EFFECTS OF EARLY AND RECENT ADVERSE EXPERIENCES ON ADRENAL RESPONSE TO PSYCHOSOCIAL STRESS IN DEPRESSED ADOLESCENTS

Abstract

Background

As observed in depressed adults, there is considerable variability in the degree and direction of hypothalamic-pituitary-adrenal (HPA) dysfunction in depressed adolescents. The variability in HPA findings may be attributed to experiential factors.

Methods

A modified version of a standard psychosocial stressor used in adults, the Trier Social Stress Test (TSST), was administered to 30 adolescents with major depressive disorder and 25 healthy adolescent volunteers. Cortisol concentrations were measured in saliva samples collected before and after the stressor. Information also was gathered on early and recent adverse experiences with standard interviews.

Results

Participants from both groups had increased cortisol secretion in response to TSST. Compared with controls, depressed subjects showed more elevated and prolonged cortisol secretion in response to TSST. The combination of early-life adversity and high levels of chronic stress during adolescence was the most powerful predictor of enhanced adrenal response to the TSST.

Conclusions

These results support previous findings on the role of experiential factors on HPA response to stress and in the development of mood disorders. Dissection of the heterogeneous pathophysiology of adolescent depression will assist in developing more specific interventions for different subgroups of patients.

INTRODUCTION

Depressive illness is a major public health problem, and is the leading cause of disability worldwide in the 15- to 44-year age group (1). Elevated risk for the disorder begins in the early teens and continues to rise in a linear fashion throughout adolescence, with lifetime rates estimated to range from 15% to 25% by late adolescence (2–4). Prospective epidemiological and clinical studies consistently reported that juvenile depression is a recurrent illness, with episodes continuing in adult life. In addition to recurrent depressive episodes into adult life, there are ongoing psychosocial difficulties, including disruption in interpersonal relationships, early pregnancy, lower education, poor occupational functioning, unemployment and elevated risk for suicidal behavior (4–6). Some studies also reported high rates of medical problems and healthcare service utilization (5,6). Given the growing economic and psychosocial burden associated with juvenile depression, determination of risk factors for the disorder and its pathophysiology are of paramount importance toward the development of effective treatment and preventive strategies.

Over the past three decades, compelling evidence has accrued indicating that stressful experiences play a prominent role in the development and maintenance of depression in adolescents and adults (7–10). The increased vulnerability to depressive episodes under stressful situations may be mediated by the hypothalamic-pituitary-adrenal (HPA) system. The HPA axis is one of the physiological systems that evolved in mammals to help focus and sustain emotional, cognitive, behavioral and metabolic activity in response to conditions of threat (11,12). Upon exposure to stress, neurons in the hypothalamic paraventricular nucleus (PVN) secrete corticotropin-releasing hormone (CRH) which then stimulates the production of adrenocorticotropic hormone (ACTH) from the anterior pituitary. ACTH, in turn, induces the release of glucocorticoids from the adrenal cortex. Several brain circuits modulate HPA activity. The hippocampus and prefrontal cortex (PFC) inhibit HPA axis, whereas the amygdala and monoaminergic input from the brainstem stimulate CRH neurons in the PVN. Glucocorticoids exert negative feedback on the HPA axis by regulating hippocampal and PVN neurons. Prolonged glucocorticoid exposure has adverse effects on the hippocampus and PFC, resulting in impaired inhibitory control of the HPA axis (11,12).

Although there is substantial evidence of altered HPA activity in depression, particularly in adults, there is considerable variability in the degree and direction of HPA dysfunction (13, 14). Recent research has highlighted the importance of experiential factors in explaining the heterogeneity in neurobiological correlates of depression (15). Specifically, many investigations focused on early-life adverse experiences in examining the association between depression and HPA dysregulation (16–19). There is considerable evidence that adverse experiences in early-life, such as parental loss, physical or emotional neglect, and physical or sexual abuse, lead to a higher risk for developing depression (20–21). The developing central nervous system is highly sensitive to adverse experiences, and early-life stress leads to increased stress reactivity and alterations in the aforementioned neuronal circuits that persist into adult life. The behavioral effects of sustained CRH at the extra-hypothalamic regions may lead to negative mood, contributing to the development and maintenance of depression (15). Research indicates, however, that the long-term mood and neurobiological changes associated with early-life stress can be modified by familial/genetic factors and the quality of subsequent environment (22–25).

