| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Division of Neuroscience, Oregon Regional Primate Research Center/Oregon Health Sciences University, Beaverton, Oregon 97006
Address all correspondence and requests for reprints to: Dr. Sergio R. Ojeda, Division of Neuroscience, Oregon Regional Primate Research Center, 505 NW 185th Avenue, Beaverton, Oregon 97006. E-mail: ojedas{at}ohsu.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(TGF), a tropic polypeptide
involved in the stimulatory control of LHRH secretion. The present
study shows that such lesions result in the rapid and selective
increase in Oct-2 transcripts in TGF-containing astrocytes
surrounding the lesion site. In both lesion-induced and normal puberty,
there is a preferential increase in hypothalamic expression of the
Oct-2a and Oct-2c alternatively spliced messenger RNA forms of the
Oct-2 gene, with an increase in 2a messenger RNA levels preceding that
in 2c and antedating the peripubertal activation of gonadal steroid
secretion. Both Oct-2a and 2c trans-activate the TGF
gene via recognition motifs contained in the TGF gene promoter.
Inhibition of Oct-2 synthesis reduces TGF expression in astroglial
cells and delays the initiation of puberty. These results suggest that
the Oct-2 gene is one of the upstream components of the glia to neuron
signaling process that controls the onset of female puberty in mammals. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A role for astroglia in this process was suggested by the surprisingly
profuse apposition of glial processes to the plasma membrane of LHRH
neurons (12) and the unique association between LHRH neurosecretory
axons and glial cells in the median eminence (13), the final common
pathway of hypothalamic neurosecretory neurons. Astroglial cells have
been shown to influence LHRH secretion by at least two means: 1)
through morphological interactions that regulate the access of LHRH
nerve endings to the portal vasculature of the median eminence (12, 14)
and, hence, the ability of the axons to deliver their neurosecretory
products for transportation to the anterior pituitary gland; and 2) via
production of growth factors recognized by tyrosine kinase receptors,
such as basic fibroblast growth factor (bFGF) (15, 16), insulin-related
growth factor I (17, 18), and the epidermal growth factor-related
peptides, transforming growth factor- (TGF) (19, 20) and
neuregulins (NRGs) (21, 22). Although some of these tropic molecules,
such as bFGF, are important for the early differentiation of the
neurons (15, 16), TGF and NRGs have been shown to predominantly
regulate the secretory activity of LHRH neurons (19, 21). In contrast
to bFGF and insulin-like growth factor I, which are recognized by
receptors located on LHRH neurons themselves, the stimulatory actions
of TGF and NRGs on LHRH release are exerted indirectly, via
receptors present on glial cells (19, 21, 23).
Because of these actions, TGF and NRGs have been postulated as
participants in the cell to cell signaling mechanism underlying the
astroglial control of LHRH neuronal function during sexual development
(21, 24). Both TGF and NRGs appear to activate, in a juxtacrine
manner, their cognate receptors located on astroglial cells, leading to
the production of bioactive molecules, such as PGE2, that
are able to directly stimulate LHRH release (25). The actions of TGF
and NRGs on hypothalamic astrocytes are intricately related to each
other, as activation of their respective erbB-1 and erbB-4 tyrosine
kinase receptors sets in motion at least one common intracellular
signaling pathway, by recruiting the coreceptor molecule erbB-2 (21, 22).
Although these studies implicate both TGF and NRGs as physiological
components in the neuroendocrine process controlling female sexual
maturation (21, 24, 26), nothing is known about the upstream molecules
responsible for activation of the genes encoding TGF, NRGs, and
their respective receptors at puberty. To begin addressing this issue,
we made two assumptions: 1) that at least some genes might be the same
as those that direct development of the neuroendocrine hypothalamus
during embryonic life; and 2) that these genes may become reexpressed
after hypothalamic lesions. As shown in other brain sites (27), such
lesions might recapitulate some early developmental events responsible
for the functional organization of the hypothalamus. Anterior
hypothalamic lesions lead to TGF gene expression in this region and
to sexual precocity (28). Conversely, blockade of epidermal growth
factor (erbB-1) receptors, the only known TGF recognition molecule,
prevents the advancement of puberty caused by the lesions, suggesting
that an increased TGF interaction with its receptor plays a critical
role in the process by which lesions accelerate female sexual
maturation (28).
Homeodomain genes of the POU family are attractive candidates to fill a role as transcriptional regulators of genes activated at puberty. In contrast to the Hox family of homeodomain genes, which is only expressed in the mid- and hindbrain, POU domain genes are widely expressed in the developing forebrain, particularly throughout its ventral aspect (29). That some POU domain genes may contribute to regulating specific, differentiated functions of the postnatal neuroendocrine brain is suggested by their persistent expression in discrete neuronal subpopulations of the adult hypothalamus (30). This adult expression appears to be limited to two classes of POU domain proteins: Oct-2 (31, 32, 33), which belongs to class II, and Tst-1/SCIP (34), Brn-1 (33), Brn-2 (33), and Brn-4 (35), which belong to class III. Several recent findings have provided direct evidence supporting the concept that these POU domain proteins function as transcriptional regulators of either neuropeptide or neurotransmitter genes expressed in the postnatal hypothalamus. For example, Oct-2 has been shown to repress tyrosine hydroxylase gene transcription (36), Oct-1 activates the neuron-specific enhancer of the LHRH gene (37), Tst-1/SCIP represses transcriptional activity of the LHRH gene (34), and Brn-2/Brn-4 trans-activates the gene encoding CRF (38), the neuropeptide controlling pituitary ACTH hormone secretion.
