اثر افزایشی برای خطر ژنتیکی اختلال شخصیت ضد اجتماعی با ترکیب 5 HTTLPR و 5 HTTVNTR پلی مورفیسم

Abstract As the serotonin transporter gene (SLC6A4 or 5-HTT) is a key regulator of central serotonergic activity, several association studies between Antisocial Personality Disorder (APD) and the SLC6A4 polymorphisms have been conducted in the last decade. In the present study, the role of both 5-HTTLPR and 5-HTTVNTR polymorphisms of the SLC6A4 gene in APD is investigated. A sample of 147 male inmates was analyzed. APD was assessed by Aluja's Antisocial Personality Disorder Scale, a measure that correlates 0.73 with the dimensional score of DSM-IV APD and 0.62 with factor II of the Psychopathy Checklist-Revised. Inmates presenting both 5-HTTLPR S/S + S/L and 5-HTTVNTR 12/12 had a higher risk of being classified in the APD group (Odds ratio = 3.48). The results also showed that the genotype and haplotype distribution was more dissimilar when extreme groups were compared with odds ratios up to 6.50. Our results supported that, in addition to the widely investigated 5-HTTLPR polymorphism, the 5-HTTVNTR polymorphism might be an interesting candidate for association studies with APD. Results also suggested that previous failures to replicate the association between serotonin transporter gene polymorphisms and APD, or similar phenotypes, could have been due to an under-representation of extremely high APD subjects in the samples analyzed.

. Introduction Antisocial Personality Disorder (APD) is a diagnosis applied to persons who routinely behave with little or no regard for the rights, safety or feelings of others. APD and its ultimate behavioral consequences, such as distinct types of aggression, violence and impulsivity, are associated with a decrease in serotonin (5-hydroxytryptamine, 5-HT) activity (Goodman and New, 2000 and Lesch and Merschdorf, 2000). Specifically, a serotonergic reduction has been identified in highly impulsive aggressors (Däderman and Lidberg, 2002), mentally disordered violent offenders (Lidberg et al., 2000), and patients suffering from severe personality disorders of the high aggression/low impulse control-type, such as the borderline and antisocial ones (Reif and Lesch, 2003). Following the release of serotonin, the serotonin transporter gene (SLC6A4; solute carrier family 6 [neurotransmitter transporter, serotonin], member 4), also known in the literature as 5-HTT gene (5-HT transporter), plays a key role in controlling 5-HT availability in the synapse by regulating the reuptake of 5-HT ( Lesch et al., 1994). The human SLC6A4 has been mapped to chromosome 17q11.1-q12, and several common polymorphisms of this gene have been identified, including two variable number of tandem repeat (VNTR) polymorphisms located in the promoter region (serotonin transporter gene-linked polymorphic region; 5-HTTLPR) ( Lesch et al., 1996) and intron 2 (termed Stin2 by Lesch et al., 1994 and 5-HTTVNTR by Ogilvie et al., 1996). The 5-HTTLPR polymorphism is a variable number of tandem repeat (VNTR) with a repeat unit of 22 nucleotides. It has two common alleles, designated as short (S) and long (L) alleles (corresponding to zero and two repeats, respectively) and two rarer alleles designated as very long (vL) and extra long (xL) alleles (carrying more than two repeats) (Gelernter et al., 1999). Heils et al. (1996) and Sander et al. (1998) demonstrated that the low-activity S allele of the 5-HTTLPR polymorphism conferred susceptibility to alcohol dependency in subjects with dissocial personality disorder in a German population. This finding is further supported by the results of a later study of a Finnish sample, which found that the S allele is associated with an increased risk of early onset alcoholism, antisocial personality disorder (APD), and habitually violent behavior ( Bennett et al., 2002). In a study of adoptees, it was demonstrated that male individuals