| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Andrus Gerontology Center and Department of Biological Sciences, University of Southern California (D.J.S., Y.S., C.P.A., C.E.F., I.R.), Los Angeles, California 90089-0191; Medical Department III, University of Leipzig (K.K.K.), D-04103, Leipzig, Germany
Address all correspondence and requests for reprints to: Dr. David J. Stone, Andrus Gerontology Center and Department of Biological Sciences, University of Southern California, Los Angeles, California 90089-0191. E-mail: dstone{at}almaak.usc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The hippocampus shows similar transient synaptic remodeling across the estrous cycle, in which there is a gradual increase in both dendritic spine density and synapses on dendritic spines in CA1 pyramidal neurons, followed by a rapid decrease during the 24 h between proestrus and estrus (6, 7). Although GFAP levels vary cyclicly in the hippocampus as well, there are obvious differences. Most strikingly, GFAP immunoreactivity in the hilus of the hippocampus is significantly increased on the afternoon of proestrus (8), 12 h before peak GFAP immunoreactivity in the hypothalamus. This suggests differential regulation of GFAP transcription by 17ß-estradiol (E2) in different brain regions.
To further elucidate the role of E2 in regulation
of GFAP, we examined several aspects of GFAP transcription. Regional
differences in transcription were examined in vivo by
in situ hybridization with an intron probe with which we
have observed transcriptional regulation of GFAP by glucocorticoids (9)
and increases in transcription with age (10). We also examined the
effects of astrocyte-neuron interactions on
E2-mediated GFAP transcription, because the
direction of the effect of glucocorticoids on GFAP transcription is
influenced by astrocyte-neuron interactions (9), and because the effect
of E2 on another astrocytic mRNA (apolipoprotein
E) is dependent upon cell-cell interactions (11). Finally, because
estrogen receptor (ER
) mRNA has been detected in astrocytes (12),
the role of a putative estrogen response element (ERE) in the GFAP
promoter was studied by site-directed mutagenesis and ER
binding.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
LH assay
Trunk blood was collected immediately after decapitation for LH
assay. LH levels were measured using the DELFIA rat LH assay (Wallac,
Oy, Finland). Briefly, serum was incubated in streptavidin-coated
plates with europium-(Eu)-labeled and biotinylated antirat LH
antibodies at room temperature. After washing and enhancement,
fluorescence was counted (Wallac 1230 Arcus Fluorometer; courtesy of
Stephanie Griffith, Harbor-University of California-Los Angeles Medical
Center). Rat LH reference preparation (AFP-7187B) was obtained from the
National Hormone and Pituitary Program (NIDDK, Bethesda, MD). Blood LH
levels were transiently increased 30-fold on proestrous evening
(P < 0.0001), suggesting that proestrous evening rats
were killed after the elaboration of glial processes and the decrease
in synapse number, as these precede the LH surge (2).
In situ hybridization
Sections were fixed in 4% buffered paraformaldehyde, washed in
PBS, and dehydrated in an ethyl alcohol series (30–100%). Sections
were prehybridized for 1 h at 55 C (prehybridization buffer: 0.75
M NaCl, 50% formamide, 10% dextran sulfate, and 0.05
M phosphate, pH 7.4) and hybridized with a
35S-labeled complementary RNA (cRNA) probe (9, 10).
Antisense 35S-labeled cRNA was transcribed from the
pBluescript transcription vector containing the 0.9-kb rat GFAP intron
I or 2.7 kb of rat GFAP complementary DNA including coding and
3'-untranslated region. Sections were hybridized for 3 h at 55 C
(mRNA probe) or for 18 h at 50 C (intron probe). Slides were then
covered with NTB2 emulsion (Eastman Kodak, Rochester, NY) and exposed
for 1 week (mRNA) or 3 weeks (intron) for cellular analysis. After
development, slides were counterstained with cresyl violet. Grain
density per cell was measured by computer videodensitometry for the 10
cells that most clearly had signal higher than background in the
arcuate nucleus, hilus, and outer molecular layer of the dentate gyrus
for each brain (see Fig. 1
). Background
measurements came from quantification of unlabeled cells. Unlabeled
cells had a signal intensity less than 1% of labeled cells. A detailed
analysis of frequency distributions of signal intensity for all regions
is in preparation.
