Multiomic Biological Approaches to the Study of Child Abuse and Neglect

Abstract

Childhood maltreatment, occurring in up to 20–30% of the population, remains far too common, and incorporates a range of active and passive factors, from abuse, to neglect, to the impacts of broader structural and systemic adversity. Despite the effects of childhood maltreatment and adversity on a wide range of adult physical and psychological negative outcomes, not all individuals respond similarly. Understanding the differential biological mechanisms contributing to risk vs. resilience in the face of developmental adversity is critical to improving preventions, treatments, and policy recommendations. This review begins by providing an overview of childhood abuse, neglect, maltreatment, threat, and toxic stress, and the effects of these forms of adversity on the developing body, brain, and behavior. It then examines examples from the current literature of genomic, epigenomic, transcriptomic, and proteomic discoveries and biomarkers that may help to understand risk and resilience in the aftermath of trauma, predictors of traumatic exposure risk, and potential targets for intervention and prevention. While the majority of genetic, epigenetic, and gene expression analyses to date have focused on targeted genes and hypotheses, large-scale consortia are now well-positioned to better understand interactions of environment and biology with much more statistical power. Ongoing and future work aimed at understanding the biology of childhood adversity and its effects will help to provide targets for intervention and prevention, as well as identify paths for how science, health care, and policy can combine efforts to protect and promote the psychological and physiological wellbeing of future generations.

Introduction

Child maltreatment refers to acts of both commission (abuse) and omission (neglect) that result in harm, potential harm, or threat of harm to a child, often, though not exclusively, by a child’s caregiver.¹ Definitions of child maltreatment vary, and our use of the term hinges upon the consequence for the child, not the perpetrator’s intent.

Maltreatment encompasses a vast array of lived experiences. Abuse alone can be physical, sexual, or psychological, frequently resulting in immediate, and sometimes, permanent physical injuries to the body and brain and even death for 2.5 of every 100,000 children.² Neglect consists of a failure to act, resulting in the absence of adequate housing, food, clothing, emotional support, or supervision.³ Like abuse, neglect can put a child in immediate danger resulting in physical injury or death. Children can and often do experience a combination of abuse and neglect.⁴

Children with histories of child abuse and neglect experience significant functional impairment across multiple domains, including increased risk for dysregulated cognitive, psychosocial, and biological development.⁵ Exposure to childhood maltreatment is associated with durable physiological changes, often resulting in poor mental and physical health outcomes.⁶ Poor health outcomes are in part due to the biological consequences of childhood abuse and neglect.^7,8 Homeostatic changes to physiology and epigenetic processes during critical developmental periods are cardinal features leading to the embedding of environmental exposure sequelae into biological processes.⁹ Specifically, identifying adaptations in biological markers (e.g. multi-omic approaches) after exposure to childhood maltreatment underscores the importance of scientific evidence to support child advocacy work spanning many sectors. This review aims to elucidate the role of genomics, transcriptomics, proteomics and epigenomics in the study of child abuse and neglect by exploring each of the components in the intricate weaving of ecology and biology, otherwise defined as gene by environment research.

Overview of Child Abuse and Neglect

Child Maltreatment

Child maltreatment is unfortunately very common, though its exact prevalence is unknown as estimates vary depending on the method and measures used for data collection. The CDC estimates that of the 74.2 million children in the US,¹⁰ 10.6 million, or 1 out of 7, has experienced maltreatment in the previous year.¹¹ However, child protective services identified only 656,000 child survivors of maltreatment in 2019.² Neither of these reports approximate rates of maltreatment across an entire childhood. However, a telephone survey of adults across 28 states found prevalence rates of self-reported experiences of childhood abuse alone to be as high as 34.42%, 17.94%, and 11.60% for emotional, physical, and sexual abuse, respectively.¹²

Variance in reported rates of child maltreatment is pervasive throughout the literature investigating the effects of child abuse and neglect due to a multitude of factors, including inconsistent inclusion criteria of maltreatment, societal under-reporting, variation between adult and child accounts, and the frequent reliance on incomplete data collected by social services.^1,3,13–15 Biological investigations into these data must consider both the ecologically-based hurdles in categorizing and collecting such data when utilizing it.

Adverse Childhood Experiences

The seminal adverse childhood experiences (ACE) study expanded the scope of childhood maltreatment research beyond only experiences of abuse and neglect to include exposure to “household dysfunction”. This category of experiences included exposure to substance abuse, parental mental illness, witnessing violence toward an individual’s mother, and criminal behavior in the household.⁵ Subsequent literature has further expanded investigations of ACEs to include additional measures of neglect, parental separation, homelessness, and bullying.^16,17 These additions represent the increasing realization that household unpredictability, witnessing violence, and chronic mediators of stress can have as large an effect on long-term developmental outcomes as more traditionally measured abuse and neglect.

