بررسی ارتباط بین آلل 2 تکرار از پلی مورفیسم پروموتر ژن MAOA و صفات شخصیتی روانی، دستگیری، حبس و رفتار ضد اجتماعی در طول عمر

Abstract A line of research has revealed that a polymorphism in the promoter region of the MAOA gene is related to antisocial phenotypes. Most of these studies examine the effects of low MAOA activity alleles (2-repeat and 3-repeat alleles) against the effects of high MAOA activity alleles (3.5-repeat, 4-repeat, and sometimes 5-repeat alleles), with research indicating that the low MAOA activity alleles confer an increased risk to antisocial phenotypes. The current study examined whether the 2-repeat allele, which has been shown to be functionally different from the 3-repeat allele, was associated with a range of antisocial phenotypes in a sample of males drawn from the National Longitudinal Study of Adolescent Health. Analyses revealed that African-American males who carried the 2-repeat allele were, in comparison with other African-American male genotypes, significantly more likely to be arrested and incarcerated. Additional analyses revealed that African-American male carriers of the 2-repeat allele scored significantly higher on an antisocial phenotype index and on measures assessing involvement in violent behaviors over the life course. There was not any association between the 2-repeat allele and a continuously measured psychopathic personality traits scale. The effects of the 2-repeat allele could not be examined in Caucasian males because only 0.1% carried it.

1. Introduction A significant amount of behavioral genetic research has examined the genetic basis to antisocial behaviors (Moffitt, 2005). The results of these studies, best summarized by a number of meta-analyses, indicate that approximately 50% of the variance in measures of antisocial phenotypes is attributable to genetic factors (Ferguson, 2010, Mason and Frick, 1994, Miles and Carey, 1997 and Rhee and Waldman, 2002). More recent research has begun to investigate genetic polymorphisms that might be partially responsible for producing variation in antisocial phenotypes (Caspi et al., 2002). Genes involved in neurotransmission have been identified as the most promising candidate genes for antisocial behaviors and traits (Ferguson & Beaver, 2009). Although studies have identified an association between polymorphisms in a number of neurotransmission genes and various antisocial behaviors, these associations are often plagued by the inability for follow-up studies to replicate the original findings. An important exception to the replication problem appears to be for a functional polymorphism in the monoamine oxidase A (MAOA) gene. The MAOA gene has been mapped to the X chromosome at location Xp11.23–11.4 (Levy et al., 1989) and codes for the production of the MAOA enzyme that catabolizes certain neurotransmitters, such as dopamine and serotonin (Shih, Chen, & Ridd, 1999). The MAOA gene has a polymorphism in the promoter region that is the result of a 30 base pair (bp) variable number of tandem repeats (VNTR) in the regulatory region of the gene. This polymorphism has been shown to be functional as different alleles correspond to the production of MAOA enzymes with different activity levels (Sabol, Hus, & Hamer, 1998). Recognizing differences in transcriptional efficiency, researchers commonly pool the alleles into two groups: those that correspond to low MAOA activity and those that correspond to high MAOA functioning. In most studies, the 2-repeat allele and the 3-repeat allele are pooled together to form the low MAOA activity genotype and the 3.5-repeat allele, 4-repeat allele, and 5-repeat allele are pooled together to form the high MAOA activity genotype (Caspi et al., 2002). The polymorphism in the promoter region of the MAOA gene has been the source of a considerable amount of research examining whether different alleles are associated with antisocial phenotypes. In a landmark study, Caspi et al. (2002) reported a link between low MAOA activity alleles and antisocial behaviors, but only among males who had been maltreated in childhood. The results of a meta-analysis seemed to confirm the association between MAOA and antisocial outcomes for maltreated males (Kim-Cohen et al., 2006). Although most of this research has revealed that MAOA only has significant effects when paired to a criminogenic environment, there is some evidence to indicate that the low MAOA activity alleles may have effects independent of environmental factors for some antisocial behaviors (Beaver, DeLisi, Vaughn, & Barnes, 2010). Guo and his colleagues (2008) provided evidence that the MAOA gene is related to delinquent behavior in a sample