at SciVerse ScienceDirect

Journal of Psychiatric Research 47 (2013) 1665e1672

Contents lists available

Journal of Psychiatric Research

journal homepage: www.elsevier .com/locate/psychires

Association of 5HTR1A gene variants with suicidal behavior:Case-control study and updated meta-analysisq

Thelma Beatriz González-Castro a, Carlos Alfonso Tovilla-Zárate b,*, Isela Juárez-Rojop a,Sherezada Pool García c, Alma Genis d, Humberto Nicolini d, Lilia López Narváez e

aUniversidad Juárez Autónoma de Tabasco, División Académica de Ciencias de la Salud, Villahermosa, Tabasco, MexicobUniversidad Juárez Autónoma de Tabasco, División Académica Multidisciplinaria de Comalcalco, Comalcalco, Tabasco, MexicocHospital General de Comalcalco, Tabasco, Secretaría de Salud, Comalcalco, Tabasco, MexicodGrupo de Estudios Médicos y Familiares Carracci, México, D.F., MexicoeCIGEN, Centro de Investigación Genómica, Comalcalco, Tabasco, Mexico

a r t i c l e i n f o

Article history:Received 15 February 2013Received in revised form6 April 2013Accepted 11 April 2013

Keywords:SuicideSerotonin receptorGenetic association studyMeta-analysis

Abbreviations: PRISMA, preferred reporting itemsmeta-analyses; DSM-IV, diagnostic and statistical maq This is an open-access article distributed unde

Commons Attribution License, which permits unresreproduction in any medium, provided the original au* Corresponding author. Tel.: þ52 9933581500x690

E-mail addresses: thelma.glez.castro@gmail.com (Ttovillaz@yahoo.com.mx (C.A. Tovilla-Zárate), iselajua2Rojop), shepoga70@hotmail.com (S. Pool García), genicolini_humberto@yahoo.com (H. Nicolini),(L. López Narváez).

0022-3956/$ e see front matter � 2013 The Authorshttp://dx.doi.org/10.1016/j.jpsychires.2013.04.011

a b s t r a c t

Introduction: The gene encoding the serotonin 1A receptor (5HTR1A) has been a candidate gene asso-ciated with suicidal behavior in case-control and meta-analysis studies. We carried out a meta-analysisand a case-control study on the 5HTR1A gene to examine the association of this gene with suicidalbehavior.Methods: We performed a systematic search in electronic databases to study meta-analytically the as-sociation of 5HTR1A gene with suicidal behavior; we found 9 published genetic association studiesconcerning the rs6295 polymorphism. To get a comprehensive knowledge of this association we con-ducted a case-control study on the following polymorphisms: rs1423691, rs6295, and rs878567 in aMexican population; the sample was composed of 152 suicide attempters and 264 healthy subjects.Results: The meta-analysis revealed that the rs6295 polymorphism is not associated with suicidalbehavior. Similarly, no significant association for polymorphisms rs6295 and rs878567 was found in thecase-control study. The polymorphism rs1423691 was excluded of the association analysis because casesand control groups were in HardyeWeinberg disequilibrium.Conclusion: The meta-analysis of functional rs6295 polymorphisms produced no association. Likewise,the analysis in our case-control study in a Mexican population resulted in lack of association of poly-morphisms rs6295 and rs878567 with suicidal behavior. However, further studies assessing differentpopulations, as well as larger samples are necessary to obtain conclusive outcomes.

� 2013 The Authors. Published by Elsevier Ltd. All rights reserved.

1. Background

Suicide is a significant public health issue and one of the mostcommon causes of death throughout the world (Anguelova et al.,2003; Kamali et al., 2001). Genetic studies are trying to identifypossible genetic markers that will be helpful for the early detection

for systematic reviews andnual of mental disorders-IV.r the terms of the Creativetricted use, distribution, andthor and source are credited.0..B. González-Castro), alfonso_2@yahoo.com.mx (I. Juárez-nis76@yahoo.com (A. Genis),dralilialonar@yahoo.com.mx

. Published by Elsevier Ltd. All righ

and prevention of suicidal behavior. The implication of the sero-tonergic system in suicidal behavior has been studied for severalyears (Arango et al., 2001; Tovilla-Zárate et al., 2011).

At present, several studies have investigated the association be-tween various 5HTR1Apolymorphisms and suicide and suggest thatthis gene is playing a role in suicidal behavior (Ohtani et al., 2004;Samadi Rad et al., 2012; Serretti et al., 2007, 2009; Videtic et al.,2009; Wasserman et al., 2006). Even though in our most recentmeta-analysis we could not find any association with suicidalbehavior, the analysis only included 4 studies comprising 957 pa-tients and 957 controls. Therefore,we decided to undertake anotherstudy in which a larger sample size and more specific selectioncriteria were included (Angles et al., 2012). On the other hand,concomitant to suicidal behavior we observed several clinical fea-tures such as the presence of impulsivity, substance abuse (alcoholdependence), andevenmajordepression.As a result, various geneticstudies are taking into consideration the analysis of suicidal

ts reserved.

T.B. González-Castro et al. / Journal of Psychiatric Research 47 (2013) 1665e16721666

behavior in association with other diseases (Benko et al., 2010;Czesak et al., 2012; Koller et al., 2006; Lemonde et al., 2003; Pitchotet al., 2005; Videtic et al., 2009; Wang et al., 2009; Wrzosek et al.,2011; Yoon and Kim, 2009; Zhang et al., 2008; Zupanc et al., 2010).

