Background
Child maltreatment (CM) can have detrimental psychological and biological consequences for the affected individuals and may even affect the next generation via behavioral (e.g., parenting) but probably also via genomic and non-genomic biological pathways. An elevated risk for diverse health problems in adulthood has been described in individuals with a history of CM; for example, psychopathology, cardiovascular disease, cancer, or even premature mortality [
1‐
3]. One possible underlying mechanism between CM and later physical health problems may be the dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, the body’s major stress system. Functional HPA axis activity is essential for the regulation of metabolic processes in order to maintain physiological homeostasis, but its functioning can be dysregulated upon chronic stress such as CM, in particular when it occurs during sensitive developmental periods.
A stress-induced activation of the HPA axis leads to the release of corticotropin-releasing hormone (CRH) from the hypothalamus, which causes the secretion of adrenocorticotropic hormone (ACTH), and in turn results in the secretion of glucocorticoids (mainly cortisol) from the adrenal cortex. Besides cortisol, the steroid hormone dehydroepiandrosterone (DHEA) and its sulfate ester DHEA-S are released upon stress. There is growing interest in the investigation of both cortisol and DHEA, as their interplay impacts many physiological systems via genomic and non-genomic mechanisms, but they seem to have opposing biological, neurological and immune-related functions (for a review see [
4]). Hence, taking both hormones into account may be a more sensitive index for the regulation of HPA axis activity [
4].
A stress-induced increase of
cortisol is crucial in response to environmental stressors as it results in the provision of energy by increasing the release of glucose and inhibition of non-essential functions (e.g., reproduction, growth; [
4]). However, persistently altered levels of cortisol can have health-threatening effects (for more details see [
5,
6]).
DHEA can antagonize some of the effects of cortisol [
7]. Indeed, neuroprotective, antioxidant, anti-inflammatory, and immune-modulatory effects have been described for DHEA (for details on metabolic pathways and physiological function of DHEA see e.g. [
8‐
11]). It is related to a broad range of physiological processes, as DHEA is a precursor of many other steroid hormones (e.g., testosterone, estradiol; [
8‐
10,
12]). Both cortisol and DHEA have been suggested as biomarkers for the regulation of HPA axis activity and related to psychiatric diseases such as depression, anxiety, PTSD, dementia, eating disorder, and externalizing problems during childhood, adolescence and even adulthood (for reviews see [
4,
8]).
CM as a chronic and traumatic stressor can lead to a persistent dysregulation of the HPA axis activity, resulting in altered
cortisol secretion in affected individuals [
13] and their children [
14]. However, findings on long-term effects of CM on cortisol measured in blood, urine and saliva are mixed (e.g., [
15‐
18]). By measuring cortisol in
hair as a reliable measure of chronic HPA axis activity [
19‐
21], several studies found an association of CM with lower levels of cortisol: an association of CM and decreased hair cortisol levels, independent of current major depression diagnosis was reported in a study with depressed patients (27 women and 16 men; mean age 41.7) and healthy age- and sex-matched controls [
22]. Another study with 55 healthy college students (18 to 24 years of age) also reported a negative association of adverse childhood experiences and hair cortisol [
23].
In contrast to cortisol, only few studies considered the role of
DHEA in the aftermath of CM so far. These studies investigated CM in PTSD patients with mixed results: whereas one study observed an association of CM with elevated concentrations of DHEA in blood plasma of PTSD patients [
24], another study investigating a sample of adult smokers with and without PTSD did not find a significant effect of CM on DHEA concentrations in blood serum [
18]. In the aftermath of sexual trauma, PTSD patients showed significantly reduced DHEA levels compared to age-matched controls in two studies measuring steroids in saliva [
25] and blood [
26]. As all reported studies investigated DHEA in blood, the assessment of DHEA in hair as a cumulative measure over time might clarify the existing inconsistent findings.