|
Northern blot hybridization
Total RNA was extracted from cell culture using Tri-Reagent
(Molecular Research Center, Cincinnati, OH). Total RNA (5 mg) was
electrophoresed on a 1% denaturing agarose gel (0.66 M
formaldehyde) and transferred to nylon membranes. Membranes were
hybridized with a [32P]GFAP cRNA probe (9) in buffer
(50% formamide, 1.5 x SSPE (0.15 M NaCl, 1
mM EDTA, 11.5 mM
NaH2PO4, pH 7.4) 1% SDS, 0.5% powdered milk,
0.1 mg/ml yeast RNA, and 0.3 mg/ml single stranded DNA) for 12 h
at 54 C, followed by high stringency washing [0.2 x SSC
(standard saline citrate)-0.2% SDS; 73 C]. Membranes were placed
against PhosphorImager exposure cassettes, from which relative
radioactivity of bands was analyzed with a PhosphorImager system
(Molecular Dynamics, Sunnyvale, CA).
Transfection of primary astrocyte cultures
Rat GFAP promoter constructs with a luciferase reporter (9) were
transiently transfected into pure astrocyte cultures and
astrocyte-neuron cocultures by lipofection using the DOTAP transfection
reagent (Boehringer Mannheim, Mannheim, Germany). Eighteen hours
after transfection, the culture medium was changed. Twenty-four hours
later, transfected cells were treated with E2 (final
concentration, 1 pM) in serum-free, phenol red-free medium
and lysed after an additional 24 h. Transfection efficiency was
assayed by cotransfection with a simian virus 40 promoter-driven
ß-galactosidase gene (pSV-ß-galactosidase control vector, Promega,
Madison, WI). Luciferase activity was measured in cell lysates by the
luciferase assay system (Promega). The activity was normalized to total
protein with the Coomassie protein assay (Pierce, Rockford, IL) and
ß-galactosidase (Promega). Data were expressed as a percentage of the
vehicle-treated control value (mean ± SEM).
GFAP promoter luciferase constructs
Some 1.9 kb of the GFAP 5'-upstream region were excised from
BSSK (Stratagene, La Jolla, CA) using KpnI and
XbaI sites of the cloning cassette. After directional
subcloning into the KpnI and NheI sites of the
pGL3 Basic plasmid (Promega), the insert was confirmed by
sequencing.
GFAP promoter constructs were cloned into pGL3 Basic (Fig. 4).
Construct A contains the full 1.9-kb 5'-upstream sequence. The other
constructs result from consecutive deletions of the GFAP 5'-upstream
sequence (Fig. 4A, numbering according to the transcription start
site). Fragments were cut out using internal restriction sites (Fig. 4A) in combination with KpnI, SmaI, and
ApaI sites in the pGL3 basic vector. Cut sites were blunt
ended using T4 DNA polymerase and religated. Fragment A8 was
constructed using one of the internal XhoII sites in
combination with Bgl2 in the pGL3 basic vector and religated
into the pGL3 basic vector cut with Bgl2.
|
Gel shift
Two complementary 32-bp oligonucleotides were synthesized using
an Applied Biosystems DNA synthesizer (Foster City, CA): wild-type
containing the putative ER-binding site (ERE1)
found in the rat GFAP promoter
(CCTTGACTCTGGGTACAGTGACCTTGGTGGGG, coordinates
-172 to -141), and ERE mutant (CCTTGACTCTGTGTTCAG
TAGCCTTGGTGGGG). Underlining = mutation site of
ERE. Double-stranded oligomers were constructed by annealing
complementary DNA in saline-Tris-EDTA after heating to 95 C for 5 min
and cooling slowly to 4 C. The wild-type DNA fragment was then labeled
using [32P]deoxy-ATP in a polynucleotide transferase
reaction. Human recombinant ER (functionally active ER; 6.25
pM; Panvera, Madison, WI) were incubated in binding buffer
[10 mM HEPES (pH 7.9), 2 mM MgCl2,
50 mM KCl, 1 mM EDTA, 5 mM
dithiothreitol, 10% glycerol, 0.1% Nonidet P-40, 200 mg BSA/ml, and
0.2 mM phenylmethylsulfonylfluoride] with 0.12
pM labeled wild-type DNA and 0.06–12 pM
wild-type or mutant competitor DNA without radioactive label. Twenty
minutes later, samples were loaded on a 20-cm, low cross-link, 8%
polyacrylamide gel while running at 150 V at 4 C. The gel was run at
350 V for 4 h, dried, exposed to a phosphorimaging screen, and
analyzed using ImageQuant software (Molecular Dynamics).
