Promoter and functional polymorphisms of HTR2C and suicide victims


Prof. Radovan Komel, Medical Centre for Molecular Biology, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Vrazov trg 2, SI-1000 Ljubljana, Slovenia. E-mail:


In Europe, the countries with the highest suicide rates form a so-called J-curve, which starts in Finland and extends down to Slovenia—a country with one of the world's highest suicide rates. So far, the strongest association between suicide and genes has been shown for the serotonergic system. A functional polymorphism 68G>C (Cys23Ser) and a promoter polymorphism–995G>A of serotonin receptor 2C (HTR2C) have already been investigated, but no associations with suicide were determined. In the present study 334 suicide victims and 211 controls of Slovenian origin were genotyped for the above-mentioned polymorphisms using standard methods. In the case of the polymorphism–995G>A no association with suicide was found. However, a significant association was observed between female suicide victims and polymorphism 68G>C. The significance remained when we combined alleles of female and male populations. An excess of GG genotype and allele G was observed. However, no statistically important differences were present when only males were analyzed. Haplotype analysis on female population showed marginal association of haplotype G-C with suicide. The present study speaks for the plausible implication of the HTR2C in suicide susceptibility.

Suicide is a significant world health problem and a socio-economic burden. In Europe, the countries with the highest suicide rates form the so-called J-curve, which starts in Finland and extends down to Slovenia (Marusic 2005). Slovenia, a small country with a Slavic population of only two million inhabitants, ranks among the top 10 countries with the world's highest suicide rates. In 2004 there were 512 suicides, a rate of 25.6 suicide victims per 100.000 citizens (World Health Organization 2008).

The results of existing genetic epidemiology studies suggest that the genetic contribution to suicide is quantitatively somewhat weaker than the environmental one (Brezo et al. 2008). So far, the strongest association between suicide and genes has been shown for the serotonergic system (Bondy et al. 2006).

The largest group of proteins contributing to the serotonin (5-hydroxytryptamine, 5-HT) signaling pathway is the serotonin receptor group, which is divided into seven subfamilies by convention, and includes 14 different genes. The family of 5-HT2 receptors consists of three members: 5-HT2A, 5-HT2B and 5-HT2C (Raymond et al. 2001). The gene for serotonin receptor 2C lies on chromosome X, on position 24 of the longer arm. It has five introns and six exons (Xie et al. 1996) and is, so far, the only G-protein binding receptor known to have the possibility of alternative splicing of mRNA and therefore of more protein isoforms (Raymond et al. 2001). 5-HT2C has been implicated in numerous brain functions, from regulation of locomotion, sex, instinct and appetite to modulation of mood and anxiety. Since anxiety and mood disorders, as well as impulsiveness, are important suicide risk factors (Evans et al. 2000), the changes in serotonin receptor 2C could be linked to suicide (Pandey et al. 2006).

Besides the potential for alternative splicing, the HTR2C (5-HT2C gene) has a vast number of single nucleotide polymorphisms (SNPs). It is probably the most polymorphic of the 5-HT receptors, since it has 1077 SNPs (according to, March 5, 2009). A functional polymorphism, transversion of G to C at position 68 (Cys23Ser) (Lappalainen et al. 1995), has been investigated in suicide victims. The polymorphism is located within the first hydrophobic region of the HTR2C protein. So far, no or only marginal implication in suicide or suicidal behavior has been determined (Serretti et al. 2007, 2009; Stefulj et al. 2004; Zhang et al. 2008). Studies of other polymorphisms of this receptor are rare and, again, have not shown any association with suicide (Serretti et al. 2009; Turecki et al. 2003). However, there are more studies of other psychiatric disorders, such as schizophrenia, depression and bulimia nervosa (Kumar et al. 2007; Ribases et al. 2008; Wilffert et al. 2005) which present very heterogeneous results when various HTR2C SNPs were tested.

It is of particular interest that gender-specific differences of suicidal behavior suggest an X-linked genetic factor (Stefulj et al. 2004). The background for this speculation could be the ratios of females to males with suicidal ideation (2:1), suicide attempts (4:1) and suicide completion (1:3) (Bondy et al. 2006; Stefulj et al. 2004). Additionally, a higher binding capacity of serotonin type 2 receptors in men than in women has already been shown in the frontal and cingulate cortices and these results are in line with the hypothetical high presynaptic activity of the female serotonin system and the consequent apparent 'down-regulation’ of 5-HT2 in women compared with men (Biver et al. 1996). Since gender differences exist in impulsiveness, aggression or sexual behavior and serotonin is a recognized inhibitory neurotransmitter of these functions (Biver et al. 1996) this is a good indicator for the X-linked genetic background for suicide.

