GLIAL-NEURONAL INTERACTIONS IN THE

MAMMALIAN BRAIN
Glenn I. Hatton
Advan in Physiol Edu 26:225-237, 2002.

You might find this additional info useful...

This article cites 26 articles, 9 of which can be accessed free at:

http://advan.physiology.org/content/26/4/225.full.html#ref-list-1

This article has been cited by 3 other HighWire hosted articles

Central blockade of oxytocin receptors during mid-late gestation reduces amplitude of
slow afterhyperpolarization in supraoptic oxytocin neurons
R. Teruyama, D. L. Lipschitz, L. Wang, G. R. Ramoz, W. R. Crowley, S. L. Bealer and W. E.
Armstrong
Am J Physiol Endocrinol Metab, November 1, 2008; 295 (5): E1167-E1171.
[Abstract] [Full Text] [PDF]
Cellular mechanisms involved in CO2 and acid signaling in chemosensitive neurons
Robert W. Putnam, Jessica A. Filosa and Nicola A. Ritucci

Downloaded from advan.physiology.org on April 1, 2011

Am J Physiol Cell Physiol, December 1, 2004; 287 (6): C1493-C1526.
[Abstract] [Full Text] [PDF]

Activation of erbB-1 Signaling in Tanycytes of the Median Eminence Stimulates

Transforming Growth Factor β1 Release via Prostaglandin E2 Production and Induces
Cell Plasticity
Vincent Prevot, Anda Cornea, Alison Mungenast, Gregory Smiley and Sergio R. Ojeda
J. Neurosci., November 19, 2003; 23 (33): 10622-10632.
[Abstract] [Full Text] [PDF]

Updated information and services including high resolution figures, can be found at:
http://advan.physiology.org/content/26/4/225.full.html

Additional material and information about Advances in Physiology Education can be found at:
http://www.the-aps.org/publications/advan

This infomation is current as of April 1, 2011.

Advances in Physiology Education is dedicated to the improvement of teaching and learning physiology, both in specialized
courses and in the broader context of general biology education. It is published four times a year in March, June, September and
December by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2002 by the
American Physiological Society. ISSN: 1043-4046, ESSN: 1522-1229. Visit our website at http://www.the-aps.org/.
 A P S R E F R E S H E R C O U R S E R E P O R T

This section, guest edited by Cheryl M. Heesch, presents articles inspired by presentations given
in the APS Refresher Course on Neurophysiology at Experimental Biology 2002.

GLIAL-NEURONAL INTERACTIONS
IN THE MAMMALIAN BRAIN

Glenn I. Hatton

Department of Cell Biology and Neuroscience, University of California, Riverside, California 92521

ecognition of the importance of glial cells in nervous system functioning is

R increasing, specifically regarding the modulation of neural activity. This brief

review focuses on some of the morphological and functional interactions
that take place between astroglia and neurons. Astrocyte-neuron interactions are of

Downloaded from advan.physiology.org on April 1, 2011

special interest because this glia cell type has intimate and dynamic associations with
all parts of neurons, i.e., somata, dendrites, axons, and terminals. Activation of certain
receptors on astrocytes produces morphological changes that result in new contacts
between neurons, along with physiological and functional changes brought about by
the new contacts. In response to activation of other receptors or changes in the
extracellular microenvironment, astrocytes release neuroactive substances that di-
rectly excite or inhibit nearby neurons and may modulate synaptic transmission.
Although some of these glial-neuronal interactions have been known for many years,
others have been quite recently revealed, but together they are forming a compelling
story of how these two major cell types in the brain carry out the complex tasks that
mammalian nervous systems perform.
ADV PHYSIOL EDUC 26: 225–237, 2002;
10.1152/advan.00038.2002.

Key words: calcium imaging; morphological plasticity; neurohypophysis; supraoptic

nucleus; taurine

Glial cells in the mammalian central nervous system
come in a variety of types, shapes, and sizes. There
are astroglia, ependymoglia, microglia, and oligoden-
droglia and at least some variation within each of
these types. Oligodendroglia associate chiefly with
the axonal portions of neurons as myelinating cells,
although in the olfactory bulb there are myelinated
dendrites. Microglia wander around the neurons, per-
forming their phagocytic tasks and secreting bioactive
agents as needed, not being particularly selective with
which portions of neurons they are found. Ependy-
moglia line the walls of the ventricles and the spinal
canal and contribute to the production of cerebrospi-
nal fluid. Except for select groups of neurons that
inhabit areas in and around the subependymal zones,
the ependymoglia have little in the way of direct
association with mature neurons. Most variable in
type, most intimately associated with all parts of neu-
rons, and thus most interesting functionally in their
relationships with neurons are the astroglia, or astro-
cytes. For these reasons, this brief review will focus
mainly on astrocyte-neuron interactions.
Textbooks that pay any attention at all to glial cells
usually present them, at the end of a chapter on nerve
cell types, as interesting structures that inhabit the
spaces left unoccupied by the neurons. Further treat-
ment of astrocytes in these sources usually includes
discussions of glial morphology as viewed in prepara-
tions immunostained for the intermediate filament