We now provide evidence implicating Oct-2 as a physiological component
of the cell-cell regulatory process by which glial cells up-regulate
the secretory activity of LHRH neurons at puberty. The results also
identify TGF, a key component of the signaling process by which
astrocytes influence the initiation of female puberty, as a downstream
target gene trans-activated by selective alternatively
spliced Oct-2 gene products in glial cells of the neuroendocrine
brain.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hypothalamic lesions
To accelerate the onset of puberty, immature 24-day-old female
rats were subjected to radiofrequency lesions of the anterior
hypothalamic area-posterior preoptic region (AHA-POA), as previously
described (28). At this age, the animals are in the early juvenile
phase of development (2). After the lesions (bilateral, 0.4 mm from the
midline), puberty is attained within 5–7 days (28), i.e.
about 1 week before the normal time of puberty.
Identification of POU domain genes expressed in the lesioned
hypothalamus
To determine whether hypothalamic lesions that cause sexual
precocity activate the expression of genes encoding POU domain
proteins, we used degenerate oligonucleotides complementary to highly
conserved regions in the amino-terminus of the POU homeodomain and the
carboxyl-terminus of the POU-specific domain (33) (Fig. 1
). Total RNA from the AHA-POA was
extracted (25) at different intervals after the lesion (8, 24, 48, 72,
96, and 144 h), and 100 ng were used for RT (39). Amplification of
the first DNA strand was carried out using 100 pmol of each degenerate
oligonucleotide and 35 cycles of amplification (denaturing at 94 C for
1 min, annealing at 55 C for 2 min, and extension at 72 C for 3 min).
The PCR products were electrophoresed on a 2% agarose gel, visualized
by ethidium bromide staining, isolated by electroelution and ethanol
precipitation, and subcloned into the pGEM-T vector (Promega Corp., Madison, WI). After transformation of competent XL1-blue
Escherichia coli cells (Stratagene, La Jolla,
CA) with the recombinant plasmids, the resulting colonies were streaked
onto Nytran membranes (Stratagene) and lifted onto
duplicate membranes for Southern blot detection of POU
domain-containing sequences. The probe used for the initial screening
was a random primer-labeled complementary DNA (cDNA) complementary to
the POU domain region of Tst-1/SCIP (33). The membranes were hybridized
and washed at low stringency [42 C, 5 x SSC (5 x SSC
= 0.75 M sodium chloride, 0.075 M
sodium acetate) and 1% SDS for the hybridization; 42 C, 0.5 x
SSC and 0.1% SDS for the final wash]. Clones showing different
degrees of hybridization intensity were sequenced by the
dideoxynucleotide termination method (40) using the Sequenase T7 DNA
polymerase and a kit (Sequenase version 2.0) purchased from
U. S. Biochemical Corp. (Cleveland, OH).
|
) and an antisense primer
(5'-CAT-CAA-GGC-TGG-TAA-GGG-3') complementary to a sequence in the
5'-end of exon 11. PCR amplification (35 cycles: 95 C, 45 sec; 55 C, 1
min; and 72 C, 2 min) with these primers yields a 242-bp DNA
fragment.
|
After cloning the PCR products into pGEM-T, followed by sequencing to confirm their identity, they were used as templates for the transcription of Oct-2 variant messenger RNAs (mRNAs). These in vitro transcribed mRNAs were then employed to construct reference curves for the estimation of tissue Oct-2 variants in a quantitative RT-PCR assay. The assay procedure employed has been described in detail previously (25, 42). The Oct-2 DNA templates were subcloned into the SmaI site of pSP64(poly(A)), a plasmid (Promega Corp.) containing a multiple cloning site flanked by an SP6 RNA polymerase promoter and a polyadenylated sequence. RNAs containing a polyadenylated tail were generated by SP6 RNA polymerase-directed transcription, and their yield was estimated by absorbance at 260 nm and by comparison with known amounts of RNA in ethidium bromide-stained gels.
Each assay tube contained two sets of primers: one to amplify the Oct-2
variant of interest (80 pmol each for Oct-2a and Oct-2b; 60 pmol for
Oct-2c), and another (at 2 pmol each) to amplify a segment of
cyclophilin mRNA, a constitutively expressed gene (43) used as an
internal standard for normalization of the Oct-2 variant values
obtained. Previous studies have shown that the content of cyclophilin
mRNA remains unchanged after similar hypothalamic lesions (28, 44).
Different amounts of standard mRNA were transcribed and amplified at
the same time as the unknowns. When performed as outlined (25, 39, 42),
the assay is optimized to minimize the two main sources of variability
in quantitative PCR: those due to differences in RT and primer
efficiency, and those related to tube effects and sample to sample
processing variability (45). The former source of variability is
reduced by referring the experimental values to mRNA standards
identical to the target sequences, amplified in the same assay; the
latter is minimized by coamplifying a fragment of cyclophilin, a
constitutively expressed gene that remains at similar levels throughout
postnatal development in brain (46). Aliquots of each PCR reaction were
electrophoresed on 2% agarose gels, transferred to Nytran membranes,
and hybridized to [
-32P]ATP end-labeled
oligonucleotides complementary to internal sequences in each of the
Oct-2 forms of interest. Upon exposure of the dry gels to Reflection
film (NEN, Boston, MA), the autoradiographic signals were analyzed by
computerized densitometry, as previously described (39, 42).
Hybridization histochemistry and immunohistochemistry
The cellular localization of Oct-2 mRNA, found to be the
predominant POU domain sequence amplified from lesioned hypothalami,
was detected by in situ hybridization (47), as previously
described (24, 28, 48), using an [35S]UTP-labeled Oct-2
antisense RNA transcribed from a cDNA template containing the POU
domain region of the Oct-2 gene. The template was obtained by RT-PCR
cloning of RNA derived from lesioned hypothalami.