The 1998 ACE study demonstrated that a person’s environment, particularly their exposure to “adverse” events in their childhood, not only their exposure to maltreatment, directly increased their risk for and incidence of ischemic heart disease, cancer, chronic lung disease, skeletal fractures, and liver disease in adulthood.⁵ This revelation had massive implications for public health and policy efforts, including new methods for disease prevention^18,19 and launched over two decades of investigations by neuroscientists, biologists, and geneticists aiming to explain the biological mechanisms responsible for the observed interactions between an individual’s childhood environment and their poor adult health outcomes.

Adversity: The Adverse Environment

Though environment can shape a person’s life at any age, a child’s dependence on their families and communities to meet their physical and emotional needs leaves them particularly vulnerable to maltreatment and other adverse experiences in their home and through their community.²

A reliable community consisting of supportive and nurturing family, school, or peer relationships acts as a buffer for the effects of ACEs.²⁰ This is achieved by promoting healthy bio-psycho-social development and bio-behavioral synchrony.²¹ The absence of these positive relationships does not prompt wellness. The absence of warmth and attention has been associated with cognitive impairments, dysregulations of physical health, disruptions of the body’s stress response, and neurocognitive impairments.²² Further, the absence of positive, emotionally, and physically safe environments can quickly threaten the wellbeing of the child.²³

Millions of children in the United States are raised in threatening or otherwise adverse environments.^11,24 These children are at increased risk for maltreatment and community-level ACEs.²⁵ This community risk is dependent on the convergence of many confounding and compounding sociological and environmental factors, including but not limited to systemic racism, violent community context, and socioeconomic status (SES).^26,27 Low SES alone is associated with decreased life expectancy, epigenetic age acceleration, and higher rates of ACEs.^28–30 Additionally, children living in low SES environments have an increased risk for exposure to violent community contexts and poly-victimization.⁴ Children raised in violent environments experience prolonged periods of hyperarousal, resulting in an increased risk of developing complex posttraumatic stress disorder (PTSD) and stress-related pathology.³¹

This increase in risk for ACEs is at least in part due to barriers that diminish a parent’s ability to care for their children often through severe socioeconomic deprivation (SED), substance use, racial discrimination, physical or psychological health problems, racial and ethnic violence, and parental incarceration.^32–38 Just as the genome and environment interact and affect the other, ACEs and adversity are difficult to tease apart. This relationship is most readily seen in rates of poverty-related neglect.³⁹ Rates of maltreatment, including neglect, are highest in low-income areas.⁴⁰ Though instances of neglect occur in households of all SESs, low SES in particular can also lead to physical neglect in a family that might otherwise choose to meet their children’s physical and emotional needs.³⁹

Threat and Deprivation

Distinctions between forms of child maltreatment, ACEs, and exposure to environmental adversity help contextualize scientific findings and guide public health efforts. However, the body interprets threatening, or otherwise fear-inducing, experiences with similar activations of the threat response system.⁴¹ Real or perceived danger’s intensity,⁴² timing,⁴³ and duration,⁴² and an individual’s unique risk and resilience factors appear to account for the range in observed outcomes more than legal or societal distinctions differentiating these states of threat or deprivation.²⁰

For this reason, investigations of the genomic components of the well-documented relationship of childhood adversity and poor physical and psychological health outcomes must also explore literature investigating states of threat and deprivation more broadly. This expansion also allows researchers to investigate neglect more closely. Using deprivation as a stand-in for neglect enables researchers to translate investigation to animal model systems better designed to determine the mechanisms involved in these observed gene and environment interactions.⁴⁴

Toxic Stress

A threatening environment or experience activates the stress response system (SRS), sometimes referred to as the fight, flight, or freeze response.⁴⁵ This adaptive activation of the hypothalamic-pituitary-adrenocortical (HPA) axis and the sympathetic-adrenomedullary (SAM) systems prepares the body for danger. Episodic and limited early life stress (ELS), and the subsequent activation of these regulatory systems, are critical for a child’s developing brain and body. Such healthy developmental experiences allow for a child’s SRS to appropriately adapt to their environment through a process of allostasis.^46,47

For example, a child’s first day of school may be stress-inducing but not inherently damaging in the presence of a supportive and nurturing environment. Though the child’s SRS may be activated, with support the child will naturally regulate. However, children in adverse environments without the buffering effects of a supportive or consistent environment, or children enduring maltreatment, undergo periods of increased and prolonged activation of the SRS, more likely leading to toxic stress responses.⁴⁸

Toxic stress results from an over-activation of the body’s stress response system,⁴⁸ resulting in either hypo- or hyper- active responses of the HPA axis.⁴⁹ Chronic activation of the central nervous system (CNS) and the HPA axis and SAM systems, particularly during sensitive periods, may change the developing brain’s architecture and connectivity, particularly in the prefrontal cortex (PFC), amygdala, and hippocampus.^50,51 These alterations upregulate risk for alterations to neurogenesis, synaptogenesis, and bio-behavioral synchrony,^52–54 all critical for neural development and function. These disruptions in stress reactivity, threat discrimination,⁵⁵ and memory formation,^56,57 contribute to observed outcomes for maltreatment victims, including learning difficulties, behavior problems, and stress-related physical ailments, and psychological pathology. Additionally, this overactivation produces an overabundance of inflammatory cytokines, increased blood pressure, and a “wear and tear” effect on the brain and organ systems, contributing to the observed poor health of adults with high ACE scores.

