Steriod hormones

Steroid hormones are a class of chemical compounds that are derived from cholesterol, and are primarily secreted by the endocrine glands, the adrenal, and the gonads (testes and ovaries; Kawata, 1995; Keenan & Soleymani, 2001; Mesce, 2002; Rieder & Coupey, 1999). The steroid hormones have widespread effects because they are transported through the bloodstream and trigger activity in the cells of various organs and serve to regulate a range of physiological functions. The adrenal steroids are secreted by the adrenal gland, and the sex hormones are largely produced by the gonads. But it is relevant to note that the adrenal also secretes some sex hormones.

There are actually two adrenal systems—the adrenal medulla and the adrenal cortex. The two adrenal systems are developmentally and anatomically related, although the hormonal products differ, as do the correlates for males and females (Fadalti, Petraglia, Luisi, Bernardi, Casarosa, Ferrari, Luisi, Saggese, Genazzani, & Bernasconi, 1999; Kawata, 1995). The adrenal medulla produces neurotransmitters, namely, the catecholamines, epinephrine, and norepinephrine. The adrenal cortex, on the other hand, produces androgen in both sexes. The adrenal cortex also produces mineralocorticoids and the glucocorticoids. Mineralocorticoids, such as aldosterone, help maintain osmotic balance in the body. Glucocorticoids, most notably cortisol in primates, were thus named because they serve to elevate blood glucose concentrations. Glucocorticoids influence many metabolic processes, particularly the formation of glucose. Further, glucocorticoids modulate blood pressure, activate the antiinflammatory and immunosuppressive responses, and are highly sensitive to environmental factors.

Glucocorticoid release from the adrenals is controlled by the HPA axis, which is influenced by a range of physical and social events. The HPA axis is particularly sensitive to stress, and is activated in response to physical and psychological factors that threaten homeostasis (Charmandari, Kino, Souvatzoglou, & Chrousos, 2003; Dorn & Chrousos, 1997). Thus, perturbations affect the organism through actions on hypothalamic secretagogues, corticotropin releasing hormone, and arginine vasopressin. This, in turn, triggers the release of adrenocorticotropin (ACTH) from the anterior pituitary. ACTH then causes the release of glucocorticoids (again, primarily cortisol in humans) from the adrenal cortex and into circulation. The rapid rise in cortisol levels can be adaptive because it can increase the availability of energy substrates.

The actions of cortisol are mediated by two types of receptors: mineralocorticoid receptors (MRs) and glucocorticoid receptors (GRs), also referred to as Type I and Type II receptors, respectively. These receptors are present on many cells throughout the body, including the brain and pituitary. The HPA axis is a self-modulating system, in that glucocorticoids act to suppress their own release through activation of MRs and GRs, which initiates both fast- and slow-acting negative feedback systems that inhibit ACTH release. These feedback systems affect the hypothalamus and pituitary, although other regions, such as the frontal cortex and hippocampus, also play a role. Thus, effects on the central nervous system are widespread.

The hypothalamic–pituitary–gonadal (HPG) axis governs the release of sex hormones (Buchanan, Eccles, & Becker, 1992; Rieder & Coupey, 1999). Like adrenal hormones, the gonadal hormones have pervasive neurophysiological effects. The principal male sex hormone, or androgen, is testosterone. It is mainly produced by the testes, although, as mentioned, it is also produced at lower levels by the adrenal cortex. Among females, low levels of androgen are produced by the ovaries. Another important adrenal androgen is dehydroepiandrosterone (DHEA). Both of these androgens promote the development and maintenance of reproductive function, and stimulate the production of skeletal muscles, bone, and secondary sex characteristics in the male.

Estrogens and progestins are the two classes of female sex hormones. Estrogens are secreted by the ovaries, and at lower levels by the adrenal glands and, in males, by the testes. (The most biologically active estrogen is estradiol.) Estrogens promote the development of female sex characteristics; they also stimulate physical growth and maturation. The main progestin is progesterone, a key hormone for maintaining pregnancy. Unlike androgens, estrogens, and progestins are secreted cyclically. As a result, their effects on the brain wax and wane through the cycle (Schmidt, Nieman, Danaceau, Adams, & Rubinow, 1998).