As Oct-2 mRNA was found to be expressed around the lesion site, in
astrocytes previously shown to contain TGF (28), a double
immunohistochemical procedure followed by confocal microscopy was
employed to detect Oct-2 in TGF-containing astroglial cells. The
proteins were visualized in 50-µm Vibratome (Leica Corp., Nussloch, Germany) sections of brains fixed by
transcardiac perfusion of Zamboni’s fixative (39). The sections were
incubated overnight at 4 C with a mixture of polyclonal antibodies to
Oct-2 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA)
diluted 1:1000 and a monoclonal antibody to the carboxyl-terminus of
rat TGF (Neomarkers, Fremont, CA) diluted 1:100 in 0.02
M potassium phosphate buffer, pH 7.4, containing 0.9%
sodium chloride (KPBS), as previously described (49). The next day, the
sections were rinsed in KPBS and incubated for 45 min with a mixture of
fluorescent secondary antibodies. TGF was detected with a
fluorescein isothiocyanate (FITC)-conjugated goat-antimouse IgG (1:250;
Jackson ImmunoResearch Laboratories, Inc., Westgrove, PA).
Oct-2 was detected with a Cy5-conjugated goat antirabbit IgG (1:250;
Jackson ImmunoResearch Laboratories, Inc.). After several
washes with KPBS, the sections were mounted on Fisher brand
SuperFrost-plus slides (Fisher Scientific, Pittsburgh,
PA), dried for 15 min at room temperature, and coverslipped in aqueous
mounting medium (49). Controls consisted of sections incubated in the
absence of either both primary antibodies or each one individually.
Confocal microscopy
Immunofluorescence images were acquired with a Leica Corp. TCS NT laser scanning confocal system (Rockleigh, NJ),
based on a Leica Corp. IRBE microscope with an oil
immersion x40 PL APO 1.25NA objective. FITC was excited with the
488-nm line of an Ar gas laser and imaged through a 530 ± 30-nm
bandpass emission filter. Cy5 was excited with the 647-nm line of a
HeNe laser and imaged through a 665-nm longpass emission filter. The
two fluorophores were imaged simultaneously, but penetration of signal
from each fluorphore into the wrong channel was tested and made
negligible by adjusting the laser’s intensity and the
photomultiplier’s gain.
Assessment of promoter activity
A SmaI/SmaI (S/S) DNA fragment containing
nucleotides -637 to +67 of the rat TGF gene’s 5'-flanking region
(50) was subcloned into the luciferase reporter vector pGL2-Basic and
used for transcriptional regulation assays. This fragment contains both
an octamer-like motif (5'-GTGGAAAT-3'), similar to the octamer motif
5'-ATGCAAAT-3' that mediates Oct-2-dependent
trans-activation in Ig gene promoters (51), and a herpes
simplex virus (HSV)-like motif (5'-TAAATGAGTA-3'), similar to the
octamer-related motif (5'-TAATGARAT-3') known to mediate
Oct-2-dependent transcriptional repression (52). The putative octamer
motif is at position -610 to -603; the HSV-like motif is located
downstream at position -596 to -587.
To define the role of each of these two sites in mediating the effects
of Oct-2 variant proteins on TGF promoter activity, the sequences
containing either motif were selectively deleted from the promoter
sequence by site-specific mutagenesis. The S/S TGF promoter fragment
was used as the template. After annealing of a 34- or a 37-mer single
stranded oligonucleotide encoding the desired deletion, the
oligonucleotide was extended via a T4 DNA polymerase-driven reaction to
generate a hemimethylated, double stranded DNA. The resulting plasmid
molecules were digested with DpnI to remove double stranded
methylated nonmutant DNA, and then transferred to an Escherichia
coli bacterial strain deficient in DNA repair strand selection to
allow propagation of the mutated DNA. The entire procedure followed the
protocol recommended in a mutagenesis kit (MORPH) purchased from 5
Prime-3 Prime, Inc. (Boulder, CO).
The astroglial cell line C6 was used for the gene transfer studies. The cells were seeded into 6-well plates at 250,000 cells/well. The following day, Transfectamine (Life Technologies, Grand Island, NY) was used to transfect the reporter plasmids (at 250 ng/ml). In all cases, the cells were cotransfected with the plasmid CMV·Sport-ß-gal (Life Technologies) at 10 ng/ml to correct for transfection efficiency. To express Oct-2 protein isoforms, the expression vectors pSCT-Oct-2a (51) and pSCT-Oct-2c (31) were simultaneously transfected at different concentrations (10–50 ng/ml). In all transfections, the total amount of transfected DNA was maintained by adding the necessary amount of pSCT expression vector DNA to each well. All transfections were performed for 5 h.
Forty-eight hours later, the cells were rinsed with PBS, and 160 µl
lysis buffer (Promega Corp.) were added to each well. The
lysates were then centrifuged for 5 min at 4 C, and the supernatants
were used to assay luciferase activity (35 µl) and ß-galactosidase
(10 µl). Both the luciferase assay (Luciferase Assay System,
Promega Corp.) and the ß-galactosidase assay
(Galacto-Light, Tropix, Bedford, MA) were carried out according to the
manufacturer’s instructions. In both cases, the light emitted by each
reaction was detected with an E.G.&G Berthold Autolumat luminometer.
The luciferase values obtained (relative light units) were corrected
for transfection efficiency using the corresponding ß-galactosidase
activity values detected for each well. The final values are expressed
as a percentage of the activity displayed by the different TGF
promoter constructs in the absence of exogenously added Oct-2
protein-encoding plasmids.