1043 - 4046 / 02 – $5.00 – COPYRIGHT © 2002 THE AMERICAN PHYSIOLOGICAL SOCIETY

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

225
 A P S R E F R E S H E R C O U R S E
glial fibrillary acidic protein (GFAP), and electron mi-
crographs showing the presence of such filaments.
Rarely is there any serious consideration given to
astrocytic contributions to brain function, for which
there is now much evidence, although occasionally
there is a passing reference to spatial buffering of K⫹.
Our task here is to advance physiology education a
step or two beyond what is found in the currently
used textbooks, a task that presents little difficulty in
the case of glial-neuronal interactions.
ASTROCYTES: A CLOSER LOOK
Conceptualization of astrocyte morphology from im-
munostains for cytoskeletal proteins such as GFAP
can be grossly misleading, because such stains reveal
so little of the cell’s structure. An elegant demonstra-
tion of just how little is revealed of the total volume of
astrocytes by GFAP stains was recently published (4).
Hippocampal astrocyte volumes were estimated from
injection of a fluorescent dye that filled the entire cell.
Immunostaining of the same cells for GFAP delineated
only ⬃15% of the cells’ total volumes. Driven home
by this finding is the idea that astrocytes occupy a far
greater amount of the space between the neurons
than would be projected by the commonly published
light microscopic images of these glia. Knowing this,
one can at least attempt to mentally fill in the other
large percentage (perhaps as much as ⬃85%) when
viewing such micrographs (Fig. 1). Bushong et al. (4)
FIG. 1.
A: an isolated cultured neocortical astrocyte immunostained for glial fibrillary acid
protein (GFAP). B: estimated space-filling capacity of the astrocyte in two dimensions
given by outlining plasma membrane.
VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002
226
R E P O R T
went on to show that the astrocytes in the CA1 region
of the hippocampal formation have nonoverlapping
domains, helping to correct yet another misconcep-
tion, as it is easy to assume from GFAP images that
there is considerable overlap in their occupation of
interneuronal spaces.
In addition to intermediate filaments, astrocytes also
express the microfilament actin, along with many
actin-binding proteins (15). Actin bundles are found
associated with the cytoskeleton and the plasma
membrane. This most likely accounts, in large part,
for the ability of astrocytes to extend and retract their
processes in response to activation of receptors, such
as ␤-adrenergic receptors, that influence the activity
Downloaded from advan.physiology.org on April 1, 2011
of adenylyl cyclase and the accumulation of cAMP
(see next section).
ASTROCYTE-NEURON JUXTAPOSITIONS
An obvious reason why one finds astrocytes associ-
ated with all parts of neurons is that astrocytes are
consistent inhabitants of synaptic regions, and virtu-
ally all parts of neurons are involved in synapses.
Shown in Fig. 2 is an example of this relationship
between synaptic boutons, dendritic spines, and as-
trocytes. Although this micrograph is from cerebellar
cortex, it could as well have come from many other
brain regions. It is clear that the astrocyte (blue) in
close apposition with several of the neural structures
 A P S R E F R E S H E R C O U R S E R E P O R T

Downloaded from advan.physiology.org on April 1, 2011

FIG. 2.
Electron micrograph showing a single astrocyte (blue) in intimate contact with axon terminals (At) making
asymmetric synapses with four spines (sp1, sp2, and two unmarked ones) (modified from Ref. 24, p. 165,
used by permission of Oxford Univ. Press).

would be exposed to many neurotransmitters/modu-
lators that are released and escape from the synaptic
cleft. Thus this glial cell stands to receive perhaps
several different signals, depending on the nature of
the synapses and the types of receptors expressed by
the astrocyte. Abundant evidence exists that transmit-
ters do indeed escape from the cleft in quantities
sufficient to trigger responses in neighboring cells (6).
Not only do astrocytes express transporters for up-
take of these molecules, they also have receptors for
many, if not all, of these transmitters, e.g., glutamate,
GABA, acetylcholine, etc. Furthermore, many modu-
lators, such as neuropeptides, are not even released
into the synaptic cleft, but rather into the perisynaptic
spaces (for review see Ref. 1). Dense-core vesicles,
representing these modulators, are rarely seen around
the active zones of synapses or synaptoid contacts
(see Fig. 12). Because astrocytes can respond to neu-
ral signals with shape changes, it may well be that the
relationship among the cells shown in Fig. 2 is only a
snapshot in time and that a different picture might be
seen after activation of certain of the synapses shown.
Illustrative of the shape changes produced in astro-
cytes by classical neurotransmitters are the micro-
graphs in Fig. 3. Primary cultures of astrocytes main-
tained in standard culture medium display a flattened
morphology, as seen in Fig. 3A. Treatment of such
cultures for 30 min with 100 nM epinephrine pro-
duces the changes seen in Fig. 3B, in which the
astrocytes retract from their former spread-out posi-
tions and become process bearing or stellate in shape.
Epinephrine, by definition a ␤-adrenergic agonist, was
most potent in inducing the response shown in Fig. 3.
Norepinephrine, a mixed-adrenergic agonist, was
only slightly less potent, and such responses are se-

