Role of Genotype in the Cycle of Violence in Maltreated Children

See allHide authors and affiliations

Science  02 Aug 2002:
Vol. 297, Issue 5582, pp. 851-854
DOI: 10.1126/science.1072290


We studied a large sample of male children from birth to adulthood to determine why some children who are maltreated grow up to develop antisocial behavior, whereas others do not. A functional polymorphism in the gene encoding the neurotransmitter-metabolizing enzyme monoamine oxidase A (MAOA) was found to moderate the effect of maltreatment. Maltreated children with a genotype conferring high levels of MAOA expression were less likely to develop antisocial problems. These findings may partly explain why not all victims of maltreatment grow up to victimize others, and they provide epidemiological evidence that genotypes can moderate children's sensitivity to environmental insults.

Childhood maltreatment is a universal risk factor for antisocial behavior. Boys who experience abuse—and, more generally, those exposed to erratic, coercive, and punitive parenting—are at risk of developing conduct disorder, antisocial personality symptoms, and of becoming violent offenders (1, 2). The earlier children experience maltreatment, the more likely they are to develop these problems (3). But there are large differences between children in their response to maltreatment. Although maltreatment increases the risk of later criminality by about 50%, most maltreated children do not become delinquents or adult criminals (4). The reason for this variability in response is largely unknown, but it may be that vulnerability to adversities is conditional, depending on genetic susceptibility factors (5, 6). In this study, individual differences at a functional polymorphism in the promoter of the monoamine oxidase A (MAOA) gene were used to characterize genetic susceptibility to maltreatment and to test whether the MAOA gene modifies the influence of maltreatment on children's development of antisocial behavior.

The MAOA gene is located on the X chromosome (Xp11.23–11.4) (7). It encodes the MAOA enzyme, which metabolizes neurotransmitters such as norepinephrine (NE), serotonin (5-HT), and dopamine (DA), rendering them inactive (8). Genetic deficiencies in MAOA activity have been linked with aggression in mice and humans (9). Increased aggression and increased levels of brain NE, 5-HT, and DA were observed in a transgenic mouse line in which the gene encoding MAOA was deleted (10), and aggression was normalized by restoringMAOA expression (11). In humans, a null allele at the MAOA locus was linked with male antisocial behavior in a Dutch kindred (12). Because MAOA is an X-linked gene, affected males with a single copy produced no MAOAenzyme—effectively, a human knockout. However, this mutation is extremely rare. Evidence for an association between MAOA and aggressive behavior in the human general population remains inconclusive (13–16).

Circumstantial evidence suggests the hypothesis that childhood maltreatment predisposes most strongly to adult violence among children whose MAOA is insufficient to constrain maltreatment-induced changes to neurotransmitter systems. Animal studies document that maltreatment stress (e.g., maternal deprivation, peer rearing) in early life alters NE, 5-HT, and DA neurotransmitter systems in ways that can persist into adulthood and can influence aggressive behaviors (17–21). In humans, altered NE and 5-HT activity is linked to aggressive behavior (22). Maltreatment has lasting neurochemical correlates in human children (23,24), and although no study has ascertained whetherMAOA plays a role, it exerts an effect on all aforementioned neurotransmitter systems. Deficient MAOA activity may dispose the organism toward neural hyperreactivity to threat (25). As evidence, phenelzine injections, which inhibit the action of monoamine oxidase, prevented rats from habituating to chronic stress (26). Low MAOA activity may be particularly problematic early in life, because there is insufficientMAOB (a homolog of MAOA with broad specificity to neurotransmitter amines) to compensate for an MAOAdeficiency (8).

Based on the hypothesis that MAOA genotype can moderate the influence of childhood maltreatment on neural systems implicated in antisocial behavior, we tested whether antisocial behavior would be predicted by an interaction between a gene (MAOA) and an environment (maltreatment). A well-characterized variable number tandem repeat (VNTR) polymorphism exists at the promoter of the MAOA gene, which is known to affect expression. We genotyped this polymorphism in members of the Dunedin Multidisciplinary Health and Development Study, a sample without population stratification confounds (27). This birth cohort of 1,037 children (52% male) has been assessed at ages 3, 5, 7, 9, 11, 13, 15, 18, and 21 and was virtually intact (96%) at age 26 years.