5HTR1A receptors are located both at post-synaptic and pre-synaptic levels. In the first case, they mediate the action of sero-tonin in cortical and limbic neurons; in the second case, they act asauto-receptors in serotonergic neurons in the raphe nuclei andprevent the release of serotonin by a negative feedback (Czesaket al., 2012; Yoon and Kim, 2009).

The 5HTR1A receptor gene is located on the long arm of chromo-somal region 5q11.2-13; it is intronless, codes for a 422 amino acid,and has a number of single nucleotide polymorphisms (SNPs) (Angleset al., 2012; Samadi Radet al., 2012;Videtic et al., 2009).However, onlya small number of the latter have been investigated so far as candidatevariants in suicidal behavior. One of the most common SNPs is thefunctional C-1019G variant (rs6295) (Chen et al., 2004; Kim et al.,2011; Koller et al., 2006). This polymorphism is located within the26-bp palindromic region, which bounds a single repressor, the nu-clearDEAF-1 related (NUDR)protein. TheG allele abolishes repressionby NUDR and produces higher expression of 5HTR1A, enhances thenegative feedback inhibitionof serotonergic rapheneurons exertedby5HTR1A autoreceptors, and leads to lower serotonergic neurotrans-mission (Benko et al., 2010; Lemonde et al., 2003). For these reasonsthe alleleGof C-1019Gpolymorphismmaybe considered an allele riskmanifest in suicide attempters. The polymorphism rs1423691 islocated downstream the band 5q12.3, in which the change of C to Tdoes not produce a different amino acid.With regard to the rs878567polymorphism, it is located in the same band of rs1423691. Likewise,the change of C to T does not produce a change in amino acid. Otherless studied polymorphisms of the 5HTR1A gene associated withsuicidal behavior are rs1423691 and rs878567. Currently, there areonly few reports showing an association between suicidal behaviorand these polymorphisms (Serretti et al., 2007, 2009).

In this study, we conducted an updated meta-analysis of allpublished data on the rs6295 polymorphism. Also, we investigatedthe possible association between a panel of markers of the 5HTR1Agene (rs6295, rs1423691 and rs878567) in a Mexican sample ofsuicide attempters and controls.

2. Methods

2.1. Meta-analysis

This meta-analysis followed the Preferred Reporting Items forSystematic Reviews and Meta-Analyses (PRISMA) criteria (Moheret al., 2009; Swartz, 2011).

2.2. Identification and selection of publications

The literature included in the current study was selected fromPubmed and EBSCO databases with the keywords: “HTR1A andsuicidal behavior”, “HTR1A and suicide”, “rs6295 and suicide”, “5-hydroxytryptamine-receptor 1A” and “serotonin receptor 1A”.The literature search was performed in September 2012. The datesof publication of the searched papers comprised from March 2003to August 2012. To be selected, the publications had to fulfill thefollowing criteria: (1) to be published in peer-reviewed journals, (2)to be written in English, (3) to contain independent data, (4) to becase-control association studies in which the frequencies of threegenotypes were clearly stated or could be calculated, and (5) theuse of healthy individuals as controls. Besides, we included onearticle consisting only of cases, because the sample in this studywas large and raised the detection power in the meta-analysisstudy (Benko et al., 2010).

Two investigators (González-Castro TB, Tovilla-Zárate CA)working independently screened each of the titles, abstracts, andfull texts to determine inclusion. The results were compared anddisagreements were resolved by consensus.

2.3. Data extraction

The following data were obtained for each of the studies: au-thors, year of publication, region, number of cases and controls,gender, age, psychiatric diagnosis, ethnical origin, sample size andgenotype and allele frequencies. These data were not alwaysavailable for all studies, but when needed, we contacted the authorsto ask for genotype distributions not included in the papers. Thestudy by Benko et al. (2010) did not report a control group but wemade an adjustment accordingly based on the other studies in theliterature that included a control group. Briefly, we calculated theweighted frequency of a particular genotype from studies thatincluded controls and applied it to the study not including a controlgroup. We considered the number of the “virtual” control group asequal to the number of patients in a specific study. Then, the hy-pothetical number of subjects with the particular genotype fre-quency was assigned in proportion to the percentage of the samegenotype that was obtained from the weighted analysis (Illi et al.,2009; Kishi et al., 2011). The outcomes of the meta-analysis werebuilt by taking into consideration the following categories: (a)exposed sick, (b) exposed not-sick, (c) not-exposed sick and (d) not-exposed not-sick. The “sick” term refers to subjects exhibitingsuicidal behavior and the “exposed” term to the allele of risk. Theinformation of the selected publications was extracted indepen-dently by the same two authors (González-Castro TB, Tovilla-ZárateCA) in accordance with the inclusion criteria.