Gel mobility shift assays
Nuclear protein extracts from the human lymphoma B cell line
BJAB, HeLa cells, and the glioma cell line C6 were prepared by the
abbreviated method of Andrews and Faller (53), using the cocktail of
protease inhibitors recommended by Kuhn et al. (54). The
double stranded oligodeoxynucleotides (ODNs) used as probes were
synthesized as complementary pairs on an automatic DNA synthesizer.
They were end labeled with [-32P]ATP in a reaction
catalyzed by DNA polynucleotide kinase and purified over a NICK column
(Pharmacia Biotech, Piscataway, NJ) before use. The
binding assay was performed using 2–6 µg protein, 20,000 cpm probe,
and 1 µg poly(dI-dC) in electromobility shift assay (EMSA) buffer (10
mM Tris pH 7.5, 50 mM NaCl, 1 mM
EDTA, 5 mM MgCl2, and 5% glycerol) in a
20-µl volume for 20 min at 25 C. The samples were subjected to
electrophoresis on a 4% nondenaturing polyacrylamide gel using 0.05
M Tris and 0.38 M glycine buffer, pH 8.8,
containing 2 mM EDTA as the running buffer at 4 C for
3 h at 100 V. The gels were then dried and exposed to film at -85
C.
The oligonucleotides used for binding were an IgG octamer
(5'-GGT-AAT-TTG-CAT-TTC-TAA-3') identical to the enhancer region
containing the octamer motif [nucleotides (nt) 537–554] in the mouse
heavy chain Ig gene (55), a TGF octamer
(5'-ACA-GGT-GGA-AAT-TCG-ACT-3') corresponding to the sequence located
between nt -614 to -597 in the TGF promoter (50) and that contains
the putative octamer motif, and a sequence
(5'-CGA-CTT-AAA-TGA-GTA-TTT-3') corresponding to the region -601 to
-584 in the TGF promoter that contains the putative HSV-like
motif.
To determine the ability of authentic Oct-2a and Oct-2c to bind to the
TGF octamer-like motif, the Oct proteins were transiently expressed
in C6 cells by transfection with the expression vectors pSCT-Oct2a and
pSCT-Oct-2c. In each case, the cells were seeded in 100-mm dishes at 1
million cells/dish. The next day the plasmids (1 µg/ml) were
transfected for 5 h using Transfectamine (Life Technologies, Inc.) according to the procedure recommended by the
manufacturer. The cells were collected 24 and 48 h later for
extraction of nuclear proteins and binding to the oligonucleotides
containing the IgG and TGF octamer sequences (see above).
Targeted disruption of Oct-2 synthesis
The antisense ODN (Oct-2 ODN) used to disrupt the synthesis of
Oct-2 proteins was directed against the sequence surrounding the first
ATG codon (31) in the rodent Oct-2 gene (5'-GGC-AGC-ATG-GTT-CAT-TCC-AGC
-3'). Although there is a second downstream ATG codon, the sequences
surrounding the first codon conform more precisely to the Kozak
consensus sequences for a translation initiation site (31). Two ODNs
were used as controls. One of them contains the same nucleotide
composition of the antisense ODN, but in a scrambled order. This
scrambled sequence does not bear similarity with any sequence deposited
in GenBank to date. The other control ODN was directed against the
5'-end of the unrelated homeodomain gene TTF-1 (56) and had the
sequence 5'-TGG-ACT-CAT-CGA-CAT-CGA-CAT-GAT-3'. Hypothalamic astrocytes
were employed to verify the effectiveness and specificity of the Oct-2
ODN in reducing Oct-2 protein levels. They were also used to examine
the consequences of ODN-mediated inhibition of Oct-2 synthesis on
TGF mRNA levels.
Astrocytes were purified from 1- to 2-day-old rats, as previously described (21, 23). After their initial seeding, the cells reached confluence in 8–10 days. At this time, the astrocytes were isolated from other contaminating cells by first shaking the flasks at 250 rpm for 6 h, replacing the medium, and then shaking again for another 18 h. Thereafter, they were replated in 6-well plates (800,000 cells/well) for ribonuclease (RNase) protection assay and in 100-mm dishes for Western blots. The cells were used for the experiments upon reaching 80–90% confluence.
To facilitate penetration of the ODNs into the cells, each ODN was combined with the synthetic cationic lipid DOTAP(Boehringer Mannheim, Indianapolis, IN), as recommended (57). After incubating a 20-fold concentrated mixture for 10 min at room temperature, the mixture was diluted with culture medium to a final concentration of 10 µM ODN/13 µM DOTAP and added to the cells for 72 h. At this time, the dishes were processed for collection of nuclear proteins, followed by Western blots, and the six-well plates were processed for RNA extraction and RNase protection assay. The RNase protection assay and the densitometric analysis of the hybridization signals were carried out as previously described (58). The proteins were extracted as outlined above for EMSA, size fractionated by electrophoresis in a 8% SDS-polyacrylamide gel, and transferred to nitrocellulose. The membranes were probed with antibodies to Oct-2 or Oct-1 (both from Santa Cruz Biotechnology, Inc.), diluted at a 1:1,000 (0.1 µg/ml), followed by incubation with a goat antirabbit horseradish peroxidase-linked second antibody (1:10,000). The conjugates were detected using the Super Signal Ultra chemiluminescent system from Pierce Chemical Co. (Rockford, IL).