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

227
 A P S R E F R E S H E R
FIG. 3.
Primary astrocyte cultures immunostained for GFAP.
A: flattened confluent control culture. Scale bar, 20
␮m. B: process-bearing appearance was produced by
30-min treatment of this culture with 100 nM epineph-
rine, a pure ␤2-agonist. Effect was blocked by IPS339,
a ␤2-antagonist.
lectively blocked by ␤-antagonists (11). ␤-Adrenergic
receptors are linked to the activation of adenylyl cy-
clase and the accumulation of cAMP. It is this second-
messenger pathway that is involved in astrocyte
shape changes (3). Therefore, the astrocyte high-
lighted in Fig. 2 may change its relationship with any
or all of the synapses upon release of a sufficient
amount of a particular transmitter from one or more
of the terminals. Data from both in vivo and in vitro
studies support the idea that these morphological
changes in astrocytes are all completely reversible.
MODEL SYSTEM ILLUSTRATES
MORPHOLOGICAL PLASTICITY
To illustrate the myriad ways in which such astrocyte
shape changes play a role in brain function, a brief
VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002
228
C O U R S E R E P O R T
review is in order of the model system in which
astrocyte morphological plasticity was first described
as part of a physiological process. More extensive
recent reviews of this phenomenon are available for
the interested reader (9, 17, 26). Figure 4 presents an
artist’s three-dimensional rendition of a rat brain, cut
away at the level of the hypothalamic supraoptic
(SON) and paraventricular (PVN) nuclei, revealing the
parvocellular PVN projections to the median emi-
nence and the projections of the magnocellular neu-
rons, via the pituitary stalk, to the neurohypophysis. It
is the magnocellular hypothalamo-neurohypophysial
system, in particular the supraoptico-neurohypophy-
sial portion, in which the glial-neuronal plasticity and
functional interactions were first discovered and have
Downloaded from advan.physiology.org on April 1, 2011
been extensively studied. A brief overview of this
system follows.
Production of the hormones oxytocin (OT) and vaso-
pressin (VP) is the responsibility of the two SON
neuron populations. Released in response to physio-
logical challenges such as dehydration, parturition,
and suckling of the young, OT and VP have well
known actions on peripheral tissues of the kidney,
uterus, myoepithelial cells of the mammary gland, and
certain vascular beds. Under basal conditions (i.e., in
the absence of such physiological challenges), SON
neurons are spontaneously active, although relatively
quiescent. Physiological activation of the system leads
FIG. 4.
Diagram of rat brain cut away at the level of the su-
praoptic (SON) and paraventricular (PVN) nuclei
showing the principal axonal projections of the mag-
nocellular neurons to the neurohypophysis.
 A P S R E F R E S H E R
to the development of two distinct, largely intrinsi-
cally generated firing patterns that characterize the
two cell types (for review see Ref. 2). As discussed
below, astrocyte signaling appears to have roles in
both the maintenance of quiescence under basal con-
ditions and the production and maintenance of the
activated states.
Neurons in the rat SON are large (15–25 ␮m in diam-
eter) and densely packed compared with hypotha-
lamic neurons in general. Also compared with that
reference group, the SON displays a higher degree of
organization, a factor contributing to its model system
status. Depicted diagrammatically in Fig. 5 are some
of the essential features of the SON involved in its
plastic capabilities. Positioned immediately lateral to
the optic chiasm/tract, the SON consists of somatic
and dendritic zones and a row of astrocytes that
makes up the ventral glial lamina (VGL). SON somata
generally give rise to a dorsally projecting dendrite
from which the parent axon usually arises (not
shown) and one to three ventrally projecting den-
FIG. 5.
Diagram of the SON and the pial-glial limitans in coro-
nal plane. Profiles representing somata are lateral to
the myelinated fibers of the optic tract (OT) in the
somatic zone (SZ). Ventral to the SZ are the parallel-
projecting dendrites (unfilled small circles), depicted
in cross section, and constituting the dendritic zone
(DZ). Mingling with only the most ventral dendrites
are the astroglial cell nuclei (larger filled circles),
whose ventrally projecting processes (shown in Fig.
6) fill the clear space between the basal lamina (small
arrows) and the most ventral dendrites. These glial
cell bodies and processes constitute the ventral glial
lamina (VGL). Dorsally projecting processes from
these glia fill most of the space not occupied by the
somata and dendrites. Ventrolaterally projecting den-
drites are not included here. Pia mater is indicated by
open arrows.
VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002
229
C O U R S E R E P O R T
drites. Because these latter dendrites turn to run ros-
trocaudally in parallel with one another, they are seen
in cross section in this coronal plane (Fig. 5). Filled
ovoid shapes representing the astrocyte nuclei are
seen at the dorsal aspect of the VGL. As is the case in
virtually all areas where brain parenchyma meets ex-
traparenchymal tissue, there is a basal lamina (small
arrows) separating the astrocytes of the VGL from the
pia mater (open arrows). Under basal conditions, the
VGL astrocytes project long processes dorsally
throughout the dendritic zone (DZ) and the somatic
zone (SZ), wrapping the dendrites and insinuating
themselves between adjacent somata (see Figs. 7 and
9). Astrocytic membrane that is not involved in the
dorsal projections fills the space between the DZ and
Downloaded from advan.physiology.org on April 1, 2011
the pial surface (Fig. 6A). Visualized with GFAP im-
munostaining, dorsally projecting processes from the
VGL astrocytes are shown wending their way through
the OT neuronal dendrites and somata (Fig. 6B). Un-
stained are the VP neurons that comprise the other
50% of SON nerve cells. Also unseen in this GFAP
immunostain is the remaining ⬃85% of each astrocyte
(see Fig. 1), which fills the spaces that appear to be
open between the neighboring dendrites or cell bod-
ies.
Electron micrographs reveal the close relationships of
the SON neurons to their associated astrocytes.
Readily seen in Fig. 7A are the fine astrocytic pro-
cesses that are interposed between these magnocel-
lular neurons of a prepartum pregnant animal. Strik-
ingly different is the picture one sees in the animal
during lactation (Fig. 7B), in which large areas of
neural membrane are closely apposed to (but not in
contact with) other neural, rather than glial, mem-
brane. Correlated with this glial withdrawal is an
increase in the number of multiple synapses (i.e., one
terminal forming synapses with two or more postsyn-
aptic elements). An example of such a synapse from a
lactating rat SON is shown in Fig. 8A, in which the
terminal is making both an axodendritic and an axo-
somatic synapse. Terminals may also make multiple
synapses exclusively with somata or with dendrites
(Fig. 9B), and as many as seven postsynaptic elements
have been observed being synaptically contacted by
one terminal. That multiple synapses increase in num-
ber whereas single conventional synapses decrease,
that there is no ingrowth of new terminals, and that
the time course of multiple synapse formation has
 A P S R E F R E S H E R C O U R S E R E P O R T