with the S variant were more likely to have higher symptom counts for conduct disorder or aggression ( Cadoret et al., 2003). However, not all studies have demonstrated this association between the S allele and aggression. In an American study, no association between antisocial alcoholism and the 5-HTTLPR polymorphism was demonstrated ( Hill et al., 2002). Furthermore, a study of African-American cocaine-dependent individuals showed no evidence of an association between the 5-HTTLPR polymorphism and impulsive-aggressive traits ( Patkar et al., 2002). Also, in a study with Chinese inmates, carriage of the S allele was associated with extremely violent criminal behavior, but not with APD ( Liao et al., 2004). The 5-HTTVNTR is a VNTR polymorphism with a repeat unit of 17 nucleotides, presenting two common alleles (corresponding to 10 and 12 repeat units) and two rarer alleles (corresponding to 11 and 9 repeat units) (Gelernter et al., 1999). This polymorphism has been the subject of considerably less research than the 5-HTTLPR. Recently, the 5-HTTVNTR polymorphism was also found to be significantly associated with Attention Deficit Hyperactivity Disorder (ADHD) in two studies, the frequency of allele 10 being higher in patients than in healthy controls (Zoroglu et al., 2002 and Kent et al., 2002). In an article related to the topic of the present study, the results of Davidge et al. (2004) revealed a significantly reduced frequency of the 5-HTTVNTR 10 repeat allele in children displaying the high-aggression phenotype compared with normal controls. The use of single polymorphic markers to explore the contribution of the SLC6A4 locus on APD could be more informative if we also incorporated the analysis of combined effects at both, genotype and haplotype levels. Recent studies on the structure of human genetic variation have revealed that the transmission of neighboring genetic variants is strongly influenced by the genomic architecture of their specific loci. In recent years, linkage disequilibrium (LD) patterns around the human genome have been investigated theoretically and empirically ( Pritchard and Przeworski, 2001 and Wall and Pritchard, 2003). Following an international initiative, the International HapMap Consortium has obtained a haplotype map of the human genome ( Internacional HapMap Consortium, 2005 and Thorisson et al., 2005) which is a useful tool to guide the design and analysis of genetic association studies. In the present study, we investigated the role of both 5-HTTLPR and 5-HTTVNTR polymorphisms of the SLC6A4 gene in APD in a sample of male inmates. APD is estimated to affect 3% of males and 1% of females in normal populations ( Sutker et al., 1993). As these percentages are considerably higher among prison inmates, especially for males ( Fazel and Danesh, 2002 and Moran, 1999), the optimal target population to conduct a study of the genetic basis of APD is male inmates. Related to this point, taking into account the low prevalence of APD in the general population, the availability of a sufficient number of subjects with extreme scores in APD could only be expected to be reached in samples such as inmates. As far as we know, no study has ever considered the possible role of the 5-HTTVNTR polymorphism in APD. Moreover, we wanted to investigate the combined effect of both polymorphisms. If indeed a certain allelic combination showed an effect, this would underline the functionality of single polymorphisms and the need to investigate both polymorphisms in clinical studies exploring potential influences on APD. In addition, we incorporate available information from HapMap on the transmission architecture of the SLC6A4 genomic region, to be integrated in a model to better explain the results obtained. In this way, the study constitutes the first attempt to combine both single polymorphisms and haplotypes in order to explore APD differences.

