University of Groningen
Gene-environment interactions in disruptive behaviors
Ruisch, Hyun
DOI:
10.33612/diss.136546089
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Ruisch, H. (2020). Gene-environment interactions in disruptive behaviors. University of Groningen. https://doi.org/10.33612/diss.136546089
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Chapter
1
General introduction and thesis outline
The development and functioning of our central nervous system shape our mental abilities and hence our cognitive skills, emotional responses, and behavior. Neurodevelopment is a complex process that involves both genetic and environmental factors and that can be disrupted in several ways. Even before birth, certain adversities may already affect brain development in utero. Moreover, the way we respond to the environments and events that we are exposed to throughout life may be moderated by our genetic make-up (i.e. GxE interaction). Therefore, inter-individual variability in behavior is thought of as resulting from variation in both genetic and environmental factors, and their interaction.
In this thesis, I present a total of six studies that describe my research on the interaction between genes and environment in relation to disruptive behavior problems in children and adolescents. In addition to GxE interactions, I focus on main effects of environmental factors and the genetics of disruptive behavior separately. Further, the relation between genetic factors, GxE interactions, and structural connectivity of the brain is addressed. In this first chapter, I will briefly summarize background information on disruptive behavior disorders and related phenotypes, findings from genetic, environmental, GxE, and diffusion tensor imaging studies of disruptive behavior, and explain our analytical strategies to investigate GxE interactions. Subsequently, I will provide an overview of the aims of this thesis and an outline for each chapter.
Disruptive behavior disorders
Oppositional-defiant disorder (ODD) and conduct disorder (CD) are part of the ‘Disruptive, Impulse-Control, and Conduct Disorders’ group according to the latest revision of the Diagnostic and Statistical Manual of Mental Disorders 5 (DSM 5), with an onset during childhood or adolescence (1). Prevalences of ODD and CD are estimated to range between approximately 2 and 16% in children and adolescents, depending on demographic factors such as age, sex, and studied population (1–3). Disruptive behavior symptoms mainly include angry, defiant, and vindictive behaviors in the case of ODD, whereas CD is characterized by aggressive, deceitful, and serious rule-breaking behaviors. Children diagnosed with ODD and CD are at risk of poor academic outcome, social problems, and delinquency in later life, which also causes considerable burden to affected families and society at large (4,5). A subgroup of children with CD (and to some extent also of children with other disruptive behavior disorders, such as Intermittent Explosive Disorder) display high levels of callous-unemotional (CU) traits (as reflected in the specifier for CU traits in the DSM 5 [‘displaying limited prosocial emotions’], which distinguishes those with a callous and unemotional interpersonal style across multiple settings and relationships (1)). CU traits have been linked with more persistent and severe symptomatology and entail a lack of remorse/guilt, limited empathy, and a shallow affect (1,6). Furthermore, frequent comorbidity of ODD and CD with attention-deficit/hyperactivity disorder (ADHD) is observed (up to 50% of children with ADHD also meet criteria for ODD and/or CD (7,8)). From a developmental viewpoint, ODD symptoms may sometimes progress into more severe CD symptomatology (3). However, although most children with CD also meet diagnostic requirements for ODD, the opposite is not true (6).
Aggression and disruptive behaviors
Aggression can be defined as any behavior aimed at causing harm to others. Although throughout evolution, aggression can be considered as an adaptive behavior that is essential for survival (9), aggression among humans causes considerable societal problems, including consequences of aggression-related behavioral disorders such as ODD and CD (4,5,9). The taxonomy of aggression is complex, with different perspectives to subtyping (e.g. reactive versus proactive, or physical versus relational aggression). Reactive aggressive behaviors result from a reaction to a perceived threat or frustration and are frequently observed in both ODD and CD. Proactive or instrumental aggression is driven by reward anticipation, which makes it conceptually more closely linked with CD symptomatology and CU traits (10). Despite being considered as different subtypes of aggression, reactive and proactive aggression often co-occur within affected individuals to various degrees (11). As such, aggression represents a range of behaviors that can be observed across healthy individuals as well as multiple psychiatric disorders, with disruptive behavior disorders representing the prototypical conditions characterized by significant aggressive behaviors. Throughout this thesis, I investigate a number of related phenotypes that reflect aggression and disruptive behavior, namely CD symptomatology (i.e. antisocial behavior), CU traits, ODD symptomatology and aggressive behavior as a more general trait (6).
Continuous (quantitative) traits in the population
Although psychiatric disorders were historically conceptualized as categorical disease entities, according to more current insights, they can also be conceived as the more extreme outcomes along continuously distributed (i.e. quantitative) traits in the population (12). Genetic evidence for this has recently been shown for major psychiatric disorders and traits, including neurodevelopmental disorders such as ADHD, autism spectrum disorders and tic disorders, and mood, anxiety and psychotic disorders (13–15). In addition, the high rates of comorbidity among psychiatric disorders in general points to a shared etiology. In this respect, genetic overlap has indeed been shown across several psychiatric traits and disorders (12,16,17). In case of disruptive behaviors, for children it is not necessarily abnormal to display some degree of disruptive behaviors such as an occasional temper tantrum or fight. However, when such behaviors are becoming more frequent and severe, problems may arise and psychiatric referral and evaluation may be considered. In this respect, the DSM emphasizes the clinical relevance of symptomatology – i.e. symptoms must result in a significant functional (e.g. academic or social) impairment and/or burden/distress for the affected person – to discriminate between pathological and physiological - i.e. ‘normal’ - ranges of behavior (1). Throughout this thesis, the phenotypes related to disruptive behavior and aggression will therefore be conceptualized as continuous traits (e.g. dimensional measures of ODD and CD symptomatology) rather than as categorical disorders in a case versus control design.