Intracerebroventricular infusion of Oct-2 antisense ODN
These in vivo experiments were performed using the
same ODN found to be effective in inhibiting Oct-2 synthesis in
vitro. The ODN was chronically infused into the third ventricle of
the brain via an infusion cannula (Plastic One, Inc., Roanoke, VA)
connected to an Alzet miniosmotic pump (model 2002, Alzet Corp., Palo
Alto, CA) implanted sc (21, 28). These pumps have a flow rate of 0.5
µl/h and a capacity of 200 µl, which results in a delivery period
of 14 days. Each pump was loaded with 5 µg/µl of either the
antisense ODN or the scrambled sequence diluted in artificial
cerebrospinal fluid (59). After preincubating the pumps at 37 C for
4 h, the assembly was implanted into intact 25-day-old juvenile
female rats, i.e. at an age that antedates the prepubertal
increase in hypothalamic Oct-2 mRNA content (see Results).
Animals were monitored daily for vaginal opening from day 30 onward.
Once vaginal opening occurred, vaginal lavages were obtained daily to
estimate the time of first ovulation (2). The animals were killed on
the day of first diestrus (defined by a predominance of leukocytes in
the vaginal lavage), following an estrous type of vaginal cytology.
This change accurately defines the occurrence of the first ovulation
(26). In all cases, ovulation was visually confirmed by the detection
of corpora lutea.
Phases of peripubertal sexual development
The developmental changes in hypothalamic Oct-2a and Oct-2c gene
expression were examined at ages corresponding to key stages of sexual
maturation in the rat (2). The juvenile period extends from postnatal
day 21 to days 28–30. It is at the end of this period that the
central, ovarian-independent mechanisms that set into motion the onset
of puberty, become activated (2). Puberty itself proceeds in sequential
stages, which have been defined according to specific criteria (2).
During late juvenile development, the animal’s vagina is not patent,
and the uterine weight is 60 mg or less, without accumulation of
intrauterine fluid. During early puberty (early proestrous phase), the
animals have an enlarged uterus and detectable intrauterine fluid (an
indication of estrogen secretion). Thereafter, the uterus becomes
ballooned with fluid, and its weight increases to more than 200 mg. The
first preovulatory discharge of LHRH and gonadotropins takes place on
the afternoon of this day, which corresponds to a phase termed late
proestrus. Ovulation occurs in the early morning of the next day (first
estrus). At this time, the vagina becomes canalized with a cytology
showing a predominance of cornified cells. The first diestrous then
follows due to the formation of the first corpora lutea. During this
phase of puberty, vaginal cytology shows a predominance of leukocytes,
and the ovaries contain fresh corpora lutea.
Statistics
Changes in Oct-2a, Oct-2b, and Oct-2c mRNA levels at different
intervals after a puberty-inducing lesion and during different stages
of sexual development were analyzed by one-way ANOVA followed by the
Student-Newman-Keuls multiple comparison test for unequal
replications.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Almost 50% of the PCR products analyzed
contained the sequence of Oct-2 (Fig. 1), a class 2 member of the POU
domain gene family (29). A minor fraction of the cDNAs analyzed
contained sequences corresponding to the POU genes Tst-1/SCIP (three
clones), Brn-1 (one clone), and Brn-3 (one clone). The rest of the
cDNAs did not contain POU domain sequences. The fraction of
Oct-2-containing clones remained relatively constant at all intervals
examined, with the exception of 8 h. At this time, one of five
cDNAs was Oct-2 positive.
Alternatively spliced forms of the Oct-2 gene are expressed in the
lesioned hypothalamus
Alternative splicing of the Oct-2 gene in B lymphocytes and the
central nervous system generates several isoforms (31, 32, 41, 60, 61).
Three of these forms, termed Oct-2a, Oct-2b, and Oct-2c (31), appear to
be particularly abundant in the embryonic nervous system and to persist
in discrete regions of the adult brain (31, 32). To determine which of
them is expressed in the normal hypothalamus and in the hypothalamus
subjected to a puberty-inducing lesion, we designed primers to amplify
the alternatively spliced sequences contained within the 3'-region of
the Oct-2 gene (31, 32, 41). Figure 2 shows the structural arrangement
of the gene in this region and the location of the primers. Although a
common 5'-primer corresponding to a sequence in exon 9 was used to
amplify Oct-2a and Oct-2b, a 5'-primer corresponding to a downstream
region in the same exon was used to amplify Oct-2c. The 3'-primers used
were specific for each isoform. The 3'-primer that amplified Oct-2a
recognizes a sequence located at the beginning of exon 11, the
3'-primer for Oct-2b is complementary to a sequence in the 74-bp
insertion that creates exon 10 in this mRNA, and the 3'-primer for
Oct-2c recognizes a sequence contained in the DNA segment spliced
between exons 10 and 11 in this form (Fig. 2).
Because in the original screening for POU domain genes the fraction of
PCR products containing the Oct-2 sequence was relatively constant
between 48–96 h after the lesion, an intermediate time (72 h) was
selected for identification of Oct-2 alternatively spliced forms. Total
RNA from lesioned hypothalami collected at this postlesion interval was
reverse transcribed and amplified with the above-described primers.