Downloaded from advan.physiology.org on April 1, 2011

FIG. 6.
A: electron micrograph showing the extensive astroglial membrane in the ventral glial
lamina of the SON. Dendrites (d), some of which are in bundles, are cut in cross section.
As, astrocyte nucleus. B: astrocyte (GFAP, green)-neuronal (red) appositions in the SON,
under conditions of low demand for the release of peptides. N, magnocellular OT neuron;
As, astrocytic process. Scale bars, 2 ␮m.

been observed to be too rapid for axonal sprouting to
occur have forced the interpretation that multiple
synapses in this system arise from single ones. How
this might happen is easily imagined from the situa-
tion shown in Fig. 8B, where a conventional synapse
is evident with the soma on the left (arrowhead).
Retraction of the fine astrocytic process (arrow)
would allow a second synapse to be made with the
soma on the right.
Another plastic change that accompanies activation of
this system is an increase in gap junctional intercellu-
lar communication among neurons, as inferred from
dye and tracer coupling studies as well as from in situ
hybridization for connexin protein mRNA (12, 16).
Such coupling among SON neurons is dendroden-
dritic and is observed exclusively between cells ex-
pressing the same peptide (i.e., either OT or VP).
Because dendrites that are wrapped by glia and iso-
lated from one another the way they are under basal
conditions (e.g., Fig. 9A) could not readily form gap
junctions, the glial retraction must play at least a
permissive role.
FUNCTIONAL SIGNIFICANCE OF
MORPHOLOGICAL PLASTICITY
Among the functional implications of these altered
glial-neuronal relationships are the following possibil-
ities. 1) There is reduced uptake and spatial buffering

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

230
 A P S R E F R E S H E R C O U R S E R E P O R T

Downloaded from advan.physiology.org on April 1, 2011

FIG. 7.
Electron micrographs of SON neurons illustrating the
changes in somasomatic direct apposition that occur
with activation of this system. A: day 1 postpartum;
small direct apposition (open arrow). B: lactating an-
imal; large areas of apposition (filled arrows). Scale
bar, 2 ␮m.

of K⫹, allowing higher concentrations of K⫹ to be

achieved in the extracellular space. This would lead
to enhanced neuronal activity. 2) Glial uptake of glu-
tamate from perisynaptic zones would be reduced,
with the consequence that this excitatory transmitter
would have prolonged action and might reach differ-
ent sites than when the glia are interposed (22). 3)
Enhanced multiple synapse and gap-junctional com-
munication establishes some basis for the synchro-
FIG. 8.
Electron micrographs of SON neurons showing types
of synapses common in lactating but rare in virgin
rats. A: axodendritic and axosomatic multiple syn-
apse. B: axon terminal making synaptic contact with
one soma and separated from another only by a thin
astrocytic process. Scale bar, 1 ␮m.