Little information is available on how developmental factors interact with early-life experiences and other modifying factors, such as the subsequent environment, to produce HPA changes. In a pilot study, Kaufman and colleagues measured pituitary-adrenal responses to CRH in three groups of children: children who were abused and also met criteria for depression, non-abused children with depression and children with no history of abuse or psychiatric disorder (17). Abused children with depression had significantly more ACTH secretion in response to CRH challenge compared with non-abused depressed children and controls. Non-abused children with depression did not differ significantly from controls. In exploratory analyses within the abused cohort of depressed children, only those children who were living under persistent adverse conditions exhibited enhanced ACTH response. Heim et al. (18) studied 49 women classified into four groups: women with no history of childhood abuse or psychiatric disorder, women with current depression but no evidence of childhood abuse, women with history of childhood abuse but no depression, and women with depression and history of childhood abuse. They were administered a standard mild psychosocial stressor, and pituitary-adrenal responses were measured. Childhood abuse, regardless of depression status, increased ACTH and cortisol secretion in response to the stressor. The effect of childhood abuse on HPA activity was further enhanced with trauma in adulthood (23).

Besides the two studies in children and adults (17,23), the interaction between early-life stress and subsequent environment on HPA activity has not been evaluated. The current study extends these findings by systematic evaluation of early-life and recent stress on HPA response to a psychosocial stressor in depressed and non-depressed adolescents. Previous studies in depressed and maltreated youth used pharmacological/neuroendocrine challenges to assess the HPA response (14,16,17). These tests do not take into account the supra-hypothalamic factors. Given that supra-hypothalamic circuits exert important influences on HPA functioning (11,12), psychological stress challenges offer the advantage of reflecting the endogenous activity of the entire HPA system (26).

METHODS AND MATERIALS

Participants

With approval from the Institutional Review Board, 30 adolescents with depression and 25 controls were recruited. The depressed adolescents met criteria for major depressive disorder (MDD), with a minimum duration of four weeks and a score of ≥15 on the first 17-items of the Hamilton Depression Rating Scale (HDRS,27). Adolescents with a current or prior history of mania, hypomania, substance use disorder symptoms, schizophrenia, schizoaffective disorder or autism were excluded from the study. Subjects also were excluded if there was a family history of bipolar disorder. All participants were free from psychotropic agents for at least eight weeks. Controls were free from any type of psychopathology in their lifetime. Controls were not included in the study if any first-degree relative had history of a major psychiatric disorder. Participants were medically healthy and free from alcohol/illicit drug use, as determined by physical examination, laboratory investigations and urine drug screens.

Diagnostic Evaluation

The diagnosis of MDD and other psychiatric disorders was based on a semi-structured instrument, the Schedule for Affective Disorders and Schizophrenia for School-Age Children - the Present and Lifetime Version (K-SADS-PL,28). The K-SADS-PL was administered to the adolescent and parent, and summary scores were tabulated. The Children’s Global Assessment Scale (CGAS,29), a global psychosocial functioning measure, also was completed. The participants completed the Beck Depression Inventory (BDI,30). The mother was interviewed regarding major psychiatric disorders in all first-degree relatives of the adolescent subject using the Family History-Research Diagnostic Criteria (FH-RDC,31).

Information on Early-Life Adversity

Information on early-life adversity was obtained from the adolescent and parent with a semi-structured interview, the Childhood Adversity Interview (32). Seven types of adversity including separation/loss, life-threatening illness/injury, physical neglect, emotional abuse/assault, physical abuse/assault, witnessing violence, and sexual abuse/assault were assessed. Information from both informants was combined for summary ratings for each domain (1 = none, 5 = most severe). Adversity score was tabulated from the sum of all seven ratings.

Information on Recent Adversity

Information on chronic stress and major stressful life events during adolescence was obtained through a semi-structured interview (33). Ten content areas (including family relationships, independence from the family, close friendships, romantic relationships, social life, school, work, finances, health of subject, and health of family members) were assessed. The adolescent was interviewed on the quality of relationships and performance in each domain within the past 6 months, and ratings were given for the magnitude of stress using objective criteria (1 = not at all stressful, 5 = extremely stressful).