The purpose of this study was to investigate the association of two polymorphisms in HTR2C, one in the promoter region with a potential influence on gene expression,–995G>A, and the second, a functional polymorphism, in the coding region, 68G>C, on a very suicide prone population.

Materials and methods

Research subjects

334 suicide victims and 211 control persons of Slovenian origin were enrolled in the study which was approved by the Slovenian Medical Ethics Committee. Samples were collected in the course of autopsy at the Institute of Forensic Medicine, Faculty of Medicine University of Ljubljana. Venous blood was taken from all subjects and stored at −70°C before DNA isolation. The control group consisted of 75 women and 136 men, with mean age of 47.06 ± 19.2 years (± SD). The mean age of the suicide victims group was 49.06 ± 17.8 years and consisted of 85 female and 249 male subjects (ratio 1:2.9). Violent and nonviolent suicide methods (Asberg et al. 1976) have been used and were distributed as follows: hanging (n = 145, 43.4%), shooting (n = 50, 15%), jumping from a height (n = 37, 11.1%), CO/medication poisoning (n = 38, 11.4%), drowning (n = 27, 8.1%), cutting/inflicting superficial wounds (n = 15, 4.5%), lying under a train (n = 16, 4.8%) and, together, self-burning, suffocation, electrocution and planned car accident (n = 6, 1.7%).


Genomic DNA was extracted from venous blood samples using the Promega commercial kit. Target DNA sequences were amplified with polymerase chain reaction (PCR) in a total volume of 25 μl. The reaction mix contained 30–100 ng of gDNA, 0.3–0.4 μm primers and 0.625 U Taq DNA polymerase in PCR Master Mix (Promega Corporation, Fitchburg, WI, USA). The PCR cycle for 68G>C consisted of 10 min preceding denaturation at 95°C followed by 35 cycles of denaturation at 95°C for 1 min, annealing at 54°C and elongation at 72°C each for 45 s and completion with a final extension step at 72°C for 7 min. Primers for polymorphism 68G>C were forward, 5′-TTGGCCTATTGGTTTGGGAAT- 3′ (bold–degenerated nucleotide used to create a restriction site), and reverse, 5′-GAAGTAATTGGTGGCATTGTGC−3′. PCR products were subsequently digested with HinfI (New England Biolabs, Hitchin, Hertfordshire, UK) according to the manufacturer's protocol. For allele C two bands of 210 and 18 bp were detected and, for allele G, only one uncut product of 228 bp. The PCR protocol and PCR product digestion for polymorphism–995G>A were as described by Turecki et al. (2003). Restriction fragments were visualized on 12% polyacrylamide gel for 68G>C and on 2.5% agarose gel for–995G>A polymorphism. Both electrophoreses were run in 1 × TBE buffer, and products were visualized with ethidium bromide.

Polymorphism–995G>A was also analyzed by high resolution melting. The PCR protocol consisted of 45 cycles of 20 s denaturation at 94°C, 10 s annealing at 70°C and 10 s elongation at 72°C. The 10 μl PCR volume contained 50 ng of gDNA, master mix LightCycler® FastStart DNA MasterPLUS HybProbe (Roche Applied Science, Rotkreuz, Switzerland), bovine serum albumin, 0.3–0.4 μm primers and fluorescent dye LCGreenTM Plus (Idaho Technology LTD, Salt Lake City, UT, USA). Primer sequences for the amplification were forward, 5′-TGTGTGTGTGTGTGTTTGGG−3′, and reverse, 5′-CCTAGAGAGTTGGGGGAACC−3′. PCR amplification produced fragments of 121 bp. After PCR completion, DNA melting on High Resolution Melter HR−1TM (Idaho Technology Inc., Salt Lake City, UT, USA) was carried out. The melting slope was 0.3°C/s, and data were recorded from 75°C to 95°C.

Data analysis

P values of χ2-test and 95%-confidence interval were calculated with SPSS v. 12.0.0 (Statistical Package for the Social Sciences, Chicago, IL, USA). For P values calculation, where more than 20% of the cells in contingency table had expected frequency less than 5, Fischer's exact test (FET) was applied. We used freely accessible software on the www (Kirkman 1996;, March 15, 2009). Significance level was set to 0.05, but after Bonferroni correction only values lower than 0.025 were considered as significant. For global significance the results were confirmed with 10 000 permutations in program Haploview v. 4.1 (Barrett et al. 2005; Cambridge, MA, USA). Due to the X-chromosomal location of the HTR2C males and females were analyzed separately.