of adolescents and young adults independent of environmental factors. Unlike the vast majority of research examining the effects of the MAOA gene, Guo et al. examined the effects of the 2-repeat allele against the effects of the 3-repeat allele and 5-repeat allele and against the effects of the 3.5-repeat allele and the 4-repeat allele. The results of their analysis indicated that carriers of the 2-repeat allele were at a statistically significant greater risk for engaging in serious and violent delinquency in adolescence and early adulthood. The effects were particularly marked for males. Guo et al. also conducted a functional analysis of the alleles and found that the 2-repeat allele, in comparison with 3-repeat and 4-repeat alleles, had the lowest level of promoter activity. The results of the study by Guo et al. suggest that pooling together the 2-repeat and 3-repeat alleles may be incorrect and that the 2-repeat allele should be examined in isolation because of its functional significance. Besides this single study, though, research has yet to fully explore this possibility and thus whether the 2-repeat allele is truly a marker for antisocial phenotypes remains to be determined. The current study examines this possibility by testing for an association between the 2-repeat allele and psychopathic personality traits, the odds of being arrested, the odds of being incarcerated, and lifetime antisocial behavior in a sample of American males.

. Findings Since prior research has revealed that the distribution of the 2-repeat allele varies by race (e.g., Reti et al., 2011 and Widom and Brzustowicz, 2006), the analysis begins by examining the allelic distributions by race. As Table 1 shows, the 2-repeat allele was carried by 0.1% of Caucasian males and by 5.2% of African-American males. To check the consistency of these estimates, all of the analyses were recalculated using self-reported race instead of interviewer-rated race. Importantly, research has revealed that self-reported race correlates almost perfectly with race identified via genetic markers (Tang et al., 2005). The results were virtually identical when using self-reports of race, where 0.1% of Caucasian males and 5.5% of African-American males were carriers of the 2-repeat allele. These estimates are directly in line with those reported in other samples (e.g., Reti et al., 2011 and Widom and Brzustowicz, 2006). Because only 0.1% of Caucasian males had the 2-repeat allele, the remaining analyses were conducted using only African-American males. Missing data on some of the outcome measures reduced the final analytical sample size to between N = 167 and N = 174 and resulted in one (1) carrier of the 2-repeat allele to be dropped from the analysis. Table 1. Descriptive statistics for selected Add Health study variables and scales. African-Americans (%) Caucasians (%) MAOA 2R 5.2 0.1 3R 52.3 37.0 3.5R 0.5 1.1 4R 42.0 60.0 5R 0.0 1.8 Arrest Yes 52.6 No 47.4 – Incarceration Yes 35.0 – No 65.0 – Composite index Zero 45.1 – One 16.2 – Two 34.7 – Three 4.0 – Mean SD Psychopathic personality traits 53.93 9.62 Note: except for MAOA, the descriptive statistics are not presented for Caucasians because none of the analyses using the antisocial variables/scales include data from Caucasians. Table options The next set of statistical models examined the association between the 2-repeat allele and the psychopathic personality traits variables. Since the psychopathic personality traits scale was coded continuously and approximated normality, ordinary least squares (OLS) regression was estimated. Importantly, all of the analyses in this study were estimated using the “cluster” command in STATA10.0 to correct for the clustering of observations in families (any cases missing a family ID number were dropped from the analyses). The results of this model revealed a non-significant association between the 2-repeat allele and scores on the psychopathic personality traits scale (b = 6.11, SE = 4.33, Beta = 0.15, p = 0.16, R2 = 0.02, N = 170). To check the sensitivity of the results, the analyses were recalculated using negative binomial regression as well as z-transforming the psychopathic personality traits scale. The pattern of results was identical to those reported using OLS regression with the untransformed psychopathic personality traits scale. The next series of statistical models used the ever arrested and ever incarcerated variables as outcome measures. Binary logistic regression models were estimated because both outcome measures were coded dichotomously. The results revealed that carriers of the 2-repeat allele were significantly more likely to be arrested