Statistics
Data were analyzed by ANOVA, and statistical significance in
pairwise comparisons between means was assessed with Dunnett’s
post-hoc tests. All statistics were run on either StatView
4.5 or SuperANOVA (both from Abacus Concepts, Berkeley, CA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, A and B), the outer molecular layer of the dentate gyrus (Fig. 1, C and D), and the hilus of the hippocampal formation. In the arcuate
nucleus, both GFAP intron (Fig. 2A) and
mRNA (Fig. 2B) were increased (by 1.7- to 3-fold) on the afternoon of
proestrus, after the LH surge. In contrast, in the outer molecular
layer of the dentate gyrus, GFAP intron and mRNA were decreased
(30–40%) on the afternoon of proestrus (Fig. 2, C and D); here, the
ascending phase of mRNA levels lagged slightly behind intron levels. In
the hilus, where GFAP protein levels respond to estrogen and
progesterone (8) no changes in either intron or mRNA levels were
detected during the estrous cycle (Fig. 2, E and F).
|
; dose-response curve), whereas in the
presence of neurons GFAP mRNA was decreased by 50% when treated with 1
pM E2 (Fig. 3B).
|
), which suffices for transcriptional
regulation by glucocorticoids (9) and contains two putative EREs.
Transfected cells were treated with 1 pM
E2 for 24 h. As was observed for mRNA
levels, the full-length promoter showed an
E2-mediated increase in activity in pure
astrocytic cultures. When cocultured with neurons, the direction of the
response to E2 was reversed from induction to
inhibition of promoter activity (Fig. 3C). The sex of neonatal rats
from which astrocytes were derived did not alter the direction of the
GFAP response to E2 (data not shown).
GFAP promoter construct analysis: role of
ERE1 in the rat GFAP promoter
Using sequential deletions in the GFAP promoter (see
Materials and Methods and Fig. 4A), we examined the activity
of putative EREs. ERE1 is contained in construct
A3, whereas ERE2 is contained only within the
full-length promoter and deleted in construct A7. Constructs A1
(minimal promoter) and A2 (-106 bp) that do not contain ERE sequence
did not respond to E2. However, in longer
constructs (A3 to construct A7, which contain
ERE1), GFAP transcription was induced by
E2, which suggests that
ERE1 in the rat GFAP promoter is functional.
Moreover, constructs A3–A7 showed consistently greater responses to
E2 (1.6- to 1.8-fold) than the full-length
promoter (1.4-fold). Thus, ERE2 and/or other
sequences that are upstream of -1546 bp may inhibit the activity of
ERE1.
Site-directed mutagenesis of ERE1 within the A3 construct
(4 bp; see Materials and Methods) and ER gel shift assays
further show the functionality of ERE1. The
E2-mediated activation of the GFAP promoter was
abolished by mutation of ERE1 in the A3 construct
transfected into cultured astrocytes (Fig. 4B).
The binding of ER to the ERE1 site in the GFAP promoter
was assayed by gel shift. Oligonucleotide (32 bp) containing the
putative ERE1 site in the GFAP promoter was incubated with
human recombinant ER or BSA for a control. Complexes of ER
protein and DNA were resolved on a polyacrylamide gel (Fig. 4C).