Genetic studies of disruptive behavior
Alterations of monoaminergic neurotransmission have repeatedly been associated with disruptive behavior in both human and animal studies (e.g. (18,19)) and initially, genetic association studies of disruptive behavior focused on polymorphisms in monoaminergic candidate genes such as MAOA, SLC6A4, COMT and DRD4 (20). However, findings for these candidate genes have been inconsistent, which may point to important limitations in the earlier literature, such as a priori weak candidate genes, underpowered samples, and inappropriate correction for multiple testing (21,22). Although rare monogenic disorders involving disruptive behavior (aggression) are known (e.g. Brunner syndrome (23,24)), disruptive behavior traits in the general population involve genetic variation across a multitude of genes (20,25). Therefore, approaches that consider genetic variation across genes or pathways (i.e. biologically related sets of genes) or investigate genetic variants across the genome (i.e. genome-wide association studies (GWASs)) provide more suitable and sophisticated ways for investigating complex human traits such as disruptive behaviors (20,22). Box 1 provides an overview of different levels of complexity in the analysis of genetic variation. In this respect, gene-sets related to monoaminergic and neuroendocrine signaling have been linked to aggression (26) and more recently, some other susceptibility genes (e.g. RBFOX1 (27) or MECOM (28)) have emerged. Although disruptive behavior GWAS sample sizes have already increased considerably, only few susceptibility loci have been discovered that survive correction for multiple testing at the genome-wide significance level (see also Box 1) (29–32). In addition to GWASs and gene-level analyses, the effect of multiple variants may also be aggregated into a polygenic risk score (PRS; see also Box 1). In this case, the summary statistics of a GWAS are used to calculate a PRS for the GWAS-phenotype for each individual of an independent sample. As such, the predictive ability of the PRS is largely dependent on the sample size of the discovery GWAS. Using PRS, the polygenicity of a trait (e.g. (14)) or evidence for a shared genetic etiology among different traits and/or disorders (e.g. (17)) can be investigated. Calculation of the scores can also be restricted to biological pathways of interest (e.g. glutamatergic signaling (33)). Currently, there has only been one study applying PRS in relation to disruptive behavior (29). Therefore, in Chapters 6 and 7, I have used PRS to investigate evidence of a shared genetic background among disruptive behavior subtypes and neural phenotypes.
Family-based studies have estimated the heritability of disruptive behavior at approximately 50% (20,25) but GWASs have provided estimates of single nucleotide polymorphism (SNP; see also Box 1) based heritability between 5 and 50% (29,32). The ‘missing’ heritability compared to family-based studies may be due to small genetic effects that are still undetected with the current GWAS sample sizes and/or could be related to factors that may not be (sufficiently) captured by available genotyping arrays such as effects of rare alleles, structural variation (e.g. copy number variants), epistasis or gene-environment interactions (34).
Box 1: Different levels of complexity in the analysis of genetic variation
Genetic variation can be considered at multiple levels of complexity. At the variant-level, the effects of individual genetic variants, such as single nucleotide polymorphisms (SNPs) or length variations, are considered. A SNP refers to the substitution of a single nucleotide at a certain position in the genome that is observed in at least 1% of the population (e.g. in 30% of the population, a thymine nucleotide is observed at a particular SNP-position whereas in 70% a cytosine nucleotide appears). In addition to single variants, the effect of multiple variants may be combined across a gene (to obtain gene-level statistics) or sets of genes (e.g. to investigate the effect of aggregated genetic variation across a biological pathway). While these approaches can be applied in a hypothesis-driven design, genetic variation can also be investigated
hypothesis-free, across the whole genome. In the case of complex genetic traits such as human behavioral
characteristics, it is thought that many common variants with small individual effects contribute to the phenotypic variation observed in the general population. To detect such small effects, a genome-wide
association study (GWAS) investigates all SNPs across the genome in relation to a certain disease or trait.
Using commercially available genotyping chips (and subsequent imputation and necessary quality control procedures), several millions of SNPs are incorporated into a GWAS. Because of ‘hypothesis-free’ testing all SNPs across the genome, the threshold for statistical significance must be adjusted for multiple testing. The most commonly accepted threshold for genome-wide statistical significance is therefore 𝑃𝑃 < 5.00E-08 (i.e. 5 ⋅ 10−8). GWAS results can be visualized through a ‘Manhattan’ plot that shows the negative logarithm of the
P-value for all SNPs against their chromosomal position across the genome, similar to the example below. Example of GWAS Manhattan plot by Demontis et al. (2019) (35). In
this example, SNPs surpassing the threshold for genome-wide significance (i.e. the dashed line) are shown as diamonds. Note that at each ‘peak,’ many more SNPs (i.e. dots) show up, but given the statistical correlation between the P-values of nearby SNPs (due to linkage disequilibrium), only one independent index-SNP is considered at each of the 12 genome-wide significant loci.
In addition to identifying susceptibility loci as such, results of a GWAS can be used for calculating polygenic risk scores (PRS). PRS are calculated by aggregating the GWAS effect sizes across multiple SNPs into a single cumulative genetic risk score for each individual in an independent ‘target’ dataset. The PRS can then be used to predict a (similar or different) phenotype in the target dataset. This way, evidence for shared genetic etiology between two traits or disorders (i.e. between the phenotype of the GWAS and the phenotype predicted in the target dataset) can be investigated. For example, if a PRS based on a GWAS for aggression would predict ADHD, this would point to a shared genetic etiology between aggression and ADHD. Because the PRS aggregates multiple small effects of individual SNPs into a single score, less power is required for the target sample compared to the base GWAS sample. Of note, PRS do not typically include all available SNPs in the score calculation. Commonly, a GWAS ‘P-value threshold’ for inclusion of SNPs into the score is used, ranging from genome-wide or ‘suggestive’ significance levels to more liberal thresholds such as 0.05 and higher.
