Table 4.
Childhood adversity and studies leveraging epigenome-wide methylation arrays in samples of adults.
| Author | Sample | Age(s) at methylation assessment | Gender | Ancestry | Tissue type | Adversity measurement | Adversity-related findings |
|---|---|---|---|---|---|---|---|
| Han et al.,87 |
N = 1130; 72% depressed Characterization of sample: Participants resided in the Netherlands; enrolled in the Netherlands Study of Depression and Anxiety |
18–65 years (M = 42 years) |
65% female | Not reported | Whole blood | Retrospectively assessed trauma via the Netherlands Mental Health Survey and Incidence Study | Childhood trauma was positively associated with epigenetic aging within the depressed group; differences between depressed and control groups for DNA methylation age were replicated in postmortem brain samples |
| Houtepen et al.,174 |
N = 85 for discovery sample (buccal); N = 45 for replication sample (blood) Characterization of sample: Enrolled in ALSPAC |
18–69 years (M = 33 years) in discovery sample; 19–45 years (M = 28 years) in replication sample | 51% female in discovery sample, 80% female in replication sample | 100% Caucasian in discovery sample, 38% Caucasian in replication sample | Buccal cells in discovery sample; Whole blood in replication sample | Retrospectively assessed physical, sexual, emotional abuse, emotional and physical neglect via the CTQ short form; age of onset via the Early Trauma Inventory | Within the discovery (buccal cell) sample, although no sites survived corrections for multiple testing, CM was associated with increased methylation at KITLG locus; KITLG methylation mediated link between CM and cortisol reactivity; within the replication (blood) sample, CM was associated with increased KITLG methylation and cortisol reactivity only in Caucasians; there was no influence of age of onset on methylation |
| Houtepen et al.,82 | N = 780 in ALSPAC, 66% exposed to adversity; N = 552 in NSHD, 68% exposed to adversity | 47 years in ALSPAC; 53 years in NSHD | 100% female | Not reported | Whole blood in ALSPAC; Buccal cells in NSHD | In NSHD, 5 ACEs prospectively measured via interviews and questionnaires by participants’ mothers, including parental physical illness, parental mental illness, parental death, parental separation, and childhood illness; suboptimal maternal bonding and childhood maltreatment were retrospectively self-reported; in ALSPAC 5 additional ACEs retrospectively assessed via questionnaire: physical, sexual, and emotional abuse, physical and emotional neglect for a total of 11 ACEs |
Although no individual CpG sites replicated across cohorts, after correction a total of 97 DMRs were associated with ACE measures in ALSPAC and 134 DMRs were associated with ACE measures in NSHD; even after adjusting for smoking nine differentially methylated regions were associated across both cohorts such that cumulative ACE score associated with methylation variance, as was parental mental illness, parental physical illness, and parental death |
| Khulan et al.,175 |
N = 83 men separated from parents in childhood during wartime and N = 83 non-separated controls Characterization of sample: Enrolled in Helsinki Birth Cohort Study |
M = 64 years for separated group; M = 62.9 years for non-separated group |
0% female | Born in Finland | Whole blood | Childhood wartime parental separation according to the Finnish National Archives’ register | No association of childhood parental separation and methylation; methylation was associated with the development of depressive symptoms; hypomethylated genes included those with roles in brain development, brain function |
| Labonté, Suderman, et al., 2012b176 |
N = 41; n = 25 completed suicide and exposed to severe CM; n = 16 controls with no suicide and no CM; n = 20 completed suicides with no CM, used for validation Characterization of sample: Quebec Suicide Brain Bank |
M = 37.3 years in CM/suicide group; M = 40.6 years in non-CM/suicide group; M = 40.9 years in control group |
0% female | 100% Caucasians of French-Canadian descent | Brain (hippocampal tissue) | Retrospectively assessed sexual abuse, physical abuse and severe neglect via psychological autopsy, including structured interviews/chart reviews using the CECA adapted for psychological autopsies | CM associated with genome-wide promoter epigenetic alterations; 362 differentially methylated promoters identified in individuals with CM compared with controls--248 showed hypermethylation and 114 showed hypomethylation; observed functional clusters of differentially methylated genes were validated against non-CM suicide completers and were shown to be involved in cellular/neuronal plasticity, with potential candidate marker Alsin (ALS2) observed to be differentially methylated |