Adrenarche

Adrenarche is the developmental prelude to sexual maturation (Rieder & Coupey, 1999; Weber, Clark, Perry, Honour, & Savage, 1997). It is a maturational process that begins in childhood, at about 6–8 years of age in humans. It has its onset about the same time as the preadolescent rise in BMI. Adrenarche is marked by a physiological increase in androgen secretion from the adrenal cortex, which takes place several years before the onset of puberty. The main hormonal product of adrenarche is the androgen, DHEA, and its sulfate (DHEAS).

It has been proposed that the HPA axis may be involved in triggering adrenarche (Weber et al., 1997). In particular, it has been suggested that pituitary release of ACTH and/or adrenal release of cortisol might play a role. Consistent with this, Weber et al. (1997) found that individuals with familial glucocorticoid deficiency were more likely to manifest a lack of adrenarche. Conversely, children who have an early onset of adrenarche, as measured by the early (before age 8) appearance of Tanner stage II–III pubic hair, show signs of heightened HPA activity (Dorn, Hitt, & Rotenstein, 1999). When compared to age-matched controls, girls with premature adrenarche show significantly higher concentrations of cortisol, as well as estradiol, thyroid-stimulating hormone, and adrenal androgens.

There is also a gradual increase in plasma insulin associated with the onset of adrenarche, and an increase in insulin-like growth factor I (IGF-I) levels (Guercio, Rivarola, Chaler, Maceiras, & Belgorosky, 2003). IGF-I is a hormone that promotes neurogenesis and neuromaturation. It has also been proposed that IGF-I might be a trigger to the commencement of adrenarche, although it does not necessarily remain elevated after the onset. In particular, growth hormone (GH) and IGF-I may be involved in the mechanism of adrenarche during prepuberty in girls.

In recent years it has been discovered that adipose tissue synthesizes a hormone called leptin that has been implicated in body weight and fat regulation (Faloia, Camilloni, Giacchetti, & Mantero, 2000; Weise, Eisenhofer, & Merke, 2002). Thus, adipose tissue is not solely devoted to energy storage and release, but rather is now viewed as an endocrine organ, as well as a potential target for steroid hormone activity. Leptin has received a great deal of attention as a neuromaturational trigger. Leptin levels increase during childhood in both sexes, and the rise is most pronounced in association with adrenarche. Although the role of leptin in adrenarche is not fully understood, there is a positive correlation between serum leptin and BMI in children. Another recent and potentially relevant finding is that adipose tissue produces cortisol from its inactive precursor, cortisone.

Gonadarche and postpubertal development

Gonadarche, the onset of puberty, is marked by activation of the HPG axis, resulting in a very rapid escalation in testosterone in males and estradiol in females (Buchanan et al., 1992; Nottelmann, Susman, Dorn, Inoff–Germain, Loriaux, Cutler, & Chrousos, 1987). The magnitude of the rise in gonadal hormones is striking (Mann, Johnson, Gimpel, & Castracane, 2003; Veldhuis, 1996). From infancy to age 12, the level of gonadal hormones is relatively stable. Then, between 12 years and early adulthood, pulsatile secretion of luteinizing hormone (LH), the hormone that stimulates gonadal growth and gonadal hormone secretion, increases more than 30-fold in boys and 100-fold in girls (Veldhuis, 1996). In girls, estradiol doubles from about 25 pg/ml in the 10–12 year period, to 50 pg/ml between 13 and 15 years. Then it skyrockets to an average of about 200 pg/ml between 16 and 20. Beyond age 20, estradiol levels never return to prepubertal values, although there is a very gradual decline to around 100 pg/ml during menopause and 75 pg/ml following menopause. Thus perimenopausal hormone changes are much less abrupt and of smaller magnitude than those occurring during normal adolescence. In males, the rise in testosterone is even more dramatic. Across the normal life span, testosterone increases from about 25 ng/dl prior to 12 years of age to over 200 ng/dl from 13 to 15 years, to 600 ng/dl between 16 and 20. Although testosterone declines very gradually after this peak in late adolescence, it remains well above 200 for most of the rest of the male life span.

For both sexes, compared to other life-course developmental changes in hormone levels, hormone changes during adolescence are unparalleled with respect to both magnitude and speed. In addition to pulsatile surges in LH, and rapid increases in testosterone and estradiol, there are also progressive postpubertal increases in other hormones, including allopregnanolone (a steroid hormone thought to originate in the ovaries and adrenal cortex), progesterone, and DHEA levels, in both boys and girls (Fadalti et al., 1999). In the absence of disease, endocrine changes are relatively mild following the early 20s, and extending throughout the adult life span.