Environmental risk factors related to disruptive behavior
In addition to genetics, environmental factors contribute to disruptive behavior. Early adversities related to the prenatal environment may already negatively affect fetal development and as such be involved as risk factors for disruptive behavior (36). According to a systematic review of prenatal and early postnatal environmental risk factors in relation to a range of disruptive behavior disorders, a multitude of different factors were implicated, such as maternal smoking and alcohol use during pregnancy, maternal stress and anxiety during pregnancy, perinatal complications, and parenting factors (37). In addition, commonly prescribed medication such as paracetamol (acetaminophen) or antidepressants during pregnancy may adversely affect neurodevelopment and predispose offspring to behavior problems (38–40). Nevertheless, although in most studies, the effects of individual risk factors are considered, it is also important to recognize that a substantial part of these environmental factors are to some degree related to each other (41,42). Therefore, in Chapter 3 of this thesis, I aimed to investigate a range of environmental factors during pregnancy and the first years of life in relation to disruptive behavior, while also taking into account their relatedness.
Several studies have pointed to smoking during pregnancy as a risk factor for behavioral problems (e.g. (43,44)) since tobacco smoke contains notoriously toxic substances (e.g. (45– 47)) and smoking during pregnancy has been linked with a range of somatic complications (e.g. (47–49)). However, it has also been shown that the association of smoking during pregnancy with offspring disruptive behavior to at least some extent represents a gene-environment correlation (rGE – see also Box 2) (50,51). More specifically, this implies that exposure to maternal smoking during pregnancy is not randomly distributed but is dependent on the genetic make-up of the mother. This same genetic liability is then passed on to the child, leading to disruptive behavior (50,51). Similarly, rGE's have been reported in the relation between prenatal environmental factors and offspring ADHD (50,52,53). Considering these issues regarding prenatal risk factors for disruptive behavior, I conducted a meta-analysis of the existing literature on substance use during pregnancy (Chapter 2 of this thesis). In addition to quantitatively assessing evidence, I aimed to point out the most important and current issues in the literature, and address some of these in the subsequent chapters about environmental factors (Chapter 3) and gene-environment interplay (Chapters 4 to 7).
In addition to environmental adversities early in life, factors in childhood and adolescence such as abuse, a harsh parenting style, or stressful life events more generally, have been implicated as risk factors of disruptive behavior (6,54,55). Similar to maternal stressors during pregnancy (36,56–58), the effect of such experiences during childhood and adolescence in relation to disruptive behavior may be related to variability in an individual’s stress (threat) response (6,59). In this respect, the relation between environmental factors and disruptive behavior may not only be confounded through rGE's, but may also be moderated by genetic variation that affects our response or vulnerability to the environment, i.e. gene-environment interactions (e.g. (54); see also Box 2).
Gene-by-environment (GxE) interactions in disruptive behavior
Genetics and environment do not only contribute independently to liability for disruptive behavior but, as mentioned, also interact with each other through gene-environment (GxE) interactions. In GxE interactions, the effect of certain environments is moderated by the genetic make-up of individuals, thereby explaining inter-individual differences in susceptibility to environmental risk factors that may contribute to heterogeneity in disruptive behavior (see also Box 2). This concept can also be considered the other way around, such that the effect of certain genetic variation may only emerge in the context of certain environments. A classic example of GxE interaction is the metabolic disorder phenylketonuria, which is caused by mutations in the phenylalanine hydroxylase (PAH) gene and results in deficiency of the PAH-enzyme and hence reduced metabolism of the amino acid phenylalanine. Toxic build-up of phenylalanine results in severe brain damage causing intellectual disability, seizures, behavioral abnormalities, and other neurological and psychiatric problems. However, if dietary intake of phenylalanine is restricted, brain damage may be prevented and development can be normal (60). Thus, in this case, the effect of the genetic mutation only manifests itself in the context of a certain environmental exposure.
Most GxE research on disruptive behavior to date has focused on candidate gene-based GxE interactions (20,61). Although some GxE’s have been reported across multiple studies - for example MAOA x maltreatment (54) and SLC6A4 x negative life events (62) - methodological constraints (e.g. ethnical heterogeneity, sample size) limit the interpretability of these results, similar to the interpretation problems for candidate gene studies of genetic main effects (21,22,63). Therefore, alternative approaches that may overcome limitations of candidate gene-based GxE’s have been suggested, such as the use of susceptibility loci discovered through GWAS (64), polygenic risk scoring (61,63), or hypothesis-free GxE discovery studies (e.g. (65,66)). Such GxE studies are, however, still scarce in the case of disruptive behavior (61).
Given the complex interplay between genetic and environmental factors – including GxE’s and rGE’s as described above (see also Box 2) – that is thought to have an important role in disruptive behavior, this topic was a main focus of the research presented in this thesis. In Chapters 4 to 7, I investigated GxE’s, taking into account rGE’s, whereas in Chapter 3, I studied environmental main effects, while also adjusting for rGE.
Box 2: Gene-environment (GxE) interactions and gene-environment correlations (rGE’s)
Gene-environment (GxE) interactions stratify effects of environmental risk factors by genotype, or vice
versa. More specifically, this means that an individual’s genetic make-up affects his or her response to the environment, which may provide insight into why some individuals appear to be more susceptible to certain adversities than others, and therefore may also help explain the observed heterogeneity for the phenotype-of-interest. The figure below shows a GxE interaction in which the gene is shown as a moderator that affects the relation between the environment and disruptive behavior.
Gene-environment correlations (rGE’s) point to non-independency (i.e. some degree of correlation)
among genotype and environment. More specifically, this means that an individual’s exposure to environmental factors does not occur at random, but is dependent on the genetic make-up of that individual. Such rGE’s may therefore also confound observed associations between environmental factors and the phenotype-of-interest. The figure below shows an rGE (bold arrow) in which the gene is shown as a
confounder that affects the relation between the environment and disruptive behavior.
