INTRODUCTION

The human serotonin 5-HT2A receptor (5-HT2A) binding kinetics are altered in major depression, suicidal and aggressive behaviors, and schizophrenia. Greater 5-HT2A binding has been reported in post-mortem brain tissue from individuals with major depression and suicide victims by some but not all studies (Stanley and Mann 1983; Mann et al, 1986; Hrdina and Du 2001; Kato et al, 1996; Ono et al, 2001; Hrdina et al, 1993; Arora and Meltzer 1989; Yates et al, 1990). Pandey et al (1995) found that higher prefrontal cortical 5-HT2A binding was associated with greater gene expression in youth suicide.

The platelet is a readily accessible source of 5-HT2A receptors with characteristics similar to the brain receptor (Cook Jr et al, 1994; Elliott and Kent, 1989). Greater platelet 5-HT2A binding has been reported in association with major depression and a history of suicidal acts (Hrdina et al, 1995; Hrdina et al, 1997; Pandey et al, 1995; Sheline et al, 1995). Our own study found higher platelet 5-HT2A binding to be related to suicidal behavior and aggression history, but not to a diagnosis of mood disorder (McBride et al, 1994).

The 5-HT2A 20-kb gene consists of three exons and two introns (Chen et al, 1992) and is located on chromosome 13q14–q21 (Sparkes et al, 1991). A significant association between T102C and schizophrenia is reported (Williams et al, 1996; Williams et al, 1997). The C allele was associated with diagnosis of schizophrenia with an odds ratio of 1.3 (p=0.008). These findings are confirmed by a meta-analysis of previous studies (Williams et al, 1997). A more recent meta-analysis demonstrated a stronger association with schizophrenia in Caucasians compared with other ethnic groups (Abdolmaleky et al, 2004). No association of genotype or allele frequencies is found with suicide or mood disorders (Arranz et al, 1997; Bondy et al, 2000; Crawford et al, 2000; Mahieu et al, 1997; Minov et al, 2001; Tsai et al, 1999; Ono et al, 2001; Correa et al, 2002; Geijer et al, 2000). Turecki et al (2003) found that higher 5-HT2A binding in the brain is associated with the T102C polymorphism in the 5-HT2A gene independently of suicide, although others disagree (Hrdina and Du, 2001). The T and C alleles may show differential expression (Polesskaya and Sokolov 2002; Bray et al, 2004).

To clarify the relationship of the T and C alleles to 5-HT2A receptor number and to mood disorders, we investigated the association of the T102C polymorphism with platelet 5-HT2A receptor binding indices and with clinical phenotype in subjects with mood disorders, with and without a history of suicide attempts, and a control group of healthy volunteers.

PATIENTS AND METHODS

Subjects

Patients (N=152) were recruited from the subjects presenting to the research clinics of a university-affiliated psychiatric hospital for evaluation and treatment of a mood disorders. All subjects met DSM-IV criteria for a current major depressive episode (N=90, 59.2% of the patients), or bipolar disorder with depressed mood (N=25, 16.4%), or dysthymia (N=2, 1.3%), or remitted major depression (N=35, 23.0%). Healthy controls (N=63) were recruited through advertisements. Only Caucasian subjects of European origin were included to reduce ethnic variation and risk of genetic stratification (Gelernter et al, 1993). Subjects were free from drugs known to affect the serotonin system for at least 14 days (with a median of 30 days). The drug-free interval was longer for drugs with a longer half-life (6 weeks for fluoxetine and 4 weeks for an oral antipsychotic). The duration of the drug-free period of the patients was established by a combination of drug screen, observation in the hospital, and a history obtained from the subject's family and the referring physician. Written informed consent was obtained as approved by the Institutional Review Board. Demographic and clinical data from the two groups are presented in Table 1.