Unlabeled 32-bp DNA containing wild-type ERE1 or mutated
oligonucleotide DNA (mERE1; containing a 4-bp mutation in
the putative ERE1-binding site; see Materials and
Methods) was used as a competitor in the binding reaction to
demonstrate that ER binds DNA specifically. Wild-type DNA at
equimolar and higher concentrations effectively competed for the
labeled ER protein/DNA complex (Fig. 4C, lanes 3–6).
mERE1 competed less effectively for binding of ER (Fig. 4C, lanes 7–10). Titration of the competition in a concentration range
from 2- to 100-fold revealed a 11.5-fold higher affinity of the
wild-type DNA for ER. These results show that ER protein binds
specifically to the wild-type ERE1-binding site in the GFAP
promoter and that the 4-bp mutations in the ERE1 reduced
the binding of ER in this competition assay.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
These results give further insight into the effect of estrogen on GFAP expression in response to brain lesions in the cerebral cortex and hippocampus. GFAP levels in the hippocampus are increased after deafferenting entorhinal cortex lesioning (ECL) (16, 17). Castration also increased GFAP mRNA, and this effect is additive with the effect of ECL and is reversed by the administration of E2 or testosterone (15, 18). Moreover, after a brain stab wound, gonadal steroids (E2, progesterone, or testosterone) decrease GFAP immunoreactivity at 72 h postlesion up to 2 mm away from the wound (19). E2 also increases reactive synaptogenesis in response to ECL (20, 21). Because inhibition of GFAP expression supports dendritic outgrowth (22), we suggest that decreased GFAP expression may be one mechanism by which E2 increases reactive synaptogenesis to the dentate gyrus in response to ECL. How these diverse steroids similarly enhance responses to brain lesions is very puzzling.
These data also give insight into the mechanism of neuroendocrine control of the LH surge. Because synaptic retraction in the arcuate nucleus precedes the LH surge, and because blocking the synaptic retraction on proestrus also blocks the LH surge (2), astrocyte process reorganization in the arcuate nucleus could be an early trigger of the cascade that leads to the LH surge. Estrogenic control of the LH surge, and thus ovulation, would be mediated by astrocytes through GFAP transcription as a primary locus of control.
The region-dependent effect of estrogen on GFAP expression parallels
the distribution of ER and ERß. In the arcuate nucleus, where
estrogen induces GFAP expression, ER is the prominent ER (23),
whereas in the hippocampus, where estrogen inhibits GFAP expression,
ERß is predominant (24). The number of astrocytes expressing these
receptors in these regions, however, is not known, and it is not clear
which nuclear ER mediates the E2 responses
observed in vivo or in vitro.
E2 regulation of GFAP transcription in
vitro
The induction of GFAP transcription by E2 is
consistent with the presence of two palindromic, ERE sequences in the
upstream region of the rat GFAP promoter: ERE1 (at -149
bp) and ERE2 (at -1830 bp; Fig. 4A). These EREs differ in
1–2 bases from the perfect palindrome consensus ERE in the
vitellogenin promoter of frog and chick (14).
We demonstrated the functional significance of ERE1 (at
-149) in the rat GFAP promoter. A 4-bp mutation in ERE1
abolished the E2-mediated increase in GFAP
expression. We also showed that human recombinant ER protein binds
to the wild-type ERE1-binding site in the GFAP promoter and
that the 4-bp mutation in this putative binding site reduces the
binding of ER by more than 10-fold. The ERE in the near upstream
region of the human GFAP promoter (at -150 bp) was also cited in an
abstract as functional on the basis of site-directed mutagenesis and
ER gel shift assays (25). Note that this ERE in the human GFAP
promoter is identical to ERE1 in the rat GFAP promoter
(26). As mentioned previously, ER has been localized to glial cells
in vitro (12), but ERß may also be involved. Currently, we
are examining the distribution of ER and ERß in rat brain
astrocytes.
Treatment with E2 induced the greatest increase in GFAP mRNA at a very low physiological concentration of 1 pM, which approximates plasma E2 levels after ovariectomy. Others also found that 1 pM E2 is sufficient to induce elaboration of astrocytic processes (27). The concentration of E2 in the microenvironment of astrocytes is not known; however, because E2 levels in CSF are less than 1/10th of blood levels (28), it is likely that E2 levels within the brain are lower than those in the periphery.