Diffusion tensor imaging (DTI) studies investigating brain structural connectivity in disruptive behavior
Behavioral phenotypes such as disruptive behavior, aggression and ADHD, have been linked to variation in brain function and structure (e.g. (67–70)). Given that brain function is dependent on brain structure and connectivity, studies have started to investigate brain structural connectivity – i.e. measures reflecting microstructural characteristics of the white matter nerve tracts that connect different regions of the brain – in relation to disruptive behavior phenotypes. The microstructure of nerve tracts in the brain can be investigated by using diffusion tensor imaging (DTI; see also Box 3) (71). DTI measures the diffusion of
Environment Disruptive behavior
Gene
Environment Disruptive behavior
water molecules in multiple directions across the brain and fits a diffusion tensor to each voxel of the brain imaging volume. Subsequently, different metrics (scalar quantities) can be derived from the tensor that mathematically describe the ‘shape’ of the tensor, which reflects the underlying tissue microstructure (see Box 3 for an overview of DTI-metrics). Commonly investigated DTI-metrics include fractional anisotropy (FA) and mean diffusivity (MD). Although MD is defined as the arithmetic average of diffusion across different directions, diffusivity perpendicular to the tract (i.e. radial diffusivity) may represent a more sensitive measure of the tract’s (myelin) integrity (72). Another metric that, in contrast to FA and mean/radial diffusivity, is also sensitive to white matter regions with crossing fibers is the mode of anisotropy (73,74).
Globally, there are three types of cerebral nerve fibers (myelinated axons) that constitute the white matter tracts that connect different parts of the brain. First, there are commissural fibers that connect brain regions between the left and right hemispheres (e.g. the corpus callosum). Commissural connections can be homotopic (i.e. connecting the same regions) or heterotopic (i.e. connecting different regions). Second, association fibers connect regions within the same hemisphere and can be generally divided into short (i.e. connecting adjacent cortical areas) and long range (i.e. connecting different lobes). Third, projection fibers connect cortical brain areas with lower parts of the brain and spinal cord and can be efferent (i.e. carrying information from the cortex down to other neurons such as the corticospinal tract) or afferent (i.e. carrying information towards the cortex such as the optic fibers in the visual system) (71). As for (antisocial) disruptive behavior, a multitude of projection, association, and commissural white-matter tracts have been implicated and both associations with higher and lower FA and MD have been reported (70). Initially, the uncinate fasciculus was theorized as a promising candidate tract, as it connects areas important for emotion processing, such as the amygdala, with the frontal regions. However, the implication of various white matter tracts suggest that multiple neural circuits connected by different white matter tracts, rather than only a single circuit, play a role in disruptive behavior. This points to the importance of using both hypothesis-free (i.e. whole-brain voxel wise studies) and hypothesis-driven approaches (70). Further, the heritability of the DTI-metrics FA and RD has been shown to be high (75,76), and some susceptibility loci have been reported in a GWAS of whole-brain average FA (77). As such, genetic variation may contribute to variability in brain structural connectivity, which in turn may underlie vulnerability to disruptive behavior. Because findings from existing DTI-studies in relation to disruptive behavior have been inconsistent and no studies have investigated genetic liability for disruptive behavior in relation to structural connectivity yet, I aimed to study the relation between PRS for disruptive behavior and white matter microstructure across the brain (Chapter 7).
Box 3: Diffusion tensor imaging (DTI) and brain structural connectivity Diffusion tensor imaging (DTI) can be used to investigate structural connectivity in the brain. DTI
characterizes the diffusion of water molecules to obtain information about the local tissue microstructure (71). Because of its oriented structure (i.e. across a tract, all nerve fibers are oriented along the primary tract direction), white matter yields a characteristic anisotropic diffusion profile. This means that along the direction of a nerve tract, diffusion occurs with less restriction than perpendicular to it (because of the axolemma and myelin), resulting in a different magnitude of diffusivity across different directions (see the below Figures B and D). In contrast, when more or less random barriers exist, the diffusivity will be approximately equal in all directions, which is more similar to the microstructure in grey matter areas (Figures A and C).
Diffusion tensor shape and tissue microstructure. In A and C, the diffusion is approximately equal in all directions due to
random barriers (e.g. in grey matter). The resulting symmetrical (spherical) shape of the diffusion tensor is considered isotropic. On the other hand, in B and D, the diffusion is relatively free of barriers along the direction of the nerve fiber tract although, in the directions perpendicular to the tract, the axolemmas and myelin restrict the diffusion. The resulting asymmetrical (ellipsoidal) shape of the tensor is considered anisotropic. A diffusion path example is shown as a dashed red line in C and D.
A commonly used DTI-metric to quantify the degree of anisotropy is fractional anisotropy (FA). FA is defined as a unitless coefficient on the theoretical interval between 0 to 1, where 0 indicates complete isotropic and 1 indicates complete anisotropic diffusion. Diffusivity is defined as a scalar factor between the flux due to molecular diffusion and the concentration gradient that drives the diffusion (according to Fick’s laws). Mean diffusivity is defined as the arithmetic average of diffusivity along the primary direction (i.e. 𝜆𝜆1; by definition 𝜆𝜆1≥ 𝜆𝜆2≥ 𝜆𝜆3) and perpendicular directions (𝜆𝜆2 and 𝜆𝜆3). Axial diffusivity is equal to 𝜆𝜆1, whereas
radial diffusivity (RD) is equal to 𝜆𝜆2+𝜆𝜆3
2 . RD is therefore more sensitive to changes in tissue integrity, as decreased axon/myelin integrity would result in increased diffusion perpendicular to the tract (72). Mode of
anisotropy (MO) distinguishes between linear and planar diffusion, and therefore is sensitive to brain areas
with crossing fibers, which complicates the interpretation of FA and RD (e.g. when there are two dominating directions of crossing fibers, FA will be reduced and RD will be increased, but this is not caused by abnormalities in white matter microstructural integrity) (73,74).
A B
Data used in this thesis
Throughout this thesis, I used data from two different cohort studies, i.e., from the ‘Avon Longitudinal Study of Parents And Children’ (ALSPAC) and ‘NeuroIMAGE’ samples (see Box 4).