Results According to the LD pattern of the SLC6A4 locus as derived from HapMap genotype data, the 5-HTTLPR and 5-HTTVNTR polymorphisms were located on the low LD region and block B2, respectively, which seems to indicate an independent distribution of these two polymorphic loci. In contrast, LD values were larger (D′ = 0.61 and r2 = 0.171), indicating a trend for co-transmission of certain alleles such as LPR_S and VNTR_12 (haplotype S_12) in this cohort. The block structure also revealed that polymorphisms located at block B2, such as the 5-HTTVNTR locus, in addition to their putative functional role, could be considered markers for the total SLC6A4 coding region. Table 1 shows the genotype and allele distributions for the whole sample, inmates presenting APD (AAPDS score ≥ 15) and those without APD (AAPDS score < 15). No departure from Hardy–Weinberg equilibrium was observed (P = 0.075 and 0.448 for 5-HTTLPR and 5-HTTVNTR, respectively). According to the results obtained, the LPR_S allele was associated with APD following a dominant model (S/L + S/S vs. L/L; odds ratio (OR) = 2.25; 95% confidence interval (CI) 1.01–5.06; P = 0.047) and the VNTR_12 allele was associated with APD following a recessive model (12/12 vs. 12/10 + 10/10; OR = 2.56; 95% CI: 1.02–6.39; P = 0.045). Table 1. Genotype distribution in groups considered. Inmates grouped according to AAPDS score (n(%)) a Inmate groups compared < 15 vs. ≥ 15 < 15 vs. ≥ 27 All inmates < 15 ≥ 15 ≥ 27 OR (95%CI) P value OR (95%CI) P value Single locus b 5-HTTLPR (rs4795541) L/L 46 (32) 15 (47) 31 (28) 8 (23) 1.00 (reference) 1.00 (reference) L/S 66 (46) 9 (28) 56 (51) 16 (46) 2.25 (1.01–5.06)c 0.047 3.0 (1.04–8.47)c 0.039 S/S 31 (22) 8 (25) 23 (21) 11 (31) L 158 (55) 39 (61) 118 (54) 32 (46) 1.00 (reference) 1.00 (reference) S 128 (45) 25 (39) 102 (46) 38 (54) 1.32 (0.76–2.30) 0.312 1.35 (0.79–2.41) 0.298 5-HTTVNTR 12/12 53 (37) 7 (22) 45 (42) 24 (65) 2.56 (1.02–6.39) 0.045 6.61 (2.21–19.15) 0.0005⁎ 12/10 76 (54) 22 (69) 54 (49) 9 (24) 1.00 (reference)d 1 (reference)d 10/10 13 (9) 3 (9) 10 (9) 4 (11) 12 182 (64) 36 (56) 144 (66) 57 (77) 1.68 (0.89–3.17) 0.106 2.60 (1.15–5.86) 0.015 10 102 (36) 28 (44) 74 (34) 17 (23) 1 (reference) 1.00 (reference) Haplotypes Haplotype countse S_12 112.5 (39) 19.8 (30) 92.7 (41) 39.6 (51) 1.67 (0.90–3.08) 0.094 2.22 (1.12–4.48) 0.011 L_10 87.7 (30) 23.1 (35) 64.6 (28) 15.8 (20) 1.00 (reference)e 1.00 (reference)e L_12 73.2 (25) 17.1 (26) 56.1 (25) 20 (26) S_10 18.5 (6) 6.0 (9) 12.5 (5) 2.5 (3) Haplotype S_12 carrier statusf None S_12 copies 51 (35) 16 (49) 35 (31) 9 (24) 1.00 (reference) 1.00 (reference) One S_12 copy 72 (49) 13 (39) 59 (52) 18 (47) 2.08 (0.87–5.00) 0.102 2.49 (0.80–7.77) 0.114 Two S_12 copies 23 (16) 4 (12) 19 (17) 11 (29) 2.17 (0.63–7.43) 0.218 4.88 (1.19–19.94) 0.027 S_12 present 95 (65) 17 (51) 78 (69) 29 (76) 2.10 (0.94–4.72) 0.071 3.15 (1.11–8.92) 0.030 GRF statusg None GRF 36 (26) 13 (42) 23 (22) 7 (21) 1.00 (reference) 1.00 (reference) One GRF 58 (42) 12 (39) 46 (43) 6 (18) 2.17 (0.85–5.49) 0.103 0.93 (0.24–3.56) 0.914 Two GRF 43 (32) 6 (19) 37 (35) 21 (62) 3.48 (1.16–10.45) 0.026 6.50 (1.78–23.64) 0.004⁎ GRF present 18 (68) 83 (78) 27 (79) 2.60 (1.11–6.09) 0.024 2.78 (0.93–8.33) 0.061 AAPDS, Aluja's Antisocial Personality Disorder Scale; 5HTT, serotonin transporter, OR, odds ratio; 95%CI, 95% confidence interval; LPR, linked polymorphic region; rs4795541, code number at dbSNP data base (http://www.ncbi.nlm.nih.gov/SNP) for 5-HTTLPR; VNTR, variable number of tandem repeat. WHAP software is available at http://pngu.mgh.harvard.edu/~purcell/whap/index.shtml. P values < 0.05 were considered as statistically significant (in italics). a Those columns show the number of subjects and the associate percentage by genetic comparison. It should be remarked that those percentages were computed considering each AAPDS group separately. For instance, you get 100% if you sum the percentages for the three 5-HTTLPR groups (L/L, L/S, S/S) for the whole sample (32 + 46 + 22), the same for the < 15 group (47 + 28 + 25), and so forth. Note that < 15 and ≥ 15 