Table 1 Demographic and Clinical Description of Mood-Disordered Patients and Healthy Controls (N=215)

Clinical Phenotype Assessment

DSM-IV (APA 1994) Axis I and II diagnoses were made using the Structured Clinical Interview (SCID-I and -II) (Spitzer et al, 1989). Lifetime aggression was rated using the Brown–Goodwin Aggression Inventory (BG) (Brown and Goodwin 1986), lifetime hostility, using the Buss–Durkee Hostility Inventory (BDHI) (Buss and Durkee 1957), and impulsivity, using the Barratt Impulsiveness Scale (BIS) (Barratt 1994). Depression was evaluated by the Beck Depression Inventory (BDI) (Beck et al, 1961) and the 17-item Hamilton Depression Rating Scale (HAM-17) (Hamilton 1960). Hopelessness was measured by the Beck Hopelessness Scale (BHS) (Beck et al, 1985). For healthy volunteers, any Axis I diagnosis by the SCID-NP (nonpatient version) and any history of first-degree relative with a mood or psychotic disorder were ruled out. Subjects and healthy volunteers had a physical examination and routine laboratory screening tests to detect neurological disease and any active physical disease that could affect their mental status or serotonin function.

DNA Isolation and Extraction

Venous blood samples were collected in EDTA tubes. The samples were first centrifuged (150g) for 15 min at room temperature to obtain platelet-rich plasma (PRP). After removal of PRP, the remaining blood fraction was centrifuged at 750g for 15 min to obtain the buffy coat, which was further purified by layering on top of a 4-ml Ficoll®/hypaque (Pharmacia) gradient. The tubes were centrifuged at 750g for 15 min in a Sorvall GLC-2B centrifuge. The white cell interface layer was transferred into new plastic tubes and 3 ml of PBS buffer added. Then, the fraction was further centrifuged at 11 000g in a Sorvall RC-5B centrifuge for 5 min at 4°C. A measure of 3 ml of PBS buffer were used to wash the white cell pellet, which was then centrifuged at 11 000g for 5 min. The supernatant was discarded, and the pellet stored at −20°C pending DNA extraction.

DNA extraction from lymphocytes pellets was performed as described by Higuchi (1992). Thawed pellets were resuspended in 3 ml PBS buffer. The suspension was centrifuged at 11 000g for 5 min at 4°C, and the supernatant discarded. Pellets were resuspended in a 500 μl of polymerase chain reaction (PCR) buffer (50 mM KCl, 10 mM Tris-HCl (pH 8.3), 2.5 mM MgCl2, 0.1 mg/ml gelatin, 0.45% NP40, 0.45% Tween 20) containing 12 μg Proteinase K. After incubation at 55–60°C for 1 h, the proteinase K was inactivated by heating at 95°C for 10 min. The samples were diluted with 10 mM Tris-EDTA buffer at 1:5 dilution and pipetted (2 μl) for PCR. Genomic DNA fractions were stored at −20°C.

Polymerase Chain Reaction

Genotyping of the 102T/C polymorphism of the 5-HT2A receptor gene (a synonymous allele, dbSNP:6313) was carried out by PCR and MspI restriction enzyme digestion, as described previously (Higuchi 1992) with modifications. Briefly, the oligonucleotide primers sense Mann7 (5′-CAAGGTGAATGGTGAGCAGAAA-3′) and antisense Mann8 (5′-TGGCAAGTGACATCAGGAAATAGT-3′) were used to amplify a PCR fragment of 425 bp length. PCR was carried out in a 25 μl volume, containing 100 ng DNA, 1 pmol of each primer, 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 2 mM MgCl2, 0.01% gelatin, 200 μM of each dNTP, and 1 U of Red Tag DNA Polymerase (Sigma, St Louis, MO). Samples were processed in a GeneAmp PCR System 2400 (Perkin-Elmer). In all, 30 temperature cycles were carried out, consisting of 30 s at 94°C, 30 s at 55°C, and 40 s at 72°C, followed by a final extension step at 72°C for 4 min. After the amplification, each aliquot of 10 μl PCR product was digested with MspI restriction enzyme (Gibco BRL, Rockville, MD) in a volume of 15 μl at 37°C overnight. The digested PCR products were separated on a 1.2% agarose gel. The 102C allele resulted in DNA fragments of 176 and 249 bp, whereas the 102T allele PCR fragments remained uncut.