Box 4: ALSPAC and NeuroIMAGE
Part of this thesis (Chapters 2 to 5) uses data from the Avon Longitudinal Study of Parents And Children (ALSPAC). ALSPAC is a prospective, longitudinal birth cohort study, which recruited 14,541 pregnant women in Avon (UK) with expected delivery dates from April 1991 to December 1992 and their subsequently born offspring. The mothers were between age 16 and 45 during recruitment and represented approximately 85% of all pregnant women in the catchment area. When the children reached the age of 7, the initially recruited sample was enriched with eligible subjects who initially failed to join the study, resulting in an additional enrollment of 713 children. ALSPAC thus comprises a general population cohort with longitudinally collected data regarding a range of phenotypic (such as different internalizing and externalizing disorders during childhood and adolescence) and environmental measures (such as perinatal factors and childhood problems). In addition, biological samples, epigenetic data, and individual genome-wide genotyping data are available. Further details are reported elsewhere (78–80).
Two research chapters of this thesis (Chapters 6 and 7) use data from the NeuroIMAGE study. NeuroIMAGE is the Dutch follow-up study of the International Multicenter ADHD Genetics (IMAGE) case-control study and includes 331 families with at least one child with ADHD and at least one biological sibling and 153 control families, resulting in a total of 412 ADHD-cases, 227 unaffected siblings, 262 healthy controls, and 81 children with ‘subthreshold’ levels of ADHD-symptoms (i.e. more symptoms than allowed for healthy controls and less symptoms than required for cases). The ADHD diagnosis was ascertained according to DSM-IV-TR criteria using information from a semi-structured diagnostic interview and rating scales. Inclusion criteria were European Caucasian descent, IQ ≥ 70, and no diagnosis of autism, epilepsy, learning disorders, neurological diseases, or genetic syndromes. In addition to extensive phenotypic (including different measures of ADHD symptomatology and disruptive behavior comorbidity) and environmental data (including pregnancy-related data and childhood data), NeuroIMAGE contains whole-brain functional and structural (including diffusion-weighted) magnetic resonance imaging (MRI) data and individual genome-wide genotyping data. Further details are reported elsewhere (81).
Aims and outline of this thesis
The aims of this thesis were to investigate environmental factors, genetic factors, neural correlates and in particular GxE interactions in relation to disruptive behavior. Therefore, I investigated a range of pre-, peri- and postnatal environmental factors, different types of genetic variation, white-matter microstructure across the brain, and a number of continuous behavioral phenotypes capturing slightly different aspects of disruptive behavior. Through the use of different and complementary analytical strategies, I aimed to study GxE’s in disruptive behavior in detail and contribute novel and relevant information to the existing literature.
In Chapter 2, I performed a meta-analysis of maternal substance use during pregnancy in relation to offspring conduct problems. Specifically, I systematically and quantitatively synthesized the existing evidence for maternal smoking, alcohol use, cannabis use, and caffeine intake during pregnancy as prenatal environmental risk factors for disruptive behavior in the offspring. In Chapter 3, I further investigated a range of pregnancy-related and perinatal factors in relation to childhood disruptive behavior in the well-powered UK population cohort ‘ALSPAC’ (Box 4). I aimed to address important issues in the existing literature, such as the high rates of comorbidity among disruptive behavior phenotypes, potential relatedness of environmental risk factors, and confounding by genetic factors. Subsequently, in Chapter 4, I investigated sex-stratified GxE’s in relation to disruptive behavior in ALSPAC, and this based on the top findings of two GWASs of disruptive behavior phenotypes and including maternal smoking during pregnancy and childhood maltreatment as environmental adversities. In addition, I aimed to replicate one of the most studied GxE’s in aggression and disruptive behavior, involving a polymorphism in MAOA. In Chapter 5, I performed a genetically hypothesis-free GxE study in ALSPAC using four environmental risk factors that were identified previously (maternal smoking, paracetamol use and life events during pregnancy, and childhood maltreatment) and a genome-wide, gene-based GxE design. In addition to discovering novel GxE’s, I attempted to elucidate underlying biological pathways by performing bioinformatics pathway analysis on the genes that emerged from the GxE analyses.
In Chapter 6, I used data from NeuroIMAGE (Box 4) to investigate evidence of shared genetic etiology between children’s aggressive behavior and CU traits. Using individual-level polygenic risk scoring, I investigated genetic sharing as such and also studied a number of previously implicated gene sets (related to dopaminergic, serotonergic, glutamatergic, and neuroendocrine signaling) as shared biological pathways of interest.
I investigated both the prediction of CU traits by the PRS as main effect, and PRS-based GxE’s involving smoking during pregnancy and childhood traumatic life events in relation to CU traits. In Chapter 7, I investigated how aggression-PRS relate to brain structural connectivity, using both the individual genome-wide genotyping and individual-level whole-brain DTI-data from NeuroIMAGE (Box 4). Using the results obtained in the previous chapter as a starting point, I studied both genetic main effects and GxE interactions in relation to DTI metrics of white matter microstructural integrity. To comprehensively investigate global effects across the brain, I performed a voxel-wise whole brain DTI study.
In the general discussion (Chapter 8), the results obtained throughout the research chapters 2 to 7 are summarized and discussed in the light of existing literature about GxE’s in disruptive behavior, and followed by some suggestions for future research.
References
1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed. Washington, DC: American Psychiatric Association; 2013.
2. Boylan K, Vaillancourt T, Boyle M, Szatmari P. Comorbidity of internalizing disorders in children with oppositional defiant disorder. Eur Child Adolesc Psychiatry. 2007;16(8):484–94.
3. Olsson M. DSM diagnosis of conduct disorder (CD)--a review. Nord J Psychiatry. 2009;63(Cd):102–12.
4. Fergusson DM, Horwood LJ, Ridder EM. Show me the child at seven: The consequences of conduct problems in childhood for psychosocial functioning in adulthood. J Child Psychol Psychiatry Allied Discip. 2005;46(8):837–49.
5. Reef J, Diamantopoulou S, Van Meurs I, Verhulst FC, Van Der Ende J. Developmental trajectories of child to adolescent externalizing behavior and adult DSM-IV disorder: Results of a 24-year longitudinal study. Soc Psychiatry Psychiatr Epidemiol. 2011;46(12):1233– 41.