Sequence analysis of the products (Fig. 3) demonstrated that all three Oct-2
forms (a, b, and c) were expressed in the lesioned hypothalamus. With
the exception of a few conservative substitutions (one Ser to Thr in
each cDNA, and one Thr to Ala in Oct-2a), the amino acid sequence
encoded by these cDNAs was identical to those reported for human Oct-2a
and mouse Oct-2b and Oct-2c (31, 32) (Fig. 3).
|
illustrates the main features of this
assay. PCR amplification of different amounts of in vitro transcribed
Oct-2c yielded a concentration-dependent increase in the amount of PCR
product detected by hybridization to an oligonucleotide complementary
to an internal sequence in the amplified fragment (upper right
panel). Regression analysis of these hybridization signals
demonstrated that the amplification reaction is linear under the
conditions employed (left panel). The lower right
panel depicts a representative profile of Oct-2c mRNA expression
at different intervals after a puberty-inducing lesion. In each sample,
the Oct-2 fragment of interest is coamplified with a fragment of the
constitutively expressed cyclophilin gene (not shown). The signal
provided by the cyclophilin product is then used to normalize the Oct-2
mRNA values obtained in each sample (for other details, see
Materials and Methods).
|
, upper panel). In contrast, Oct-2b mRNA levels increased
slightly after the lesion to became significantly elevated only at
144 h (Fig. 5, middle panel), a time at which most
lesioned animals were already undergoing puberty (as assessed by the
accumulation of uterine fluid, vaginal opening, or ovulation).
Surprisingly, the pattern of Oct-2c mRNA expression was different from
that of either Oct-2a or Oct-2b (Fig. 5, lower panel). The
mRNA levels remained unchanged for the first 24 h after the
lesion, increased abruptly between 48 and 72 h, and returned to
basal levels by 96 h.
|
). In fact, both Oct-2a and Oct-2b mRNA
levels appeared to decrease within 8–24 h after the lesion, remaining
at low values thereafter.
|
), i.e. preceding the first
endocrine manifestations of the pubertal process. Plasma steroid levels
during this particular phase of development remain at basal,
prepubertal values (not shown). A secondary increase (P
< 0.05) in Oct-2a mRNA occurred between the late juvenile phase (day
28) and the morning of the day of the first preovulatory gonadotropin
surge, with values declining in the afternoon (when the gonadotropin
surge occurs) to levels similar to those seen at the initiation of
puberty, i.e. late juvenile-early proestrous phases (Fig. 7, upper panel). Similar, but less well defined, changes were
observed in the medial basal hypothalamus (lower panel).
|
).
The Oct-2 gene is expressed in reactive astrocytes surrounding the
lesion site
Hybridization histochemistry of the brain from animals subjected
to hypothalamic lesion 72 h earlier revealed the presence of Oct-2
mRNA in the frontal cortex (Fig. 8A, arrows) and the piriform cortex (arrowheads), a
localization previously reported by others (31, 32). As previously
noted (32, 33), we detected strong hybridization to the suprachiasmatic
nucleus (Fig. 8B), a rostral hypothalamic structure not compromised by
the lesion. In addition to this normal distribution, an abundance of
Oct-2 transcripts was observed in the area surrounding the lesion site
(Fig. 8, A and C). The intensity of the hybridization signal in this
region was markedly greater than the uniformly low level of
hybridization observed in the same region of the AHA of intact animals
(Fig. 8D). Sections incubated with a sense probe did not show specific
hybridization in any of these regions (not shown). The distribution
pattern of the Oct-2 hybridization signal around the lesion site (Fig. 8, A and C) and the predominant association of silver grains with small
nuclei in this area (Fig. 8E) suggested that most of the expression
occurs in astrocytes surrounding the lesion site.
|
in reactive astrocytes
72–96 h after similar lesions of the AHA-POA (28). Double
immunofluorescence followed by confocal microscopy revealed the
presence of Oct-2 immunoreactivity in TGF-containing astrocytes
surrounding the lesion site (Fig. 9). In
most cases the immunoreactive material appeared to be confined to the
nucleus of astrocytes (Fig. 9, arrowheads), but it was also
detected in GFAP-negative cells (Fig. 9, middle and
left panels, red color).
|
gene transcription in glial cells via
an octamer-like motif, reactive astrocytes surrounding the site of a
puberty-inducing lesion express both the TGF and Oct genes.
Moreover, the temporal pattern of changes in Oct-2 variant mRNA levels
observed after hypothalamic lesions and during the onset of normal
puberty correlated well with the pattern of TGF gene expression
observed in both situations (24, 28). The possibility that TGF may
be a target gene for Oct-2 action was, therefore, considered.
Examination of the TGF promoter sequence (Fig. 10) revealed the presence of both an
octamer-like sequence between nt -610 and -603, and a downstream
sequence between nt -596 and -587 similar to the TAATGARAT element in
the HSV promoter, which has been shown to mediate inhibitory
transcriptional effects of Oct-2 proteins (52, 63).
|
promoter (-637 to +67) that contains both the octamer
and the HSV-like motifs, and expression vectors containing either the
Oct-2a or Oct-2c coding regions resulted in stimulation of the
promoter’s activity (Fig. 10A). The effect appeared to be additive, as
judged by the significantly greater (P < 0.01)
promoter response to increasing concentrations of Oct-2c in the
presence of a constant dose of Oct-2a (Fig. 10A). The potency of the
Oct-2 variants to increase TGF promoter activity was consistent with
that previously reported for Oct-2 isoforms on other genes in neuronal
cells (64). Deletion of the octamer-like motif abolished the
stimulatory effect of both Oct-2 proteins (Fig. 10B). In contrast,
deletion of the HSV-like motif did not reduce the ability of the Oct-2
proteins to increase TGF transcriptional activity (Fig. 10C).
Oct-2a and Oct-2c bind to the canonical octamer motif of Ig genes,
but not to the TGF octamer-like motif
Nuclear protein extracts from the B lymphocyte cell line BJAB,
which is rich in Oct-2 proteins, strongly bound to the IgG octamer
(Fig. 11, left panel). The
heaviest (slowest migrating) protein-DNA complex corresponds to the
well characterized binding of Oct-1 to this octamer sequence (60, 65).