| Lawn et al.,85 | N = 989 in ALSPAC, 23% exposed to CM; N = 773 in NSHD, 7% exposed to CM | Two timepoints in ALSPAC: 29 and 47 years; 53 years in NSHD | 100% female | Not reported | Whole blood in ALSPAC; Buccal cells in NSHD | In NSHD, 5 ACEs prospectively measured via interviews and questionnaires by participants’ mothers, including parental physical illness, parental mental illness, parental death, parental separation, and childhood illness; suboptimal maternal bonding and childhood maltreatment were retrospectively self-reported; in ALSPAC 5 additional ACEs retrospectively assessed via questionnaire: physical, sexual, and emotional abuse, physical and emotional neglect for a total of 11 ACEs | Sexual abuse associated with higher DNA methylation age in ALSPAC at both timepoints (no sexual abuse data available in NSHD); cumulative adversity was not associated with DNA methylation age in ALSPAC or NSHD |
| Lutz et al.,84 | N = 78 depressed adults who died by suicide; 35% with severe CM | Postmortem | Not reported | Not reported | Anterior cingulate cortex | Severe physical and sexual abuse up to age 15 validated from medical charts, coroner files, and child protective services reports | Hyper- and hypomethylation was detected in the group exposed to abuse compared to the control group; the three most significantly differentially methylated regions intersected with genes directly related to myelin and oligodendrocytes: LINGO3, POU3F1, and ITGB1 |
| Marinova et al.,177 |
N = 45; n = 30 former indentured child laborers, n = 15 controls Characterization of sample: Grew up in rural Switzerland |
M = 75.9 years in experimental group, M = 72.8 years in control group | 47% female in experimental group; 53% female in control group | Not reported | Buccal cells | Retrospectively reported former indentured child laborer status | Differential methylation between the two groups, with the strongest difference in SKAP2 |
| Marzi et al.,178 |
N = 2232 twins; 28% with severe victimization experiences Characterization of sample: Enrolled in Environmental Risk (E-Risk) Longitudinal Study in England and Wales |
18 years | Not reported | Not reported | Whole blood | Prospectively assessed exposure to domestic violence; physical, sexual, emotional abuse and emotional and physical neglect; bullying; family violence; cyber victimization; exposure to crime via dossiers that included information from home visit staff, mothers, children, family doctors, and child protection agencies; child completion of Juvenile Victimization Questionnaire | Methylation associated with adversity exposure overlapped with tobacco smoking, thus could not be differentiated |
| Mehta et al.,186 |
N = 169; n = 108 with no PTSD (31% with CM), n = 61 with PTSD (52% with CM) Characterization of sample: Urban, low-income |
M = 43.25 years in total sample; M = 44.23 years in CM with no PTSD group; M = 39.56 years in CM + PTSD group; M = 43.69 years in no CM + PTSD group |
72% female | 89% African American; 11% other | Whole blood | Physical, sexual, emotional abuse, emotional and physical neglect via the CTQ | Gene-expression profiles of PTSD patients with CM were nearly nonoverlapping with CM-exposed controls; these gene expression changes were associated with methylation changes in the same loci in the CM group; functional annotation analyses supported enrichment of central nervous system development and in immune-related tolerance induction pathways in the PTSD group with CM whereas apoptosis and growth rate networks were enriched in the PTSD group without CM |
| O’Donnell et al.,83 |
N = 188; n = 99 control group, n = 89 Nurse Family Partnership intervention group; 27% exposed to CM; Characterization of sample: Resided in Canada |
27 years | 53% female in control group; 48% female in NFP group | 85% Caucasian in control group; 75% Caucasian in NFP group | Whole blood | Child abuse and neglect retrieved from substantiated reports from child protective services from child ages birth through 15 | CM, and participation in NFP, were associated with variation in methylation at age 27; the magnitude of the association between CM and methylation was reduced when accounting for smoking; no significant association between epigenetic age acceleration and CM or participation in NFP |
| Prados et al.,179 |
N = 189; n = 96 borderline personality disorder patients with high CM; n = 93 depressed patients with low CM Characterization of sample: Participants resided in France and Switzerland |
M = 36.7 years | 78% female | Not reported | Whole blood | Physical, sexual, emotional abuse, emotional and physical neglect via the CTQ | Differential methylation observed in individuals with BPD and high CM compared to individuals with depression and low CM. Patterns of methylation also differed with respect to the severity of CM |