groups comprised the total sample and, therefore, subjects considered in the ≥ 27 column were also included in ≥ 15 group. b 5-HTTLPR and 5-HTTVNTR genotypes were not available in four and five subjects, respectively. c Dominant model for 5-HTTLPR S allele, subjects carrying genotypes 5-HTTLPR_L/S and 5-HTTLPR_S/S were considered together. d Recessive model for 5-HTTVNTR 12 allele, subjects carrying genotypes 5-HTTVNTR_12/10 and 5-HTTVNTR_10/10 were considered together. e Haplotypes were inferred for all subjects by the WHAP software. Haplotype pairs and their corresponding probabilities were estimated for each individual. Haplotypes are for 5-HTTLPR and 5-HTTVNTR. Logistic regression analysis was performed by WHAP software. Comparison was for haplotype S_12 vs. all other haplotypes. f Logistic regression analysis was performed by SPSS software adjusting for haplotype probability. g Carriers of genotype risk factor (GRF) were defined as those carrying genotypes 5-HTTLPR L/S, 5-HTTLPR S/S and 5-HTTVNTR 12/12. GRF status was available for 138 subjects. ⁎ Statistically significant after Bonferroni correction for multiple testing. Table options We also explored the combined effects of these two polymorphisms by analyzing the contribution of haplotypes and combined genotypes to the association observed. As Table 1 shows, haplotype S_12 seems to confer a degree of risk of being in the APD group, as this haplotype was more prevalent in those presenting APD (41% vs. 30%) and a trend for significance was observed when haplotype S_12 distribution was compared with all other haplotypes considered together (P = 0.094). We then went on to test for differences in haplotype S_12 carrier status. Even though the broad distribution of S_12 carrier status did not show statistical significance, haplotype S_12 once again showed borderline significance for APD risk (carriers of one or two copies of haplotype S_12 were more prevalent in those classified as APD; 69% vs. 51%; P = 0.071). We also tested for combined genotype contribution to the association described. According to the results obtained by the single locus analysis, 5-HTTLPR S/S + S/L and 5-HTTVNTR 12/12 genotypes were considered genotype risk factors (GRF) for APD. Therefore, inmates were classified into three groups, depending on GRF doses. Table 1 also presents the results for this combined analysis. Inmates presenting two GRFs had a higher risk of being classified in the APD group than those presenting no GRF (OR = 3.48; 95% CI 1.16–10.45; P = 0.026). Fig. 2 shows the AAPDS mean scores for: 1) the total inmates sample, 2) inmates with an AAPDS score equal to or higher than 15, and 3) inmates scoring lower than 15. Mean scores reveal significant differences among genotypes for the total inmate sample. It should be noted, however, that differences remain significant only for those in the highest scoring group, but not for the lowest. This pattern suggests that the relationship of these two polymorphisms with APD is mainly produced by subjects with the highest scores. Following this hypothesis, extreme groups, defined by 25th quartile (AAPDS score < 15; n = 33) and 75th quartile (cut-off scores ≥ 27; n = 38) of AAPDS score, were compared. Most of the odds ratios for being in the highest score group assigned to each risk factor were statistically significant ( Table 1), with the largest odds ratios being for 5-HTTVNTR 12/12 vs. 5-HTTVNTR 12/10 + 10/10, presence of Haplotype S_12 vs. no copies of Haplotype S_12 and two GRF vs. no GRF. The odds ratio with associated P values lower than 0.006 (0.05/9) may be considered significant after correction for multiple testing. Mean AAPDS scores in inmates. Inmates grouped according to risk genotype for ... Fig. 2. Mean AAPDS scores in inmates. Inmates grouped according to risk genotype for 5-HTTLPR (triangles), 5-HTTVNTR (squares) and combined genotype risk factors (circles). a) Mean scores in all inmates (black symbols), b) inmates scoring higher than or equal to 15 (white symbols) and c) inmates scoring lower than 15 (grey symbols). Bars indicate 95% confidence interval.