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

231
 A P S R E F R E S H E R C O U R S E R E P O R T

FIG. 9.
Electron micrographs of SON dendritic zones of animals under basal (A) and activated (B)
conditions. In A, dendrites tend to be separated from one another and wrapped by glial
processes (arrows). In B, dendritic membrane tends to be in apposition with that of other

Downloaded from advan.physiology.org on April 1, 2011

dendrites, forming dendritic bundles. Those dendrites with similar numbers are in
bundles. A multiple synapse is indicated (arrows). As, astrocyte nucleus. Scale bar, 2 ␮m.

nous firing of thousands of OT neurons that precedes ing dendrites. Because OT and VP have been found to
the milk ejection reflex. have receptor-mediated actions on these neurons (5,
13, 20), the glial presence or absence is likely to be of
Weaning of pups terminates lactation and returns the some functional significance.
SON to its prelactation morphology. Similar changes
are induced by dehydration produced by either water SON astrocytes contain the amino acid taurine (7). In
deprivation or saline drinking, and these, too, are the example shown in Fig. 10, the taurine-containing
reversible upon rehydration. Dehydration-induced
morphological plasticity in the DZ is illustrated in Fig.
9. In the well hydrated rat SON, the dendrites are
mostly separated from one another and are even
wrapped by processes of nearby astrocytes (Fig. 9A).
In contrast, the dendrites of a dehydrated animal form
bundles (i.e., direct appositions with no intervening
glial processes) upon the retraction of the astrocytic
processes from their previous interdendrite positions
(Fig. 9B). As in the case of the morphological plastic-
ity in vitro (shown in Fig. 3), changes in the dendritic
zone in situ have been observed after 20 –30 min of
osmotic stimulation (27). Plastic changes of the type
shown here, occurring in response to physiological
stimulation, also have implications for the fate of the
OT and VP that are released from the dendrites (25, FIG. 10.
21). Very likely, dendritic release of peptide would be Electron micrograph of a taurine-immunoreactive as-
tonically inhibited under basal conditions (see be- trocyte completely surrounding an axon (ax) and its
low), but any peptide that was released would imme- bouton (b) making contact with a supraoptic dendrite
diately be in contact with astrocyte membrane (Fig. (D). Visible in this astrocyte are glial filaments (gf).
Note that, although the astrocyte is replete with tau-
9A). In contrast, peptide released from dendrites un- rine, the dendrite and the afferent process contain
der the stimulated conditions (Fig. 9B) would have little or none. GSSP, gold-substituted silver periodate.
immediate access to the membrane of the neighbor- Scale bar, 290 nm.

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

232
 A P S R E F R E S H E R C O U R S E R E P O R T

astrocyte is engulfing a portion of a SON dendrite and
a presynaptic bouton. Release of taurine from SON
astrocytes (8) occurs in the well hydrated animal and
has an inhibitory effect on the firing of VP neurons, in
particular via glycine receptor activation (14). Strych-
nine blockade of these glycine receptors in a water-
loaded rat results in increased firing of VP neurons.
Acute dehydration reduces taurine release, and retrac-
tion of the glial processes accompanies chronic dehy-
dration, thereby further removing possible inhibitory
influences on the activation of the system in response
to the physiological challenge.
PLASTICITY IN THE NEUROHYPOPHYSIS
Axons from the magnocellular neurons terminate in
neurosecretory terminals and tend to occupy a large
proportion of the basal lamina, through which se-
creted peptides must pass to enter the fenestrated
capillaries leading to the general circulation (Fig.
11A). With activation of the system, in this case par-
turition, the pituitary astrocytes quickly become pro-
cess bearing (comparable to those in Fig. 3B), release
the engulfed terminals, and retract from their posi-
tions along the basal lamina (Fig. 11B). Neurosecre-
tory terminals are then found to abut most of the basal
lamina along the perivascular space. Note the termi-
nals devoid of dense-core vesicles (arrowheads) in Fig.
11B. Neurohypophysial astrocytes remain retracted
throughout lactation. Studies of dehydration-induced
plastic responses in these astrocytes have revealed

Downloaded from advan.physiology.org on April 1, 2011

the neurohypophysis (Fig. 4), where there is a popu-
lation of pituitary astrocytes that display the same
type and degree of plasticity as those in the SON.
These astrocytes also contain large amounts of tau-
rine, which they release under normal to low osmotic
conditions (18). Plasticity and signaling phenomena
in the neurohypophysis parallel those in the SON.
Under basal conditions, the flattened, spread-out as-
trocytes (comparable to those in Fig. 3A) engulf the
that the terminals abutting the basal lamina enlarge,
whereas those associated with the glia become
smaller. Accompanying these plastic changes are de-
creases in expression of certain extracellular matrix
proteins, presumably playing a permissive role in al-
lowing this reorganization of cellular processes (19).
Of special interest in this bidirectional signaling story
are the synaptoid contacts that occur in the neurohy-

FIG. 11.
Electron micrographs of neurohypophysial astrocytes from prepartum (A) and postpar-
tum rats (B). Note the neurosecretory axon profiles (*) surrounded by glial cytoplasm in
A. The astrocyte in B engulfs no axon profiles, some of which contact the basal lamina
(arrowheads) and are devoid of secretory granules. For clarity, the membranes of the
astrocytes have been highlighted.

