nerability to early parturition and/or low-birth-weight babies. The induction of
CRH gene expression by cortisol possibly accelerates the normal parturition process
and fetal development. The over-expression of CRH becomes a signal of danger, per-
haps, and the pregnancy begins to terminate. The positive feedback placental CRH
system is one possible mechanism regulating the timing of parturition (Figure 8.4).
In cases of preterm labor, maternal cortisol concentrations are signi´¬ücantly
higher than in mothers who carry to term (e.g. Hobel et al., 1999; Erickson et al.,
2001). Additionally, plasma CRH levels are signi´¬ücantly higher and CRH-binding
protein levels are signi´¬ücantly lower during the last 10 weeks of gestation in those
who deliver preterm compared to those who carry to term. When women with var-
ious physical and psychosocial risk factors who delivered preterm were compared
to those women who delivered at term with the same risk factors, the cortisol and
CRH concentrations in the preterm groups were signi´¬ücantly higher (Erickson et al.,
2001). This suggests that stress during pregnancy does not necessarily result in exces-
sive cortisol and CRH concentrations; additional vulnerability factors apparently
240 J. Schulkin et al.
Placental CRH (pg/g)
Figure 8.4 Human placental CRH in pregnancy complicated by pre-eclampsia and uncomplicated
pregnancies. Adapted from Goland et al. (1995)
Engages in risk-
Free CRH levels
Figure 8.5 Levels of CRH in women who were risk and non-risk takers in mid pregnancy and who
had or did not have a preterm delivery. Adapted from Erickson et al. (2001)
contribute to the expression of physical and psychosocial stressors in peripheral or
placental neuroendocrine markers (Figure 8.5).
Pregnancy and neonatal development
Events in utero can have long-term physiological and behavioral consequences
(e.g. Barker, 1997; Seckl, 1997; Welberg and Seckl, 2001). For example, elevated
241 Glucocorticoid facilitation of CRH in the placenta and the brain
levels of salivary free cortisol concentrations in the mother during pregnancy can
have a negative impact on infant motor and mental development (Buitelaar et al.,
2003; Huiink et al., 2002), and may facilitate avoidance and shyness behaviors
(Trautman et al., 1995).
Psychosocial stress has been linked to low-birth-weight babies and elevated
levels of CRH (e.g. Hobel et al., 1999). While the data are not entirely consistent the
bulk of the data suggests that diverse forms of psychological events can negatively
impact reproduction in humans and other species. In one prospective study in
humans, levels of stress and anxiety in the third trimester were associated with
low-birth-weight babies and decreased age related births (Wadhwa et al., 1993).
Natural disasters, such as earthquakes have been linked to the gestational length
(Glynn et al., 2001). There is also evidence that levels of CRH in the maternal cir-
culation are associated with decreased habitation to stimuli by the fetus (Sandman
et al., 1999).
More recent studies have noted relationships between maternal psychosocial
factors. Individual differences in salivary cortisol responses in human neonates
within the ´¬ürst 24 h of life were examined, and two ´¬ündings were noted. First,
salivary cortisol appears to remain moderately stable during the ´¬ürst 24 h of
postnatal life. Second, maternal age and socioeconomic status appear to in´¬‚uence
hypothalamicÔÇ“pituitaryÔÇ“adrenal (HPA) axis regulation in that newborns who
exhibit high-cortisol responses in the ´¬ürst 24 h of life had mothers who were older.
In addition, this subset of neonates also had higher autonomic responses at birth
and were rated by nurses as more distressed (Erickson et al. unpublished observa-
tions). Finally, neonates with both low- and high-baseline cortisol concentrations
had older mothers (more than 30 years) while neonates who had moderate corti-
sol levels had younger mothers. The ´¬ündings from this study are suggestive of
maternal in´¬‚uences and contemporaneous relations between cortisol and negative
effect in the opening hours of life.
Physiological effects of the prenatal environment include changes in program-
ming of the central nucleus of the amygdala (CeA), and a vulnerability in the infant
towards perceiving events as fearful (Welberg and Seckl, 2001). Importantly, preg-
nant rats treated with DEX for the entire period, or for the last third of gestation
had infants that had lower body weights at birth and when these infants were
tested 6ÔÇ“8 months later they had diminished exploratory behavior in an open ´¬üeld;
treatment for the ´¬ünal third of pregnancy also resulted in de´¬ücits in forced swim
test. Importantly, in these neonates amygdala CRH mRNA was elevated in both the
DEX treated groups (Welberg and Seckl, 2001). The enzyme 11 beta-hydroxy-
steroid dehydrogenase type 2 may be particularly important for some of the effects
of adult programming that perhaps result from glucocorticoid activation (Welberg
et al., 2000).
242 J. Schulkin et al.
Glucocorticoids readily cross from the peripheral systemic circuitry into the
brain. This has implications for the effects of increased circulating cortisol on the
fetus and mother during pregnancy. Stress can (but not necessarily; see Erickson
et al., 2001) increase glucocorticoid concentrations in the mother during preg-
nancy, and thus exposes the fetal developing brain to higher levels of glucocorticoids
affecting early brain development. There are data in humans that repeated high
levels of glucocorticoids by DEX treatment might impact size at birth (French et al.,
1999). Finally, cortisol can easily cross into the brain of the mother, potentiating
the extra-hypothalamic positive feedback system and increasing CRH expression
in the maternal amygdala (Welberg and Seckl, 2001).