Several factors have been hypothesized to trigger gonadarche. Central among these are two hormones implicated in adrenarche; namely, leptin and IGF. There is some evidence that a marked rise in leptin levels occurs just before the onset of puberty, and then leptin decreases to baseline levels after the initiation of puberty (Faloia et al., 2000). This suggests a role for leptin in triggering the onset of puberty. Further evidence to support this notion comes from experimental studies of rodents, which indicate that exogenous leptin administration advances the onset of pubertal development.

There are data indicating that IGF also regulates pubertal onset (Guercio et al., 2003; L'Allemand et al., 1996). In particular, it appears that an increase in IGF-1 may be involved in the stimulation of gonadotropin-releasing hormone activity. Conversely, it also appears that rising sex hormones contribute to increased GH and IGF-1 in children with normal puberty, as well as precocious puberty. The pubertal increase in sex steroids may therefore induce the neuromaturational events that activate the systems that control GH and IGF-1 secretion.

Adrenal hormones also change during puberty. Developmental research has shown that levels of the adrenomedullary hormone, epinephrine, and its metabolite metanephrine, decrease with pubertal maturation, and are higher in males than females (Weise et al., 2002). Weise and colleagues also found that decreases in epinephrine were most strongly linked with pubertal stage and leptin, whereas metanephrine was best predicted by estradiol and leptin. Further, epinephrine and metanephrine were inversely correlated with DHEAS, estradiol, testosterone, leptin, and insulin. In contrast, plasma norepinephrine, which is mainly derived from sympathetic nerve terminals, increases with advancing puberty and rising testosterone levels among boys. These findings suggest that sympathoadrenal hormones may be implicated in sexual maturation. In particular, the adrenal medulla might have a significant role in linking leptin and body weight to the timing of sexual maturation.

Puberty and the HPA axis

There is now mounting evidence of a pubertal increase in activity of the HPA axis. Most cross-sectional and longitudinal studies of salivary and urinary cortisol find an increase with age during the adolescent years. Cross-sectional studies of normal children reveal a gradual rise in salivary and urinary cortisol during middle childhood, then a marked increase that begins around 13 years of age and continues through adolescence (Kenny, Gancayco, Heald, & Hung, 1966; Kenny, Preeyasombat, & Migeon, 1966; Kiess, Meidert, Dressendorfer, Scheiver, Kessler, & Konig, 1995; Lupien, Wilkinson, Briere, Menard, Ng, Kin, & Nair, 2002; Wingo, 2002). Recent longitudinal studies have also revealed increases in cortisol release during adolescence, with the most significant augmentation occurring at 13 years (Wajs–Kuto, De Beeck, Rooman, & Caju, 1999; Walker & Bollini, 2002). Studies that have examined pubertal stage indicate that the changes are strongly linked with sexual maturation (Kenny et al., 1966; Kiess et al., 1995; Tornhage, 2002).

It is of interest, however, that studies that measure cortisol in plasma do not always find an increase in cortisol levels during the course of adolescence (Fadalti et al., 1999; Knutsson, Dahlgren, Marcus, Rosberg, Bronnegard, Stierna, & Albertsson–Wiklund, 1997). For example, a recent longitudinal investigation of serum cortisol in 28 adolescents found no developmental change (Knutsson et al., 1997). These discrepancies may be due to differences in subjects' responses to blood versus saliva sampling, or may reflect differential subject attrition based on sampling procedure. It has been shown that research participants are more likely to decline blood sampling, and those with greater psychiatric symptoms show higher attrition (Davies, Collins, Gregory, & Clarke, 1996; Keks, Copolov, Mackie, Stuart, Singh, McGorry, & Coffey, 1991). Thus, blood sampling may alter HPA activity, and thereby obscure the relation between cortisol levels and other variables. Consistent with this assumption, it appears that group differences in cortisol levels, including those induced by ACTH administration, are more readily detected in salivary assays than in plasma (Kirschbaum et al., 1999). Finally, the discrepancy between the developmental course of plasma and salivary cortisol may be a consequence of the fact that salivary cortisol is an index of plasma-free cortisol, and thus may be a better measure of neuroactive cortisol.

