. 37
( 51 .)


most sensitive to circulating corticosterone within the range of concentrations found
across the circadian day. Data adapted from Watts and Sanchez-Watts (1995a). (c) The
organization of an open-loop feedback system. Unlike the closed-loop system depicted in
(a), there is no in¬‚uence from ¬nal target, the thymus, to the control mechanisms in the
brain. (d) The well-documented relationship between thymus weight and circulating
corticosterone. Data adapted from Watts and Sanchez-Watts (1995a)
206 A. G. Watts

set-point limits (Figure 7.2(a)). For the HPA axis, the negative-feedback action of
corticosterone is regarded as the major inhibitory signal that acts to reduce the
value of all the appropriate variables (gene expression, peptide levels, and secretory
rates) within the system. In reality, however, negative-feedback loops offer little
more than the application of a simple ˜push“pull™ principle to neuroendocrine
control. Although this model offers reasonable explanations for simple re¬‚ex
neuroendocrine mechanisms (e.g. the AVP secretory response to hemorrhage or
elevated plasma osmolality), it fails to account satisfactorily for those complex syn-
thetic and secretory features of HPA neuroendocrinology that have a more antici-
patory nature. Examples of anticipatory events within the HPA axis include the
daily variations in glucocorticoid secretion timed by the suprachiasmatic nucleus
(SCH) that precede activity and feeding, or the conditioning or habituating effects
of repeated stimuli on the glucocorticoid responses.

How is glucocorticoid inhibition manifest on ACTH secretogogue gene expression?
This question has been most commonly addressed using unstressed adrenalec-
tomized (ADX) rats maintained with ad lib food/water/saline, and given a regimen
of constant exogenous corticosterone for 5“7 days. Experiments of this type gen-
erate the simple inverse log10 function between circulating corticosterone and CRH
mRNA levels in the PVHmp (Watts and Sanchez-Watts, 1995a) that has formed
the basis of the classic ˜negative-feedback™ model of gene control (Figure 7.2(b)).
For comparison with this closed-loop feedback model, Figure 7.2(d) shows the
familiar inverse relationship between circulating corticosterone and thymus
weight. This is an open-loop arrangement (Figure 7.2(c)) because, in contrast to
the closed-loop feedback actions of corticosterone on the CRH neuron, there is no
target-derived negative-feedback signal from the thymus to the HPA axis.
In male rats the inverse relationship between circulating corticosterone and
CRH mRNA levels has the greatest dynamic range between corticosterone concen-
trations of 10 to about 150 ng/ml (Figure 7.2(b)), which is that found during the
normal daily variations (Swanson and Simmons, 1989; Watts and Sanchez-Watts,
1995a; Watts et al., 2004). Virtually no further reduction in CRH mRNA occurs if
concentrations increase beyond 200 ng/ml. Although a comparable dose“response
curve has not been obtained for AVP mRNA in the PVHmp, AVP gene products are
apparently even more sensitive to circulating corticosterone than CRH; only very
low circulating levels are required to reduce mRNA to undetectable levels.

Cell speci¬city
At this point it is important to note that corticosterone regulates CRH gene expres-
sion in a cell-speci¬c manner. Although glucocorticoids are sometimes thought of
207 Glucocorticoids and the ups and downs of neuropeptide gene expression

as downregulating CRH gene expression in the brain, in fact the only cell type where
this actually occurs is the CRH neuroendocrine motor neuron in the PVHmp. In
every other cell type that synthesizes CRH, glucocorticoids either upregulate gene
expression or have no effect whatsoever (Beyer et al., 1988; Swanson and Simmons,
1989; Frim et al., 1990; Makino et al., 1994; Watts and Sanchez-Watts, 1995a).
An interesting comparison in this regard is the neurons in the lateral part of the
central nucleus of the amygdala (CEAl). Here, corticosterone concentrations over
the daily range increase CRH mRNA in a manner that is virtually the inverse of
that seen in the PVHmp (Figure 7.3; Watts and Sanchez-Watts, 1995a). These data
emphasize that rather than a ¬xed component of the CRH gene regulation deter-
mining how glucocorticoids function, variations in the local cellular environment,
including intracellular factors (e.g. signal transduction pathways) together with
extracellular factors (e.g. the nature of afferent inputs) play major roles in deter-
mining how glucocorticoids interact with the CRH gene.


Mean gray level of CRH mRNA
(% mean ADX value)




1.0 1.5 2.0 2.5
Log10 plasma corticosterone
concentration (ng/ml)
Central nucleus of the amygdala Paraventricular nucleus
Figure 7.3 CRH mRNA levels in the paraventricular nucleus (solid circles, solid line) and lateral
part of the central nucleus of the amygdala (open circles, dashed line) from
adrenalectomized male rats with various doses of exogenously applied corticosterone.
Note that corticosterone has very different effects on the accumulation rates of
CRH mRNA depending on the cell type examined. (Data adapted from Watts and
Sanchez-Watts 1995a)
208 A. G. Watts

How do glucocorticoids regulate CRH gene expression?

The rat PVH is the classic example of a hypothalamic motor nucleus that controls
many different neuroendocrine, autonomic, and behavioral actions. Before we dis-
cuss what we know of the mechanisms engaged by glucocorticoids to regulate
CRH gene expression in the PVH, it is worth brie¬‚y describing the overall organi-
zation of this complex cell group. I will do this by ¬rst describing the PVH in terms
of its different structural compartments, and then describe CRH neuroendocrine
motor neurons in terms of their functional compartments to provide the backdrop
for considering in more detail how glucocorticoids regulate CRH and AVP gene
expression in these neurons.

