29 kb gene length. Genotyping was performed in 228 MD patients and 253 healthy control subjects. Results: Six SNPs were found at linkage disequilibrium in both patients and control subjects, suggesting a haplotype block structure. Single marker association analyses showed only one SNP significantly associated with MD. Several haplotypes were associated with MD. When all six locus haplotypes were divided into two groups, above or below a 5% threshold, the compound haplotype group below a 5% frequency resulted as associated with the disease (31.6% vs. 18.0% in control subjects, p ⬍ 10⫺5). A “sliding window” analysis attributed the strongest disease association to a haplotype configuration localized between introns 7 and 8 (p ⬍ 10⫺5). Conclusions: Haplotype analysis indicates that TPH-1 associates with MD. The most common TPH-1 variants appear to carry no risk, while some of the less frequent variants might contribute to genetic predisposition to MD. Key Words: Gene variant, haplotype, linkage disequilibrium, major depression, tryptophan hydroxylase, SNP

M

ajor depression (MD) is the most common psychiatric illness, with a lifetime prevalence of about 15% to 20% (Fava and Kendler 2000). The heritability of depression is estimated to 30% to 40% (Fava and Kendler 2000; Malhi et al 2000). Alterations in the serotonin (5-HT) system have been related to the origin of depression (Ressler and Nemeroff 2000). There is extensive evidence for the presence of 5-HT nerve terminals and/or 5-HT receptors in neuroendocrine regions such as hippocampus, hypothalamus, and brainstem (Chaouloff 1993; Graeff 1993). Serotonin turnover is particularly active in cortical and limbic areas involved in emotional aspects of behavior (Westenberg 1996; Whitaker-Azmitia et al 1990). Serotonergic neurotransmission is related to anxiety in animal models as well as in humans (Cloninger 1987; Griebel 1995; Handley 1995). Clinically, reduced 5-HT uptake has been associated with depression and anxiety (Faludi et al 1994; Iny et al 1994; Owens and Nemeroff 1994). Several genes of the serotonergic system have been variously correlated with depression by gene polymorphism association studies (Malhi et al 2000). Among these is the tryptophan hydroxylase (TPH) gene, coding for the rate-limiting enzyme in the biosynthesis of 5-HT (Cooper and Melcer 1961). Tryptophan hydroxylase variants might be associated with pathogenesis events involving dysfunction of the 5-HT system. As such, TPH is one of the major candidate genes for psychiatric and behavioral disorders, in particular depression and suicidal behavior (Roy et al 1997). Until recently, only one gene encoding TPH was

From the Department of Clinical Neuroscience (RG, GZ, EGJ, MA, RL), Psychiatry Section, Karolinska Institute and Hospital; and Center for Molecular Medicine (RG, GZ, RL), Karolinska Institute, Stockholm, Sweden. Address reprint requests to Rosario Leopardi, M.D., Ph.D., Karolinska Institute, Department of Clinical Neuroscience, Psychiatry Section, Karolinska Hospital, Bldg R5, 17176 Stockholm, Sweden; E-mail: rosleo@ mbox.ki.se. Received April 21, 2005; revised July 12, 2005; accepted July 27, 2005.

0006-3223/06/$32.00 doi:10.1016/j.biopsych.2005.07.034

known, which we refer to as TPH-1. The gene is localized on the human chromosome 11p15.3-p14, it is about 29 kilobase (kb) long, and includes 11 exons (Craig et al 1991). Nielsen et al (1997) identified two informative single nucleotide polymorphisms (SNPs) in TPH-1, A218C (rs1800532) and A779C (rs1799913), both located in the TPH-1 gene intron 7. Single nucleotide polymorphism A779C (rs1799913) has been associated with 5-hydroxyindol acetic acid (5-HIAA) concentrations in cerebrospinal fluid (CSF) (Jonsson et al 1997). Based on intron 7 polymorphism, TPH-1 variants were shown to be a potential liability factor for suicidal behavior in a sample of Finnish alcoholics (Nielsen et al 1998) and in a group of Swedish patients with unipolar disorder, substance abuse, and cluster B personality disorders (Geijer et al 2000). Further association studies of TPH intron 7 variants have either partially confirmed the findings in other populations (Mann et al 1997; Paik et al 2000) or not found any association (Abbar et al 1995; Bennett et al 2000; Ono et al 2000). Association between the TPH-1 SNP A218C (rs1800532) and bipolar disorder has also been reported (Bellivier et al 1998). However, such findings have not been confirmed either for bipolar disorder or for other mood disorders (Furlong et al 1998; Serretti et al 2002). Also, a recent meta-analysis concluded that results from association studies between TPH-1 and bipolar disorder are not reliable (Lohmueller et al 2003). Recently, a second isoform of TPH, thereby named TPH-2, has been described (Walther and Bader 2003). Its gene is located on chromosome 12 and shows 71% homology to TPH-1 on the amino acid level (Walther and Bader 2003). Tryptophan hydroxylase-1 and TPH-2 are expressed in nearly equal amounts in several human brain regions such as frontal cortex, thalamus, hippocampus, hypothalamus, and amygdala. Tryptophan hydroxylase-2 is predominantly expressed in the brainstem, whereas it is absent from peripheral tissues such as heart, lung, kidney, duodenum, and adrenal gland, where only TPH-1 is expressed (Zill et al 2004a, 2004b). Haplotype analysis of TPH-2 showed recently a significant association with MD (Zill et al 2004a), suggesting that variants of this gene could represent risk factors in the development of MD or others psychiatric disorders. Single marker analyses in case-control studies are prone to generate conflicting results in common disease association studBIOL PSYCHIATRY 2006;59:295–300 © 2006 Society of Biological Psychiatry