Consequences to the Brain and Body

It is clear that the impact of childhood adversity on the individual are immediate, pervasive, and extend far beyond the immediate physical and psychological injuries incurred at the time of exposure. Children who have experienced maltreatment and other adverse events are more likely to experience academic problems, conduct disorders,⁵³ ADHD,⁵⁸ anxiety, aggression, depression,⁵⁹ suicide risk,⁶⁰ high-risk sexual behaviors,⁶¹ and substance use.⁶² Further, adults with high ACE scores are more likely to develop PTSD,⁶³ depression,⁶⁴ heart disease,⁶⁵ chronic obstructive pulmonary disease (COPD),^66,67 and cancer.⁶⁸ While an individual’s biology is impacted by their environment regardless of their lifestyle choices,⁶⁹ individuals with high ACE scores are more likely to be exposed to and adopt poor health behaviors such as substance use, smoking, poor diet, and lack of exercise than individuals with no ACEs.^5,70 Often these behaviors, which contribute to disease development, are utilized as a means of coping with the psychological distress often associated with ACEs.⁷¹ This adoption of maladaptive coping mechanisms, particularly in the presence of toxic stress, may lead to accelerated biological aging as seen in both telomere length (TL) and mitochondrial DNA copy number (mtDNAcn) and in other biomarkers as outlined below.⁷² This chain-like reaction of a person’s ecology and biology highlights the intertwined mechanisms resulting in a person’s vulnerability to developing poor health outcomes in adulthood.

Risk Factors

While exposure to childhood adversity and ACEs is one of the most potent risk factors for increasing a person’s risk for developing poor health outcomes, not every child is at an equal risk of exposure to ACEs,⁷³ and not every child with ACEs is equally likely to develop poor physical or psychological health.

Many factors contribute to a child’s risk of adversity and maltreatment, beyond those community risk factors discussed in the adversity section of this review. Individual characteristics which place a child at greater risk of ACEs include a child’s age, sex, race,ethnicity, physical and psychological health, and intellectual ability.^58,73–77

Additionally, many elements of the exposure itself contribute to a child’s risk and the consequences observed, including the type of ACE,⁷⁸ the accumulation of multiple ACEs,⁷⁹ the duration of these experiences,⁸⁰ and their timing.⁸¹ As a person’s ACE score increases, so does their risk of developing poor physical and psychological health.⁸² Experiencing multiple types of adverse experiences over an extended period of time also significantly increases a child’s risk of toxic stress.⁸³ Additionally, this observed cumulative effect predicts depression and suicide risk⁸⁰ and alcohol use.⁸⁴ However, more research is required to investigate the mechanisms responsible for this correlation.

Critical Periods

The developing brain and body are particularly vulnerable to environmental impacts during critical developmental periods.^81,85 Though more research is needed to distinguish these critical periods further, it is clear that the consequences of ACEs and toxic stress depend partially on the child’s age at the time of exposure. The importance of the timing of ACEs is observed in both developmental neuroscience and genetic research.^43,85 Notably, neurobiological research indicates sensitive periods for the hypothalamus,^49,86 hippocampus,⁸⁷ PFC,⁸⁸ that when combined with alterations to the development of the amygdala, and nucleus accumbens might contribute to the development of adolescent depression.⁸⁹ While maltreatment affects risk for all psychiatric disorders, perhaps the most obvious example of understanding how trauma directly effects psychiatric symptoms is with posttraumatic stress disorder (PTSD). By definition, PTSD requires trauma exposure, and it thus provides the ability to examine the impact of environment (trauma) and biology (genetics) on the brain and behavior. Additionally, research into the neurobiology of fear circuitry in PTSD development has indicated that ACEs affect amygdala reactivity, volume, and structure, and homeostasis of the HPA-axis. The analyses of these effects have identified three sensitive periods for which consequences of neural development are most observed: before age 3, around age 10, and the onset of puberty.^81,85 Below we will further review findings at the genomic, transcriptomic, proteomic, and epigenomic levels related to the impact of childhood maltreatment on biomarkers related to long-term outcomes on body, brain, and behavior.

