Association between a genetic variant in the serotonin transporter gene (SLC6A4) and suicidal behavior in patients with schizophrenia

The lifetime risk of committing suicide among patients with schizophrenia is approximately 5% [1, 2]. Also attempted suicide is common, with estimates ranging from 20% to 40% [3]. The heritable component of suicide attempt is partly related to psychiatric disorders but also partly independent of them [4–6].

The serotonin (5-hydroxytryptamin; 5-HT) system has a central role in the circuitry of cognition and emotions and exerts significant effects on anxiety, mood, impulsivity, sleep, ingestive behavior, reward systems, and psychosis [7]. Pharmacological agents targeting central 5-HT system has substantial effects on emotional behaviors [8–10]. Moreover, several reports imply that 5-HT may be involved in pathophysiological events associated with suicide [11–15]. In fact, postmortem examination of suicide victims shows significantly lower serotonin transporter binding in the prefrontal cortex [16].

Epidemiological and genetic studies indicate that there is a genetic component to suicidal behavior [17]. For example, a single nucleotide polymorphism (SNP) (rs1800532, A218C) in the tryptophan hydroxylase 1 gene (TPH1), which encodes a rate-limiting enzyme involved in the development of 5-HT, was previously found to be associated with suicide or suicidal attempt [18, 19]. The same SNP was also investigated by us in the Scandinavian case–control material used in the present study [20]. However, no association was detected between TPH1 A218C and suicide attempt among the patients, although an association was found to schizophrenia.

Another gene in the serotonin system, reported several times to be involved in the risk of attempting suicide, is the serotonin transporter gene (SLC6A4 or 5-HTT) [11, 12, 19, 21]. The human SLC6A4 gene (Figure 1), located on chromosome 17q11.2, is about 39,500 base pairs long and contains 14 exons, of which exons 1a and 1b are alternatively transcribed. The gene encodes a membrane protein that plays a major role in the regulation of synaptic serotonin concentrations by recycling serotonin from the synaptic cleft.

Results from studies of 5-HTT knockout mouse have provided important clues regarding downstream effects of altered 5-HT function on behavior and the development of the brain, such as regulation of cell proliferation, migration and differentiation of neuronal tissue [22, 23]. Interestingly, early alterations in 5-HT availability during the development of the 5-HT system might impact the distribution and density of specific 5-HT receptors, thereby altering the postsynaptic effects of 5-HT on target neurons and the reactivity of these neurons to specific stimuli [24].

In recent years, five genome-wide studies analyzing suicidal behavior in patients with bipolar disorder, alcoholism and major depression identified candidate regions on chromosome 2p11-12, 2p25 and 6q25-26 [6, 25–29]. As far as we know, no genes associated with suicidal behavior have been reported in these regions yet. Neither are we aware of any genome wide studies analyzing suicidal behavior in patients with schizophrenia.

In the present study, we wanted to further explore the possible involvement of the serotonin system in suicide among patients with schizophrenia, by analyzing genetic variations in the SLC6A4 gene using a Scandinavian case–control sample. Scandinavian countries are generally considered to be well suited for genetic studies, as the populations are ethnically homogeneous and only recently have been subject to non-Caucasian immigration. Seven SNPs located in the SLC6A4 gene (Figure 1), were selected and genotyped in 837 patients suffering from schizophrenia and related disorders, of which 738 had information on suicidal behavior, and in 1,473 control individuals.

We also performed genotype association analyses. However, due to the low frequency of the minor allele of rs16965628 (0.05) only one individual was CC homozygous, and consequently the same association signal was captured with the allele and the genotype test. We did not find any evidence of heterogeneity in the association signal between the three countries, nor between male and female patients for this marker (P-value = 0.75 and 0.17, for country and gender as modifiers, respectively). In other words, all countries and both genders contributed to the observed association between rs16965628 and suicidal behavior.

For completeness, the allele and genotype frequencies of the seven polymorphisms in patients with suicidal behavior were also contrasted against the frequencies in healthy controls. The results were similar to the comparison within the patient group, and the rs16965628 C-allele was significantly lower among individuals with suicidal behavior than in the controls (p-value = 0.01 after correction for multiple testing, data not shown).

We analyzed linkage disequilibrium (LD) between the polymorphisms and found two blocks. All markers, except rs16965628, were located in the observed LD blocks (Additional file 1: Figure S1). Haplotype association analyzes for the two blocks, with respect to suicidal attempt and schizophrenia susceptibility, were performed with UNPHASED. The haplotype and single marker analyzes gave similar results, i.e. there were no significant haplotype association with neither phenotype (p > 0.106, data not shown).