The study offers three advantages for testing gene-environment (G × E) interactions. First, in contrast to studies of adjudicated or clinical samples, this study of a representative general population sample avoids potential distortions in association between variables (28,29). Second, the sample has well-characterized environmental adversity histories. Between the ages of 3 and 11 years, 8% of the study children experienced “severe” maltreatment, 28% experienced “probable” maltreatment, and 64% experienced no maltreatment (27). (Maltreatment groups did not differ onMAOA activity, χ2(2) = 0.38,P = 0.82, suggesting that genotype did not influence exposure to maltreatment.) Third, the study has ascertained antisocial outcomes rigorously. Antisocial behavior is a complicated phenotype, and each method and data source used to measure it (e.g., clinical diagnoses, personality checklists, official conviction records) is characterized by different strengths and limitations. Using information from independent sources appropriate to different stages of development, we examined four outcome measures (27). Adolescent conduct disorder was assessed according to criteria of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV); convictions for violent crimes were identified via the Australian and New Zealand police; a personality disposition toward violence was measured as part of a psychological assessment at age 26; symptoms of antisocial personality disorder were ascertained at age 26 by collecting information about the study members from people they nominated as “someone who knows you well.” A common-factor model fit the four measures of antisocial behavior well (27), with factor loadings ranging from 0.64 to 0.74, showing that all four measures index liability to antisocial behavior.

Using moderated regression analysis, we predicted scores on a composite antisocial index comprising the four measures of antisocial behavior (27) (Fig. 1). The main effect of MAOA activity on the composite index of antisocial behavior was not significant (b = 0.01, SE = 0.09, t = 0.13, P = 0.89), whereas the main effect of maltreatment was significant (b = 0.35, SE = 0.07, t = 4.82,P < 0.001). A test of the interaction betweenMAOA activity and maltreatment revealed a significantG × E interaction (b = –0.36, SE = 0.14, t = 2.53, P = 0.01). This interaction within each genotype group showed that the effect of childhood maltreatment on antisocial behavior was significantly weaker among males with high MAOA activity (b = 0.24, SE = 0.11, t = 2.15,P = 0.03) than among males with low MAOAactivity (b = 0.68, SE = 0.12, t = 5.54, P < 0.001).

Figure 1

Means on the composite index of antisocial behavior as a function of MAOA activity and a childhood history of maltreatment (27). MAOA activity is the gene expression level associated with allelic variants of the functional promoter polymorphism, grouped into low and high activity; childhood maltreatment is grouped into 3 categories of increasing severity. The antisocial behavior composite is standardized (z score) to a M = 0 and SD = 1; group differences are interpretable in SD unit differences (d).

We conducted further analyses to test if the G ×E interaction was robust across each of the four measures of antisocial behavior that made up the composite index. For all four antisocial outcomes, the pattern of findings was consistent with the hypothesis that the association between maltreatment and antisocial behavior is conditional, depending on the child's MAOAgenotype (G × E interactionP = 0.06, 0.05, 0.10, and 0.04, respectively). For adolescent conduct disorder (Fig. 2A), maltreated males (including probable and severe cases) with the low–MAOA activity genotype were more likely than nonmaltreated males with this genotype to develop conduct disorder by a significant odds ratio (OR) of 2.8 [95% confidence interval (CI): 1.42 to 5.74]. In contrast, among males with high MAOAactivity, maltreatment did not confer significant risk for conduct disorder (OR = 1.54, 95% CI: 0.89 to 2.68). For adult violent conviction (Fig. 2B), maltreated males with the low–MAOAactivity genotype were more likely than nonmaltreated males with this genotype to be convicted of a violent crime by a significant odds ratio of 9.8 (95% CI: 3.10 to 31.15). In contrast, among males with highMAOA activity, maltreatment did not confer significant risk for violent conviction (OR = 1.63, 95% CI = 0.72 to 3.68). For self-reported disposition toward violence (Fig. 2C) and informant-reports of antisocial personality disorder symptoms (Fig. 2D), males with the low–MAOA activity genotype who were maltreated in childhood had significantly elevated antisocial scores relative to their low-MAOA counterparts who were not maltreated. In contrast, males with high MAOA activity did not have elevated antisocial scores, even when they had experienced childhood maltreatment.