The second, less retarded, and, in our hands, less well resolved
complex contains both Oct-2b and Oct-2a proteins (51, 60). There was
also a fainter, faster migrating complex previously noticed by others
(66). In contrast to this pattern of migration, binding of BJAB nuclear
proteins to either the TGF octamer (Fig. 11, left panel)
or the TGF HSV-like motif (not shown) resulted in the formation of a
single, smaller protein-DNA complex of a size comparable to that of the
fastest migrating BJAB-IgG octamer complex (Fig. 11, left
panel).
|
, middle panel). The same extracts did not form similar
protein-DNA complexes when bound to the TGF octamer, but, instead,
gave rise to a slower migrating band (Fig. 11, right panel,
arrow). Both IgG-octamer- and TGF-octamer-binding proteins were
more prominently expressed at 24 than at 48 h after
transfection.
Disruption of Oct-2 synthesis reduces TGF gene expression in
astrocytes and delays the onset of puberty
In two separate experiments, hypothalamic astrocytes exposed for
72 h to the antisense Oct-2 ODN (10 µM) exhibited an
approximately 50–70% reduction in the content of a protein
corresponding in size to that of the Oct-2a complex (41) present in the
human B lymphocyte cell line BJAB (Fig. 12B). The selectivity of this effect
was shown by the inability of Oct-2 ODN to alter the levels of an
approximately 40-kDa protein nonspecifically recognized by the Oct-2
antibodies (Fig. 12B). More importantly, the treatment failed to reduce
the content of the related POU domain protein Oct-1, a ubiquitously
expressed transcriptional regulator (65) (Fig. 12A). An ODN with an
identical base composition as Oct-2 ODN, but arranged in a scrambled
order (Oct-2 SCR), altered neither Oct-1 nor Oct-2 protein content
(Fig. 12, A and B). Likewise, an antisense oligonucleotide directed
against the 5'-end of TTF-1, an unrelated homeodomain gene highly
expressed in the embryonic diencephalon (67) and the postnatal
hypothalamus (Lee, B. J., et al., unpublished
observations) failed to affect Oct-1 or Oct-2 protein levels (Fig. 12, A and B).
|
gene
expression, as evidenced by the reduction in TGF mRNA levels
detected by RNase protection assay (Fig. 12C). The specificity of this
effect was demonstrated by the inability of Oct-2 ODN to alter the
steady state levels of cyclophilin mRNA, which is constitutively
expressed in brain and other tissues (43, 68), and by the
ineffectiveness of Oct-2 SCR to alter TGF mRNA levels. Densitometric
analysis of the hybridization signals followed by normalization to the
cyclophilin signal (derived from a shorter film exposure) demonstrated
a 6-fold reduction in TGF mRNA levels after exposure of the cells to
Oct-2 ODN (Fig. 12D).
Chronic infusion of the Oct-2-ODN into the third ventricle of female
rats, initiated before the initial, ovarian-independent increase in
hypothalamic Oct-2a mRNA levels significantly (P <
0.01) delayed the onset of puberty, as assessed by the age at first
ovulation (Fig. 13). By 38 days of age,
100% of the animals infused with Oct-2 SCR and 90% of the untreated
control rats had ovulated compared with only 30% of the Oct-2
ODN-infused rats.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Unexpectedly, the analysis of octamer proteins in isolated brain cells by EMSAs revealed that Oct-2 proteins are much more abundant in astrocytes than neurons (32, 60). A logical inference from these findings is that Oct-2 gene products may not only regulate transcription of neuron-specific genes, but also transregulate gene subsets expressed in astrocytic networks of the postnatal brain. To date, neither the downstream genes targeted by Oct-2 proteins in glial cells nor the physiological functions that may be affected in these cells by Oct-2-dependent regulatory events have been identified.
The present results indicate that the TGF gene expressed in
astroglial cells is one of the downstream genes subjected to Oct-2
regulatory control. More importantly, our results place this regulatory
action within the context of a complex physiological event, as they
indicate that the Oct-2 gene is one of the components of the
developmental process by which the neuroendocrine brain controls the
advent of female reproductive competence. We demonstrate that both an
injury of the neuroendocrine brain that induces sexual precocity and
the natural onset of puberty itself result in differential activation
of three alternatively spliced products of the Oct-2 gene expressed in
brain (31), namely the Oct-2a, -b, and -c isoforms. Although
accumulation of Oct-2a transcripts is maximal within 8 h after the
lesion, Oct-2b mRNA levels do not increase significantly until 4–5
days later. In striking contrast, Oct-2c mRNA content remains at basal
levels at the time when Oct-2a mRNA content is already maximally
elevated and increases markedly, but transiently, between 48–72 h
after the lesion. During normal peripubertal development hypothalamic
Oct-2a mRNA levels increase during the second half of the juvenile
period, a time that heralds the initiation of puberty, and then again
on the day of the first preovulatory surge of gonadotropins that
triggers the first ovulation. In close similarity to the profile
observed after a puberty-inducing lesion, hypothalamic levels of Oct-2c
transcripts increase only on the day of first proestrus,
i.e. many hours after the initial, juvenile increase in
Oct-2a mRNA content. Thus, both injury-induced puberty and normal
puberty are associated with activation of Oct-2 gene expression and a
pattern of alternative splicing events that results in the
differential, but temporally correlated, appearance of Oct-2a and 2c
mRNA forms in the hypothalamus. Although the factors responsible for
the regulation of these alternatively spliced forms are not known, our
results demonstrate that the activation of Oct-2 gene expression caused
by a puberty-inducing lesion occurs in reactive astrocytes surrounding
the lesion site.
