RESEARCH ARTICLE Neuropsychiatric Genetics Genetics of Aggressive Behavior: An Overview Kim Veroude,1 Yanli Zhang-James,2,3 Noelia Fernandez-Castillo,4,5,6 Mireille J. Bakker,1 Bru Cormand,4,5,6 and Stephen V. Faraone2,3,7* 1Department of Cognitive Neuroscience, Donders Institute for Brain, Cognition and Behaviour, Radboudumc, Nijmegen, The Netherlands 2Departments of Psychiatry and of Neuroscience and Physiology, SUNY Upstate Medical University, Syracuse, New York 3Departments of Neuroscience and Physiology, SUNY Upstate Medical University, Syracuse, New York 4Departament de Genetica, Facultat de Biologia, Universitat de Barcelona, Catalonia, Spain 5Institut de Biomedicina de la Universitat de Barcelona (IBUB), Catalonia, Spain 6Centro de Investigacion Biomedica en Red de Enfermedades Raras (CIBERER), Spain 7K.G. Jebsen Centre for Research on Neuropsychiatric Disorders, University of Bergen, Bergen, Norway Manuscript Received: 22 April 2015; Manuscript Accepted: 5 August 2015 The Research Domain Criteria (RDoC) address three types of aggression: frustrative non-reward, defensive aggression and How to Cite this Article: offensive/proactive aggression. This review sought to present Veroude K, Zhang-James Y, Fernandez- the evidence for genetic underpinnings of aggression and to Castillo N, Bakker MJ, Cormand B, determine to what degree prior studies have examined pheno- Faraone SV. 2016. Genetics of Aggressive types that fit into the RDoC framework. Although the constructs Behavior: An Overview. of defensive and offensive aggression have been widely used in the animal genetics literature, the human literature is mostly Am J Med Genet Part B 171B:3–43. agnostic with regard to all the RDoC constructs. We know from twin studies that about half the variance in behavior may be explained by genetic risk factors. This is true for both dimen- seems logical that during this period of time people who had the sional, trait-like, measures of aggression and categorical defi- variants of genes that promoted aggression were more likely to nitions of psychopathology. The non-shared environment seems survive than other people. These variants have persisted in the to have a moderate influence with the effects of shared environ- human genome and partly explain why some people exhibit ment being unclear. Human molecular genetic studies of ag- aggressive behaviors. gression are in an early stage. The most promising candidates are Although the word “irascibilem” comes from the Latin “iras- in the dopaminergic and serotonergic systems along with hor- cibilem”, meaning “to attack,” in current language aggression monal regulators. Genome-wide association studies have not yet means much more. In the genetics literature aggression has been achieved genome-wide significance, but current samples are too small to detect variants having the small effects one would expect for a complex disorder. The strongest molecular evidence for a Kim Veroude, Yanli Zhang-James, Noelia Fernandez-Castillo, and genetic basis for aggression comes from animal models compar- Mireille J. Bakker are joint first authors. Dr. Zhang-James, M.J. Bakker and Dr. K. Veroude state no financial ing aggressive and non-aggressive strains or documenting the interests or potential conflicts of interest. effects of gene knockouts. Although we have learned much from Grant sponsor: NIMH; Grant number: R01MH094469; Grant sponsor: these prior studies, future studies should improve the measure- Spanish “Ministerio de Economı´a y Competitividad”; Grant number: ment of aggression by using a systematic method of measure- SAF2012-33484; Grant sponsor: AGAUR; Grant number: 2014SGR932; ment such as that proposed by the RDoC initiative. Grant sponsor: European Community’s Seventh Framework Programme Ó 2015 Wiley Periodicals, Inc. (FP7/2007-2013); Grant numbers: n˚602805, n˚603016; Grant sponsor: Centro de Investigacion Biomedica en Red de Enfermedades Raras (CIBERER). Key words: aggression; genetics; twin; GWAS; candidate à genes; mutations Correspondence to: Stephen V. Faraone, Ph.D., Departments of Psychiatry and of Neuroscience and Physiology, SUNY Upstate Medical University, 750 E. Adams St., Syracuse, New York 33136. INTRODUCTION E-mail: [email protected] Article first published online in Wiley Online Library During the early stages of human evolution, aggression was proba- (wileyonlinelibrary.com): 8 September 2015 bly an adaptive trait, as it is for many animals in the wild today. It DOI 10.1002/ajmg.b.32364 Ó 2015 Wiley Periodicals, Inc. 3 4 AMERICAN JOURNAL OF MEDICAL GENETICS PART B operationalized in many ways. As a categorical disorder it has been environment, is attributed to influence of the non-shared envi- studied as conduct disorder (CD), oppositional defiant disorder