2.4. Data analysis

For the meta-analysis procedures, we used the EPIDAT 3.1 pro-gram (http://dxsp.sergas.es). This software is freely available forepidemiologic analysis of tabulated data. Datawas analyzedwith therandom-effects model following the reports in the literature (Angleset al., 2012; Tovilla-Zárate et al., 2011). Sample heterogeneity wasanalyzedwith the Dersimonian and Laird’s Q test. The result of the Qtest was complemented with graphs to help visualize those studiesfavoring heterogeneity. The results of the meta-analysis areexpressed as odds ratios (ORs). To address the problem of publica-tion bias, the Egger’s test and funnel plots were calculated with theEPIDAT 3.1 software. This plotting standardizes the effect of each ofthe published studies on the vertical axis and its correspondentprecision on the horizontal axis. Finally, a chi-squared (c2) analysiswas used to calculate the HardyeWeinberg equilibrium to evaluategenotype distribution. Studies deemed for inclusion in the system-atic review were scored for methodological quality using theNewcastle-Ottawa Assessment Scale (NOS). A score of six was takenas the cut-off point to distinguish higher from lower quality studies.Quality assessment was done by the same two authors (González-Castro TB, Tovilla-Zárate CA) based on the NOS instrument.

2.5. Case-control study

The group of patients consisted of 152 unrelated suicideattempters. They were consecutively recruited in the outpatientservice from the General Hospital of Comalcalco in the state ofTabasco, Mexico. In addition, 264 healthy volunteers attending theBlood Donor Center were included as controls. All subjects cameexclusively from Comalcalco. The participants in this studydescended from Mexican parents and grandparents to reduceethnic variation and stratification effects.

186 potentially relevant articles(EBSCO 88 Pubmed= 98)

Excluded based on titlen= 104

Screening of 49 references based on their abstracts

Excludedn= 6 review articles

n= 16 no outcome of interest

12 relevant studies retrieved for detailed evaluation

Excluded n= 4

9 studies in the meta-analysis including our study

Fig. 1. Flow-chart showing the different stages of the meta-analysis.

T.B. González-Castro et al. / Journal of Psychiatric Research 47 (2013) 1665e1672 1667

2.6. Clinical evaluation

Patients with attempted suicide were evaluated by a trainedpsychiatrist or clinical psychologists with at least a master’s degreelevel. Initially, a clinical interview helped us to determine the mostrelevant symptoms and the diagnostic was stabilized for DSM-IV inwhich disorders pertaining to Axis I, II and III were considered.DSM-IV lifetime diagnoses of mental disorders among the patientswere classified as: schizophrenia spectrum disorder (30.2%), anxi-ety disorder (50.0%), and undiagnosed (19.8%). We defined suicideattempt as a self-harm behavior with at least some intent to endone’s life (Mann, 2003). Subjects were excluded when self-injurybehaviors were determined to have no suicidal intention or idea-tion based on the assessment of suicidal intention, using the Scalefor Suicide Ideation in the Spanish version (Beck et al., 1979).

On the other hand, each control was interviewed using sys-tematic forms to obtain a detailed medical and psychiatric historyin order to exclude subjects who had relatives with a lifetime his-tory of a mental disorder or suicidal behavior.

2.7. Ethics statement

Written informed consent was obtained from all subjects afterthey were given a detailed and extensive description of the study;they did not receive any economical remuneration. The study wasapproved by the local Ethics and Research Committee DAMC-UJAT(UJAT-DAMC-2012-02). The study was carried out in accordancewith the ethical standards laid down in the 1964 Declaration ofHelsinki.

2.8. Genotype assays

Peripheral blood samples were drawn from all subjects; DNA ofleukocytes was extracted using a modified version of the protocolby Lahiri and Nurnberger (1991). The polymerase chain reaction(PCR) end-point method was used in all patients to analyze thefollowing 5HTR1A genotypes: rs6295(C-1019G), rs1423691, andrs878567. The complete protocol appeared recently in a previousmeta-analysis (Tovilla-Zárate et al., 2011). All genotyping was per-formed blind to patient outcome. As a quality control in our gen-otyping analyses we used random blind duplicates.

2.9. Statistical analysis

The presence of HardyeWeinberg equilibrium was tested usingPearson’s goodness-of-fit chi-squared test. Genotype and allelefrequencies between groups were compared using Chi-squared andFisher’s Exact tests. Tests for association using multi-marker hap-lotypes were performed utilizing Thesias version 3.1 software. Thepower to detect associations was analyzed with the Quanto 1.2software. By way of an example: for rs6295 with minor allelefrequency ¼ 0.3, dominant model, effect size ¼ 1.5, we obtained apower of 0.43. Significance was set at p < 0.05. The Haploview 4.2program (Barrett et al., 2005) was employed to calculate the linkagedisequilibrium (LD) of the markers.

3. Results

3.1. Meta-analysis study

As for the literature search, a total of 186 studies were identified,but only 9 were used including our case-control study. Fig. 1 showsthe selection process of the meta-analysis (Benko et al., 2010;Lemonde et al., 2003; Samadi Rad et al., 2012; Serretti et al., 2007,2009; Videtic et al., 2009; Wrzosek et al., 2011; Yoon et al., 2009).

Our meta-analysis consisted of 2366 cases and 2943 controls(Table 1). We enlarged a previous meta-analysis on rs6295 (C-1019G) (Videtic et al., 2009) with 1217 cases and 1986 controls thatincluded samples of 7 Caucasian populations, 1 Asian population,and our case-control association study.