5-HT2A Receptor Binding Assay

The assay was performed as described previously (Khait et al, 1999). Briefly, platelet membranes were prepared as follows. PRP obtained as described above was centrifuged at 16 000g for 10 min at 4°C. The resultant pellet was resuspended in 5 ml of normal saline and spun again. We have shown (Khait et al, 1999) that a prolonged storage of platelet pellets affects binding. Therefore, in the present study storage time was limited to less than a month. At the time of the binding assay, the pellet was lysed by suspension in hypotonic solution (Tris-EDTA buffer: 5 mM Tris-HCl, 0.1% EDTA, pH 7.5), and the suspension was homogenized using a Brinkmann homogenizer (setting 6) for 10 s. The mixture was centrifuged twice at 16 000g for 10 min at 4°C, once with hypotonic buffer, and finally with the incubation buffer (Tris-HCl 50 mM (pH 7.4) containing 120 mM NaCl, 5 mM KCl, and 2 mM MgCl). The pellet was then resuspended in the incubation buffer to form the final membrane suspension for binding studies. LSD binding assay was performed as described by Geaney et al (1984). Four-tenths (0.4) ml of the membrane suspension (containing 20–150 μg protein) was incubated with [3H]LSD (NEN Life Science Products, Boston, MA) in a final volume of 0.5 ml for 4 h at 37°C. After incubation, the reaction was terminated by the addition of 4 ml ice-cold Tris buffer (50 mM of Tris-HCl, pH 7.7) and rapidly filtered through Whatman GF/B filters using a Brandel 24-channel cell harvester. The filters were prewashed with buffer before filtration and washed three times with 4 ml of buffer after filtration. All filters had been soaked in 2.5% polyethylenimine (Fluka Chemie) and dried prior to filtration, before being counted in a Packard Liquid Scintillation Analyzer. Filters were placed in scintillation vials with scintillation cocktail for 48 h. Saturation binding isotherms were performed using a range of ligand concentrations (0.1–2.0 nM). The specific binding of [3H]LSD was defined as the difference between total and nonspecific binding determined by the addition of 300 nM spiperone. Binding indices, Bmax and KD, were calculated by the EDBA/LIGAND program (Biosoft, UK). Total bound did not exceed 5% of the total radioactivity. Protein content was determined by Folin-reagent procedure.

Analyses Controlling for Seasonal Effects and Relationship of Binding to Genotype

Seasonal variation of platelet 5-HTR2A density or Bmax is an additional source of variance that has been reported in healthy volunteers (Spigset and Mjorndal, 1997; Spigset et al, 1998), and by us in subjects with a major depressive episode (Khait et al, 1999; Khait et al, 2002). In order to control for this variance, an adjusted Bmax was calculated for each subject by the following procedure. For each subject group, a monthly average max(m), where m=1, 2,…12 months was calculated and a monthly adjustment coefficient A(m) = max(m)/B0, where B0 is the group average, was assigned to the month m. Seasonally adjusted Bmax for each subject was calculated by dividing raw Bmax by the coefficient A(m) corresponding to the month m of the sample blood drawing. With this correction, the monthly averages for seasonally adjusted Bmax (max (m))are all the same and equal to the whole group average.

RESULTS

Descriptive

Demographic and clinical data from mood-disordered subjects and normal controls are presented in Table 1. There were no significant differences in age distribution between the mood-disordered group and controls (t=0.32, df=213, p=0.748). There was a trend for more females in mood-disordered subjects compared with controls (χ2=3.39, df=1, p=0.065). Therefore, we controlled for sex in analyses where appropriate. As a group, the patients were moderately depressed and, as expected, scored significantly higher in all rating scales compared with controls (p<0.0001; see Table 1). The observed genotype distributions for patients and controls were not significantly different from Hardy–Weinberg equilibrium (data not shown).

Relationship Between Clinical Parameters and 102T/C Genotype

We found no significant difference in either genotype distribution or allele frequencies between the mood-disordered group and the control group (Table 2). Within the mood-disorder group, suicide attempters had significantly more heterozygotes compared with nonattempters (χ2=7.64, df=2, p=0.020). There was no significant difference in severity of depression, lifetime aggression, hostility or impulsiveness across genotype in the depressed subject group or the controls, and in each of the subgroups, that is, suicide attempters and nonattempters. The same analysis for females only did not change the results (data not shown).