296 BIOL PSYCHIATRY 2006;59:295–300

R. Gizatullin et al

ies (Cardon and Bell 2001). This is due, at least in part, to the fact that common psychiatric disorders are predicted to associate with common alleles (Cardon and Bell 2001; Lohmueller et al 2003). The use of haplotypes as more specific risk markers than single alleles is currently being explored (Clark 2004; Judson et al 2000). Haplotype blocks have become central to this type of approach (Gabriel et al 2002). However, reliable construction of such blocks would require parent-offspring transmission data, which are often not available. An alternative concept directed toward molecular pathogenetic studies is the gene-based haplotype, i.e., a combination of alleles located within a gene unit, independent of genetic heritability (Hoehe 2003). Gene-based haplotypes do not correspond to haplotype blocks, as one haplotype block may contain more than one gene and one gene unit may span over several haplotype blocks. Potentially, detailed gene-based haplotypes could be the most precise markers possible for a given gene, as they would contain all the variations in the gene (Hoehe 2003). We have initiated studies aimed at applying the use of gene-based haplotypes as a starting point for the identification of risk haplotypes within a classical case-control study design. However, approaches to construct gene-based haplotypes carrying risk for disease are complex (Glazier et al 2002; Hoehe 2003; Judson et al 2000). Extensive genotyping and/or sequencing would be required for entire genes in every subject within a chosen study population, often a prohibitive effort. Thus, it would be desirable to devise strategies that reduce complexity. To this end, we have introduced a hypothesis-based screening strategy. We report here results obtained applying our screening strategy to the analysis of the TPH-1 gene in a group of 228 patients suffering from MD and in 253 healthy control subjects. This approach led us to identify six SNPs and to define TPH-1 haplotypes, spanning about 23 kb from the promoter region to exon 8, which are associated with MD.

Methods and Materials Human Subjects The study was approved by the Ethics Committee of the Karolinska Hospital. Both control and patient groups consisted of unrelated Caucasian individuals of North European descent

living in the Stockholm County. All subjects were interviewed by specially trained physicians, using the Structured Clinical Interview for DSM-III-R (Spitzer et al 1986) or DSM-IV (First et al 2002). The control group was recruited as reported earlier (Gustavsson et al 1999). Briefly, subjects were either re-examined healthy individuals, mainly staff and medical students, or subjects drawn from the general population for biological psychiatric studies performed at the Karolinska Institute. The control group included 253 individuals (mean age ⫾ standard deviation: 42.6 ⫾ 13.7 years, 98 women and 155 men). At the time of blood collection and diagnostic interview, none of the individuals in this group had suffered from any psychiatric condition. The patient group was composed of 228 subjects (mean age ⫾ standard deviation: 46.8 ⫾ 8.6 years, 166 women and 62 men). Patients were recruited from a major insurance company while on long-term (over 3 months) sick leave. Letters were sent to patients on sick leave because of a nonpsychotic psychiatric diagnosis. Patients were then approached by telephone and invited to participate in a clinical study. All patients were ambulatory, and none had received inpatient care for their current illness. Patients meeting DSM-IV criteria for major depressive disorder at any time during the current sick leave period were included in the present study. Genotyping Venous blood was drawn and immediately frozen in aliquots at ⫺70°C or below until analyzed. Genomic DNA was prepared from whole blood by use QIAamp DNA Blood Mini kit (Qiagen GmbH, Hilden, Germany). The extracted DNA was stored at 4°C until analyzed. The DNA (50 ng/reaction) was amplified by polymerase chain reaction (PCR), carried out in a T3 Thermocycler (Biometra GmbH, Göttingen, Germany) in a total volume of 25 ␮L. The reaction buffer was composed of 1.5 to 2.5 mmol/L magnesium chloride (MgCl2), 67 mmol/L Tris-HCl pH 9.2 (Sigma, Stockholm, Sweden), 16.6 mmol/L ammonium sulfate [(NH4)2SO4], .1% vol/vol Tween 20 (Amersham Pharmacia Biotech, Uppsala, Sweden), 200 ␮m deoxyribonucleotide triphosphates (dNTPs), 20 pmol of each primers (MWG Biotech AG, Ebersberg, Germany), and .75 units of Taq DNA polymerase (Roche Diagnostic GmbH, Mannheim, Germany). The PCR products were then digested overnight with appropriate restriction enzymes, subjected to electrophoresis on 2%