We measured the HardyeWeinberg equilibrium in all geno-typed populations, both groups, cases and controls, were in equi-librium (p > 0.05) with the exception of the studies by Benko et al.(2010) with p < 0.001 in the control group and Samadi Rad et al.(2012) with a p < 0.0001 in control and patient groups. Also,Lemonde et al. (2003) reported disequilibrium (p < 0.0001) in thepatient group only. Similarly, whenwe calculated all populations ina combined way we encountered p ¼ 0.60 in the controls andp ¼ 0.0002 in the cases. As a result, we measured the HardyeWeinberg equilibrium after heterogeneity was discarded and ob-tained p ¼ 0.33 in controls and p ¼ 0.68 in suicide attempters.

The pooled OR derived from all studies showed a non-significantassociation of the G allele in the C-1019G polymorphism with sui-cidal behavior (Random effects: OR: 1.33; 95% CI: 0.87e2.03;p(Z) ¼ 0.72) and presence of heterogeneity in the studies(Q ¼ 203.50, df ¼ 9; p ¼ <0.0001). The Egger’s test provided noevidence for publication bias (t ¼ �1.62, df ¼ 8; p ¼ 0.14) (Fig. 2A).Given that some studies were not in HardyeWeinberg equilibriumwe conducted an analysis in which the study giving rise to het-erogeneity was discarded. Even then we encountered no associa-tion between rs6295 and suicidal behavior (OR: 0.93; 95% CI: 0.83e1.04; p(Z) ¼ 0.1.3) (Fig. 2B).

We carried out a sub-analysis in the Caucasian populations andfound no association [Egger’s test; Random effects model; OR: 1.47;95% CI: 0.91e2.39 with heterogeneity (Fig. 3A) and OR: 0.94; 95%CI: 0.83e1.08 without heterogeneity (Fig. 3B)]. We could notperform an analysis in the Asian population because we only had asingle study on the association of rs6295 with suicidal behavior.

Table 1Descriptive characteristics of 9 studies on the role of the C-1019G polymorphism of the 5HTR1A gene in suicidal behavior.

Study (Year) Sample size, n(Cases-Controls)

Location Diagnosis Genotypes Alleles Gender (Males) Mean age

GG GC CC G C Cases Controls Cases Controls

Videtic et al. (2009) 323/190 SlovenianCaucasians

Suicide victims 91 160 72 342 182 48 49

Benko et al. (2010) 725/1103 HungarianCaucasians

Suicide attemptersin general population

193 368 163 754 1590 129 30 0

Serretti et al. (2009) 111/289 GermanCaucasians

Suicide attempters 26 55 30 107 291 43 123 39 45

Samadi Rad et al. (2012) 191/218 IranianCaucasians

Suicide victims 56 53 82 217 342 47 45

Lemonde et al. (2003) 102/116 QuebecCaucasians

Suicide victimswith major depression

17 30 55 64 196 53 55 32 34

Serretti et al. (2007) 259/312 GermanCaucasians

Suicide attempters 61 132 66 254 317 39 48

Serretti et al. (2007) 92/163 ItalianCaucasians

Suicide attemptersand victims

17 50 25 84 152 39 48

Wrzosek et al. (2011) 38/112 PolandCaucasian

Suicide attemptersin alcohol dependence

10 16 12 36 113 >18 >18

Yoon and Kim (2009) 181/176 KoreanAsians

Suicide attempterswith major depression

93 79 9 120 89 82 80 40 40

González-Castro(see Section 3.2)

152/264 MexicanLatinAmericans

Suicide attempters 26 58 68 265 303 56 97 25 31

T.B. González-Castro et al. / Journal of Psychiatric Research 47 (2013) 1665e16721668

3.2. Case-control study

The mean age of the 152 patients (56 males, 96 females) was25.5 (9.56) years old (range: 14e56 years). The possibility ofchildhood abuse was not evaluated. The mean age of the 264controls (97 males, 167 females) was 33.1 (13.0) years of age (range:14e61 years). Socio-demographic features of suicide attempters arepresented in Table 2.

The HardyeWeinberg equilibrium was measured in all geno-typed populations. The polymorphisms rs6295 and rs878567 werein equilibrium in both groups (p > 0.05). We encountered twoexceptions: the rs1423691 polymorphismwas not in equilibrium inthe control group (p ¼ 0.0003) or in the group of patients(p ¼ 0.006). In consequence, only the rs6295 (C-1019G) andrs878567 polymorphisms could be analyzed. The LD (Linkage-Disequilibrium) in our Mexican sample taken as a whole wascomplete (r2 ¼ 1.0), as well as when taken separately in controlsand suicide attempters for variants rs6295, rs1423691 and

Fig. 2. Odds ratios and forest plots for all the models in overall studies. (A) G allele v

rs878567 (Fig. 4). Also, we measured the LD with the MEX panelanalysis, but there were no references of any of the followingpolymorphisms: rs6295, rs1423691 and rs878567.

We calculated the genotype counts and frequency distributionsfor polymorphisms rs6295 (C-1019G) and rs878567 of the 5HTR1Agene in the 152 suicide attempters and control group. The resultsare shown in Table 3.

The analysis for the rs6295 (C-1019G) polymorphism yielded nosignificant differences in genotype (c2 ¼ 0.89, p ¼ 0.64, df ¼ 2) orallele (c2 ¼ 0.78, p¼ 0.37, df ¼ 1) frequencies between controls andcases. The same results were obtained for the rs878567 poly-morphism: no association was found in genotype (c2 ¼ 2.83,p ¼ 0.24, df ¼ 2) or allele (c2 ¼ 2.73, p ¼ 0.09, df ¼ 1) frequenciesbetween controls and cases.