After obtaining information on chronic stress in each domain, participants were probed systematically about the occurrence and timing of acute life events. Narrative summaries of the event and surrounding context were presented to a group of trained raters. The raters were blind to the participant’s diagnostic status and reaction to the stressor. Consensus group ratings were given for the degree of stress (1 = not at all stressful, 5 = extremely stressful) for each event, and whether the event was a positive, neutral or negative experience under the given circumstances. Only events that were considered negative were included in the analyses, and a summary score of stress impact from the negative events was tabulated.

Baseline Adrenal Activity and Adrenal Response to Psychosocial Stress

Baseline saliva samples were collected at 30-minute intervals for two hours (5 samples). For stress induction, a standardized psychosocial stress protocol that was shown to reliably induce HPA activity in children and adults was employed (34,35). The task involved a 5-minute preparation period for a public-speaking task that was completed subsequently over a 5-minute period and followed by an arithmetic task for 5 more minutes in front of an audience. Details of the protocol are provided in the On-line Supplement. Saliva samples were obtained immediately after the task and at 10-minute intervals for 60 minutes (7 samples). The time of HPA assessment was in the late afternoon/early evening, and was based on prior data suggesting a delayed circadian phase in adolescents (36).

Cortisol Assay and Primary Variables of Interest

A radioimmunoassay procedure was employed for the cortisol assay (37). Pre-stress cortisol value was computed by determining the mean of last two baseline samples. The primary dependent variable was net peak value, which was calculated by subtracting the pre-stress value from the highest value following the stress procedure. Net post-stress adrenal response was the secondary dependent variable. This measure/metric was determined by subtracting the pre-stress value from each post-stress sample and these summary scores were used to calculate the area under the curve (AUC) using the trapezoidal rule.

Statistical Analysis

Analysis of variance with repeated measures (with Greenhouse-Geisser correction) was utilized for computing cortisol changes in response to the stressor. Linear multiple regression analyses were performed for determining predictors of HPA response. Prior to computing regression analyses, the tolerance/variance inflation factor (VIF) was examined for effects of inter-correlations among the potential predictors, and these effects were accounted for in the regression model (38; also see On-line Supplement). The cortisol values were log-transformed prior to analysis. For group comparisons, the chi-square was used to analyze categorical variables and Student’s t test for continuous variables.

RESULTS

Demographic and Clinical Characteristics of the Sample

Demographic and clinical features are outlined in Table 1. The groups did not differ significantly with respect to age, gender or race/ethnicity. Depressed adolescents scored significantly higher on the BDI and HDRS, but lower on CGAS, than controls. Participants with depression scored significantly higher on early-life adversity and recent chronic stress. However, they did not differ significantly on the number of negative life events or magnitude of stress.

Table 1.

Demographic and clinical characteristics (mean ± SD) of the sample

	Control (n = 25)	Depressed (n = 30)	p
Age (years)	15.0 ± 1.5	15.1 ± 1.4	NS
Gender (M/F)	12/13	13/17	NS
Ethnicity (AA/AS/CC/HS)	3/5/11/6	3/5/14/8	NS
BDI score	3.0 ± 3.4	23.4 ± 8.0	.0001
HDRS score	0.7 ± 0.8	20.1 ± 4.9	.0001
CGAS score	84.5 ± 6.1	53.9 ± 8.6	.0001
Early-life adversity score	9.6 ± 2.0	11.9 ± 2.8	.002
Chronic stress score (past 6 months)	20.1 ± 3.8	27.3 ± 4.4	.0001
Number of negative life events (past 6 months)	2.0 ± 1.4	2.7 ± 2.1	NS
Stress impact score from the negative events	4.0 ± 3.0	6.2 ± 5.2	NS
Anxiety disorder(s) diagnosis	--	9 (30.0%)	--
Disruptive disorder(s) diagnosis	--	4 (13.3%)	--
Parental depression	--	13 (43.3%)	--

NS = non-significant

AA = African-American; AS = Asian; CC = Caucasian; and HS = Hispanic

BDI = Beck Depression Inventory

HDRS = Hamilton Depression Rating Scale

CGAS = Children’s Global Assessment Scale

Baseline Cortisol Secretion and Cortisol Response to Psychosocial Stress

Serial cortisol secretory patterns prior to, and following, the stress procedure are depicted in Figure 1. The stressor was effective in inducing cortisol secretion in both groups (main effect of time: F_2.92 = 67.47, p ≤ .0001). Depressed adolescents exhibited higher (main effect of group: F_1,53 = 9.41, p ≤ .005) and more prolonged (group × time interaction: F_2.92 = 5.34, p ≤ .005) cortisol response to the stressor than controls.