Table 1. Primers for PCR Amplification and Pyrosequencing Marker

SNP IDa

SNP Positionb

1

rs4537731

124889

2

rs 684302

116360

3

rs211105

111308

4

rs1800532

103821

5

rs1799913

103260

6

rs7933505

101992

Primer Sequence TTTATGGCATTGAAGTAAGAGCAC TTTTGGCTCCTGGCACTTAAC AGAGAGATGGAGCAAAACACTAC CCAGTCCTTCCAAATCTGATAC CAAGGCAAGATTTATATGAGTT CTCAGGAAAACAGAAGGGTAGGGT AATGGCATCTACCTTATGGGTTC CTTTATTTTTCTCCATGGGACTCA ATTGGATTTCGATTTGATTG GGCAAAACTAGGTTCAGC CAGCGTGACAAACTTGTACC CCCCAAAGCTTTTGTTGTGCGT TCAGATTCCACATTGCCGTTGAAC

Annealing Temperature

Restriction Enzyme

55°C

SaulIIA

55°C

AluI

55°C

CfrI

58°C

NheI

55°C

Pyrc

55°C

BauI

PCR, polymerase chain reaction; SNP, single nucleotide polymorphism; NCBI, National Center for Biotechnology Information. a SNP ID number from the NCBI SNP database (http://www.ncbi.nlm.nih.gov/SNP/). b SNP and primer positions are shown based on NCBI clone AC124058. c For detection of SNP rs1799913 pyrosequencing was used.

www.sobp.org/journal

Primer Positionb 125089–125066 124772–124792 116248–116270 116459–116438 111050–111071 111481–111458 103993–103971 103655–103678 103335–103316 103196–103213 103283–103264 101746–101767 102194–102171

BIOL PSYCHIATRY 2006;59:295–300 297

R. Gizatullin et al agarose gels (Roche Diagnostic GmbH), and visualized after an ethidium bromide staining. Table 1 shows primer sequences, PCR annealing temperatures, and restriction enzymes used for each SNP. Single nucleotide polymorphism rs1799913 was analyzed by pyrosequencing, using a Pyrosequencer PSQ 96 and a PSQ 96 SNP Reagent Kit (Pyrosequencer, Uppsala, Sweden) according to the manufacturer’s instructions. Polymerase chain reactions were carried out in 50 ␮L volume containing 100 ng genomic DNA, PCR buffer (1.5 mmol/L MgCl2, 10 mmol/L Tris-HCl pH 8.3, 50 mmol/L potassium chloride [KCl], .1% vol/vol Tween 20), 200 ␮m dNTPs, 10 pmol of each primer (the reverse primer was biotinylated), and 1.5 units of Taq DNA polymerase (Roche Diagnostic GmbH). Thermal cycling was performed with an initial denaturation for 5 minutes at 96°C, followed by 50 cycles of denaturation for 30 seconds at 96°C, primer annealing for 30 seconds at 55°C, and synthesis for 30 seconds at 72°C. A final primer extension was conducted for 5 minutes at 72°C. The PCR products (140 base pair [bp]) were run on 2% agarose gels (Roche Diagnostic GmbH) and visualized after an ethidium bromide staining. A total 45 ␮L of PCR product was used for pyrosequencing, and 15 pmol of the forward sequencing primer was applied to detect the polymorphisms. Statistical Analyses Genotype and allele frequencies, as well as Hardy-Weinberg equilibrium, were calculated using Microsoft Excel macro PHARE version 2.1 (developed by David G. Cox, Lyon, France), which can be downloaded at http://bioinformatics.org/macroshack/ programs/PHARE/description.html. For association analyses of individual genotypes and alleles, chi-square analyses on 2 ⫻ 2 or 2 ⫻ 3 contingency tables were carried out (http://www.georgetown.edu/ faculty/ballc/webtools/ web_ chi.html). The significance level for all statistical tests was .05. For pairwise linkage disequilibrium (LD) and haplotype analyses, the Arlequin program version 2.0 was used (University of Geneva, Geneva, Switzerland, http://lgb.unigene.ch/arlequin). A total of 100,000 permutations were performed in each analysis. Lewontin’s D= values were used to illustrate the extent of LD, and the corresponding p-values are also shown. Bonferroni correction was used for multiple testing, using the total number of SNPs as correction factor (Zill et al 2004a).