We performed a second analysis in the sample of suicideattempters and control subjects with respect to gender. The geno-type and allele distribution is shown in Table 4. The analysis of thers878567 polymorphism in the female gender yielded a significant

s C allele with heterogeneity and (B) G allele vs C allele without heterogeneity.

Standard error

Funnel plot Ef

fect

estim

atio

n

0 0.5 0.1 0.15 0.2 0.25Standard error

Funnel plot

Effe

ctes

timat

ion

0 0.5 0.1 0.15 0.2 0.25

0

A B

Fig. 3. Egger’s funnel plot indicating publication bias for studies on suicidal behavior. (A) G allele vs C allele with heterogeneity and (B) G allele vs C allele without heterogeneity.

T.B. González-Castro et al. / Journal of Psychiatric Research 47 (2013) 1665e1672 1669

association in the allele frequency (c2 ¼ 5.54, p¼ 0.01, df¼ 1). Withregard to the rs6295 (C-1019G) polymorphism we observed thesame results as on the first analysis, no association was found ingenotype and allele frequencies both in males and females. In thisMexican sample, haplotype analysis in relation to attempted sui-cide did not reveal any significant association either (Table 5).

4. Discussion

In this study we analyzed the possible association between the5-HTR1A gene and suicidal behavior based on a case-control studyand an updated meta-analysis. We have already published a meta-analysis of the 5-HTR1A gene (Angles et al., 2012), but we consid-ered necessary to perform another meta-analysis with updatedpublished results that consisted of a larger sample size and morespecific selection criteria, in which various of the 5HTR1A poly-morphisms were studied in relation to suicidal behavior.

The hypothesis that the presence of a single G allele in the C-1019G polymorphism could result in an elevated risk for suicideattempts was investigated by means of a meta-analysis. Thus, weperformed an analysis with all reports published to date. Thestudies presented heterogeneity; however, we conducted a further

Table 2Socio-demographic features of Mexican controls and suicide attempters.

Socio-demographic features Controls Cases c2 df p

Gender [n (%)]Males 97 (36.7) 56 (36.8) 0.004 1 0.98Females 167 (63.3) 96 (63.2)

Marital status [n (%)]Single 117 (44.3) 77 (50.6) 4.63 3 0.20Married 124 (46.9) 63 (41.4)Separated/divorced 10 (3.8) 9 (5.9)Widowed 13 (5.0) 3 (2.1)

Socio-economic level [n (%)]High 3 (1.3) 2 (2.4) 7.31 2 0.02a

Middle 171 (64.7) 78 (50.3)Low 90 (34.0) 72 (47.3)

a p Value significant.

analysis without heterogeneity. We also carried out a sub-analysisin Caucasian populations. Our findings of the G allele in the C-1019G (rs6295) polymorphism support previous results, showingno association between suicidal behavior and the rs6295 (C-1019G)polymorphism when genotypes or allele frequencies were consid-ered. Even when we undertook an explorative analysis includingonly Caucasian samples, no significant associationwas found either.These results are in agreement with our previous meta-analysis ofthe 5-HTR1A gene stating the lack of association (Angles et al.,2012), though this time our sample consisted of 1217 cases and1986 controls more and it also included our case-control study in aMexican population. In addition, the literature search for the pre-sent analysis was performed three years later than the previous

Fig. 4. Linkage disequilibrium in 5HTR1A markers (rs6295, rs1423691 and rs878567)in suicide attempters and control group.

Table 3Genotypes, allele counts and frequency distributions for rs6295 (C-1019G), rs1423691 and rs878567 polymorphisms of the 5HTR1A gene in suicide attempters and controlgroup in a Mexican population.

Genotypes Cases Controls c2 df p Alleles Cases Controls c2 df p

rs6295 rs6295GG, n (%) 26 (0.17) 55 (0.21) 0.89 2 0.64 G, n (%) G: 120 (0.39) G: 225 (0.43) 0.78 1 0.37CG, n (%) 68 (0.45) 115 (0.43) C, n (%) C: 184 (0.61) C: 303 (0.57)CC, n (%) 58 (0.38) 94 (0.36)

rs878567 rs878567TT, n (%) 38 (0.25) 79 (0.30) 2.83 2 0.24 T, n (%) T: 146 (0.48) T: 285 (0.54) 2.73 1 0.09CT, n (%) 70 (0.46) 127 (0.48) C, n (%) C: 158 (0.52) C: 243 (0.46)CC, n (%) 44 (0.29) 58 (0.22)

T.B. González-Castro et al. / Journal of Psychiatric Research 47 (2013) 1665e16721670

one. The above evidence suggests that the rs6295 (C-1019G) poly-morphism is not associated with suicidal behavior. Future studiescomprising larger samples of suicidal behavior are important todetermine this result conclusively.