Omic Approaches to Understanding Effects of Childhood Maltreatment

Introduction to Gene x Environment Interactions

To elucidate specific mechanisms that underpin the consequences of child abuse and neglect, researchers have hypothesized an interaction between genetic determinants and perilous environmental risk factors. Specifically, these associations aim to identify accurate etiological pathways and address the long-standing “nature vs. nurture” debate. Such associations are best defined by genotype-environment correlations and genotype-environmental interactions (GxE).

Genotype-environmental correlations are characterized by exposures that are associated with genetic propensities.⁹⁰ The field classifies such correlations as either passive, active, or evocative gene-environment correlations.⁹¹ Each condition seeks to explain the directionality of the association between genetic architecture and environmental conditions. Passive gene-environmental correlations refer to the association between heritable traits and environmental context that supports the inherited genotype.⁹² Jaffee et al. 2004, highlights this by showing that physical maltreatment is causally related to the development of a child’s antisocial behavior. Active gene-environment correlation occurs when an individual possesses a heritable inclination to select environmental exposure, e.g. genetically-inclined extroverted individuals may seek out more risky social environments relative to shy individuals. Evocative gene-environment correlations occur by an individual’s ability to evoke reactions from other people based on their genetic propensities.⁹³ Children who exhibit externalizing symptoms during childhood and adolescence may have an increased risk for experiencing maltreatment due to maladaptive parenting, caused by evocative-genotype associations.⁹⁴ This genetic overlap does not abstain parents from seeking adequate parental support for rearing children with challenging behaviors. Additionally, genotype-environmental correlations suggest exposure to childhood maltreatment has a potential role in the onset of mood disorders, ADHD, and schizophrenia.^95,96,97

Conversely, genetic correlations can introduce bias into GxE. GxE addresses genetic sensitivities, susceptibilities, and resiliencies to environmental factors.⁹⁸ The effect of genetic influences and environmental risk factors is contingent upon one’s genetic architecture and degree of environmental exposure. In GxE, the phenotypic outcome (dependent variable) is reliant on the presence of both environmental exposure (independent variable) and genetic composition (independent variable).

Single nucleotide-polymorphisms (SNPs) are isolated genetic variants among single base pairs within the genome.⁹⁹ SNPs are undoubtedly the most common form of genetic polymorphisms and range in physiological contribution from dampened to heightened risk for disease.¹⁰⁰ Such variations must occur in at least 1% of the population to truly be considered a ‘common’ SNP.¹⁰¹ Identifying candidate genes and SNPs in the context of childhood maltreatment is still fairly unexplored and the results from published findings vary.

Inconsistencies in study findings are attributed to poor replicability in GxE studies due to differing sample sizes.⁹⁸ Below we review some of the more well-powered ‘candidate’ GxE studies, however, until such findings are replicated in much larger sample sizes and are found to be significant at a ‘genome-wide level’ we must be skeptical of the observations. Van der Auwera et al. (2018), did not find any genome-wide significant associations between candidate genes for major depressive disorder and exposure to childhood trauma.¹⁰² In contrast, others have found that FKBP5 SNP rs9296158 showed a significant main effect on depressive symptoms. The interaction term between rs9296158 and childhood maltreatment was predictive of adult depressive symptomatology.¹⁰³ Additionally, the interaction term between CRHR1 SNPs rs7209436 and rs110402, and exposure to childhood trauma provide evidence that exposure to childhood trauma may be related to suicidal behavior and depressive symptoms.^104,105 Notably the FKBP5 gene regulates glucocorticoid receptor activation, which is both upstream and downstream of the CRH peptide and its CRHR1 receptor in regulating the stress response. In addition to influencing affective disorders, exposure to childhood trauma disrupts stress homeostasis at the molecular level, resulting in some individuals with histories of exposure to childhood trauma showing unique sensitivities to cortisol. These findings were only found among carriers of the CRHR1 risk haplotypes with histories of childhood trauma exposure compared to non-trauma exposed individuals, further supporting genetic consequences of early trauma exposure.¹⁰⁶ In summary, while the effects have been variable across studies, some of the most robust findings related to GxE risk with regards to stress related disorders across different cohorts involve polymorphisms in genes directly related to HPA-stress axis regulation.

Genomics

Among people who experience childhood abuse or neglect, there is great variability in the physiological, neurobiological, and behavioral effects experienced after this early life exposure. This is sometimes understood in terms of vulnerability or resilience. Environmental factors such as type of neglect or abuse can affect prognosis after childhood maltreatment, but genomic factors can also have a significant impact on vulnerability. Genomic, epigenomic, transcriptomic, and proteomic research in this area aims to understand why some individuals are more likely than others to develop certain psychiatric traits or disorders such as neuroticism, suicidality, major depressive disorder (MDD), or PTSD after experiencing childhood abuse or neglect. Understanding the molecular mechanism responsible for the phenotypic presentation of child abuse and neglect is important to develop better therapeutic interventions.