Results Choice of TPH -1 SNPs The objective of this study was to identify TPH-1 gene variants associated with MD. To this end, we aimed to chart TPH-1 risk

haplotypes. The first part of this study was designed to choose SNPs that would allow us to construct such haplotypes. We reasoned that any allele belonging to a haplotype block is expected to be in homozygosity when all other alleles belonging to the block are. We chose two TPH-1 intron 7 SNPs (indicated as numbers 4 and 5 in Table 1) reported in the literature to be associated with a number of psychiatric disorders. The entire patient group was genotyped for these two SNPs, and subjects that were homozygous for both markers were subgrouped into the four possible genotype combinations (CC/CC, CC/AA, AA/CC, AA/AA). From these four subgroups, three subjects were chosen at random. If there were fewer than three subjects in any subgroup, DNA samples from all subjects in that subgroup were used. DNA sequencing was carried out in eight subjects over a total of about 10 kb, including all exons, all intron regions within about .5 kb from the exon boundaries, and in some promoter regions. No novel polymorphism was identified (data not shown). For further analyses, we selected all SNPs from the sequenced regions where all subjects in any of the four subgroups were homozygous for at least one allele. Four SNPs were identified that met our criteria (indicated as numbers 1, 2, 3, and 6 in Table 1). Thus, a total of six loci were included (Table 1). The population analyzed in this study was tested for these six SNPs only. All SNPs were at Hardy-Weinberg equilibrium in the control population. Single Locus Analysis Results of single locus association analyses are summarized in Table 2. There was a highly significant allelic association between the SNP 5 A allele and MD (p ⫽ .001). Moreover, a genotypic association was found between this SNP and the disorder (p ⫽ .003). Both these associations were preserved after correction for multiple testing (p ⫽ .008 and p ⫽ .021, respectively). We also observed genotypic associations between SNPs 1 and 2 and MD (p ⫽ .019 and p ⫽ .012, respectively). However, after Bonferroni correction these associations were no longer significant. LD Analysis Linkage disequilibrium data indicated as Lewontin’s D= values and corresponding p-values were calculated for all SNP pairs in both cases and control subjects. Both groups share almost a homogeneous linkage pattern as shown in Tables 3 and 4. All SNPs are in high LD with each other in the control group. In the patient group, SNP 1 was found to be in strong LD only with SNP

Table 2. Allele and Genotype Distributiona Allelic Tests

SNP

Polymorphism

Control Subjectsd

1 2 3 4 5 6

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

.51(A) .57(C) .75(T) .60(C) .60(C) .60(G)

b

c

Genotypic Tests Control Subjects

d

e

MD

MD

p-Value

1-1

1-2

2-2

1-1

1-2

2-2

p-Valuee

.51(A) .51(C) .78(T) .55(C) .50(C) .57(G)

.9735 .0518 .3452 .0996 .0013 .4344

68 93 143 89 86 86

122 102 96 126 132 128

63 58 14 38 35 39

76 81 139 75 55 81

81 69 78 100 117 97

71 78 11 53 56 50

.0190 .0119 .6120 .0680 .0035 .1057

SNP, single nucleotide polymorphism. Bold numerals highlight significant SNPs (determined by a significant p-value). Same SNP numbering as in Table 1. c Allele 1/allele 2. d Major allele frequency is shown. e Not corrected for multiple testing. a

b

www.sobp.org/journal

298 BIOL PSYCHIATRY 2006;59:295–300

R. Gizatullin et al

Table 3. D= and p-Values for TPH-1 SNP Combinations in Healthy Control Subjectsa SNPsb 1 2 3 4 5 6

1

.519 .782 .589 .540 .551

2

3

4

5

6

⬍10⫺5

⬍10⫺5 ⬍10⫺5

⬍10⫺5 ⬍10⫺5 ⬍10⫺5

⬍10⫺5 ⬍10⫺5 ⬍10⫺5 ⬍10⫺5

⬍10⫺5 ⬍10⫺5 ⬍10⫺5 ⬍10⫺5 ⬍10⫺5

.720 .880 .804 .865

.857 .747 .687

.837 .948

.892

TPH-1, tryptophan hydroxylase-1; SNP, single nucleotide polymorphism. a Upper diagonal: p-values for pairwise LD; lower diagonal: D=-values for each SNP pair combination. b SNP numbering, as listed in Table 1, follows the physical location on the gene. D= values ⬎ .5 are shown in bold.

3. As for the other SNPs in this group, all show significant LD with p-values ⬍ 10⫺5. Six-Marker Haplotype Analysis A six-marker haplotype analysis was carried out across the TPH-1 gene. Table 5 shows the estimated haplotype frequencies in both groups, with Bonferroni corrected p-values. Haplotype analyses showed the existence of five common haplotypes, all with a frequency above 5% in both groups. About 75% of the entire study population carried such common haplotypes. Haplotype 3 (GCTCCG) shows a significant association with the control group (␹2 ⫽ 7.1, df ⫽ 1, corrected p ⫽ .047). Haplotype 2 (GCGCCG) differs from haplotype 3 only at SNP 3 (rs211105, G/T) and attained significant association before but not after Bonferroni’s correction. Taken together, haplotypes 2 and 3 account for about one third of the total study population, and their compound association with the control group is much stronger than haplotype 3 alone (␹2 ⫽ 16.18, df ⫽ 1, corrected p ⫽ .0003), suggesting that the shared five-SNP haplotype GCCCG (configuration 1-2-4-5-6) might be protective. Haplotype 4 also differs from haplotype 3 at one SNP (number 1) but did not attain significant association. Two other common haplotypes, namely numbers 1 (ATTAAA) and 5 (GTTAAA), which also differ from each other at only one SNP, account totally for about one third of the entire study population and show no association. Individual frequencies for the remaining haplotypes were too low to allow meaningful association analyses (see Discussion). To avoid further data fragmentation, all haplotypes below 5% in both populations were accumulated (Table 5, number 6). This