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

233
 A P S R E F R E S H E R C O U R S E R E P O R T

pituitary astrocyte can provide appropriate signals to

the nerve terminal, and release of peptide from the
terminal can influence the glial cell. For instance, VP
released from the neurosecretory granules is likely to
trigger a sizeable intracellular calcium rise, as has
been shown in cultured pituitary astrocytes, mediated
by a V1 receptor (10). As discussed below, rises in
intracellular free calcium in astrocytes can lead to
release of neuroactive substances, such as glutamate,
which in this case could enhance secretion of pep-
tide. Interruption of this apparently positive feedback
relationship would come while the intracellular cal-
cium stores in the astrocyte reloaded.
FIG. 12.
High-power electron micrograph of a neurosecretory
axon containing both dense-core vesicles and clear
DIRECT ASTROGLIAL-NEURONAL SIGNALING

Downloaded from advan.physiology.org on April 1, 2011

microvesicles, making a synaptoid contact with a pi-
tuitary astrocyte (P). Scale bar, 224 nm.
pophysis between neurosecretory terminals, contain-
ing both clear microvesicles and dense core vesicles,
and the astrocytes (Fig. 12). Synaptoid contacts were
reported long ago to increase in frequency with de-
hydration (28), suggesting that they were involved in
the activation of this system. Such contacts are called
synaptoid rather than synaptic, because there is only
a presynaptic specialization and not a postsynaptic
one as there is in a true synapse. Nonetheless, the
juxtaposition is obviously favorable for two-way sig-
naling between the cells. Release of taurine from the
Because astrocytes have receptors for many neuro-
transmitters and neuromodulators, neuronal-glial sig-
naling is well established. New and accumulating ev-
idence that astrocytes engage in receptor-mediated
release of neuroeffector molecules establishes the glial-
neuronal signaling pathway. In addition to the astro-
cytic release of the inhibitory amino acid taurine men-
tioned earlier, astrocytes have now been shown to
release glutamate in response to a variety of signaling
molecules that induce significant rises in astrocyte
intracellular calcium (e.g., VP, bradykinin, ATP).
HPLC was used to show that the bradykinin-induced
calcium rise in astrocytes results in glutamate release
(Fig. 13A). Data from a variety of studies designed to

FIG. 13.
A: bradykinin (BK; 10 nM) releases measurable quantities of glutamate from astrocytes, as
measured by HPLC. B: cultures of forebrain astrocytes loaded with the fluorescent indi-
cator fura 2 and ratio imaged for calcium response to bradykinin (10 nM) stimulation.
Ordinate, ratio responses to two excitation wavelengths, 340 and 380 nm. Yellow-red
pseudocolor ⬇1 ␮M [Ca2ⴙ].

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

234
 A P S R E F R E S H E R C O U R S E R E P O R T

eliminate other possibilities (e.g., release via reversed

glutamate transporter) suggest that this release is exo-
cytotic and calcium dependent. In earlier studies,
calcium rises in astrocytes were evoked using meth-
ods that left open the question of what mechanism
actually led to release. For example, either mechani-
cal stimulation or application of certain neuropep-
tides, such as bradykinin, could evoke large calcium
responses in cultured astrocytes (Fig. 13B). More re-
cent experiments have established that a rise in inter-
nal free calcium from a resting level of ⬃85 to ⱖ140
nm is sufficient to release glutamate. Data from one
illustrative series of experiments (23) are presented
below and were kindly provided by Dr. V. Parpura,
FIG. 15.
University of California, Riverside. Neuronal inward currents (bottom trace) and glial