6. Blair RJR, Leibenluft E, Pine DS. Conduct Disorder and Callous–Unemotional Traits in Youth. N Engl J Med. 2014;371:2207–16. 7. Kutcher S, Aman M, Brooks SJ, Buitelaar J, Van Daalen E, Fegert J, et al. International consensus statement on
attention-deficit/hyperactivity disorder (ADHD) and disruptive behaviour disorders (DBDs): Clinical implications and treatment practice suggestions. Eur Neuropsychopharmacol. 2004;14(1):11–28.
8. Reale L, Bartoli B, Cartabia M, Zanetti M, Costantino MA, Canevini MP, et al. Comorbidity prevalence and treatment outcome in children and adolescents with ADHD. Eur Child Adolesc Psychiatry. 2017;26(12):1443–57.
9. Georgiev A V., Klimczuk ACE, Traficonte DM, Maestripieri D. When violence pays: A cost-benefit analysis of aggressive behavior in animals and humans. Evol Psychol. 2013;11(3):678–99.
10. Kempes M, Matthys W, De Vries H, Van Engeland H. Reactive and proactive aggression in children. A review of theory, findings and the relevance for child and adolescent psychiatry. Eur Child Adolesc Psychiatry. 2005;14(1):11–9.
11. Smeets KC, Oostermeijer S, Lappenschaar M, Cohn M, van der Meer JMJ, Popma A, et al. Are Proactive and Reactive Aggression Meaningful Distinctions in Adolescents? A Variable- and Person-Based Approach. J Abnorm Child Psychol. 2017;45(1):1–14. 12. Martin J, Taylor MJ, Lichtenstein P. Assessing the evidence for shared genetic risks across psychiatric disorders and traits. Psychol
Med. 2018;48(11):1759–74.
13. Taylor MJ, Martin J, Lu Y, Brikell I, Lundström S, Larsson H, et al. Association of Genetic Risk Factors for Psychiatric Disorders and Traits of These Disorders in a Swedish Population Twin Sample. JAMA Psychiatry. 2019;76(3):280–9.
14. Abdulkadir M, Mathews CA, Scharf JM, Yu D, Tischfield JA, Heiman GA, et al. Polygenic Risk Scores Derived From a Tourette Syndrome Genome-wide Association Study Predict Presence of Tics in the Avon Longitudinal Study of Parents and Children Cohort. Biol Psychiatry. 2019;85(4):298–304.
15. Bralten J, Van Hulzen KJ, Martens MB, Galesloot TE, Arias Vasquez A, Kiemeney LA, et al. Autism spectrum disorders and autistic traits share genetics and biology. Mol Psychiatry. 2018;23(5):1205–12.
16. Bulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR, et al. An atlas of genetic correlations across human diseases and traits. Nat Genet. 2015;47(11):1236–41.
17. Cross-Disorder Group of the Psychiatric Genomics Consortium. Identification of risk loci with shared effects on five major psychiatric disorders: A genome-wide analysis. Lancet. 2013;381(9875):1371–9.
18. Bevilacqua L, Doly S, Kaprio J, Yuan Q, Tikkanen R, Paunio T, et al. A population-specific HTR2B stop codon predisposes to severe impulsivity. Nature. 2010;468(7327):1061–8.
19. Nautiyal KM, Tanaka KF, Barr MM, Tritschler L, Le Dantec Y, David DJ, et al. Distinct Circuits Underlie the Effects of 5-HT1B Receptors on Aggression and Impulsivity. Neuron. 2015;86(3):813–26.
20. Waltes R, Chiocchetti AG, Freitag CM. The neurobiological basis of human aggression: A review on genetic and epigenetic mechanisms. Am J Med Genet Part B Neuropsychiatr Genet. 2016;171(5):650–75.
21. Vassos E, Collier DA, Fazel S. Systematic meta-analyses and field synopsis of genetic association studies of violence and aggression. Mol Psychiatry. 2014;19(4):471–7.
22. Kendler KS. What psychiatric genetics has taught us about the nature of psychiatric illness and what is left to learn. Mol Psychiatry. 2013;18(10):1058–66.
23. Brunner H, Nelen M, Breakefield X, Ropers H, Van Oost B. Abnormal Behavior Associated with a Point Mutation in the Structural Gene for Monoamine Oxidase A. Science (80- ). 1993;262(5133):578–80.
24. Zhang-James Y, Faraone S V. Genetic architecture for human aggression: A study of gene–phenotype relationship in OMIM. Am J Med Genet Part B Neuropsychiatr Genet. 2016;171(5):641–9.
25. Veroude K, Zhang-James Y, Fernàndez-Castillo N, Bakker MJ, Cormand B, Faraone S V. Genetics of aggressive behavior: An overview. Am J Med Genet Part B Neuropsychiatr Genet. 2016;171(1):3–43.
26. van Donkelaar MMJ, Hoogman M, Shumskaya E, Buitelaar JK, Bralten J, Franke B. Monoamine and neuroendocrine gene-sets associate with frustration-based aggression in a gender-specific manner. Eur Neuropsychopharmacol. 2020;30:75–86.
27. Fernàndez-castillo N, Gan G, Donkelaar MMJ Van, Vaht M, Weber H, Retz W, et al. RBFOX1, encoding a splicing regulator, is a candidate gene for aggressive behavior. Eur Neuropsychopharmacol. 2020;30:44–55.
28. van Donkelaar MMJ, Hoogman M, Pappa I, Tiemeier H, Buitelaar JK, Franke B, et al. Pleiotropic Contribution of MECOM and AVPR1A to Aggression and Subcortical Brain Volumes. Front BehavNeurosci. 2018;12(61).
29. Tielbeek JJ, Johansson A, Polderman TJC, Rautiainen M-R, Jansen P, Taylor M, et al. Genome-Wide Association Studies of a Broad Spectrum of Antisocial Behavior. JAMA Psychiatry. 2017;74(12):1242–50.
30. Rautiainen M, Paunio T, Repo-Tiihonen E, Virkkunen M, Ollila H, Sulkava S, et al. Genome-wide association study of antisocial personality disorder. Transl Psychiatry. 2016;6(9):e883.