Table 5. Estimated Haplotype Frequencies Frequencies (%)

1 2 3 4 5 6d

Haplotypesa

Control Subjects

Patients

␹2

pc

ATTAAA GCGCCG GCTCCG ACTCCG GTTAAA All rare <5%

26.83 18.56 18.08 12.01 6.56 17.96

26.62 12.71 11.92 9.57 7.56 31.62

.01 6.19 7.07 1.43 .32 24.01

NS NS .0468 NS NS <10ⴚ5

a

Alleles are ordered according to their physical location on the gene, as listed in Table 1. b p-values calculated for each haplotype versus all others by the ␹2 test. Significant p-values are shown in bold. c p-values corrected by Bonferroni method. Significant p-values are shown in bold. d All haplotypes with estimate frequencies ⬍5% in both patient and control samples.

compound low-frequency haplotype group was found to be associated with the disorder (31.6% vs. 18.0% in control subjects; ␹2 ⫽ 24.01, df ⫽ 1, corrected p ⬍ 10⫺5). Three-Locus Haplotype “Sliding Window” Analysis To define TPH-1 gene regions that may carry stronger associations, we analyzed haplotypes formed by three-locus combinations with a “sliding window” approach, i.e., SNP combinations 1-2-3, 2-3-4, 3-4-5, and 4-5-6. Table 6 shows only the three-locus haplotypes with significant associations after Bonferroni correction for each configuration. Haplotypes GTG (configuration 1-2-3) and TGC (configuration 2-3-4) were infrequent, however, less so among patients than control subjects (␹2 ⫽ 8.1, df ⫽ 1, corrected p ⫽ .027; and ␹2 ⫽ 7.1, df ⫽ 1, corrected p ⫽ .047, respectively). Combination 4-5-6 provided the most informative data. Haplotype CCG, which was the overall most common three-locus haplotype, displayed lower frequencies in patients than control subjects (43% vs. 54% in control subjects, ␹2 ⫽ 12.1, df ⫽ 1, Bonferroni corrected p ⫽ .003). Haplotype CAG displayed an opposite pattern with higher frequencies in patients (7.5% vs. 3.2% in control subjects, ␹2 ⫽ 9.0, df ⫽ 1, corrected p ⫽ .016). Haplotype AAG was also significantly associated with MD (3.4% in patients vs. .4% in control subjects, ␹2 ⫽ 13.1, df ⫽ 1, corrected p ⫽ .002). The latter two risk haplotypes differ by one SNP, such Table 6. Three-Locus Haplotype analysisa Frequencies(%)

Table 4. D= and p-Values for TPH-1 SNP Combinations in MD Patientsa SNPsb 1 2 3 4 5 6

1

.386 .841 .491 .379 .470

2

3

4

5

6

Configurationsb

⬍10⫺5

⬍10⫺5 ⬍10⫺5

⬍10⫺5 ⬍10⫺5 ⬍10⫺5

⬍10⫺5 ⬍10⫺5 ⬍10⫺5 ⬍10⫺5

⬍10⫺5 ⬍10⫺5 ⬍10⫺5 ⬍10⫺5 ⬍10⫺5

1-2-3 2-3-4 4-5-6

.527 .742 .561 .704

.907 .745 .733

.841 .814

.806

TPH-1, tryptophan hydroxylase-1; SNP, single nucleotide polymorphism; MD, major depression. a Upper diagonal: p-values for pairwise LD; lower diagonal: D=-values for each SNP pair combination. b SNP numbering, as listed in Table 1, follows the physical location on the gene. D= values ⬎ .5 are shown in bold.

www.sobp.org/journal

Three-Locus Haplotypes

Control Subjects

Patients

␹2

pc

GTG TGC AAG CAG CCG

.94 1.34 .40 3.22 54.07

3.67 4.15 3.44 7.52 42.95

8.07 7.08 13.07 9.00 12.12

.027 .047 .002 .016 .003

SNP, single nucleotide polymorphism. Only haplotypes with significant associations are shown. b Totally four configurations are made by the six SNPs. Each configuration contains eight different combinations of three-locus haplotypes (4 ⫻ 8 ⫽ 32). Numbers in each configuration indicate which SNPs are included. The position of each SNP is according to their physical location on the gene, as listed in Table 1. c p-values after Bonferroni correction. a

R. Gizatullin et al that their compound frequency corresponds to the two-locus haplotype AG in configuration 5-6. This haplotype alone had a frequency of 11.0% in patients versus 3.6% in control subjects, thereby associating strongly with MD (␹2 ⫽ 20.0, df ⫽ 1, corrected p ⬍ 10⫺5).