The genome is comprised of all the inherited genetic material in an organism, which in humans generally refers to DNA. Child abuse and neglect is a classification of exposure, but it is correlated with higher rates of psychopathology later in life. One overarching question is how the genetic material inherited from parents to offspring may dictate this psychopathology – this is called heredity. Twin studies are often used to study heritable traits by quantifying the genetic contributions to the development of a certain illness or characteristic. A twin study by Sartor et al. (2012), using the Australian National Health and Medical Research Council volunteer twin panel, compared heritability of PTSD and MDD in individuals who had experienced childhood maltreatment categorized as either exposure to high-risk trauma or low-risk trauma.¹⁰⁷ Results showed that 46% of variance in PTSD and 27% of variance in MDD was attributable to genetic factors. Notably, with this familial inheritance model, where additive genetic (A), shared environmental (C), and nonshared environmental (E) influences between siblings are examined, they suggest that there is complete correlation between the genetic factors that contributed to MDD and PTSD. In other words, heritable influences on high-risk trauma exposure, PTSD, and MDD, can be traced to the same sources, such that genetic risk may not be disorder specific. PTSD research provides models for understanding severe stress, but there are distinctions between molecular markers in individuals with PTSD due to childhood maltreatment versus other trauma.¹⁰⁸

In 2020, a large genome-wide association study (GWAS) by Dalvie et al. was published using the UK Biobank (UKBB) as a discovery sample with a replication sample from the Psychiatric Genomics Consortium PTSD group (PGC-PTSD), with >150,000 subjects in total. Many studies focus on behavioral phenotypes, but this study was able to assess genomic factors that could influence exposure to childhood maltreatment, potentially as an example of evocative or active gene-environment correlation as outlined above. The main GWAS analysis tested genotypes for association to self-reported childhood maltreatment within each study and meta-analysis was conducted across studies. Two genome-wide significant loci rs142346759 and rs10262462, associated with childhood maltreatment in the discovery dataset remained significant in meta-analysis. These GWAS hits were annotated to genes FOXP1 and FOXP2, both of which have been implicated in language, cognitive, and emotional disorders. A variant on chromosome 12 was identified but its biological pathway is unclear. A significant genetic overlap was found with positive correlation between childhood maltreatment and depressive symptoms, MDD, and neuroticism suggesting there may be shared underlying mechanisms of predisposition. The SNP-based estimate of heritability of reported childhood maltreatment was ~6%. Notably, in analysis within maltreatment types, the heritability of sexual maltreatment was found to be 0% which is not supportive of a GxE effect. It is important to note that work from several groups have shown that developmental timing of childhood trauma leads to different biological pathways mediating adult risk,^43,109 thus, these variables will be important to include in future studies of GxE contributions to trauma-related outcomes.

The focus in genomic research has shifted towards large-scale genome-wide studies, such as the Dalvie et al. manuscript above, which can be used to form polygenic models of genomic influence. This follows a period of GxE candidate gene studies which focused on single genetic polymorphisms and their correlation to behavioral phenotypes. Retrospectively there is concern among investigators that what appeared to be positive interactions in these studies could be driven by confounders such as ethnicity, gender, age, SES, etc.¹¹⁰ Keller calls for greater control of covariates in models of analysis for candidate gene studies. While these shortcomings must be carefully considered, there is a large body of research focused on candidate genes as moderators for the effects of childhood abuse and neglect. The following examples of such studies still provide important insight into how biological effects of trauma may be related to molecular (genetic, epigenetic, proteomic) factors, though they must be taken with some caution until replicated in larger samples.

A number of studies have examined polymorphisms in genes that encode proteins critically important for monamine neurotransmitter regulation (e.g. serotonin, dopamine, noradrenaline), which are critical for mood and emotion processing and are targets of common antidepressant and antianxiety medication treatments. Using the Dunedin Multidisciplinary Health and Development study, Caspi et al., (2002) tested the effect of a functional polymorphism in the promotor of monoamine oxidase A (MAOA) on four measures of antisocial behavior. This polymorphism was previously found to be associated with aggression and increased brain dopamine and norepinephrine in mice and humans.¹¹¹ The main effect of MAOA activity on antisocial measures was not significant, but there was a significant GxE effect. The effect of childhood maltreatment on antisocial behavior was significantly weaker among males with high MAOA activity. Polymorphic variation in the gene encoding the serotonin transporter protein has also been studied for its moderation of the effects of ELS. The short (s) allele in the promoter region 5-HTTLPR is associated with the development of depression in adults with a history of child maltreatment.¹¹² This analysis was repeated in a subsequent study of children who had been taken from their parents’ care due to abuse or neglect allegations by the Connecticut Department of Children and Families (DCF). They found the s allele conferred vulnerability to depression, and the s/s genotype conferred the highest depression scores.¹¹³