Table 2 102T/C Genotype and Allele Frequencies and 5-HT2A Receptor Indices in Mood-Disordered Patients and Healthy Controls

5-HT2A Receptor Binding Indices and Clinical Parameters

There were no significant effects for age (F=0.027, df=213,1; p=0.871) or gender (t=−0.582, df=213; p=0.561) on the season-adjusted 5-HT2A Bmax. Therefore, we combined binding data for both genders. Table 2 presents the platelet 5-HT2A Bmax and KD for each group and subgroups of suicide attempters and nonattempters. No group differences were observed.

In this sample, Bmax did not correlate significantly with lifetime aggression score as measured by BG (r=0.011, p=0.89), hostility on the BDHI (r=0.102, p=0.229), HDRS (r=0.125, p=0.126), or impulsivity on the BIS (0.084, p=0.328).

5-HT2A Bmax and 102C/T Genotype

A linear regression model, constructed with season-adjusted Bmax as the dependent variable, and genotype and diagnosis (mood disorder or healthy control) as independent variables, was significant for the whole model (F=2.56, df=5, 209; p=0.028). The association of genotype with Bmax was significant (F=3.53, df=2,213; p=0.031), and the interaction between diagnosis and genotype approached significance (p=0.089). Diagnosis alone was not significant (p=0.251). The reason that the interaction term approaches significance is that the genotype effect on binding was significant for the control group (F=7.56, df=2,60; p=0.001) and not significant for the mood-disordered group (F=1.276, df=2,149; p=0.282) or subgroups of mood disorders (bipolar, major depressive disorder). Adding sex to the model did not change the results, and there was no effect of sex on Bmax (F=0.001, df=1,214, p=0.98). As shown in Figure 1, control subjects with the TT genotype had higher Bmax than subjects with the other genotypes (92.7 vs 69.7 fmol/mg in the TC genotype and 76.3 in the CC genotype). Mood-disordered subjects with the TT genotype had a Bmax (88.2 fmol/mg) closer to that of the TC genotype (86.9 fmol/mg) and the CC genotype (79.0 fmol/mg) (F=1.26, df=2,149; p=0.282), although the rank order was the same as for controls. A post hoc test for multiple comparisons confirmed the difference in the control group between the TT and the TC (23.0 fmol/mg) as significant after Bonferroni correction (p=0.001), whereas the difference between TT and CC (16.4 fmol/mg) approached significance (p=0.074), while the difference between CC and TC (6.55 fmol/mg) was negligible (p=0.99) (see Figure 1). No effect of genotype was found on KD.

Figure 1
figure 1

Platelet 5-HT2A binding density and C102T genotype in mood-disordered subjects and healthy volunteers.

DISCUSSION

Our results confirm previous reports that genotype frequencies of 102T/C 5-HT2A receptor gene polymorphism in mood disorders do not differ from healthy volunteers (Arranz et al, 1997; Bondy et al, 2000; Crawford et al, 2000; Mahieu et al, 1997; Minov et al, 2001; Tsai et al, 1999; Ono et al, 2001; Correa et al, 2002; Geijer et al, 2000). Nevertheless, we found an excess of the TC genotype in suicide attempters compared with nonattempters. This is an exploratory finding requiring replication. We found a significant association between platelet 5-HT2A receptor density (Bmax) and the TT genotype of this polymorphism. This effect appears stronger or is perhaps confined to normal volunteers compared with depressed subjects. There was no sex effect on the findings. The apparent absence or much weaker nature of this relationship in depressed subjects is not explained by a difference in statistical power because we had more depressed subjects (N=63 controls vs N=152 mood-disordered subjects). The finding that affinity (KD) did not differ significantly between healthy controls and mood-disordered patients indicates that the genetic effect involves Bmax or receptor number. The TT genotype is associated with more 5-HT2A receptors in healthy volunteers but apparently not in the mood disorders group even though there was the same rank order of binding by genotype (TT>TC>CC). This finding is consistent with Polesskaya and Sokolov (2002), who studied the relationship of the C and T alleles to the expression of the HTR2A gene in post-mortem brain tissue and found the expression level with the T allele was higher compared with the C allele in schizophrenia and a control group. They did not include a sample with mood disorders.