Figure 1.

Mean (and SEM) salivary cortisol values at baseline and following psychosocial stress in controls (n = 25) and depressed adolescents (n = 30). Asterisks depict significant group differences in post-stress adrenal response after controlling for pre-stress values.

Prediction of Stress Responsiveness in Adolescents

In order to determine the predictors of adrenal response, a multiple linear regression procedure was used. Net peak cortisol secretion was the primary dependent variable (see Table 2). Demographic variables did not predict adrenal response. Among clinical variables, anxiety disorder diagnosis predicted net peak cortisol response. Among experiential factors, early-life adversity and chronic stress during adolescence were significant predictors. The same pattern occurred when net AUC was used as the dependent variable.

Table 2.

Predictors of net peak salivary cortisol concentrations in response to the TSST in depressed adolescents and controls (N = 55)

Predictor variables	T*	R2	Adj. R2	ΔR2	ΔF	Δp	Std. β	t	p
Model 1: demographic factors		.01	−.04	.01	0.24	NS
age	0.83						−.00	−0.03	NS
gender	0.98						−.09	−0.67	NS
ethnicity	0.83						.08	0.54	NS
Model 2: clinical factors**		.14	.04	.14	1.35	NS
age	0.83						−.02	−0.24	NS
gender	0.97						−.04	−0.62	NS
ethnicity	0.79						.04	0.65	NS
depression diagnosis	0.12						.08	1.63	NS
HDRS score	0.12						.06	1.19	NS
anxiety disorder	0.82						.16	2.47	.02
Model 3: experiential factors**		.34	.21	.20	4.44	.008
age	0.75						−.01	−0.12	NS
gender	0.90						−.06	−0.94	NS
ethnicity	0.76						.03	0.46	NS
depression diagnosis	0.09						.01	0.26	NS
HDRS score	0.11						.02	0.39	NS
anxiety disorder	0.69						.12	2.00	.05
early-life adversity	0.57						.21	3.80	.001
chronic stress (past 6 months)	0.41						.15	2.86	.01
impact of negative stress events	0.68						.01	0.12	NS

^*T = Tolerance (1−R²). A Tolerance value of less than 0.20 indicates significant inter-correlation among variables

^**Ridge regression was performed to adjust for multicollinearity

HDRS = Hamilton Depression Rating Scale

Model 1: k = 0, F = 0.24, df = 3,51, NS; Model 2: k = 1.0, F = 1.35, df = 6,48, NS

Model 3: k = 1.0, F = 2.57; df = 9,45, p ≤ .05

The significant predictors were then examined further for their unique contribution to net peak cortisol response (see Table 3). Although depression was not a significant predictor in the previous analysis, it was still included because of significant differences in primary predictor variables and HPA response between depressed and control groups. Because both early-life adversity and chronic stress predicted adrenal response, an interaction term combining these two variables also was included in the regression model. The interaction of early-life adversity and chronic stress during adolescence was the best predictor of adrenal response to the stressor. Graphical representation of the combined effects of early-life adversity and recent chronic stress on net peak cortisol secretion is depicted in Figure 2. To illustrate, early-life adversity and chronic stress scores were stratified (values above the mean + 1 SD in controls were considered as high).

Table 3.