Cell-cell interactions and transcriptional inversion
Cell-cell interactions are involved in the regulation of GFAP
transcription by estrogen. Monotypic astrocytes responded to
E2 by induction of both mRNA levels and rates of
transcription as assayed by exogenous rat GFAP promoter activity after
transient transfection. In contrast, coculture of astrocytes with
neurons reversed these responses to inhibition. Similarly, the
transcriptional response of GFAP to glucocorticoids was inverted from
positive to negative by the presence of neurons in cocultures (9), the
latter of which was the direction of in vivo
glucocorticoid-induced responses throughout the brain (29). The ability
of neurons to alter the estrogen response in astrocytes is also region
dependent. Hypothalamic glial cells in culture responded to
E2 treatment with increases in
GFAP-immunoreactive processes when cocultured with hypothalamic
neurons, but not when in monotypic culture. However, coculture with
cerebellar neurons failed to change the estrogen response of
hypothalamic astrocytes from that of monotypic cultures (30).
In vivo, GFAP transcription is subject to both negative and
positive physiological influences by hormones and cytokines (this
report and Ref. 26). The ability of neurons in vitro to
switch the direction of E2- and
glucocorticoid-mediated GFAP expression suggests that local changes in
neuronal activities and/or astrocyte-neuron interactions in
vivo could accomplish the same result. Ongoing studies in the lab
implicate a tetra-phorbolacetate response element sequence
(AP-1-binding site) in the transcriptional inversion induced by both
E2 and glucocorticoids (Rozovsky, I., and C.
E. Finch, in preparation). These results are not surprising in light of
recent findings of differential transcriptional activation by the ER
and ERß at AP-1 sites (31), in which ER activates transcription
and ERß inhibits transcription. The effect of induced mutation of the
AP-1 site in the GFAP promoter on GFAP expression is currently under
investigation in this lab.
|
Footnotes |
|---|
Received December 17, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and -ß mRNA in the
rat central nervous system. J Comp Neurol 388:1–19[CrossRef][Medline]
and ERß at AP1 sites. Science 277:1508–1510This article has been cited by other articles:
 |
|
 |
|
  A. M. Wong, I. Rozovsky, J. M. Arimoto, Y. Du, M. Wei, T. E. Morgan, and C. E. Finch Progesterone Influence on Neurite Outgrowth Involves Microglia Endocrinology, January 1, 2009; 150(1): 324 - 332. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Parkash and G. Kaur Neuronal-glial plasticity in gonadotropin-releasing hormone release in adult female rats: role of the polysialylated form of the neural cell adhesion molecule J. Endocrinol., August 1, 2005; 186(2): 397 - 409. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M. Dhandapani, V. B. Mahesh, and D. W. Brann Astrocytes and Brain Function: Implications for Reproduction Experimental Biology and Medicine, March 1, 2003; 228(3): 253 - 260. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. B. Cashion, M. J. Smith, and P. M. Wise The Morphometry of Astrocytes in the Rostral Preoptic Area Exhibits a Diurnal Rhythm on Proestrus: Relationship to the Luteinizing Hormone Surge and Effects of Age Endocrinology, January 1, 2003; 144(1): 274 - 280. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Rozovsky, M. Wei, D. J. Stone, H. Zanjani, C. P. Anderson, T. E. Morgan, and C. E. Finch Estradiol (E2) Enhances Neurite Outgrowth by Repressing Glial Fibrillary Acidic Protein Expression and Reorganizing Laminin Endocrinology, February 1, 2002; 143(2): 636 - 646. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. M. Wise, D. B. Dubal, M. E. Wilson, S. W. Rau, and M. Bottner Minireview: Neuroprotective Effects of Estrogen--New Insights into Mechanisms of Action Endocrinology, March 1, 2001; 142(3): 969 - 973. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. B. Dubal, P. J. Shughrue, M. E. Wilson, I. Merchenthaler, and P. M. Wise Estradiol Modulates bcl-2 in Cerebral Ischemia: A Potential Role for Estrogen Receptors J. Neurosci., August 1, 1999; 19(15): 6385 - 6393. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Mong, E. Glaser, and M. M. McCarthy Gonadal Steroids Promote Glial Differentiation and Alter Neuronal Morphology in the Developing Hypothalamus in a Regionally Specific Manner J. Neurosci., February 15, 1999; 19(4): 1464 - 1472. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