| Robakis et al.,180 | N = 54 pregnant women recruited from psychiatric clinic | M = 32.3 years | 100% female |
65% Caucasian 11% East Asian 4% South Asian 6% Hispanic 2% African American 11% Multiracial |
Buccal cells | Retrospectively assessed physical, sexual, emotional abuse, emotional and physical neglect via the CTQ | CM associated with methylation density in 1580 regions, 162 regions when using FDR < .05; genes included those with metabolic, cellular process, and regulatory functions; CM not associated with mean methylation density over entire captured region of OXTR |
| Roberts et al.,181 |
N = 34; 50% with high CM exposure, 15% with medium CM exposure, 35% with no CM exposure Characterization of sample: Enrolled in the Growing Up Today Study |
23–29 years (M = 26.3 years in no CM group; M = 25.4 years in medium CM group, M = 25.2 years in high CM group) |
100% male | 92% white in no CM group, 100% white in medium CM group, 88% white in high CM group | Sperm | Retrospectively assessed physical, sexual, emotional abuse via the CTQ and Conflict Tactic Scales | Differential methylation between those exposed to CM and those that were not; adulthood trauma exposure and mental health partially mediated the association between CM and methylation |
| Smith et al.,182 | N = 110; n = 25 with PTSD without CM, n = 25 with PTSD and CM, n = 26 with CM, n = 34 without CM |
M = 41.3 years in control group; M = 44.1 years in PTSD group |
40% female in control group; 37% female in PTSD group | 100% African American | Whole blood | Retrospectively assessed sexual, physical, and emotional abuse via the CTQ; stressful life events (e.g., interpersonal stressors, crime, divorce) in the last year or ever via the Stressful Events Questionnaire | No change in global methylation levels in participants with CM or with increased total life stress; methylation was inversely associated with total life stress at one CpG site; no CpG site was associated with CM; in an examination of 60 genes previously identified in the literature, CpG sites in BDNF and CXCL1 were associated with both PTSD and total life stress; differences in methylation profiles as a function of CM did not survive experiment-wide correction for multiple testing |
| Suderman et al.,183 |
N = 40; 30% exposed to CM Characterization of sample: From 1958 British Birth Cohort |
45 years | 100% male | Not reported | Whole blood and lymphoblast cell lines |
Retrospectively assessed verbal, emotional, physical, and sexual abuse via a questionnaire including items derived from the Parental Bonding Instrument, the British National Survey of Health and Development and the US National Comorbidity Survey |
Differential methylation associated with CM – both hyper- and hypomethylation; of the differentially methylated genes that perform some regulatory function, most were hypomethylated in CM sample |
| Tamman et al.,86 | N = 212 veterans | 22–93 years (in larger sample of 1,135 veterans) | 100% male | 100% European American | Saliva | Retrospectively assessed trauma, including child sexual abuse, via the Trauma History Screen | Childhood sexual abuse associated with increased DNA methylation age; greater number of lifetime traumas associated with increased DNA methylation age |
| Zannas et al.,184 |
N = 392 Characterization of sample: Participants enrolled in the Grady Trauma Project |
18–77 years (M = 41.3 years) |
71% female | 100% African American | Whole blood | Retrospectively assessed physical, sexual, emotional abuse, emotional and physical neglect via the CTQ | Cumulative lifetime stress (including childhood events) was associated with acceleration of epigenetic aging; childhood abuse exposure alone was not significantly associated with markers of epigenetic aging |
| Zhang et al.,185 | N = 518, 29% with childhood adversity, 52% with diagnosis of alcohol dependence | Mean age ranged from 33–43 years across groups stratified by alcohol dependence and childhood adversity | 56% female | 54% African American; 46% European American | Whole blood | Exposure to childhood adversity assessed using four questions from the Semi-structured Assessment for Drug Dependence and Alcoholism | Childhood adversity associated with hypermethylation of seven individual CpG sites (ALDH1A1, CART, CHRNA5, HTR1B, OPRL1, PENK, and RGS19) in European American participants with and without alcohol dependence; associations of childhood adversity and methylation in African American participants were not replicated across those with and without alcohol dependence |
CM childhood maltreatment, CTQ Childhood Trauma Questionnaire, ALSPAC Avon Longitudinal Study of Parents and Children, NSHD MRC National Survey of Health and Development, PTSD post-traumatic stress disorder, NFP Nurse Family Partnership, BPD borderline personality disorder, FDR False Discovery Rate.