The power of our sample was calculated with a α-level of 0.025. For single marker analyses in our sample we had a power of 0.80 to detect a small effect size of w = 0.145, that corresponded to a difference of ∼15% in allele frequency between the groups. Power calculations were performed with an on-line program G*Power (Faul et al. 2007; Kiel, Germany).

Haplotype analysis and linkage disequilibrium (LD) for the two markers, using the parameters |D′| and r2, was calculated with THESIAS program v. 3.1 (Tregouet & Garelle 2007; Paris, France).

Genotyping results were tested for Hardy-Weinberg equilibrium (; Yale University, New Haven, CT, USA, December 10, 2009) where applicable.


The results for genotype and allele frequency distributions for separated female and male populations are shown in Table 1. Genotype distributions were in Hardy-Weinberg equilibrium (data not shown).

Table 1.  Genotype and allele frequency distributions for polymorphisms–995G>A and 68G>C in the gene for serotonin receptor 2C in female and male populations of controls and suicide victims
Locus polymorphism HTR2C -995G>AGenotypeAllele
Female controls, n (%)54 (73.0)17 (23.0)3 (4.0)125 (84.5)23 (15.5)
Female suicide victims, n (%)50 (58.8)30 (35.2)5 (6.0)130 (76.5)40 (23.5)
 P(FET) = 0.170χ2 = 0.556, df = 1, P = 0.227
Male controls, n (%)/113 (83.1)23 (16.9)
Male suicide victims, n (%)/210 (84.7)38 (15.3)
    χ2 = 0.166, df = 1, P = 0.771
Locus polymorphism HTR2C 68G>CGenotypeAllele
Female controls, n (%)41 (57.7)28 (39.4)2 (2.7)110 (77.5)32 (22.5)
Female suicide victims, n (%)65 (79.3)15 (18.3)2 (2.4)145 (88.4)19 (11.6)
 P(FET) = 0.008χ2 = 5.14, df = 1, P = 0.028
 95% CI = 0.175–0.728, OR = 0.357  
Male controls, n (%) /105 (82.7)22 (17.3)
Male suicide victims, n (%) /208 (87.0)31 (13.0)
    χ2 = 1.27, df = 1, P = 0.277

Genotype and allele frequency distributions for the female group were statistically significant for polymorphism 68G>C. Comparison of genotype frequencies showed a much higher frequency of homozygotes GG and lower frequency of heterozygotes GC in the group of female suicide victims than in the control group. Similarly, allele G was more abundant among female suicide victims than in controls (∼11% more). Genotype and allele frequency distributions in female population for promoter polymorphism–995G>A were not statistically significant.

In male population no statistically significant differences in allele frequency distributions between suicides and controls were found for the polymorphisms 68G>C or–995G>A. Also, no statistically significant differences were found when violent and nonviolent suicide methods were analyzed separately (data not shown).

After combined analysis for male and female populations statistical significance of the allele distributions for polymorphism 68G>C remained, and higher frequency of G allele was detected in suicide victims (χ2 = 7.25, df = 1, P = 0.005, 95%CI = 0.370 − 0.859, OR = 0.564). Analysis of allele frequencies of promoter polymorphism–995G>A gave non-significant results (χ2 = 0.579, df = 1, P = 0.256).

Haplotype frequency distributions for the two polymorphisms in female population are shown in Table 2. Marginal statistical significance for the distributions of haplotype G-C between the suicide victims and controls were determined (P = 0.024, OR = 2.065, 95%CI = 1.105 − 3.865) when compared to the haplotype G-G. However, we have to point out that LD between the two polymorphisms was not reliable. Despite |D′| being equal to one, the values of r2 were low, and we could best explain this by gene conversion (Andolfatto & Nordborg 1998). Haplotype distributions in male populations were not statistically significant.

Table 2.  Estimated haplotype frequencies for HTR2C polymorphisms 68G>C and–995G>A in female suicide victims and controls
Haplotype*Suicide victimsControls
  1. *Rare haplotypes (frequency below 0.1) were excluded from the analysis.



The present study speaks for a possible role of serotonin receptor 2C functional polymorphism in suicide susceptibility, and is therefore a kind of counterevidence to previous studies having reported negative results (Serretti et al. 2007, 2009; Stefulj et al. 2004; Turecki et al. 2003).

For the only functional polymorphism of HTR2C, 68G>C (Cys23Ser), it was already discovered that the protein forms with cysteine and with serine have distinct functional characteristics (Lappalainen et al. 1999; Okada et al. 2004), but these differences in amino acid sequence apparently constitute too small contribution to exert a significant effect in complex and heterogeneous psychiatric disorders (Lappalainen et al. 1999).