Significant predictors of net peak salivary cortisol concentrations in response to the TSST in depressed adolescents and controls (N = 55)

Predictor variables	T*	R2	Adj. R2	ΔR2	ΔF	Δp	Std. β	t	p
Model 1: clinical factors		.16	.13	.16	4.95	.02
depression diagnosis	0.84						.15	1.08	NS
anxiety disorder	0.84						.32	2.27	.03
Model 2: experiential factors		.40	.35	.24	10.03	.0001
depression diagnosis	0.51						.20	1.29	NS
anxiety disorder	0.77						.21	1.68	.10
early-life adversity	0.63						.38	2.76	.008
chronic stress (past 6 months)	0.47						.34	2.13	.04
Model 3: interaction**		.32	.25	.32	4.65	.001
depression diagnosis	0.49						.02	0.54	NS
anxiety disorder	0.76						.08	2.18	.03
early-life adversity	0.03						.12	3.80	.001
chronic stress (past 6 months)	0.04						.09	2.82	.01
early adversity × chronic stress	0.01						.13	5.38	.001

^*T = Tolerance (1−R²). A Tolerance value of less than 0.20 indicates significant inter-correlation among variables

^**Ridge regression was performed to adjust for multicollinearity

Model 1: k = 0; F = 4.95; df = 2, 52; p ≤ .05

Model 2: k = 0; F = 8.35; df = 4, 50; p ≤ .0001

Model 3: k = 2.0; F = 4.65; df = 5, 49; p ≤ .001

Figure 2.

Net peak salivary cortisol concentrations in response to psychosocial stress in adolescents with low and high levels of early-life adversity and chronic stress in the past 6 months [low early-life adversity--low recent chronic stress (n = 27); low early-life adversity--high recent chronic stress (n = 9); high early-life adversity--low recent chronic stress (n = 5); and high early-life adversity--high recent chronic stress (n = 14)]. The high early-life adversity--high recent chronic stress group differed significantly from the two low early-life adversity groups (F_3,51 = 9.11, p ≤ .0001). The other groups did not differ significantly from each other.

DISCUSSION

Depressed adolescents manifested higher and more prolonged adrenal response to a psychosocial stressor than healthy volunteers. The results suggest that in adolescents, as in children and adults, experiential factors influence HPA regulation (16–18). HPA response to the stressor was highest in those who had a combination of early-life adversity and high levels of chronic stress during adolescence (17,23).

Consistent with the findings in humans, animal research has demonstrated that early-life adverse experiences not only lead to behavioral consequences, but to also long-term dysregulation of the HPA axis (39–41). However, it is not clear when during development the behavioral and neuroendocrine abnormalities begin to manifest and persist. In contrast to findings in adults, depressed children with early-life adversity did not manifest altered pituitary-adrenal responses if there was no ongoing adversity (17,23). In the current study, the effect of early-life adversity on adrenal response was moderated by chronic stress during adolescence. As observed in this study and other investigations, early-life adversity, ongoing stressors and psychopathology explain only a portion of the variance in HPA responsiveness. Factors that likely contribute additional variance in HPA responsiveness include genetic predisposition, temperament/personality traits, physical activity, coping styles and social support (42–44). A better understanding of the risk and protective factors in at-risk individuals will be helpful in developing more effective treatment and preventive interventions (4,24,43,44).

The variability in HPA regulation in depressed youngsters and adults, with and without early-life adverse experiences, suggests that these two subgroups of patients might benefit from different treatment strategies. For instance, patients with elevated HPA activity might respond well to CRH antagonists (45). Also, data from clinical and preclinical studies suggest that treatment with antidepressant agents reduces physiological responsivity to stress (13,46). However, in a recent report, adult patients with chronic major depression and childhood trauma responded poorly to antidepressant medication, whereas a better response to psychotherapy was observed (47). HPA activity measures were not reported in this study (47). Integration of neurobiological and psychosocial measures in intervention studies will be helpful in determining which subgroups of patients respond optimally to antidepressant agents, psychotherapy or combination treatment (15,48,49).

In conclusion, the findings presented in this report highlight the importance of early-life and recent experiential factors in explaining the variability in neuroendocrine regulation in adolescent depression. Further research is needed to understand how genetic, maturational, neurobiological and subsequent psychosocial factors interact to confer vulnerability to depression across the life-span in individuals exposed to early-life adversity. Persistently elevated HPA activity might increase vulnerability to a variety of medical and mental disorders (12,50). Assessment of resiliency factors in at-risk persons also is crucial for developing more effective treatment and prevention methods.