In addition, we performed a case-control study in a Mexicanpopulation. To our knowledge this is the first study addressing thegenetic association between various polymorphisms (rs6295 andrs878567) of the 5-HTR1A gene with suicidal behavior in a Mexicanpopulation. Our case control study consisted in looking for the asso-ciation of polymorphisms rs6295 and rs878567 with attemptedsuicide, but when we analyzed genotype and allele frequencies ofpolymorphisms rs878567 and rs6295 we did not encounter any as-sociation. The possible role of 5-HTR1A receptors in the pathophys-iology of suicidal behavior has been assessed in case-control studiesbased on familial andpostmortemdata, butwithhighly controversialresults (Arango et al., 2001; Cheetham et al., 1989; Lowther et al.,1997; Stockmeier, 1997). With regard to the post-mortem studies,some of them found a reduction in 5-HTR1A receptor distributionvolume and a decrease in the total number of 5-HTR1A receptors(given as an index) in the dorsal raphe nucleus of suicide victimscomparedwith the control group. These studies also reported that 5-HTR1A receptor binding capacity in the medium raphe nucleus waslower in suicide victims compared with control subjects; they alsosuggested that this reduction in 5-HTR1A auto receptors could be anadaptivemechanism,whose function is to increase serotonin activity(Huang et al., 2004; Lowther et al., 1997; Stockmeier, 1997).

The most studied polymorphism of the 5HTR1A gene in relationto suicidal behavior is C-1019G (rs6295). Lemonde et al. (2003)reported an association of the G allele of the C-1019G poly-morphism with a sample consisting of French-Canadian suicide

Table 4Genotypes, allele counts and frequency distributions for rs6295 (C-1019G) and rs87856Mexican population by gender.

5HTR1A genotypes Males

Cases Controls c2 df p

rs6295GG, n (%) 11 (19.6) 24 (24.7) 0.96 2 0.61CG, n (%) 20 (35.7) 37 (38.1)CC, n (%) 25 (44.7) 36 (37.2)

rs878567TT, n (%) 21 (37.5) 31 (32.1) 0.74 2 0.69CT, n (%) 21 (37.5) 43 (44.3)CC, n (%) 14 (25.0) 23 (23.6)

Alleles Cases Controls c2 df p

rs6295G, n (%) 42 (37.5) 85 (43.8) 1.16 1 0.28C, n (%) 70 (62.5) 109 (56.2)

rs878567T, n (%) 63 (56.2) 105 (54.1) 0.12 1 0.71C, n (%) 49 (43.8) 89 (45.9)

a p Value significant.

victims, but not with suicidality among depressed patients. Simi-larly, the study by Serretti et al. (2007) showed an association of theG allele with suicide attempt only in female groups. Data bySawiniec et al. (2007) dealing with suicide attempts and, recently,Samadi Rad et al. (2012) in suicide victims also support the hy-pothesis of the association between the G allele and suicidalbehavior. Alternatively, the studies of Videtic et al. (2009), Benkoet al. (2010), Wrzosek et al. (2011), and Yoon et al. (2009) showeda lack of association of either the C allele or the G allele in thispolymorphism (rs6295) of 5HTR1Awith suicidal behavior. In 2009,a new and larger study of Serretti et al. (2009) showed no associ-ation of genotype, allele, or haplotype frequency between thers6295 (C-1019G) polymorphism of 5HTR1A and suicidal behavior.Our analysis of rs6295 (C-1019G) is in agreement with their results.However, no conclusive outcomes have been yet attained. Finally,with respect to the rs878567 polymorphism, we could not find anassociationwith suicide attempters in our case-control study in theMexican sample, nor in the Italian and German population samplesof Serretti et al. (2007, 2009).

The discrepancies in the results of association studies may beexplained as follows. First, there are differences in diagnosis in thepopulation of patients. Lemonde et al. (2003) and Yoon et al. (2009)recruited patients with major depression; Wrzosek et al. (2011)worked with alcohol dependent patients, and Benko et al. (2010)studied attempted suicides in the general population. Second,there are differences in genetic heterogeneity in the populations.Several studies have shown that allele frequencies of the variouspolymorphisms of the 5-HTR1A gene may be ethnic-dependent.One of the strengths of our study is that it only included subjectsfrom Comalcalco in the state of Tabasco, Mexico, with parents and

7 polymorphisms of the 5HTR1A gene in suicide attempters and control group in a

Genotypes Females

Cases Controls c2 df p

rs6295GG, n (%) 15 (15.6) 31 (18.5) 0.44 2 0.80CG, n (%) 48 (50.0) 78 (46.7)CC, n (%) 33 (34.4) 58 (34.8)

rs878567TT, n (%) 17 (17.7) 48 (28.7) 5.62 2 0.06CT, n (%) 49 (51.0) 84 (50.2)CC, n (%) 30 (31.3) 35 (21.1)

Alleles Cases Controls c2 df p

rs6295G, n (%) 78 (40.6) 140 (41.9) 0.083 1 0.77C, n (%) 114 (59.4) 194 (58.1)

rs878567T, n (%) 83 (43.2) 180 (53.9) 5.54 1 0.01a

C, n (%) 109 (56.8) 154 (46.1)

Table 5Haplotypes for 5HTR1Amarkers -rs6295 (G/C), rs1423691 (C/T) and rs878567 (C/T)-in controls and suicide attempters.

5HTR1A haplotypes Suicide attempters(frequency)

Controls(Frequency)

Or (CI95%) p

Reference TC 0.55 0.52CG 0.38 0.41 0.89(0.67e1.19) 0.46CC 0.04 0.04 0.83(0.45e1.54) 0.56TG 0.01 0.01 0.66(0.17e2.53) 0.55

T.B. González-Castro et al. / Journal of Psychiatric Research 47 (2013) 1665e1672 1671

grandparents of Mexican descent. Our sample is relativelyhomogenous.