In addition, CM may also impact the affected individual’s
offspring: children of mothers with a history of CM showed higher salivary cortisol concentrations compared to children of control mothers [
14]. An effect of maternal traumatic experiences in adulthood on the offspring’s HPA axis regulation was confirmed in infants of mothers living with PTSD, with lower cortisol concentrations in saliva [
27] and blood serum [
28] of the offspring.
Parental CM may affect the offspring’s health and development by the dynamic interplay between environmental (e.g. parenting behavior, health behavior like smoking), and biological mechanisms (including genetic, epigenetic and non-genomic imprinting via the endocrine milieu during pregnancy). The offspring’s biological constitution may be influenced during the prenatal period already, which is one of the most sensitive periods for development [
29]. On the endocrine level, a dysfunctional activity of the maternal HPA axis (e.g., due to a history of CM) may influence the developing HPA axis and endocrine milieu of the fetus during gestation with persisting effects on the offspring’s HPA axis regulation and health. Indeed, during pregnancy maternal, fetal, and placental endocrine systems strongly interact in maintaining intrauterine homeostasis, ensuring maturation of vital organs in the developing fetus and the timing of parturition [
30]. Whereas cortisol is essential for fetal development, as it stimulates the differentiation and maturation of vital organ systems (in particular the fetal lungs) before birth (for reviews see [
30‐
33]), DHEA (and its sulfated form DHEAS) is critical for the physical and neural development, as it is the main source for estrogen synthesis in the placenta, which plays a pivotal role in the timing of parturition and thus for the health and survival of the newborn [
30,
32]. Furthermore, DHEA is essential for brain development and is involved in neurite growth, neurogenesis, neuronal survival, and apoptosis (for reviews see [
8,
9]).
Although steroidogenesis in maternal, placental and fetal compartments is interdependent (for a review see [
34]), the fetal HPA axis represents a separate biological system that is worth studying, as cortisol and DHEA, whether of fetal or maternal origin, are likely to influence development, and potentially have long-term effects on HPA function.
The analysis of cortisol and DHEA in scalp hair presents a new and promising methodological approach to measure these endocrine steroids in a non-invasive manner. Contrary to measuring acute circulating levels of cortisol and DHEA in body fluids such as blood, urine, or saliva, the assessment of cortisol and DHEA concentrations in hair allows for the retrospective analysis of the total exposure to these steroids over time [
19‐
21]. Furthermore, the analysis of steroids in hair overcomes some of the methodological limitations when using body fluids [
21] such as circadian fluctuations throughout the day.
The aim of this study was to investigate the influence of CM on the regulation of the HPA axis during pregnancy by measuring cortisol and DHEA in hair samples of postpartum women shortly after parturition. The analyzed 3 cm hair segment of these mothers was assumed to reflect the last trimester of pregnancy. Following previous findings on CM-related reductions in hair cortisol levels [
22,
23], we expected a negative association of the severity of maternal CM with
cortisol concentrations in the last trimester of pregnancy measured in hair of postpartum mothers. Furthermore, we also explored an association of CM and
DHEA levels in the last trimester of pregnancy measured in the hair of postpartum women. In addition, we tested a potential association of the mothers’ CM experiences and her offspring’s HPA axis in an explorative analysis. Cortisol and DHEA were measured in the newborn’s hair, which were assumed to reflect fetal HPA axis activity before parturition.
Discussion
In contrast to previous findings, hair cortisol levels were not associated with the amount of CM experienced in our sample of postpartum women. Prior studies reported a negative association between CM and hair cortisol [
22,
23]. However, our results are in line with another study reporting no such association [
50]. However, none of these studies investigated steroids during pregnancy. Differing results may be explained by the physiological and endocrine alterations in the last trimester of pregnancy, which might mask the replicated finding of lower cortisol in the aftermath of CM. In short, CRH is now also secreted by the fetal adrenal gland, the placenta and decidua, and consequently contributes to an exponential increase of maternal plasma CRH. Due to elevated CRH as well as a longer half-life of cortisol in plasma (as a result of estrogen-induced reduction of cortisol catabolism in the liver) cortisol levels increase with a peak during the third trimester of pregnancy (for more details see [
33]). The one study investigating hair cortisol in relation to CM in pregnant women confirmed our results of no association in Caucasian pregnant women with relatively low levels of CM [
51].