Part 2 Glucocorticoid induction of CRH in the
brain, and fear-related behaviors
Glucocorticoid and other steroid receptors are part of a major class of DNA-binding
factors that regulate gene transcription (Schulkin et al., 1998). Glucocorticoids are
lipophillic, pass through the bloodÔÇ“brain barrier, and bind to intracellular high-
and low-af´¬ünity corticosteroid receptors to form homodimers which then regulate
gene expression by binding directly to DNA. These corticosteroidÔÇ“receptor com-
plexes regulate transcription of numerous genes in most organs of the body and
brain, including several inducible transcriptional factors (Bremner et al., 1997).
Both adrenal steroids (glucocorticoids and mineralocorticoids) compete for access
to the receptor sites (De Kloet, 1991); both hormones can in´¬‚uence CRH expres-
sion in the PVN and increase the level of CRH in the CeA (Watts and Sanchez-Watts,
1995). The effects on the developing amygdala have implications for increased fear
and anxiety responses (see review by Korte, 2001) and, therefore, may impact on
both infant temperament and mental health later in life.
Glucocorticoids and CRH are typically associated with the HPA axis. However,
when discussing these hormones in the context of stress responses, it is important
to take into consideration their activities within extra-hypothalamic regions. In
regions such as the amygdala, glucocorticoids can potentiate activity, while inhibit-
ing activity in other regions such as the hippocampus and PVN, and these gluco-
corticoid effects in´¬‚uence physiological, cognitive and behavioral domains.
Glucocorticoids are part of both positive and negative feedback systems regulat-
ing CRH expression. Within the context of positive feedback regulation, CRH and
CRH mRNA expression in the CeA is increased by peripherally administered cor-
ticosterone, while at the same time CRH gene expression in the PVN are decreased
(Swanson and Simmons, 1989; Makino et al., 1994a; Watts and Sanchez-Watts,
1995). Regulation of the CRH receptors in the hypothalamus and amygdala may
also have different sensitivities to corticosterone (Makino et al., 1995). Glucocorticoid
administration can also lead to increased CRH expression in the bed nucleus of the
stria terminalis (Makino et al., 1994a, b; Watts and Sanchez-Watts, 1995), a struc-
ture that has been described as extended amygdala, and has also been associated
with anxiety responses (Davis et al., 1997) (Figure 8.6).
Glucocorticoids are secreted under a number of experimental conditions in
which fear, anxiety, novelty, and uncertainty are experimental manipulations (Mason,
1975; Breier, 1989). Across a number of species, including humans, glucocorticoids
are secreted when there is loss of control, or the perception of loss of control (worry
is associated with the loss of control) (e.g. Breier, 1989). Conversely, circulating
244 J. Schulkin et al.
CRH mRNA in CeA (dpm/mg)
(a) Control Week
CRH mRNA in BNST (dpm/mg)
(b) Control Week
CRH mRNA in PVN (dpm/mg)
(c) Control Week
Figure 8.6 CRH mRNA levels in three regions of the brain (a) CeA, (b) bed nucleus of the stria
terminalis (BNST), and (c) PVN of the hypothalamus across weeks 1 and 2 in adrenally
intact rats implanted with corticosterone (see Makino et al., 1994a, b for more details).
dpm: disintegrations per minute
245 Glucocorticoid facilitation of CRH in the placenta and the brain
peripheral glucocorticoids are reduced when there is perceived control. Predicting
the onset of an aversive signal reduces the level of circulating glucocorticoids (Mason,
1975). Within the clinical literature, one of the most consistent ´¬ündings in
depressed patients is elevated levels of cortisol and an enlarged adrenal cortex (e.g.
Sachar et al., 1970). These ´¬ündings are congruent with those of Richter (1949) who
observed an enlarged adrenal gland in stressed, fearful, wild rats when compared to
unstressed laboratory analogs.
The CRH is now well known to be both a peptide that regulates pituitary and
adrenal function and an extra-hypothalamic peptide hormone linked to a number of
behaviors, including behavioral expressions of fear (Koob et al., 1993; Kalin
et al., 1994). CRH cell bodies are widely distributed in the brain (Palkovits et al.,
1983; Swanson et al., 1983). The majority of CRH neurons within the PVN are clus-
tered in the parvicellular division. Other regions with predominant CRH-containing
neurons are the lateral bed nucleus of the stria terminalis and the central region of
the CeA. To a smaller degree, there are CRH cells in the lateral hypothalamus, pre-
frontal and cingulate cortex. In brainstem regions, CRH cells are clustered near the
locus coeruleus (BarringtonsÔÇ™ nucleus) (Valentino et al., 1994; 1995), parabrachial
region and regions of the solitary nucleus. Central CRH activation is consistently
and reliably linked to the induction of fear in animal studies (Kalin et al., 1994; Koob
et al., 1993). Intraventricular infusions of CRH, for example, are known to facilitate
fear-related socially derived contextual responses, in addition to activating greater
metabolic activation of the amygdala (Strome et al., 2002).
Central infusions of CRH induce or potentiate a number of fear-related behav-
ioral responses (Takahashi et al., 1989), and infusion of CRH antagonists both
within the amygdala and outside of it reduce fear-related responses (Koob et al.,
1993). Startle responses are enhanced by CRH infusions (Swerdlow et al., 1989).
CRH injected into the lateral ventricles increases freezing to fearful stimuli and
potentiates acoustic startle in rats (Liang et al., 1992; Koob et al., 1993). Conversely,