Discussion We report on TPH-1 gene variants associated with MD in a Caucasian population of North European descent. Tryptophan hydroxylase-1 has been associated with a number of psychiatric disorders including mood disorders, although association with MD has not been confirmed (Furlong et al 1998; Serretti et al 2002). Most reports on TPH-1 have been association studies with individual markers. We are not aware of any TPH-1 association study using SNP-based haplotypes in the context of MD. Our results will have to be reproduced in replica populations to be confirmed. Analyses of individual SNPs show that SNP 5 (rs1799913) was the only one associated with depression after Bonferroni correction for multiple tests (p ⫽ .001). This SNP, also known in the literature as SNP A779C, has been associated with cerebrospinal fluid (CSF) 5-hydroxyindoleacetic acid (5-HIAA) concentrations (Jonsson et al 1997), and several reports have shown an association with suicidal behavior (Geijer et al 2000; Nielsen et al 1998). Association of two more SNPs, numbers 1 (rs4537731) and 2 (rs684302), lost significance after Bonferroni correction. It should be mentioned that the Bonferroni correction is very conservative, particularly in our case, where measurements are not independent, since all SNPs are in LD. However, single marker association studies are known to be prone to generate weak associations and/or inconsistent reproducibility (Cardon and Bell 2001; Lohmueller et al 2003). This is due to several reasons, one of which being the fact that parent-offspring transmission of genetic variants follows certain constraints, such that most individual alleles carry only limited information on gene variants (Clark 2004). Alleles are mostly organized in haplotype blocks, i.e., limited arrays of allelic combinations inherited with minimal recombination, and rather stably maintained in populations (Gabriel et al 2002). Haplotypes are currently the focus of intense research effort (Clark 2004). Haplotype blocks have become key to an understanding of gene variant inheritance, as well as to reduce informational complexity when dealing with whole-genome scanning strategies (Gabriel et al 2002). However, the reconstruction of haplotype blocks requires parent-offspring transmission data, which are often not readily available in clinical studies. An alternative concept has recently been proposed, namely the gene-based haplotype (Hoehe 2003). A gene-based haplotype is a combination of alleles located within a gene unit. Potentially, gene-based haplotypes are the most precise markers possible for a given gene, as they can contain all the variations in the gene sequence (Hoehe 2003). While the concept of gene-based haplotype does not necessarily relate to gene function, ideally it might serve as background information to formulate gene-based functional haplotypes, defined as the individual gene-unit sequences that determine structure, function, and regulation of the gene and its product (Hoehe 2003). We therefore implemented a SNP screening strategy as a first step to define the complete sequence of risk gene variants. As noted in the introduction, two SNPs in TPH-1 intron 7 have been repeatedly associated with a variety of psychiatric disorders (Mann et al 1997; Nielsen et al 1997, 1998; Paik et al 2000). We

BIOL PSYCHIATRY 2006;59:295–300 299 reasoned that if these two SNPs belonged to a risk haplotype, this would partially explain the relative inconsistency of the literature data. This would be in agreement with a recent meta-analysis of the psychiatric genetics literature, which indicated that common alleles found often, albeit inconsistently, associated with disease do carry a real risk for disease (Lohmueller et al 2003). As expected from our SNP selection strategy, all SNPs were in significant LD both in patients and control subjects. Specifically, all SNPs were in strong LD with each other in the control group. In the patient group, SNP 1 was found to be in strong LD only with SNP 3. As for the other SNPs in the patient group, all showed significant LD with p-values ⬍ 10⫺5. The LD patterns and their statistical significance level expressed in p-values were essentially similar in both groups, suggesting a haplotype block pattern. Haplotype analyses showed the existence of five common six-locus haplotypes, all with a frequency above 5% in both groups (Table 5). About 75% of the entire study population carried such common haplotypes. Three of these five haplotypes were less common in patients than control subjects, whereas two of them were more equally distributed. However, when all of the less common haplotypes below an arbitrary 5% threshold were compound (about one fourth of all haplotypes), this group of haplotypes resulted strongly associated with the patient group (corrected p ⬍ 10⫺5, see Results). The data suggest a scenario whereby the most common TPH-1 haplotypes range from neutral to protective, while the less frequent haplotypes could contribute to genetic predisposition for MD. Conceivably, given the diversity of these haplotypes, a causal polymorphism might be shared by a number of low-frequency haplotypes or several distinct causal variants might exist in this sample. However, such data should be taken with caution, as frequency estimates of rare haplotypes are highly susceptible to error (Tishkoff et al 2000). We also performed a haplotype sliding window analysis along the gene to look for regions that may carry particularly strong associations. The SNP combination 5– 6 was the most informative. These SNPs encompass a fairly short region, less than 2 kb long, spanning from the well-studied intron 7 polymorphisms (see introduction) into intron 8. Haplotypes in this region associate strongly with depression (corrected p ⬍ 10⫺5, see Results). This suggests that SNPs 5 and 6 together might be strong determinants of genetic risk for depression. As discussed above, SNP 5 individually showed a significant association with MD in this study, which is in agreement with this set of data. This region is therefore a candidate for the location of a possible haplotypelinked functional polymorphism. Tryptophan hydroxylase has recently become the focus of renewed attention, as a novel enzyme isoform (TPH-2) has been discovered (Walther and Bader 2003). While the two genes have little overall sequence homology, on the amino acid level the proteins share about 70% identity (Walther and Bader 2003). Tryptophan hydroxylase-2 has been shown to be the predominant isoform expressed in the brainstem, the major locus of 5-HT–producing neurons (Zill et al 2004a, 2004b). However, TPH-1 and TPH-2 messenger RNA (mRNA) are expressed in nearly equal amounts in human brain regions such as frontal cortex, thalamus, hippocampus, hypothalamus, and amygdala (Zill et al 2004a, 2004b). There is extensive evidence for the presence of 5-HT nerve terminals and receptors in neuroendocrine regions such as the hippocampus and the hypothalamus (Chaouloff 1993; Graeff 1993). Also, 5-HT turnover is particularly active in cortical and limbic areas involved in emotional aspects of behavior (Westenberg 1996; Whitaker-Azmitia et al 1990). www.sobp.org/journal