ronment, e2 ¼ 1–rMZ, which also includes measurement error (ODD) and antisocial personality disorder (APD). These catego- [Holzinger, 1929; Falconer, 1960]. It is important to note here ries are convenient for diagnosticians because other work suggests that the non-shared (unique) environment includes all experi- aggression to be a quantitative trait that is better operationalized on ences that contribute to differences between children in the same dimensions of externalizing behavior, rule breaking, psychopathy family, i.e. a common event (for example parents’ divorce) can and violence. affect siblings differently. A dimensional view of aggression is consistent with the ap- Twin studies have investigated aggression from different per- proach taken by the NIMH Research Domain Criteria (RDoC) spectives, e.g. as a personality trait [Miles and Carey, 1997], as Initiative [Sanislow et al., 2010]. RDoC seeks to focus researchers antisocial behavior [Rhee and Waldman, 2002] or as a symptom of on the fundamental mechanisms underlying psychopathology. In childhood and adolescent psychopathology. Previous reviews of doing so, it has been creating a dimensional taxonomy of behavior twin studies and adoption studies on aggression have estimated that, hopefully, corresponds better to underlying mechanisms than heritability up to 0.50, with an additional large role for non-shared does a system of discrete diagnoses. environmental influences and a small influence of the shared In the RDoC nomenclature, aggression is categorized into three environment [Viding et al., 2008; Tuvblad and Baker, 2011]. areas: frustrative non-reward, defensive aggression and offensive Genetic effects seem to predominantly account for phenotypic (or proactive) aggression. Frustrative non-reward refers to behav- correlations between different forms of aggression, such as reactive iors that correspond to the withdrawal or prevention of reward. (defensive) and proactive (offensive) aggression, although few This derives from human and animal studies showing that aggres- studies have examined this [Rhee and Waldman, 2011]. To update sion occurs after repeated, failed attempts to obtain rewards even these prior reviews, we conducted a systematic search for studies after sustained efforts. Defensive aggression refers to behaviors in the period January 2009 until February 2015. PubMed and caused by the perception of an immediate threat, which have PsycINFO were searched for peer-reviewed papers to identify the goal of eliminating the threat. Offensive (or proactive) aggres- studies of twins with characteristics of externalizing behavior sive behaviors are instrumental behaviors aimed at achieving a and psychopathy, regardless of age. We used the following search positive goal, often in the face of competition or in the context of strategy: aggressà OR antisocial behavà OR aggressive traità OR social hierarchies. behavior problemà OR behaviour problemà OR problem behavià The long-term goal of RDoC is to map RDoC phenotypes to OR CD OR conduct disorderà OR conduct problemà OR crime underlying mechanisms. In this review, we sought to present the OR criminalà OR delinquenà OR disruptive behavà OR ODD OR evidence for genetic underpinnings of aggression and to determine oppositional defiant disorderà OR antisocial personality OR to what degree prior studies have examined phenotypes that fit psychopathy OR sociopathy AND heritabilitÃ. neatly, or at all, into the RDoC framework. We focus the review A total of 254 records were retrieved. Neither books nor on three types of genetic studies: twin studies, human association unpublished articles were retrieved from the references. Titles studies of aggression and animal model studies. and abstracts were read by at least two of the authors (MJB and KV); article selection is summarized in Figure 1. Articles were TWIN STUDIES OF AGGRESSION retained if they: 1) included constructs related to aggression, i.e. aggressive traits, externalizing/impulsive-antisocial behavior and This section outlines recent findings from twin studies on aggres- violent criminality/offences/delinquency or diagnostic categories sion and related psychopathology, i.e. ODD, CD and APD. Studies ODD/CD/APD 2) reported univariate heritability estimates 3) had using the classical twin design estimate heritability by comparing been published in peer–reviewed journals from January 2009 the covariation between monozygotic (MZ; identical) and dizy- onwards. Reference lists from the identified articles were manually gotic (DZ; fraternal) twins [Plomin et al., 1994; Boomsma et al., searched for relevant publications. Articles were excluded if they 2002]. MZ twins are assumed to share 100% of their genetic were not written in English, were a case-report, were review