Related to the previously mentioned genetic modulation of stress, a variation of the corticotropin-releasing hormone receptor gene (CRHR1) which is implicated in physiological stress in the HPA axis, is associated with neuroticism.¹¹³ In the Environmental Risk (E-Risk) Longitudinal Twin Study, the TAT haplotype formed by 3 CRHR1 SNPs was associated with a protective effect against depression.¹¹⁴ Furthermore, in a study of maltreated children, CRHR1 variation moderated the association of maltreatment and neuroticism differentially depending on type of maltreatment and the number of types of maltreatment.¹¹⁵

Expanding on previous gene-by-environment findings, Ressler et al. (2010) examined a gene-by-gene-by-environment interaction and found that s 5-HTTLPR x CRHR1 haplotype x child maltreatment was associated with heightened depressive symptoms. In another study, MAOA activity was also found to moderate the HTTLPR x environment effect on depressive symptoms.¹¹⁶ Understanding the compounding or additive nature of SNP variation on psychiatric symptoms helps support a polygenic model of childhood abuse and neglect.

Transcriptomics and Proteomics

Identifying gene variants and characterizing gene regulation is important in identifying biological pathways related to psychiatric symptomology, but to better understand the molecular mechanisms of these pathways, additional techniques are required. Studying the transcriptome, which includes all the RNAs expressed in a given cell or tissue, is one way to study such mechanisms.¹¹⁷ Proteomics – the study of proteins, is another important step in parsing out the mechanisms of genomic influence. Although blood is sometimes used, because it is more accessible, brain tissue is also relevant for psychiatric transcriptomic analysis. Because only two-thirds of brain-expressed genes are detected in the blood, tissue type can have a large impact on research outcomes. Tissue quality can also affect transcriptomic analysis and RNA quality of human brains is lower than from experimental models.¹¹⁷ RNA sequencing (RNAseq) is implemented for transcriptomic analysis of differential gene expression and splicing of mRNAs and can be used to study both protein coding and noncoding genes.

Human brain tissue is sometimes available from postmortem subjects. A Canadian study used the Quebec Suicide Brain Bank to explore differential gene expression in subjects who had been exposed to childhood abuse or neglect compared to control subjects. They included 12 suicide victims with a history of childhood abuse, 12 suicide victims with negative history of childhood abuse and 22 control subjects.¹¹⁸ McGowen et al. found that expression of glucocorticoid receptor mRNA in hippocampal tissue was significantly reduced in suicide victims with history of childhood abuse relative to non-abused suicide victims or controls. Additional psychopathology such as substance use disorders or mood disorders did not have an effect on glucocorticoid receptor (GR) expression.¹¹⁸ These findings suggest that changes in GR expression are closely associated with a history of early life adversity. That said, it is also important to note that changes in GR and cortisol may also be a result of homeostatic responses to stress, and not necessarily causal in these relationships.

A data set from the National Institute of Mental Health Center for Collaborative Genomic Studies on Mental Disorders was used by Minelli et al. (2018) to compare gene expression in subjects who experienced childhood trauma with and without MDD diagnosis. Sample groups of 367 MDD patients and 344 controls were classified as having four types of trauma – sexual abuse, physical abuse, emotional abuse, and emotional neglect. Gene expression, measured from blood samples with RNAseq, was used in a likelihood ratio test along with childhood trauma. An association between neglect and Mediator Complex Subunit 22 (MED22) was significant, although no associations were observed for the other three types of childhood trauma. MED22, which was downregulated in MDD patients, encodes for a protein that contributes to coordination of transcription and cell lineage development.¹¹⁹ This analysis did not reveal robust results that could suggest biological pathways differentially expressed in MDD patients with history of child neglect. Shortcomings from this study may be a result of limited sample size as well as in analyzing gene expression in blood as opposed to brain tissue.

Due to aforementioned limitations in tissue accessibility, research on ELS, abuse, and neglect in mouse models aims to identify molecular factors that translate to childhood abuse and neglect in humans.⁸³ One effective mouse model of early life neglect utilizes maternal separation and early weaning (MSEW).¹²⁰ Using this model, researchers found a dysregulation of genes related to mature oligodendrocytes (Enpp2, Mbp) and translation initiation factors in the prefrontal cortex.¹²¹ Exon-level analysis found myelin-related genes were also dysregulated. Proteomic analysis in PFC tissue found a downregulation of oligodendrocyte proteins CNP, MBP, OMG, and PTN4. By connecting transcriptomic and proteomic analysis, Bordner et al. concluded that MSEW may have caused disruption of the myelination of oligodendrocytes, which is important for proper PFC function. While much work remains to be done, the expansion of transcriptomic and proteomic work from postmortem human brain samples in psychiatric disorders^122,123 should allow further mechanistic work to examine the role of childhood maltreatment in biomarkers for adult brain-based disorders.