Structural compartmentalization of the PVH
In keeping with its diverse functional roles, the PVH contains a number of struc-
turally distinct compartments. These are most easily seen in the rat, where there
is clear spatial segregation between compartments (Swanson and Kuypers, 1980),
but is less obvious in other species. Based upon its efferent projections, the PVH is
divisible into at least two major structural compartments: neuroendocrine motor
neurons that project to the neurohypophysis, and parvicellular pre-autonomic
neurons that project to the hindbrain and spinal cord (Figure 7.4). Each of these
major compartments can be further subdivided, ¬rst in terms of their projections,
and then again by way of the different chemical phenotypes that represent great




(a) (b) (c)
Figure 7.4 The organization of the rat PVH. (a) The cytoarchitectonics shows a Nissl stained coronal
section through PVH. The majority of CRH neuroendocrine neurons are located in the
dorsal part of the mp (mpd) PVH. Also shown are the posterior magnocellular (pm), the
periventricular (pv), and dorsal parvicellular (dp) and ventral parts of the mp (mpv) PVH
both of which project to the hindbrain and spinal cord. (b) The efferents show neurons
retrogradely labeled after injections of two ¬‚uorescent tracers into the vasculature (dark
cells in the pm, mpd, and pv), and the cervical spinal cord (white cells in the dp and mpv).
(c) The CRH mRNA shows the in situ hybridization signal from CRH mRNA in the mpd
209 Glucocorticoids and the ups and downs of neuropeptide gene expression

potential for diverse neural signaling. In this way, the neuroendocrine compart-
ment consists of oxytocin and AVP-containing magnocellular motor neurons that
project to the posterior pituitary, together with six different types of parvicellular
motor neurons that project to the ME and control hormone secretion from the
anterior pituitary.
Subdividing parvicellular pre-autonomic neurons in terms of their efferent pro-
jections is more complex because they project to a variety of targets in the mid-
brain, hindbrain, and spinal cord. One signi¬cant difference, however, is whether they
project to the dorsal vagal complex or to the spinal cord (Swanson and Kuypers,
1980). Parvicellular pre-autonomic neurons contain a variety of neuropeptides,
including oxytocin, AVP, enkephalin, dynorphin, and CRH (Hallbeck et al., 2001). It
is worth noting that most PVH neurons also appear to be glutamatergic (Herman
et al., 2002). Depending on the physiological status of the animal, CRH is synthe-
sized in all the major structural compartments of the rat PVH (Swanson 1991;
Watts, 1992; 1996) and importantly, that corticosterone in¬‚uences the expression
of the CRH gene in these PVH cell types in quite different ways (Swanson and
Simmons, 1989).

Functional compartmentalization of the CRH neuroendocrine motor neuron
Corticosterone can affect the activity of the HPA axis at many different levels rang-
ing from how the CRH gene is transcribed to secretion rates of ACTH. Similarly,
the actions of corticosterone are likely to alter the function of CRH neuroendo-
crine neurons at a variety of different levels. This complexity means that particu-
lar attention has to be directed towards the behavior of dependent variables as
interpreted within the context of CRH gene control.
One way to deal with a system as complex as the CRH neuron is to partition it
into functional compartments and then examine how these compartments inter-
act. Figure 7.5 illustrates one such schema for the CRH neuron (designated by the
gray box), which can help constrain the interpretational models derived from the
behavior of particular dependent variables. Although it is somewhat arbitrary as to
where one compartment ends and another starts, this model allows us to place the
cellular processes occurring in each compartment within a wider context of the
whole neuron.
The ¬rst compartment in this model (numbered 1 in Figure 7.5) consists of the
afferent sets that project to CRH neurons and control their activity. The structural
organization of these afferents is considered later in this section. In the second
compartment (numbered 2 in Figure 7.5) afferent sets interact with CRH neurons
using appropriate sets of receptors and signal transduction pathways. In this
manner, the effects of stress, energy metabolism or indeed of any other physi-
ological process are ultimately mediated by the actions of neurotransmitters and
210 A. G. Watts


Neural afferent sets

Receptors and signal

CRH gene


CRH peptide





CRH release into hypophysial
portal vasculature

Figure 7.5 A schematic to show six possible functional compartments within a CRH neuroendocrine
neuron in the PVH. It shows that neural afferents regulate two principal processes:
membrane excitability, which controls peptide release, and peptide synthesis.
Corticosterone can regulate both processes either by modulating the activity of neural
afferents, or more directly by way of transmitter receptors, signal transduction pathways,
or gene transcription

circulating factors at this functional level. These factors are continually integrated
by CRH neurons to generate their ongoing activity patterns. Signal integration and
signal transduction mechanisms then control two fundamental processes:

(1) CRH synthesis (Compartment 3).
(2) Membrane excitability (Compartment 4).
211 Glucocorticoids and the ups and downs of neuropeptide gene expression

The third compartment contains the machinery in the endoplasmic reticulum and
Golgi complex that controls the translation of CRH mRNA into CRH peptide. The
fourth compartment controls membrane excitability by way of the neuron™s speci¬c
complement of ion channels, and so ultimately controls ¬ring rate and subsequent
release of CRH from terminals in the ME (Compartment 6). Processes in the ¬fth
compartment package peptide into vesicles and transport them along axons for stor-
age in neuroendocrine terminals in the ME for release. And the sixth compartment
contains the stimulus“release coupling mechanisms that release peptide from termi-


. 37
( 51 .)