In the present study we found an association between the SLC6A4 rs16965628 polymorphism and attempted suicide among patients with schizophrenia (P-value = 0.00092 and global P-value = 0.01). The odds ratio was 0.39, indicating that the presence of the minor C-allele protects against suicidal behavior. The rs16965628 polymorphism is located in the first intron of the SLC6A4 gene, approximately 9 kb downstream of promoter region. The most studied polymorphisms in the SLC6A4 gene are a 44-base pair insertion–deletion in the promoter region, generating the major L (LA and LG) and S-alleles, and a 17-bp variable number of tandem repeats (VNTR) in the second intron [45]. During the years conflicting results have been reported regarding genetic variants in the SLC6A4 gene and suicidal ideation [46]. The ambiguous results from different studies possibly reflect insufficient sample sizes to obtain adequate statistical power, and heterogeneity between populations [47]. However, the two latest of three meta-analyses report association between SLC6A4 variants and suicidal behavior [11, 12, 21]. In the latest and largest meta-analysis, including 39 different association studies (covering all published studies up to 2006), a significant association was found between 5-HTTLPR and suicide attempts/suicide. The L-allele was underrepresented among individuals with suicidal behavior (odds ratio was 0.88), suggesting that the investigated gene variant had a protective effect against suicide [11].

The mechanism behind this association is unknown, but the 5-HTTLPR polymorphism appears to affect SLC6A4 gene expression. That is, the transcription of the SLC6A4 gene is lower in the presence of the S-allele as compared to the L or the LA-alleles [48, 49]. As mentioned in the introduction, postmortem examination of suicide victims shows significantly lower serotonin transporter binding in the prefrontal cortex [16]. Thus it is possible that the protective effect of the L-allele is the result of elevated expression of the serotonin transporter protein, leading to an overall increase in serotonin transporter binding.

The relative abundance of allele transcripts of the serotonin transporter gene in human cell lines from HapMap CEPH trios, confirm the effects of 5-HTTLPR variation on the transcription of the SLC6A4 gene [50]. However, in these cell lines variation in the rs16965628 SNP was associated with considerably more transcriptional variation than the 5-HTTLPR polymorphism, and the rs16965628 C-allele was linked with increased transcription (the mean allele G/C transcript ratio was 0.47) [50]. Thus we speculate that the decreased frequency of suicidal behavior among schizophrenic patients that carried the rs16965628 C-allele, observed in this study, may have been caused by increased transcription of the SLC6A4 gene and a corresponding increase in serotonin transporter binding.

The gene SLC6A4 has frequently been implicated in psychiatric disorders [51]. Meta-analyses have reported association between 5-HTTLPR S-variant - and sometimes also the 17-bp VNTR in intron 2 polymorphism - and bipolar disorder [52-55], co-morbid bipolar disorder and tobacco use disorder [56], alcohol dependence [57,58], major depression [59,60], antidepressant-induced mania [61,62], a modulation effect of 5-HTTLPR on stress and depression [63], however highly discussed [64], and anxiety-related traits [65-67], almost all disorders with elevated rates of suicide or suicide attempts. Almost all of the above mentioned disorders show elevated rates of suicide or suicide attempts. In schizophrenia, the 5-HTTLPR and 17-bp VNTR in intron 2 polymorphisms have been associated with the disorder in individual studies and the 17-bp VNTR also in several meta-analyses [68–70]. However, the latest update in the SzGene database (http://www.szgene.org) indicates lack of overall allele association. In previous studies, 5-HTTLPR and schizophrenia was investigated by us using Swedish case–control samples, partly overlapping with the present. The results indicated that alleles within the gene were associated with age of onset [71] and disease [44]. Other researchers reported that a haplotype, including markers 5-HTTLPR and rs16965628, was associated with obsessive compulsive disorder [72]. Furthermore, rs16965628 was found to modulate task-related activation in ventral prefrontal cortex in patients with posttraumatic stress disorder [73]. Patients with obsessive compulsive disorder and posttraumatic stress disorder are known to have an elevated risk of committing suicide [74, 75].

In the present study, we did not find any significant association between genetic variants in SLC6A4 and schizophrenia although an association was found to suicide attempt. This may seem contradictive but previous epidemiological studies including monozygotic and dizygotic twins showed an elevated risk of attempting suicide even after controlling for psychiatric disorders, indicating that genetic factors independent of psychiatric disorders affect the risk of suicide attempt [4, 5]. Thus our results do not support that the investigated SNPs are associated with a substantial increase in the disease risk in the Scandinavian population. However, even though our sample size was large, it does not have the power to unambiguously detect weak signals, and thus a true association with an odds ratio below 1.2 cannot be ruled out [76].

The present results provide support for an association between a relatively uncommon SLC6A4 polymorphism and suicidal behavior among schizophrenic patients of Scandinavian origin. As the marker was not associated with the disease, the decreased risk for suicidal behavior appears to be unlinked to that of schizophrenia susceptibility. Although the estimated effect of the rs1695628 C-allele was substantial, its low frequency suggests that the potential contribution to variation in suicidal behavior in the population would be limited. Further studies in independent samples are needed to establish a link between rs1695628 and suicidal behavior in the Scandinavian and other populations.