31. Dick DM, Aliev F, Krueger RF, Edwards A, Agrawal A, Lynskey M, et al. Genome-wide association study of conduct disorder symptomatology. Mol Psychiatry. 2011;16:800–8.
32. Pappa I, St Pourcain B, Benke K, Cavadino A, Hakulinen C, Nivard MG, et al. A genome-wide approach to children’s aggressive behavior: The EAGLE consortium. Am J Med Genet Part B Neuropsychiatr Genet. 2016;171(5):562–72.
33. Rampino A, Taurisano P, Fanelli G, Attrotto M, Torretta S, Antonella L, et al. A Polygenic Risk Score of glutamatergic SNPs associated with schizophrenia predicts attentional behavior and related brain activity in healthy humans. Eur Neuropsychopharmacol. 2017;27:928–39.
2009;461(7265):747–53.
35. Demontis D, Walters RK, Martin J, Mattheisen M, Als TD, Agerbo E, et al. Discovery of the first genome-wide significant risk loci for attention deficit/hyperactivity disorder. Nat Genet. 2019;51(1):63–75.
36. Kim DR, Bale TL, Epperson CN. Prenatal Programming of Mental Illness: Current Understanding of Relationship and Mechanisms. Curr Psychiatry Rep. 2015;17(2):5.
37. Latimer K, Wilson P, Kemp J, Thompson L, Sim F, Gillberg C, et al. Disruptive behaviour disorders: a systematic review of environmental antenatal and early years risk factors. Child Care Health Dev. 2012 Sep;38(5):611–28.
38. Hoover R, Hayes V, Erramouspe J. Association Between Prenatal Acetaminophen Exposure and Future Risk of Attention
Deficit/Hyperactivity Disorder in Children. Ann Pharmacother. 2015;49(12):1357–61.
39. Stergiakouli E, Thapar A, Davey Smith G. Association of Acetaminophen Use During Pregnancy With Behavioral Problems in Childhood: Evidence Against Confounding. JAMA Pediatr. 2016;170(10):964–70.
40. Rotem-Kohavi N, Oberlander TF. Variations in Neurodevelopmental Outcomes in Children with Prenatal SSRI Antidepressant Exposure. Birth Defects Res. 2017;109(12):909–23.
41. Westerneng M, Witteveen AB, Warmelink JC, Spelten E, Honig A, de Cock P. Pregnancy-specific anxiety and its association with background characteristics and health-related behaviors in a low-risk population. Compr Psychiatry. 2017;75:6–13.
42. Little B, Snell L, Gilstrap L, Johnston W. Patterns of multiple substance abuse during pregnancy: implications for mother and fetus. South Med J. 1990;5:507–18.
43. Maughan B, Taylor A, Caspi A, Moffitt TE. Prenatal Smoking and Early Childhood Conduct Problems: Testing Genetic and Environmental Explanations of the Association. Arch Gen Psychiatry. 2004;61(8):836–43.
44. Murray J, Maughan B, Menezes AMB, Hickman M, Macleod J, Matijasevich A, et al. Perinatal and sociodemographic factors at birth predicting conduct problems and violence to age 18 years: comparison of Brazilian and British birth cohorts. J Child Psychol Psychiatry Allied Discip. 2015;56(8):914–22.
45. Chernoff N. Teratogenic effects of cadmium in rats. Teratology. 1973;8(1):29–32.
46. Mochizuki M, Maruo T, Masuko K, Ohtsu T. Effects of smoking on fetoplacental-maternal system during pregnancy. Am J Obstet Gynecol. 1984;149(4):413–20.
47. Ziaei S, Nouri K, Kazemnejad A. Effects of carbon monoxide air pollution in pregnancy on neonatal nucleated red blood cells. Paediatr Perinat Epidemiol. 2005;19(1):27–30.
48. Lammer EJ, Shaw GM, Iovannisci DM, Van Waes J, Finell RH. Maternal smoking and the risk of orofacial clefts: susceptibility with NAT1 and NAT2 polymorphisms. Epidemiology. 2004;15(2):150–6.
49. Pereira P, Da Mata F, Figueiredo A, de Andrade K, Pereira M. Maternal Active Smoking During Pregnancy and Low Birth Weight in the Americas: A Systematic Review and Meta-analysis. Nicotine Tob Res. 2017;19(5):497–505.
50. Rice F, Langley K, Woodford C, Davey Smith G, Thapar A. Identifying the contribution of prenatal risk factors to offspring development and psychopathology: What designs to use and a critique of literature on maternal smoking and stress in pregnancy. Dev Psychopathol. 2018;30(3):1107–28.
51. Rice F, Harold GT, Boivin J, Hay DF, Van den Bree M, Thapar A. Disentangling prenatal and inherited influences in humans with an experimental design. PNAS Proc Natl Acad Sci United States Am. 2009;106(7):2464–7.
52. Thapar A, Rice F, Hay D, Boivin J, Langley K, Van Den Bree M, et al. Prenatal smoking might not cause attention-deficit/hyperactivity disorder: Evidence from a novel design. Biol Psychiatry. 2009;66(8):722–7.
53. Leppert B, Havdahl A, Riglin L, Jones HJ, Zheng J, Davey Smith G, et al. Association of Maternal Neurodevelopmental Risk Alleles with Early-Life Exposures. JAMA Psychiatry. 2019;76(8):834–42.
54. Byrd AL, Manuck SB. MAOA, childhood maltreatment, and antisocial behavior: Meta-analysis of a gene-environment interaction. Biol Psychiatry. 2014;75(1):9–17.
55. Afifi TO, McMillan KA, Asmundson GJG, Pietrzak RH, Sareen J. An examination of the relation between conduct disorder, childhood and adulthood traumatic events, and posttraumatic stress disorder in a nationally representative sample. J Psychiatr Res. 2011;45(12):1564–72.
56. Grigoryan G, Segal M. Lasting Differential Effects on Plasticity Induced by Prenatal Stress in Dorsal and Ventral Hippocampus. Neural Plast. 2016;2016:2540462.