Our finding of elevated
DHEA levels in hair is in accordance to some studies measuring DHEA in
body fluids of PTSD patients [
24,
52,
53]: Significant amounts of the variation of elevated DHEA were explained by childhood trauma history in women with PTSD [
24]. However, other studies did not find an association of CM with DHEA, but with cortisol in blood plasma and serum of PTSD patients [
18,
54]. Studies investigating DHEA in hair are lacking, thus further studies are needed to clarify effects of CM on chronic DHEA levels.
Furthermore, this study explored the influence of maternal CM on the next generation’s HPA axis regulation via non-genomic effects by influencing the endocrine milieu in utero of the developing fetus. In contrast to previous findings, we did not observe an association of maternal CM and cortisol levels in the newborn’s hair. However, in previous studies that related maternal CM [
14] or PTSD [
27,
55] with lower cortisol concentrations, cortisol was measured in saliva of the offspring at the age of six months or even adulthood. Thus, results are not comparable to our findings regarding prenatal steroid concentrations measured in newborns' hair, as prenatal HPA axis activity differs significantly from HPA axis functioning during childhood or adulthood and does not allow retrospective conclusions about the in-utero environment of the developing fetus.
Regarding
DHEA, higher maternal CM was positively associated with elevated DHEA concentrations in the hair of our newborn sample but only as a non-significant trend (
p = .07), explaining 17% of the variation in hair DHEA. This relationship was significant in a subsample of newborns (
N = 15) whose mothers were interviewed in detail about a broader range of adverse childhood experiences including also witnessing physical violence towards parents or siblings, peer emotional violence, and peer physical violence (see Additional file
2). However, as the sample size of this subsample was small, this finding should be seen as hypothesis-generating for future studies, and thus no final conclusions should be drawn at this stage (see Additional files
3 and
4). The only other study investigating cumulative prenatal DHEA concentrations measured DHEA in the fingernails of newborns whose mothers had experienced stress during pregnancy and report a positive association [
56]. However, they did not examine maternal lifetime stress such as adverse childhood experiences.
This results hint towards a potential effect of maternal childhood experiences on the intrauterine endocrine environment of the developing offspring in late gestation, but presumably also during the whole gestational process. Indeed, CM has been associated with altered HPA axis activity during pregnancy of affected mothers [
51]; and a dysfunctional activity of the maternal HPA axis during gestation has been associated with alterations in the offspring’s HPA axis [
57,
58]. Altered levels of steroids in the mother during pregnancy might influence the fetal HPA axis directly by crossing the placental barrier, and indirectly by stimulating placental CRH production and reducing uteroplacental blood flow [
59]. Thus, the developing fetus might achieve and incorporate information about its later environment via intrauterine maternal-placental-fetal communication – a phenomenon referred to as “fetal programming” [
60,
61]. By programming of the fetal HPA axis, the offspring may adapt to the environment in which they will be born in anticipation of an exposure to similar environmental conditions in postnatal life [
57,
62]. However, results in this study regarding a potential transgenerational effect of CM were only significant on a trend level or were only present in a subsample for which also information on witnessing physical violence towards parents or siblings, peer emotional violence, and peer physical violence was available. Thus, our findings need replication in larger samples but can be used for the formulation of hypotheses and calculation of effect sizes for future studies.