Cross-sectional data from a recent study of adolescents conducted in our laboratory replicates the age-related increase in salivary cortisol documented in previous reports. Our research is intended to track the developmental course of youth with personality disorders to identify risk factors for conversion to Axis I disorders. Participants range in age from 12 to 17 years. One diagnostic group (n = 26) comprises adolescents who met DSM-IV diagnostic criteria for schizotypal personality disorder (SPD). Another group (n = 28) meets criteria for one or more other Axis II personality disorder (OPD). The participants in the “normal” comparison group (n = 26) do not meet criteria for any DSM-IV disorder.

During the first assessment of this sample, cortisol was measured in saliva. A total of five saliva samples were taken over the course of the assessment, and they were averaged for the present analyses. All subjects gave their first sample between 9:00 and 9:30 a.m. Thereafter, saliva samples were collected every hour, for a total of five samples.

The correlation of age with cortisol was positive and significant for all groups at r = .30, .42, and .46 for the normal, OPD, and SPD groups, respectively. Curve fitting tests revealed that the relation between age and mean cortisol had no significant nonlinear component. Figure 1 shows the linear trend for residualized cortisol values, by age, for the three groups of youth. (Residualized cortisol values were obtained with regression analysis to control for potential effects of psychotropic medications.) As can be seen, the developmental trends are similar for the two groups.

The linear relationship between age and residualized cortisol levels for normal and personality disordered youth.

We have found the same linear trends in cortisol for males and females. These findings indicate that the developmental increase in cortisol secretion is similar across sex and diagnostic group. Second-wave assessments are now being conducted on this sample. This will allow us to address the question of longitudinal change in cortisol secretion and its relation with indicators of adjustment.

Adolescent stress sensitivity: The behavioral and biological levels

As with adults, research on children and adolescents has shown that exposure to stressful events is associated with increased risk for psychopathology (Chen, Matthews, & Boyce, 2002; Pine, 2003). Further, this effect may be at least partially mediated by activity of the HPA axis. For example, family conflict predicts cortisol levels, as well as the number of days ill, in children and adolescents (Flinn & England, 1997). Experimental studies indicate that, in response to a laboratory stressor (Trier Social Stress Test for Children), adolescents show increases in heart rate and cortisol (Buske–Kirschbaum, Jobst, Wustmans, Kirschbaum, Rauh, & Hellhammer, 1997).

Many have assumed that the behavioral changes that occur during adolescence are partially a consequence of additional external stressors impinging on the child during this developmental transition. But is there empirical support for the notion that exposure to stress is greater during adolescence? There is indeed evidence that the rate of occurrence of stressful events increases from childhood through adolescence (Gest, Reed, & Masten, 1999). But the increase is specific to certain classes of events; namely those that are classified as “nonindependent.” Events that are independent of the individual (i.e., physical illness in the family, natural disasters, academic problems) are relatively stable from childhood through adolescence. In contrast, negative events that can be linked to the individual's behavior (e.g., conflicts with friends or family members, interpersonal problems in work or school) increase significantly through adolescence. Further, the increase in negative, nonindependent experiences is linked with both the child's maladjustment and parental adjustment problems. Thus, the stressful events that increase during adolescence may be a partial consequence of the individual's behavior.

Of course, the physical changes associated with puberty may be inherently stressful because they involve a modification in physical appearance. As mentioned, there is evidence that experiencing puberty earlier is a risk for subsequent adjustment problems, particularly in girls (Hayward, Killen, Wilson, Hammer, Litt, Kraemer, Haydel, Varady, & Taylor, 1997). Girls who mature earlier than their peers manifest more problem behaviors than those who mature at more typical ages. There is also evidence of “accentuation,” such that girls with preexisting behavioral problems who mature early showed an exacerbation of problems, whereas those who mature at the same time or later than their peers do not manifest worsening of their behavioral problems (Caspi & Moffitt, 1991). In terms of social experiences, being either early or late in pubertal maturation might have negative effects on the individual because they experience a deviation from normal developmental timing (Brooks–Gunn, 1988; Neugarten, 1979). Thus, early-onset puberty may act as a stressor. Alternatively, early onset of hormonal changes may trigger biological processes that render the child more vulnerable to stress, and thereby contribute to adjustment problems.