57. Talge N, Neal C, Glover V. Antenatal maternal stress and long-term effects on child neurodevelopment: how and why? J Child Psychol Psychiatry Allied Discip. 2007;48(3–4):245–61.
58. Van den Bergh B, Mulder E, Mennes M, Glover V. Antenatal maternal anxiety and stress and the neurobehavioural development of the fetus and child: links and possible mechanisms. A review. Neurosci Behav Rev. 2005;29(2):237–58.
59. Blair RJ. The neurobiology of psychopathic traits in youths. Nat Rev Neurosci. 2013;14(11):786–99. 60. Blau N, van Spronsen FJ, Levy HL. Phenylketonuria. Lancet. 2010;376(9750):1417–27.
61. Holz NE, Zohsel K, Laucht M, Banaschewski T, Hohmann S, Brandeis D. Gene x environment interactions in conduct disorder: Implications for future treatments. Neurosci Biobehav Rev. 2018;91:239–58.
62. Tielbeek JJ, Karlsson Linnér R, Beers K, Posthuma D, Popma A, Polderman TJC. Meta-analysis of the serotonin transporter promoter variant (5-HTTLPR) in relation to adverse environment and antisocial behavior. Am J Med Genet Part B Neuropsychiatr Genet. 2016;171(5):748–60.
63. Dick DM, Agrawal A, Keller MC, Adkins A, Aliev F, Monroe S, et al. Candidate Gene–Environment Interaction Research. Perspect Psychol Sci. 2015;10(1):37–59.
64. Psychiatric GWAS Consortium Steering Committee. A framework for interpreting genome-wide association studies of psychiatric disorders. Mol Psychiatry. 2009;14(1):10–7.
65. Van der Auwera S, Homuth G, Pistis G, Penninx B, Milaneschi Y, Mullins N, et al. Genome-wide gene-environment interaction in depression: A systematic evaluation of candidate genes. Am J Med Genet Part B Neuropsychiatr Genet. 2017;177(1):40–9. 66. Biernacka JM, Chung SJ, Armasu SM, Anderson KS, Lill CM, Bertram L, et al. Genome-wide gene-environment interaction analysis of
pesticide exposure and risk of Parkinson’s disease. Park Relat Disord. 2016;32:25–30.
67. Noordermeer SDS, Luman M, Oosterlaan J. A Systematic Review and Meta-analysis of Neuroimaging in Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) Taking Attention-Deficit Hyperactivity Disorder (ADHD) Into Account. Neuropsychol Rev. 2016;26(1):44–72.
68. Gao Y, Shuai D, Bu X, Hu X, Tang S, Zhang L, et al. Impairments of large-scale functional networks in attention-deficit/hyperactivity disorder: A meta-analysis of resting-state functional connectivity. Psychol Med. 2019;49(15):2475–85.
69. Hoogman M, Muetzel R, Guimaraes JP, Shumskaya E, Mennes M, Zwiers MP, et al. Brain imaging of the cortex in ADHD: A coordinated analysis of large-scale clinical and population-based samples. Am J Psychiatry. 2019;176(7):531–42.
70. Waller R, Dotterer HL, Murray L, Maxwell AM, Hyde LW. White-matter tract abnormalities and antisocial behavior: A systematic review of diffusion tensor imaging studies across development. NeuroImage Clin. 2017;14:201–15.
71. Jellison BJ, Field AS, Medow J, Lazar M, Salamat MS, Alexander AL. Diffusion Tensor Imaging of Cerebral White Matter: A Pictorial Review of Physics, Fiber Tract Anatomy, and Tumor Imaging Patterns. Am J Neuroradiol. 2004;25:356–69.
73. Ennis DB, Kindlmann G. Orthogonal tensor invariants and the analysis of diffusion tensor magnetic resonance images. Magn Reson Med. 2006;55(1):136–46.
74. Douaud G, Jbabdi S, Behrens TEJ, Menke RA, Gass A, Monsch AU, et al. DTI measures in crossing-fibre areas: Increased diffusion anisotropy reveals early white matter alteration in MCI and mild Alzheimer’s disease. Neuroimage. 2011;55(3):880–90. 75. Kochunov P, Jahanshad N, Marcus D, Winkler A, Sprooten E, Nichols TE, et al. Heritability of fractional anisotropy in human white
matter: A comparison of Human Connectome Project and ENIGMA-DTI data. Neuroimage. 2015;111:300–11.
76. Kochunov P, Fu M, Nugent K, Wright SN, Du X, Muellerklein F, et al. Heritability of complex white matter diffusion traits assessed in a population isolate. Hum Brain Mapp. 2016;37(2):525–35.
77. Sprooten E, Knowles EE, McKay DR, Göring HH, Curran JE, Kent JW, et al. Common genetic variants and gene expression associated with white matter microstructure in the human brain. Neuroimage. 2014;97:252–61.
78. Boyd A, Golding J, Macleod J, Lawlor D, Fraser A, Henderson J, et al. Cohort Profile: the ’children of the 90s’ – the index offspring of the Avon Longitudinal Study of Parents and Children. Int J Epidemiol. 2013;42:111–27.
79. Fraser A, Macdonald-wallis C, Tilling K, Boyd A, Golding J, Davey smith G, et al. Cohort profile: The avon longitudinal study of parents and children: ALSPAC mothers cohort. Int J Epidemiol. 2013;42(1):97–110.
80. Golding J, Pembrey M, Jones R, ALSPAC study team. ALSPAC--the Avon Longitudinal Study of Parents and Children. I. Study methodology. Paediatr Perinat Epidemiol. 2001;15(1):74–87.
81. von Rhein D, Mennes M, van Ewijk H, Groenman AP, Zwiers MP, Oosterlaan J, et al. The NeuroIMAGE study: a prospective phenotypic, cognitive, genetic and MRI study in children with attention-deficit/hyperactivity disorder. Design and descriptives. Eur Child Adolesc Psychiatry. 2015;24(3):265–81.