300 BIOL PSYCHIATRY 2006;59:295–300 Given the close interplay between the two TPH isoforms in key brain areas related to mood disorders, an understanding of their relative role in pathogenic events leading to depression will require further studies on a molecular level. A TPH-2 haplotype analysis has recently revealed an association of the TPH-2 gene with major depression (Zill et al 2004a). The haplotype analysis that we have performed provides a basis for a complete molecular characterization of THP-1 gene variants. This knowledge might contribute to a better understanding of the pathogenic role played by the TPH-1 gene variants in MD. Financial support was received from the Swedish Science Council and from the AFA insurance company (MA, RL). We thank Dr. Salim Mottagui-Tabar, Center for Genomics and Bioinformatics, Karolinska Institute for his help in the haplotype analyses. Abbar M, Courtet P, Amadeo S, Caer Y, Mallet J, Baldy-Moulinier M, et al (1995): Suicidal behaviors and the tryptophan hydroxylase gene. Arch Gen Psychiatry 52:846 – 849. Bellivier F, Leboyer M, Courtet P, Buresi C, Beaufils B, Samolyk D, et al (1998): Association between the tryptophan hydroxylase gene and manic-depressive illness. Arch Gen Psychiatry 55:33–37. Bennett PJ, McMahon WM, Watabe J, Achilles J, Bacon M, Coon H, et al (2000): Tryptophan hydroxylase polymorphisms in suicide victims. Psychiatr Genet 10:13–17. Cardon LR, Bell JI (2001): Association study designs for complex diseases. Nat Rev Genet 2:91–99. Chaouloff F (1993): Physiopharmacological interactions between stress hormones and central serotonergic systems. Brain Res Brain Res Rev 18:1–32. Clark AG (2004): The role of haplotypes in candidate gene studies. Genet Epidemiol 27:321–333. Cloninger CR (1987): A systematic method for clinical description and classification of personality variants. A proposal. Arch Gen Psychiatry 44:573– 588. Cooper JR, Melcer I (1961): The enzymic oxidation of tryptophan to 5-hydroxytryptophan in the biosynthesis of serotonin. J Pharmacol Exp Ther 132:265–268. Craig SP, Boularand S, Darmon MC, Mallet J, Craig IW (1991): Localization of human tryptophan hydroxylase (TPH) to chromosome 11p15.3-p14 by in situ hybridization. Cytogenet Cell Genet 56:157–159. Faludi G, Tekes K, Tothfalusi L (1994): Comparative study of platelet 3Hparoxetine and 3H-imipramine binding in panic disorder patients and healthy controls. J Psychiatry Neurosci 19:109 –113. Fava M, Kendler KS (2000): Major depressive disorder. Neuron 28:335–341. First MB, Spitzer RL, Gibbon M, Williams JB (2002): Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Non-patient Edition. (SCID-I/NP) New York: Biometrics Research, New York State Psychiatric Institute. Furlong RA, Ho L, Rubinsztein JS, Walsh C, Paykel ES, Rubinsztein DC (1998): No association of the tryptophan hydroxylase gene with bipolar affective disorder, unipolar affective disorder, or suicidal behaviour in major affective disorder. Am J Med Genet 81:245–247. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, et al (2002): The structure of haplotype blocks in the human genome. Science 296:2225–2229. Geijer T, Frisch A, Persson ML, Wasserman D, Rockah R, Michaelovsky E, et al (2000): Search for association between suicide attempt and serotonergic polymorphisms. Psychiatr Genet 10:19 –26. Glazier AM, Nadeau JH, Aitman TJ (2002): Finding genes that underlie complex traits. Science 298:2345–2349. Graeff FG (1993): Role of 5-HT in defensive behavior and anxiety. Rev Neurosci 4:181–211. Griebel G (1995): 5-Hydroxytryptamine-interacting drugs in animal models of anxiety disorders: More than 30 years of research. Pharmacol Ther 65:319 –395. Gustavsson JP, Nothen MM, Jonsson EG, Neidt H, Forslund K, Rylander G, et al (1999): No association between serotonin transporter gene polymorphisms and personality traits. Am J Med Genet 88:430 – 436.