compared to carriers of other alleles (b = 1.35, SE = 0.82, OR = 3.85, p = 0.098, Nagelkerke R2 = 0.03, N = 169). In particular, 2-repeat allele carriers were 3.85 times more likely to be arrested than carriers of other alleles. The predicted probabilities associated with being arrested were 0.80 for 2-repeat allele carriers and 0.51 for carriers of other alleles. In total, 8.6% of the subjects who had been arrested carried the 2-repeat allele compared with only 2.4% of the subjects who had never been arrested. An additional binary logistic regression equation also revealed a statistically significant association between the 2-repeat allele and the odds of being incarcerated (b = 1.11, SE = 0.67, OR = 3.06, p = 0.096, Nagelkerke R2 = 0.02, N = 174). Carriers of the 2-repeat allele were 3.06 times more likely to have been incarcerated than carriers of other alleles. Overall, the predicted probability of being incarcerated was 0.60 for 2-repeat carriers and 0.33 for carriers of other alleles. In total, 9.5% of the subjects who had been incarcerated carried the 2-repeat allele compared with only 3.4% of the subjects who had not been incarcerated. The last main set of analyses examined the association between the 2-repeat allele and the composite antisocial phenotype index. The results of the OLS equation revealed a statistically significant and positive association, wherein carriers of the 2-repeat allele, on average, scored higher on the composite index when compared to non-carriers of this allele (b = 0.88, SE = 0.35, Beta = 0.21, p = 0.014, N = 167). To ensure that this association was not affected by the distribution of the composite index, the model was recalculated using negative binomial regression. The substantive results were identical to those garnered when using OLS regression. 3.1. Supplemental analyses To check the robustness of the results, two lifetime antisocial behavioral outcome measures were employed. First, self-reported violence scales were created for each of the four waves of data collection. These four scales were then z-transformed, summed together, and the resulting summated scale was once again z-transformed. This scale provides an estimate of lifetime involvement in violent antisocial behaviors. An OLS regression equation was then calculated to estimate the association between the 2-repeat allele and scores on the lifetime violence scale, with the results revealing a positive and statistically significant association (b = 1.11, SE = 0.49, Beta = 0.27, p = 0.026, R2 = 0.07). Importantly, the missingness across all four waves of data resulted in losing two (2) cases with the 2-repeat allele so all of these analyses (and the following one) were based on a sample size of N = 130, of which 6.2% had the 2-repeat allele. The second sensitivity analysis was conducted by assigning all respondents who scored 1.5 standard deviations or higher on the lifetime violence scale a value of one (1) and all other scores a value of zero (0). In addition, at wave 4 respondents were asked whether they had ever been convicted or pled guilty to a crime other than a minor traffic violation. This item was then dichotomized (0 = no, 1 = yes) and summed together with the dichotomized lifetime violence scale. Scores on this lifetime antisocial behavior index ranged between zero (0) and two (2) and is similar to indexes that have been used previously (Beaver et al., 2007 and Haberstick et al., 2005). An OLS regression equation revealed a positive and statistically significant association between the 2-repeat allele and this lifetime antisocial behavioral scale (b = 0.45, SE = 0.24, Beta = 0.20, p = 0.062, R2 = 0.04). The pattern of results was confirmed when re-estimating the models using negative binomial regression. In addition, we also explored the possibility that the individual MAOA alleles might have a monotonic relationship with the outcome measures. To do so, all of the alleles were disaggregated into an ordinal variable which measured each respondents’ genotype (i.e., 1 = 4-repeat allele, 2 = 3-repeat allele, and 3 = 2-repeat allele [the 3.5-repeat allele was removed because only one subject possessed this allele]). The effects were largely consistent with those reported when MAOA was dichotomized, except that there was not an association with either of the outcome measures that were estimated using the lifetime violence scales. Follow-up analyses revealed that there were not any significant differences between the 3-repeat allele and the 4-repeat allele on any of the outcome measures, strongly suggesting that the 2-repeat allele was driving the significant associations with the outcome measures