Turecki et al (2003) studied post-mortem brain samples of suicide victims and comparison subjects and found higher 5-HT2A Bmax in suicide victims with the C allele than in controls with the C allele, excluding the polymorphism as an explanation of greater binding in suicide, although Turecki studied a population of completers rather than attempters and post-mortem brains instead of platelets. Our data, in combination with those two reports, showed that having the T allele might be associated with higher Bmax in platelets and brain. Our group demonstrated in the past that higher platelet 5-HT2A binding is related to suicidal behavior and aggression history, but not to a diagnosis of mood disorder (McBride et al, 1994). Since the number of suicide attempters in this study was relatively small (N=52), the lack of association of suicidality with higher Bmax in this sample may be a type II error due to size effect. However, not finding more 5-HT2A binding in the mood-disordered subjects is consistent with our previous report (McBride et al, 1994). Another explanation for lack of association with suicidality in the current study may be a complex mode of inheritance in the HTR2A gene, rather than simple Mendelian transmission that makes each subject different in expression of its genetic traits. For example, Kato et al (1996) found a genomic imprinting mechanism in the HTR2A gene, namely, only maternal alleles will be expressed. A trait of imprinting and anticipation was described in the inheritance of mood disorders in Caucasian and Japanese populations (Ohara et al, 1998). Given the possibility of imprinting at the C102T locus, the lack of parental data in our study is a limitation.

If this genetic effect occurs in the brain, namely an association between the TT genotype and more 5-HT2A receptors, there is the possibility of greater signal transduction via phosphoinositide turnover and second messengers inositol triphosphate and diacylglycerol. This may enhance 5-HT2A-mediated effects on brain function and effects of antidepressant and atypical antipsychotics. Electroconvulsive shock upregulates 5-HT2A receptors as does stress, perhaps mediating antidepressant and homeostatic effects. The HTR2A 102T/C polymorphism or another polymorphism in linkage disequilibrium with it, including the 1438A/G promoter polymorphism, may affect gene regulation in ways that have not yet been elucidated.

Limitations

As in all case–control studies in psychiatric genetics, we must be aware of false-positive and false-negative findings due to sample size and ethnic stratification. We included only Caucasian subjects to minimize the stratification error. Notably, the association with schizophrenia seems stronger in Caucasians (Abdolmaleky et al, 2004). Another limitation is that the peripheral 5-HT2A platelet binding is not a direct measure of the brain 5-HT2A binding, and different transcription factors may regulate the gene in neurons expressing HTR2A gene centrally compared with megakaryocytes (Mann et al, 1992; Pandey et al, 1995). The nonassociation of this genotype with binding in the depressed population may reflect other regulatory effects, associated with diagnosis, that need to be identified. This study applies the same correction for seasonal effect to both groups, whereas there might be a different pattern of seasonal variability. This could potentially distort the comparisons between the two groups. Given the known imprinting at the C102T locus (Kato et al, 1996), the lack of parental data in our study limits interpretation of the results since we could not assess such an effect. HTR2A gene transcription is regulated in the megakaryocyte, and the half-life of platelets is 9 days. A 14-day drug-free period represents 1 ½ lives and might not be long enough for drug effects to wear off. Most of our subjects were drug free for 4 weeks, reducing the likelihood of a drug effect. There are other variants in the HTR2A gene that were not studied here, for example, the 1438A/G promoter polymorphism with which the C102T is in perfect linkage disequilibrium. The other variants might be more directly related to receptor binding. We studied major depression and bipolar disorder and, although we did not detect an effect of these subgroups, a larger study is needed for adequate statistical power.

In conclusion, this study replicates the lack of association of the T102C with suicidality and mood disorders, but demonstrates an association of the TT genotype with a higher number of 5-HT2A receptors in the platelets of healthy but apparently not in mood-disordered subjects. This difference in the gene to receptor number association between healthy and mood-disordered subjects may reflect pathological mechanisms of importance that warrant further study.