www.sobp.org/journal

R. Gizatullin et al Handley SL (1995): 5-Hydroxytryptamine pathways in anxiety and its treatment. Pharmacol Ther 66:103–148. Hoehe MR (2003): Haplotypes and the systematic analysis of genetic variation in genes and genomes. Pharmacogenomics 4:547–570. Iny LJ, Pecknold J, Suranyi-Cadotte BE, Bernier B, Luthe L, Nair NP, et al (1994): Studies of a neurochemical link between depression, anxiety, and stress from [3H]imipramine and [3H]paroxetine binding on human platelets. Biol Psychiatry 36:281–291. Jonsson EG, Goldman D, Spurlock G, Gustavsson JP, Nielsen DA, Linnoila M, et al (1997): Tryptophan hydroxylase and catechol-O-methyltransferase gene polymorphisms: Relationships to monoamine metabolite concentrations in CSF of healthy volunteers. Eur Arch Psychiatry Clin Neurosci 247:297–302. Judson R, Stephens JC, Windemuth A (2000): The predictive power of haplotypes in clinical response. Pharmacogenomics 1:15–26. Lohmueller KE, Pearce CL, Pike M, Lander ES, Hirschhorn JN (2003): Metaanalysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet 33:177– 182. Malhi GS, Moore J, McGuffin P (2000): The genetics of major depressive disorder. Curr Psychiatry Rep 2:165–169. Mann JJ, Malone KM, Nielsen DA, Goldman D, Erdos J, Gelernter J (1997): Possible association of a polymorphism of the tryptophan hydroxylase gene with suicidal behavior in depressed patients. Am J Psychiatry 154: 1451–1453. Nielsen DA, Jenkins GL, Stefanisko KM, Jefferson KK, Goldman D (1997): Sequence, splice site and population frequency distribution analyses of the polymorphic human tryptophan hydroxylase intron 7. Brain Res Mol Brain Res 45:145–148. Nielsen DA, Virkkunen M, Lappalainen J, Eggert M, Brown GL, Long JC, et al (1998): A tryptophan hydroxylase gene marker for suicidality and alcoholism. Arch Gen Psychiatry 55:593– 602. Ono H, Shirakawa O, Nishiguchi N, Nishimura A, Nushida H, Ueno Y, et al (2000): Tryptophan hydroxylase gene polymorphisms are not associated with suicide. Am J Med Genet 96:861– 863. Owens MJ, Nemeroff CB (1994): Role of serotonin in the pathophysiology of depression: Focus on the serotonin transporter. Clin Chem 40:288 –295. Paik I, Toh K, Kim J, Lee C (2000): TPH gene may be associated with suicidal behavior, but not with schizophrenia in the Korean population. Hum Hered 50:365–369. Ressler KJ, Nemeroff CB (2000): Role of serotonergic and noradrenergic systems in the pathophysiology of depression and anxiety disorders. Depress Anxiety 12(suppl 1):2–19. Roy A, Rylander G, Sarchiapone M (1997): Genetics of suicides. Family studies and molecular genetics. Ann N Y Acad Sci 836:135–157. Serretti A, Cristina S, Lilli R, Cusin C, Lattuada E, Lorenzi C, et al (2002): Family-based association study of 5-HTTLPR, TPH, MAO-A, and DRD4 polymorphisms in mood disorders. Am J Med Genet 114:361–369. 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. Tishkoff SA, Pakstis AJ, Ruano G, Kidd KK (2000): The accuracy of statistical methods for estimation of haplotype frequencies: An example from the CD4 locus. Am J Hum Genet 67:518 –522. Walther DJ, Bader M (2003): A unique central tryptophan hydroxylase isoform. Biochem Pharmacol 66:1673–1680. Westenberg HG (1996): Developments in the drug treatment of panic disorder: What is the place of the selective serotonin reuptake inhibitors? J Affect Disord 40:85–93. Whitaker-Azmitia PM, Shemer AV, Caruso J, Molino L, Azmitia EC (1990): Role of high affinity serotonin receptors in neuronal growth. Ann N Y Acad Sci 600:315–330. 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.