In addition to early onset pubertal development acting as a stressor, it has been suggested by some that the influence may be in the opposite direction, such that negative life events trigger an early onset of puberty. There is evidence that factors such as nutrition, exercise (Brooks–Gunn, 1988; Malina, 1983), and interpersonal stress (e.g., Graber & Brooks–Gunn, 1996; Graber, Brooks–Gunn, & Warren, 1995) are correlated with the age at onset of puberty. It is possible that stress-induced activation of the HPA axis affects the timing of puberty by altering the gonadal hormone system. Thus, stress exposure, by augmenting HPA activity, may hasten the rise in glucocorticoid release that normally occurs at puberty, and thus trigger premature release of gonadal hormones.

This brings us back to the question of stress sensitivity: are adolescents more sensitive to stress than younger children or adults? Despite the strong evidence that adolescents experience greater negative affect than either younger children or adults (Arnett, 1999), there have been relatively few studies of developmental changes in the biobehavioral impact of stressful experiences. The published data that are available, however, suggest that adolescence may be linked with an increased vulnerability to stressful events, which is an actual change in how stress is experienced. It has been shown in studies of normal youth that the strength of the relationship between negative life events and dysphoric mood increases with age during adolescence (Rice, Harold, & Thapar, 2003). An excellent study by Silberg, Pickles, Rutter, Hewitt, Simonoff, Maes, Carbonneau, Murrelle, Foley, and Eaves (1999) explored the influence of genetic factors and environmental stressors on adolescent depression in twins. They replicated the developmental increase in depression during adolescence, especially among girls. They also observed a significant adverse effect of negative life events which was, again, more pronounced in the postpubertal girls. Thus, it appeared that the stress–behavior link grew stronger with each year during adolescence, consistent with greater stress sensitivity. Further, the analysis for genetic effects indicated that the influence of genetic factors on life events increased through adolescence, suggesting a developmentally related increase in genetic variance for life events. The authors conclude that latent genetic liabilities are expressed after puberty, a developmental phenomena that will be revisited later in this article.

With regard to the biological substrate for adolescent stress sensitivity, the postpubertal increase in cortisol secretion, reviewed earlier, offers one explanation: the HPA axis may become more responsive to external events. In conjunction with this, other neural systems involved in the biological response to environmental challenges may be subject to maturational change. Animal models yield support for this assumption. Research with rodents has revealed that the neural systems that govern the stress response mature during the postpubertal period (Kellogg, Awatramani, & Piekut, 1998). For example, stress responsiveness of cortical receptors for GABA, an inhibitory neurotransmitter, emerges over adolescence. Also, rodents manifest postpubertal maturation in the amygdala, a structure that projects to many other brain regions (e.g., cortex, hippocampus) that show age-related changes and are linked with the mature stress response.

The amygdala plays a similar role in humans, regulating the “affective” learning that takes place in relation to emotional and behavioral responses to stressful conditions. Through postmortem research, Benes (2003a, 2003b) has shown that amygdala fibers continue to develop throughout childhood, well into adolescence. These changes increase the ability of the amygdala to regulate emotional and attentional responses in the cingulate cortex. Because the cingulate region is a limbic brain area implicated in the pathophysiology of mental illness, maturation of the amygdalocingulate pathway during late adolescent years may influence the development of psychiatric symptoms.

Other evidence of heightened biological stress responsivity in human adolescents has come from psychophysiological studies. One investigation measured skin conductance responses to neutral (loud noise) and stressful (person screaming) sensory stimuli (Miller & Shields, 1980). Both adolescents and adults exhibited more pronounced responses, and more trials to habituate, following the stressful stimulus. But, compared to adults, adolescents showed more persistent skin conductance responding, taking significantly more exposure trials to habituate. In another study using a laboratory stress paradigm, it was shown that adolescents responded to stress with greater β-adrenergic activation than did younger children (Allen & Matthews, 1997). It is also of interest to note that exposure to stress may enhance sensitivity to subsequent stressors; youth who experience a higher level of stress in their lives manifest higher diastolic blood pressure in response to acute laboratory stressors (Matthews, Gump, Block, & Allen, 1997). Thus, consistent with the findings on cortisol secretion, psychophysiological research indicates that the postpubertal period may be associated with a more pronounced response to acute stress.