To evaluate our case-control study let us describe its limitations.The size of our case-control sample is small compared with studiesinvolving other diseases and may not have sufficient power todetect an association with suicidal behavior. We did not reportpsychopathologies related to suicide attempters. Also, we did notevaluate possible endophenotypes which could be increasing therisk for suicide attempts. Finally, we could not exclude the geneticheterogeneity in the sample as a whole. On a broader perspective,the number of studies in the meta-analysis is small compared toanalyses dealing with other types of diseases. Also, we did notperform a sub-analysis of suicide attempters and suicide com-pleters, or an analysis based on gender because not all the studiesspecified these variables explicitly.

In conclusion, our case-control study in a Mexican populationshowed no association of the functional rs6295 or rs878567 poly-morphismwith suicidal behavior. Themeta-analysis of the functionalrs6295 (C-1019G) polymorphism supports the outcomes of our pre-vious meta-analysis and of the present case-control study, that is, alack of association. To gain a comprehensive knowledge of the asso-ciation between 5-HTR1A polymorphisms and suicidal behavior it isnecessary to include larger samples and integral studies.

Role of funding source

The collection of data and the genotyping of subjects werecarried out thanks to the support of grants from the UJAT-DAMC-2012-02 and CONACYT CB-2012-177459.

Authors’ contributions

TZC and GCTB conceived the study, participated in its design andhelped to draft the manuscript. TZC, JRI, LNL and NH helped toperform the statistical analysis and to draft the manuscript. PGSrecruited participants and helpedwith data integration and analysis.GA, TZC and HN coordinated and supervised the integration of data.

Competing interests

The authors declare not to have any competing interests.

Acknowledgments

The authors gratefully acknowledge our research volunteerswho helped to recruit the participants in this study. The collectionof data and the genotyping of subjects were carried out thanks tothe support of grants from the University of Tabasco (UJAT-DAMC-2012-02) and CONACYT (CB-2012-177459).

References

Angles MR, Ocana DB, Medellín BC, Tovilla-Zárate C. No association between theHTR1A gene and suicidal behavior: a meta-analysis. Revista Brasileira de Psi-quiatria 2012;34:38e42.

Anguelova M, Benkelfat C, Turecki G. A systematic review of association studiesinvestigating genes coding for serotonin receptors and the serotonin trans-porter: II. Suicidal behavior. Molecular Psychiatry 2003;8:646e53.

Arango V, Underwood MD, Boldrini M, Tamir H, Kassir SA, Hsiung S, et al. Serotonin1A receptors, serotonin transporter binding and serotonin transporter mRNAexpression in the brainstem of depressed suicide victims. Neuro-psychopharmacology 2001;25:892e903.

Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD andhaplotype maps. Bioinformatics 2005;21:263e5.

Beck AT, Kovacs M, Weissman A. Assessment of suicidal intention: the scale forsuicide ideation. Journal of Consulting and Clinical Psychology 1979;47:343e52.

Benko A, Lazary J, Molnar E, Gonda X, Tothfalusi L, Pap D, et al. Significant associ-ation between the C(-1019)G functional polymorphism of the HTR1A gene andimpulsivity. American Journal of Medical Genetics Part B: NeuropsychiatricGenetics 2010;5:592e9.

Cheetham SC, Crompton MR, Czudek C, Horton RW, Katona CL, Reynolds GP. Se-rotonin concentrations and turnover in brains of depressed suicides. BrainResearch 1989;502:332e40.

Chen TJ, Yu YW, Hong CJ, Chen MC, Tsai SJ. Association analysis for the C-1019Gpromoter variant of the 5-HT1A receptor gene with auditory evoked potentialsin major depression. Neuropsychobiology 2004;50:292e5.

Czesak M, Le Francois B, Millar AM, Deria M, Daigle M, Visvader JE, et al. Increasedserotonin-1A (5-HT1A) autoreceptor expression and reduced raphe serotoninlevels in deformed epidermal autoregulatory factor-1 (Deaf-1) gene knock-outmice. The Journal of Biological Chemistry 2012;287:6615e27.

Huang YY, Battistuzzi C, Oquendo MA, Harkavy-Friedman J, Greenhill L, Zalsman G,et al. Human 5-HT1A receptor C(-1019)G polymorphism and psychopathology.The International Journal of Neuropsychopharmacology 2004;7:441e51.

Illi A, Setala-Soikkeli E, Viikki M, Poutanen O, Huhtala H, Mononen N, et al. 5-HTR1A, 5-HTR2A, 5-HTR6, TPH1 and TPH2 polymorphisms and major depres-sion. Neuroreport 2009;20:1125e8.

Kamali M, Oquendo MA, Mann JJ. Understanding the neurobiology of suicidalbehavior. Depression and Anxiety 2001;14:164e76.

Kim HK, Kim SJ, Lee YJ, Lee HJ, Kang SG, Choi JE, et al. Influence of the interactionbetween the serotonin 1A receptor C-1019G polymorphism and negative lifestressors on the development of depression. Neuropsychobiology 2011;64:1e8.

Kishi T, Okochi T, Tsunoka T, Okumura T, Kitajima T, Kawashima K, et al. Serotonin1A receptor gene, schizophrenia and bipolar disorder: an association study andmeta-analysis. Psychiatry Research 2011;185:20e6.