Figure 2

The association between childhood maltreatment and subsequent antisocial behavior as a function of MAOAactivity. (A) Percentage of males (and standard errors) meeting diagnostic criteria for Conduct Disorder between ages 10 and 18. In a hierarchical logistic regression model, the interaction between maltreatment and MAOA activity was in the predicted direction, b = – 0.63, SE = 0.33,z = 1.87, P = 0.06. Probing the interaction within each genotype group showed that the effect of maltreatment was highly significant in the low–MAOAactivity group (b = 0.96, SE = 0.27,z = 3.55, P < 0.001), and marginally significant in the high-MAOA group (b = 0.34, SE = 0.20, z = 1.72, P = 0.09). (B) Percentage of males convicted of a violent crime by age 26. The G × E interaction was in the predicted direction, b = – 0.83, SE = 0.42,z = 1.95, P = 0.05. Probing the interaction, the effect of maltreatment was significant in the low–MAOA activity group (b = 1.20, SE = 0.33, z = 3.65, P < 0.001), but was not significant in the high MAOA group (b = 0.37, SE = 0.27, z = 1.38, P = 0.17). (C) Mean z scores (M = 0, SD = 1) on the Disposition Toward Violence Scale at age 26. In a hierarchical ordinary least squares (OLS) regression model, theG × E interaction was in the predicted direction (b = – 0.24, SE = 0.15,t = 1.62, P = 0.10); the effect of maltreatment was significant in the low–MAOA activity group (b = 0.35, SE = 0.11,t = 3.09, P = 0.002) but not in the high MAOA group (b = 0.12, SE = 0.07, t = 1.34, P = 0.17). (D) Mean z scores (M = 0, SD = 1) on the Antisocial Personality Disorder symptom scale at age 26. TheG × E interaction was in the predicted direction (b = – 0.31, SE = 0.15,t = 2.02, P = 0.04); the effect of maltreatment was significant in the low–MAOA activity group (b = 0.45, SE = 0.12,t = 3.83, P < 0.001) but not in the high MAOA group (b = 0.14, SE = 0.09, t = 1.57, P = 0.12).

These findings provide initial evidence that a functional polymorphism in the MAOA gene moderates the impact of early childhood maltreatment on the development of antisocial behavior in males. Replications of this G × Einteraction are now needed. Replication studies should use valid and reliable ascertainments of maltreatment history and should obtain multiple measures of antisocial outcomes, in large samples of males and females (30). If replicated, the findings have implications for research and clinical practice. With regard to research in psychiatric genetics, knowledge about environmental context might help gene-hunters refine their phenotypes. Genetic effects in the population may be diluted across all individuals in a given sample, if the effect is apparent only among individuals exposed to specific environmental risks. With regard to research on child health, knowledge about specific genetic risks may help to clarify risk processes. Numerous biological and psychological processes have been put forward to explain why and how experiences of maltreatment are converted into antisocial behavior toward others (17, 24,31–34), but there is no conclusive evidence that any of these processes can account for the progression from childhood maltreatment to later criminal violence. Moreover, some youngsters make the progression, but others do not, and researchers have sought to understand why (35). The search has focused on social experiences that may protect some children, overlooking a potential protective role of genes. Genes are assumed to create vulnerability to disease, but from an evolutionary perspective they are equally likely to protect against environmental insult (36). Maltreatment studies may benefit from ascertaining genotypes associated with sensitivity to stress, and the known functional properties of MAOA may point toward hypotheses, based on neurotransmitter system development, about how stressful experiences are converted into antisocial behavior toward others in some, but not all, victims of maltreatment.

Until this study's findings are replicated, speculation about clinical implications is premature. Nonetheless, although individuals having the combination of low-activity MAOA genotype and maltreatment were only 12% of the male birth cohort, they accounted for 44% of the cohort's violent convictions, yielding an attributable risk fraction (11%) comparable to that of the major risk factors associated with cardiovascular disease (37). Moreover, 85% of cohort males having a low-activity MAOA genotype who were severely maltreated developed some form of antisocial behavior. Both attributable risk and predictive sensitivity indicate that these findings could inform the development of future pharmacological treatments.

Supporting Online Material

Materials and Methods

Tables S1 and S2

  • * To whom correspondence should be addressed. E-mail: t.moffitt{at}


View Abstract

Navigate This Article