Epigenomics

Despite the fact that childhood trauma is an important risk factor for a majority of psychiatric disorders, how genetic risk may interact with environmental factors on long-term health outcomes is currently unknown.¹²⁴ The growing field of epigenetics provides a conceptual framework that may add insights into the biology of childhood trauma, potentially leading to novel preventive, diagnostic, and therapeutic approaches.¹²⁵ Specifically, decoding epigenetic mechanisms in the context of childhood trauma is critical to understand how experiences in childhood are embedded in biological systems and exert influence on development and health throughout the lifespan.

The term epigenetics is often used in a broad and inclusive sense describing molecular factors that reversibly modify chromatin function without directly changing the DNA sequence. Mechanisms considered relevant under the term epigenetics include DNA and histone modifications, as well as noncoding RNA signaling, among others.¹²⁶ In the context of psychiatric disorders and childhood trauma, most studies, including this review, focus on DNA methylation (DNAm), likely due to sample requirements, costs and available technologies for human clinical studies. Studies to date are often correlational by nature and it remains challenging to determine causality for most of the published findings.

In contrast, animal models of stress and trauma exposure are able to discern causality, but often lack translational relevance for human disorders associated with trauma exposure.¹²⁷ Another important feature of epigenetic modifications in general and in the context of trauma exposure is tissue specificity. In contrast to genetic studies that can utilize almost any tissue available to determine genotype, the epigenotype and the influence of the environment on epigenetic modifications is highly tissue specific. This particular feature is the basis for a common criticism of DNAm studies utilizing blood and saliva tissue samples. Indeed, only a limited set of DNAm sites show a strong correlation between the brain and peripheral tissues, and which is likely driven, in part, by genetic factors.¹²⁸ Although this may restrict the insights that can be gained into the underlying molecular mechanisms of trauma-related psychiatric disorders, studies utilizing peripheral tissue can provide meaningful and readily accessible biomarkers to determine disease risk or treatment response.

Similar to early genetic studies, a majority of DNAm studies so far were hypothesis-driven approaches focusing on genes related to the HPA axis, serotonergic signaling, and other pathways known to be involved in trauma-related disorders.¹²⁹ The gene for the glucocorticoid receptor Nr3c1 is an early example that is well-supported by animal studies.¹³⁰ Variation of maternal care in rodent models effectively influenced DNAm at a transcription factor binding site in the Nr3c1 gene, leading to long-lasting stress-related behavioral changes. Previous research in rodents identified the neuron-specific glucocorticoid receptor (Nr3c1) as a site of epigenetic change as part of the HPA response to stress.¹³¹ Similar studies in human post-mortem brain tissue suggest a similar effect.¹¹⁸ However, multiple studies in peripheral tissue yielded overall inconsistent results with small effects sizes compared to the initial studies. Also note that despite previous literature indicating that DNAm variability may be a potential component modulating developmental stress, a recent study did not find the significant results of DNAm variability in the cumulative effects of adversity.⁴³

Similar to the glucocorticoid receptor gene, several studies investigated DNAm changes in response to stress and trauma in FKBP5, a co-chaperone of the glucocorticoid receptor and central regulator of HPA-axis function.¹³² Importantly, the activity of FKBP5 within the negative feedback mechanism of the HPA axis is also genotype-dependent, thus representing a prime example of how the environment may interact with genetic disposition on functional molecular and behavioral outcomes.¹³³ For both genes, NR3C1 and FKBP5, the field has accumulated a plethora of supporting data contributing to our understanding of the underlying mechanisms of childhood trauma, epigenetic changes and behavioral outcomes. While these two genes are prominent examples, other studies focused on a broad spectrum of monoaminergic or serotonergic genes, neurotrophic factors, genes involved in immune function and other neuropeptides.¹³⁴ Similar again to psychiatric genetics, the field soon started to embrace large-scale epigenome-wide association studies (EWAS), moving away from hypothesis driven studies. The first EWAS on childhood abuse emerged together with the availability of affordable microarray technology, profiling a limited set of DNAm sites across the human genome. However, these studies have not yet yielded consistent results at the level of individual sites or genes so far.¹²⁹ The underlying reasons are likely similar to limitations observed in candidate gene studies and include limited sample sizes, differences in tissue sources, (ancestry/ethnic) sample composition and data processing, and probe reliabilities.

Nevertheless, EWAS have the potential to assess higher-level biological processes beyond individual genes and it seems that a more coherent picture at the level of gene pathways emerged recently. Specifically, evidence for epigenetic changes in the immune system, neural, developmental and cardiovascular processes, and stress signaling have been reported more consistently.¹²⁹ Notably, a limited set of studies used post-mortem brain tissue of individuals with a history of childhood abuse. Among them, Lutz et al., 2017 found differential methylation in the anterior cingulate cortex pointing to a dysregulation of oligodendrocyte and myelin-related pathways. In addition, transcriptomic studies provided evidence for an overlap of gene expression and DNAm signatures observed in rodent models of maternal care.¹³⁵