Koller G, Bondy B, Preuss UW, Zill P, Soyka M. The C(-1019)G 5-HT1A promoterpolymorphism and personality traits: no evidence for significant association inalcoholic patients. Behavioral and Brain Functions 2006;2:7.

Lahiri DK, Nurnberger Jr JI. A rapid non-enzymatic method for the preparation ofHMW DNA from blood for RFLP studies. Nucleic Acids Research 1991;19.

Lemonde S, Turecki G, Bakish D, Du L, Hrdina PD, Bown CD, et al. Impaired repressionat a 5-hydroxytryptamine1A receptor genepolymorphismassociatedwithmajordepression and suicide. The Journal of Neuroscience 2003;23:8788e99.

Lowther S, De Paermentier F, Cheetham SC, Crompton MR, Katona CL, Horton RW.5-HT1A receptor binding sites in post-mortem brain samples from depressedsuicides and controls. Journal of Affective Disorders 1997;42:199e207.

Mann JJ. Neurobiology of suicidal behaviour. Nature Reviews Neuroscience 2003;4:819e28.

Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for sys-tematic reviews and meta-analyses: the PRISMA statement. Open Medicine2009;3:21.

Ohtani M, Shindo S, Yoshioka N. Polymorphisms of the tryptophan hydroxylasegene and serotonin 1A receptor gene in suicide victims among Japanese. TheTohoku Journal of Experimental Medicine 2004;202:123e33.

Pitchot W, Hansenne M, Pinto E, Reggers J, Fuchs S, Ansseau M. 5-Hydroxytryptamine 1A receptors, major depression, and suicidal behavior.Biological Psychiatry 2005;58:854e8.

Samadi Rad B, Ghasemi A, Seifi M, Samadikuchaksaraei A, Baybordi F,Danaei N. Serotonin 1A receptor genetic variations, suicide, and life eventsin the Iranian population. Psychiatry and Clinical Neurosciences 2012;66:337e43.

Sawiniec J, Borkowski K, Ginalska G, Lewandowska-Stanek H. Association between5-hydroxytryptamine 1A receptor gene polymorphism and suicidal behavior.Przeglad Lekarski 2007;64:208e11.

Serretti A, Calati R, Giegling I, Hartmann AM, Moller HJ, Rujescu D. Serotonin re-ceptor HTR1A and HTR2C variants and personality traits in suicide attemptersand controls. Journal of Psychiatric Research 2009;43:519e25.

Serretti A, Mandelli L, Giegling I, Schneider B, Hartmann AM, Schnabel A, et al.HTR2C and HTR1A gene variants in German and Italian suicide attempters andcompleters. American Journal of Medical Genetics Part B: NeuropsychiatricGenetics 2007;5:291e9.

Stockmeier CA. Neurobiology of serotonin in depression and suicide. Annals of theNew York Academy of Sciences 1997;836:220e32.

Swartz MK. The PRISMA statement: a guideline for systematic reviews and meta-analyses. Journal of Pediatric Health Care 2011;25(1):1e2.. http://dx.doi.org/10.1016/j.pedhc.2010.09.006.

Tovilla-Zárate C, Juárez-Rojop I, Ramón-Frías T, Villar-Soto M, Pool-García S,Medellín BC, et al. No association between COMT val158met polymorphismand suicidal behavior: meta-analysis and new data. BMC Psychiatry2011;11:151.

T.B. González-Castro et al. / Journal of Psychiatric Research 47 (2013) 1665e16721672

Videtic A, ZupancT, Pregelj P, Balazic J, TomoriM,KomelR. Suicide, stress and serotoninreceptor 1A promoter polymorphism -1019C > G in Slovenian suicide victims.European Archives of Psychiatry and Clinical Neuroscience 2009;259:234e8.

Wang S, Zhang K, Xu Y, Sun N, Shen Y, Xu Q. An association study of the serotonintransporter and receptor genes with the suicidal ideation of major depression ina Chinese Han population. Psychiatry Research 2009;170:204e7.

Wasserman D, Geijer T, Sokolowski M, Rozanov V, Wasserman J. The serotonin 1Areceptor C(-1019)G polymorphism in relation to suicide attempt. Behavioraland Brain Functions 2006;2:14.

Wrzosek M, Lukaszkiewicz J, Serafin P, Jakubczyk A, Klimkiewicz A, Matsumoto H,et al. Association of polymorphisms in HTR2A, HTR1A and TPH2 genes with

suicide attempts in alcohol dependence: a preliminary report. PsychiatryResearch 2011;190:149e51.

Yoon HK, Kim YK. TPH2 -703G/T SNP may have important effect on susceptibility tosuicidal behavior in major depression. Progress in Neuropsychopharmacology &Biological Psychiatry 2009;33:403e9.

Zhang J, Shen Y, He G, Li X, Meng J, Guo S, et al. Lack of association between threeserotonin genes and suicidal behavior in Chinese psychiatric patients. Progressin Neuropsychopharmacology & Biological Psychiatry 2008;32:467e71.

Zupanc T, Pregelj P, Tomori M, Komel R, Paska AV. No association between poly-morphisms in four serotonin receptor genes, serotonin transporter gene andalcohol-related suicide. Psychiatria Danubina 2010;22:522e7.