Downloaded from advan.physiology.org on April 1, 2011

intracellular [Ca2ⴙ] (top trace) produced by photolysis
To establish the intracellular calcium rise itself as of caged Ca2ⴙ producing two levels of intracellular
sufficient to release glutamate, the elegant experi- [Ca2ⴙ] in the astrocytes.
ment depicted in Fig. 14 was performed. Single neu-
rons were grown in culture on beds of several astro-
cium in the astrocytes. Simultaneous monitoring of
cytes (as in Fig. 13B). Both the neuron and the
the fluorescence generated by the calcium rise in the
astrocytes were loaded with the fluorescent free-cal-
glia and the inward current flow across the membrane
cium indicator fluo-3 and the caged calcium com-
of the neuron is shown in Fig. 15. Two successive
pound NP-EGTA (NPE). A patch electrode was then
pulses of UV light produce increments in the levels of
used in whole-cell configuration to dialyze out the
intracellular calcium and in the amplitude of inward
loaded compound from the neuron, after which a
current. Demonstration that this inward current was
pulse of UV light photolyzes some of the caged cal-
generated by activation of glutamate receptors on the
neuron was shown in a parallel experiment (Fig. 16).
Astrocyte fluorescence (Fig. 16A, top trace) and in-
ward current in the associated neuron (Fig. 16A, bot-
tom trace) followed a pulse of UV light, as before. In
the presence of the caged calcium (NP-EGTA) in the
glia and the absence of the N-methyl-D-aspartate
(NMDA) and non-NMDA antagonists D-2-amino-5-
phosphonopentanoic acid (D-AP5) and 6-cyano-7-nit-
roquinoxaline-2,3-dione (CNQX), these responses
were robust (Fig. 16B, left bars). When these antag-
onists were added to the medium, little or no inward
current was observed despite a still substantial cal-
cium rise in the glia (middle bars). In the absence of
both the antagonists and the caged compound, nei-
ther response was appreciable (right bars).

FIG. 14.
Method for selectively raising intracellular [Ca2ⴙ] in
astrocytes and direct measurement of effects on a
neuron placed on a small group of astrocytes. NPE,
NP-EGTA, a caged calcium compund. See text for de-
scription.
CONCLUSIONS
Taken together, these data strongly indicate that re-
ceptor-mediated intracellular calcium increases are
capable of inducing glutamate release from astrocytes
in sufficient quantities to generate sizeable currents in

VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002

235
 A P S R E F R E S H E R
FIG. 16.
Neuronal inward currents in response to raised intra-
cellular [Ca2ⴙ] in associated astrocytes are blocked by
glutamate receptor antagonists. A: astrocyte fluores-
cence (top trace) and neuronal inward current (bot-
tom trace) in response to UV pulse liberating caged
Ca2ⴙ (NPE-EGTA). B: bar graph showing responses
(means ⴞ SE) of the astrocytes (top) and the neurons
(bottom) in the presence (ⴙ) or absence (ⴚ) of caged
Ca2ⴙ and/or glutamate receptor blockade. D-AP5, D-2-
amino-5-phosphonopentanoic acid; CNQX, 6-cyano-7-
nitroquinoxaline-2,3-dione.
nearby neurons. There are, of course, clear implica-
tions in this glial-neuronal signaling for synaptic trans-
mission and neuromodulation, even away from the
synapse. That the astrocytic processes are mobile and
able to be retracted, also in response to receptor
activation, adds yet another probable level of com-
plexity to this phenomenon. Although there is much
work yet to be done in this exciting area, it seems that
the two-way communication between astroglia and
neurons is firmly established.
Thanks are due F. Mercier, V. Parpura, and T. Ponzio for helpful
suggestions on an earlier draft of the manuscript. This work was
supported by National Institute of Neurological Disorders and
Stroke Grants NS-09140 and NS-16942.
Address for reprint requests and other correspondence: G. I. Hat-
ton, Dept. of Cell Biology & Neuroscience, Univ. of California,
Riverside, CA 92521 (E-mail: glenn.hatton@ucr.edu).
Received 24 July 2002; accepted in final form 11 September 2002
REFERENCES
1. Agnati LF, Zoli M, Stromberg I, and Fuxe K. Intercellular
communication in the brain: wiring versus volume transmis-
sion. Neuroscience 69: 711–726, 1995.
2. Armstrong WE. Morphological and electrophysiological clas-
sification of hypothalamic supraoptic neurons. Prog Neurobiol
47: 291–339, 1995.
VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002
236
C O U R S E R E P O R T
3. Bicknell RJ, Luckman SM, Inenaga K, Mason WT, and
Hatton GI. ␤-Adrenergic and opioid receptors on pituicytes
cultured from adult rat neurohypophysis: regulation of cell
morphology. Brain Res Bull 22: 379 –388, 1989.
4. Bushong E, Martone M, Jones Y, and Ellisman M. Proto-
plasmic astrocytes in CA1 stratum radiatum occupy separate
anatomical domains. J Neurosci 22: 183–192, 2002.
5. Chevaleyre V, Dayanithi D, Moos F, and Desarmenien M.
Developmental regulation of a local positive autocontrol of
supraoptic neurons. J Neurosci 20: 5813–5819, 2000.
6. Dani JW, Chernjavsky A, and Smith SJ. Neuronal activity
triggers calcium waves in hippocampal astrocyte networks.
Neuron 8: 429 – 440, 1992.
7. Decavel C and Hatton GI. Taurine immunoreactivity in the
rat supraoptic nucleus: prominent localization in glial cells.
J Comp Neurol 354: 13–26, 1995.
8. Deleuze C, Duvoid A, and Hussy N. Properties and glial
origin of osmotic-dependent release of taurine from the rat
Downloaded from advan.physiology.org on April 1, 2011
supraoptic nucleus. J Physiol 507: 463– 471, 1998.
9. Hatton GI. Astroglial modulation of neurotransmitter/peptide
release from the neurohypophysis: present status. J Chem
Neuroanat 16: 203–222, 1999.
10. Hatton GI, Bicknell RJ, Hoyland J, Bunting R, and Mason
WT. Arginine vasopressin mobilises intracellular calcium via
V1-receptor activation in astrocytes (pituicytes) cultured from
adult rat neural lobes. Brain Res 588: 75– 83, 1992.
11. Hatton GI, Luckman SM, and Bicknell RJ. Adrenalin activa-
tion of ␤2-adrenoceptors stimulates morphological changes in
astrocytes (pituicytes) cultured from adult rat neurohypophy-
ses. Brain Res Bull 26: 765–769, 1991.
12. Hatton GI and Yang QZ. Peripartum interneuronal coupling
in the supraoptic nucleus. Brain Res 932: 120 –123, 2002.
13. Hurbin A, Orcel H, Alonso G, Moos F, and Rabie A. The
vasopressin receptors colocalize with vasopressin in the mag-
nocellular neurons of the rat supraoptic nucleus and are mod-
ulated by water balance. Endocrinology 143: 456 – 466, 2002.
14. Hussy N, Deleuze C, Pantaloni A, Desarmenien MG, and
Moos F. Agonist action of taurine on glycine receptors in rat
supraoptic magnocellular neurones: possible role in osmoreg-
ulation. J Physiol 502: 609 – 621, 1997.
15. Kalnins VI, Subrahmanyan L, and Opas The cytoskeleton
M. In: Astrocytes: Cell Biology and Pathology of Astrocytes,
edited by Federoff S and Vernadakis A. Orlando, FL: Academic,
1986, p. 27–59.
16. Micevych PE, Popper P, and Hatton GI. Connexin 32 mRNA
levels in the rat supraoptic nucleus: up-regulation prior to parturition
and during lactation. Neuroendocrinology 63: 39–45, 1996.
17. Miyata S and Hatton G. Activity-related, dynamic neuron-glial
interactions in the hypothalamo-neurohypophysial system. Mi-
crosc Res Tech 56: 143–157, 2002.
18. Miyata S, Matsushima O, and Hatton GI. Taurine in poste-
rior pituitary: localization in astrocytes and selective release by
hypoosmotic stimulation. J Comp Neurol 381: 513–523, 1997.
19. Miyata S, Takamatsu H, Maekawa S, Matsumoto N, Wa-
tanabe K, Kiyohara T, and Hatton GI. Plasticity of neuro-
hypophysial terminals with increased hormonal release during
dehydration: ultrastructure and biochemical analyses. J Comp
Neurol 434: 413– 427, 2001.
 A P S R E F R E S H E R
20. Moos F, Freund-Mercier MJ, Guerne Y, Guerne JM, Sto-
eckel ME, and Richard P. Release of oxytocin and vasopres-
sin by magnocellular nuclei in vitro: specific facilitatory effect
of oxytocin on its own release. J Endocrinol 102: 63–72, 1984.
21. Neumann I, Landgraf R, Bauce L, and Pittman QJ. Osmotic
responsiveness and cross talk involving oxytocin, but not va-
sopressin or amino acids, between the supraoptic nuclei in
virgin and lactating rats. J Neurosci 15: 3408 –3417, 1995.
22. Oliet S, Piet R, and Poulain D. Control of glutamate clear-
ance and synaptic efficacy by glial coverage of neurons. Science
292: 923–926, 2001.
23. Parpura V and Haydon P. Physiological astrocytic calcium
levels stimulate glutamate release to modulate adjacent neu-
rons. Proc Natl Acad Sci USA 97: 8629 – 8634, 2000.
VOLUME 26 : NUMBER 4 – ADVANCES IN PHYSIOLOGY EDUCATION – DECEMBER 2002
237
C O U R S E R E P O R T
24. Peters A, Palay S, and Webster H. The Fine Structure of the
Nervous System. New York: Oxford Univ. Press, 1991.
25. Pow DV and Morris JF. Dendrites of hypothalamic magno-
cellular neurons release neurohypophysial peptides by exocy-
tosis. Neuroscience 32: 435– 439, 1989.
26. Theodosis D. Oxytocin-secreting neurons: a physiological
model of morphological neuronal and glial plasticity. Front
Neuroendocrinol 23: 101–135, 2002.
27. Tweedle CD, Smithson KG, and Hatton GI. Rapid synaptic
changes and bundling in the supraoptic dendritic zone of the
perfused rat brain. Exp Neurol 124: 200 –207, 1993.
28. Wittkowski W and Brinkmann H. Changes of extent of
neurovascular contacts and number of neuro-glial synaptoid
contacts in the pituitary posterior lobe of dehydrated rats. Anat
Embryol (Berl) 146: 157–165, 1974.
Downloaded from advan.physiology.org on April 1, 2011