Strengths and limitations
One major limitation of this study is the relatively small sample of newborns with sufficient amount of hair for analysis due to the limited availability of sufficient hair for analysis. Studies on neonatal hair growth suggested that neonatal hair fibers at birth reflect the metabolic activity of prenatal development in the third trimester of pregnancy [
41]. Nevertheless, length of maternal and newborn hair samples differed (3 cm in mothers vs. approximately 1 cm in newborns) and might not represent the exact same time period. In addition, cortisol and DHEA concentrations measured in the newborns’ hair may be of fetal and, partly, of maternal origin: Even though fetal exposure to maternal steroids is limited due to the actions of the placental 11β-HSD2 enzyme, which transforms cortisol into its inactive form cortisone and thus provides a partial barrier [
59,
63,
64], a small percentage of maternal glucocorticoids still passes the placenta into the fetal system (for reviews see e.g. [
65]). Furthermore, methodological issues (variability associated with growth rate of hair and inconsistencies in collection of hair samples [
66,
67]) as well as incorporation of steroids in hair [
68‐
73] are still discussed. Besides incorporation via blood during formation of hair follicles, external sources such as sweat or, in the case of newborns, amniotic fluid may also affect hair that has already emerged on the scalp [
67,
71].
Further limitations are related to the psychological data assessment: Whereas the maternal steroids in the 3 cm hair segment reflected the last three months, only the perceived stress in the last month was considered as a covariate in the analyses (assessed with the PSS4, which instructs participants to refer to the past month). Furthermore, the covariate lifetime psychopathology was only assessed in self-report and not with a standardized psychiatric interview.
Strengths of this study are the assessment of steroids in hair as a cumulative measure of HPA axis activity, the consideration of two important steroids (cortisol and DHEA) and the investigation of hair samples of mother-newborn dyads extending existing literature on the CM-related transmission of HPA axis dysregulation, which most often confined analysis of steroid hormones to the offspring generation. Furthermore, the influence of parental behavior on steroid levels in the hair of newborns can be nearly excluded in this study, since hair samples were collected within six days after parturition. Remarkably, we observed an effect of CM on HPA axis activity in a sample of well-educated healthy women with relatively low psychosocial burden (only five women were worried about financial problems due to the child’s birth, four reported cramped living conditions [ratio of number of rooms/number of persons in the flat ≤0.5], and one woman experienced mild physical violence by her intimate partner within the last 8 weeks). However, results can thus only be generalized to a limited degree and studies on high-risk samples are needed.
Implications for research and conclusion
Results of this study revealed an association of increasing cumulative DHEA, but not cortisol levels in maternal hair with rising number of CM experiences. This might reflect long-term alterations in the HPA axis in response to maltreatment experiences in childhood, which extend into pregnancy and highlights the importance to investigate DHEA in addition to cortisol. With due caution, our data also provide some support for the hypothesis that CM may influence the offspring’s prenatal HPA axis, as we observed a positive association of maternal CM and elevated DHEA concentrations in the newborns’ hair, but as a non-significant trend.
Measuring cumulative cortisol and DHEA concentrations in mothers and their newborns shortly after parturition seems promising for the investigation of prenatal biological effects of maternal trauma history on her offspring’s neuroendocrine system, without the confounding effect of parental behavior.
Further studies in larger (high-risk) samples are needed and would allow further analyses (e.g., taking ongoing life circumstances, time between CM and pregnancy, as well as the specific effects of different types of CM into account), in order to better understand the underlying biological mechanisms in the interplay between CM and a potentially altered HPA axis activity. Furthermore, longitudinal studies should investigate whether altered levels of DHEA in late gestation are related to altered patterns of HPA axis activity in childhood and adulthood. Pregnancy is associated with significant physiological alterations of HPA axis functioning and the biological function as well as implications of elevated DHEA in non-pregnant individuals cannot easily be transferred into the third trimester of pregnancy. Thus, long-term effects of increased DHEA levels in late gestation on health outcomes in the mothers and their offspring needs to be addressed in prospective studies and the investigation of hair DHEA levels in non-pregnant women with a history of CM seems warranted.
Acknowledgements
We thank Sarah Wilker for comments on an earlier version of the manuscript and Traudl Hiller (medical technical assistant) for support in data collection.