Substantial work has suggested that early developmental abuse and neglect impair development and increase risk for later outcomes – with perhaps among the most clear examples being studies of maltreated children from Romanian orphanages.¹³⁶ Interestingly, epigenetic research in other cohorts has also indicated that maltreatment before the age of 3 predicts differential DNAm patterns,⁴³ which may increase the risk for posttraumatic stress severity.¹³⁷ These effects appear to decrease as a child ages, requiring more severe forms of adversity exposure for effects in DNAm to occur by even middle childhood.⁴³ Additionally, this early critical period appears to hold for exposure to sexual and physical abuse at age 3.5 by accelerating epigenetic aging by a month for boys and two months for girls using Hannum’s epigenetic clock.¹³⁸ While we do not yet fully understand what biological processes fully underlie these early sensitive periods, it is thought that multiple aspects of sensory-motor, cognitive, and emotional brain development and regulatory circuits are all greatly impacted by maltreatment during these first years of life.

The studies above highlight both the potential and the limitations of epigenetic studies, particularly those investigating DNAm, in human populations. Thought notably other epigenetic factors beyond DNAm, such as histone regulation and noncoding RNA expression, certainly contribute to observed phenotypic expressions. However those extend beyond the scope of this review. Below, we will discuss several key factors that determine the outcome and reproducibility of future epigenetic studies on childhood abuse.

The human environment is a highly complex system with a myriad of potentially synergistic or counteracting factors. Past attempts to isolate specific events in time, such as childhood abuse, seem far from the reality of this complex environment. A better approximation will require a more comprehensive assessment of the environment beyond individual questionnaires and methods to integrate complex environmental conditions in biological studies. In line with this notion are limitations on phenotypic heterogeneity and how to accurately measure childhood abuse and other factors, as different questionnaires and study designs can have a strong impact on outcome measures. This is particularly relevant for epigenetic studies with respect to the type and timing of trauma exposure. Trauma exposure is not uniform and recent studies suggest that type and timing of trauma may have a substantial impact on epigenetic programming and later psychopathology.^43,52

Given that the majority of human studies so far utilized peripheral tissue, it will be even more important to investigate the effects of tissue specificity and how changes in heterogeneous tissues like blood can influence DNAm profiles. To this end, the use of single-cell technologies will help to increase specificity of DNAm studies in these tissue types. While ideally well-characterized post-mortem brain tissue should be used to study the mechanistic underpinnings of how childhood abuse affects brain function, the paucity of such tissue has critically limited the feasibility of these studies. Similarly pivotal are improved cohort designs, with a much stronger emphasis on longitudinal studies in the future to determine the dynamic changes occurring in the epigenome. Finally, similar to childhood maltreatment occurring in a complex system of environmental factors, DNAm changes arise in the context of other biological factors, specifically genetic variation and complementary epigenetic mechanisms that determine epigenetic profiles and drive epigenetic susceptibility. Despite the challenging problems outlined above, epigenetic studies with a focus on childhood abuse are promising as they not only provide a conceptual framework for the underlying biological mechanisms but also open avenues for much needed new interventions.

Conclusions

Childhood maltreatment remains far too common, and incorporates a range of active and passive factors, from abuse to neglect, to the impacts of broader structural and systemic adversity. Despite the effects of childhood maltreatment and adversity on a wide range of adult physical and psychological negative outcomes, it is clear that not all individuals respond similarly. Understanding the differential mechanisms contributing to risk vs. resilience in the face of developmental adversity is critical to improving preventions, treatment, and policy recommendations.

Evidently, modern large-scale approaches to understanding genomic, epigenomic transcriptomic, and proteomics will transform our understanding of biomarkers and biological factors that differentiate risk vs. resilience in the aftermath of childhood trauma. This review has examined some of the different examples in recent years that have begun to examine specific stress-related pathways mediating childhood maltreatment effects, as well as new findings beginning to emerge from the largest multi-omic international consortia.

Understanding how gene (aka biology) and environment interact is one of the most important and fascinating scientific problems of our day, and it also provides a path forward to improving the lives of individuals. Elucidation of the biology of childhood adversity and its effects are certainly only part of the equation for reducing these unfortunate experiences. However, they do provide one part of the solution for understanding cycles of risk, and potentially a path for how biological research, health care, and policy can come together for a better future.

Highlights.

• Understanding the differential biological mechanisms contributing to risk vs. resilience in the face of developmental adversity is critical to improving preventions, treatments, and policy recommendations.

• This review provides an overview of childhood abuse, neglect, maltreatment, threat, and toxic stress, and the effects of these forms of adversity on the developing body, brain, and behavior.

• The review examines examples from the current literature of genomic, epigenomic, transcriptomic, and proteomic discoveries and biomarkers that may help to understand risk and resilience related to childhood trauma exposure.

• Ongoing and future work aimed at understanding the biology of childhood adversity will help to provide targets for intervention and prevention.