In the most comprehensive association study, Croatian and German researchers analyzed more than 400 suicide victims and almost 600 controls (Stefulj et al. 2004). Comparison of our results with this study showed a higher proportion of GG genotype and allele G in Slovenian suicide victims. This finding is very interesting, since Croats and Slovenes are of the same Slavic ethnicity. At this point a broader association study would be needed to confirm our results. In a more complex study of suicide attempters and completers by Serretti et al. (2007) a direct association of various HTR2C polymorphisms was not found, although anger- and aggression-related traits, which are considered as an intermediate phenotype of suicidal behavior, showed mild associations.

Our findings of genotype distribution differences between Slovenian and other ethnic populations and sexes could have a broader significance. The effect of inter-population variability of the 68G>C was investigated in an European Collaborative Project of 10 populations and was found to be significant (Lerer et al. 2001). A distinct allele distribution between sexes was shown in a study of different psychiatric patients and healthy people (Fehr et al. 2000). Further studies to confirm these suggestions would be of interest.

To further support the role of 5-HT2C in suicide we analyzed another polymorphism, a polymorphism in the promoter region,–995G>A. This polymorphism, in association with suicide, has been investigated only by Turecki et al. (2003), who found no statistically important differences in genotype and allele frequency distributions between suicides and controls, even after analyzing for violent and nonviolent suicides. The present study showed similar results. However, discrepancies in the genotype distributions between suicide victims of the two studies were observed. Namely, in the Slovenian sample we established a 26% excess of the genotype GG while 30% more heterozygotes GA were present in the French-Canadian population. This could be explained on the basis of ethnical differences between Slovenes and Canadians.

Results of the haplotype analysis suggest that gene conversion might have occurred in this gene region. Namely, our results were very similar to the results of McCarthy et al. (2005) where two promoter polymorphisms (–995G>A and –759C>T) were in strong LD according to parameter |D′| with the polymorphism 68G>C in the second exon. But, like in our case, the r2 values in McCarthy study were low, implicating recombination and gene conversion contributing to the lack of allelic association between polymorphisms and also to inconsistent associations between HTR2C and different phenotypes (McCarthy et al. 2005).

Despite the very limited positive association of 5-HT2C with suicide it appears that genetic studies of 5-HT2C may be very important. Selective 5-HT2C receptor antagonists are sometimes administered together with selective serotonin reuptake inhibitor (SSRI) drugs used for the treatment of different psychiatric disorders like depression (Cremers et al. 2007) or psychosis. (Eltayb et al. 2007) which have been associated with suicide. However, studies of deliberate self-harm (DSH) that could be associated with suicide, showed no association between impulsiveness, suicide risk, lifetime history of depression, or family history of DSH in the DSH subjects and 68G>C (Evans et al. 2000; Pooley et al. 2003).

The plausible role of 2C receptors in suicide susceptibility was further indicated in the study of Padney et al. , where higher expression of 5-HT2C receptor protein in the prefrontal cortex of suicide victims than in controls was observed. This change was related to post-translational modifications and the formation of edited forms of 2C receptors (Pandey et al. 2006).

Despite some promising results of our study there were some limitations that have to be taken in consideration. Although the sample was rather large and we were able to detect small size effect between the polymorphisms, the power reduced when we formed different subgroups (e.g. violent and nonviolent suicide victims, split males and females) and this could have led to false positive results in gender differences. In addition, the lack of data on the plausible psychiatric diagnosis of the suicide victims could be important when needed to contribute in elucidating the role of HTR2C in psychiatric disorders that are often present and interlaced with suicide. And finally, our study encompassed of the analysis of only two polymorphisms out of over one thousand (although both of them have potentially important role in the gene and are not located in intronic regions), and may be therefore less informative.

It has to be born in mind that the HTR2C contains a great variety of polymorphisms and therefore has numerous possibilities to act in different ways. Single polymorphisms may therefore not have a very straightforward or significant effect, but concerted studies of more polymorphisms or haploblocks could well show a more important influence on suicide susceptibility. In addition, the position of the gene for 5-HT2C has to be considered very carefully, since it is X-linked, and could therefore act differently from other genes located on the autosomal chromosomes. In the context of further concerted studies our results could be considered as a valuable contribution to common efforts aiming to understanding the role of serotonergic system in suicide susceptibility.


This work was supported by the Slovenian Research Agency, program grant No. P1-0104-0381.