All existing evidence indicates that the basic response of the central
nervous system to injury is similar in different regions of the brain
(70). Thus, it would be fully expected that lesions in regions other
than the hypothalamus would result in similar cellular and molecular
responses. The functional consequences of such lesions will, however,
depend on the region affected. In most cases, brain injury will result
in loss of function. Although it is still possible that lesions of the
anterior hypothalamus result in loss of inhibitory neuronal systems
controlling LHRH secretion (71), recent studies have shown that such
lesions advance puberty because they lead to the activation of
facilitatory systems involved in the stimulation of LHRH neuronal
activity (72). The TGF/erbB-1 receptor signaling module is one of
these facilitatory systems. Both the ligand and the membrane-anchored
recognition components of the system are present in hypothalamic
astrocytes (23, 24), and their synthesis is rapidly increased after a
puberty-inducing lesion (28, 73). The contributions of erbB receptor
activation to the process by which hypothalamic lesions accelerate
puberty and to normal puberty are demonstrated by the ability of an
erbB-1 receptor blocker targeted to the site of the lesion or to the
median eminence of the hypothalamus to prevent the acceleration of
sexual development induced by the lesion (28) and the initiation of
normal puberty (24), respectively. The importance of TGF, the
preferred erbB-1 ligand in brain (74), in the process of both
lesion-induced and normal puberty was initially inferred by the
activation of its expression in reactive astrocytes after the lesion
(28) and in the hypothalamus during normal puberty (24). Subsequent
studies employing genetic approaches demonstrated that general
activation of TGF synthesis in transgenic animals (75) or focal
activation in the vicinity of LHRH neurons, via genetically modified
cells (26), resulted in precocious puberty.
The similar patterns of Oct-2 and TGF expression observed in the
hypothalamus after puberty-inducing lesions and during normal puberty
and the preferential accumulation of both gene products in astrocytes
surrounding the lesion site raised the possibility that TGF may be
one of the genes subjected to Oct-2 regulatory control in astroglial
cells. Indeed, analysis of the TGF promoter revealed the presence of
both an octamer-like motif with similarity to the octamer motif present
in Ig promoters (76) and a nonoverlapping (OCTA-)
TAATGARAT motif (77) similar to the sequence that mediates Oct-2
inhibition of HSV immediate early gene promoters (52). Functional
assays showed that an intact octamer-like motif, but not the HSV-like
sequence, is required for Oct-2a and 2c to trans-activate
the TGF promoter. In fact, in the absence of the TAATGARAT sequence,
Oct-2c was more effective in stimulating TGF transcription than in
the intact promoter. These findings are in keeping with the concept
that neither Oct-1 nor Oct-2 is able to activate small nuclear RNA
(Oct-1) or mRNA (Oct-2) promoters containing the
(OCTA-) TAATGARAT site (77). No evidence for
repression of TGF gene transcription mediated by the HSV-like motif
was found in our experiments using glial cells. Repression of the HSV
immediate early gene promoters (52) and the tyrosine hydroxylase
promoter (63) by this mechanism has been described in neuronal cells.
Perhaps expression of promoters in a glial context, such as that used
in the present experiments, prevents manifestation of Oct-2-initiated,
TAATGARAT-mediated transcriptional repression. The finding that both
Oct-2a and Oct-2c are able to trans-activate the TGF gene
in this particular cellular context is, however, consistent with the
previous demonstration that both proteins act as
trans-activators of octamer-containing promoters (31).
An unexpected finding was the apparent inability of Oct-2a and 2c to
bind to the TGF octamer-like motif in the TGF promoter despite
requiring the motif to trans-activate the promoter. The
appearance of a higher mol wt complex after transient expression of
either Oct protein in C6 cells raises the possibility that
Oct-2a/c-mediated trans-activation of the TGF promoter in
a glial cell context requires the recruitment of an additional
protein(s) able to interact with sequences within and/or adjacent to
the octamer motif. Such a protein must be different from high mobility
group protein 2, which increases the DNA-binding activity of Oct-2, but
does not become part of the protein-DNA complex (78). It may, however,
bear some functional similarity to the B cell coactivator Bob1/OBF-1,
which enhances Oct-1- and Oct-2-mediated transcription via an
interaction that strictly requires the presence of an octamer motif in
the promoter (79, 80). Recruitment of cell-specific regulatory proteins
and the presence of promoter-selective activating domains are emerging
as novel mechanisms underlying the cell- and gene-specific regulation
of gene expression (79, 80, 81). Further experiments, beyond the scope of
the present study, are required to demonstrate the existence of such
mechanisms in the Oct-2-mediated control of TGF gene expression.
If Oct-2-mediated transcriptional activation of glial TGF gene
expression plays a physiological role in the process by which
hypothalamic astrocytes facilitate the advent of reproductive
competence, inhibition of such activity in vivo should delay
initiation of the pubertal process. Such was indeed the case, as
disruption of Oct-2 synthesis via third ventricular administration of
an antisense ODN significantly delayed the timing of first ovulation.
The delay, although clear-cut, was less pronounced than that elicited
by the targeted disruption of hypothalamic erbB-1 (24) or erbB-2 (21)
receptors. This is to be expected, as the antisense treatment reduced,
but did not eliminate, Oct-2 protein levels and TGF mRNA content in
hypothalamic astrocytes. Furthermore, compensatory changes, perhaps
involving cell-specific expression of transcriptional coactivators,
would be expected to occur upon reduction of Oct-2 protein synthesis.
For instance, Ig gene transcription is not affected by targeted
disruption of the Oct-2 gene (82), but is severely compromised in mice
lacking the B cell-specific transcriptional coactivator
OBF-1/Oca-B/Bob-1 (83). This factor is recruited to the octamer motif
of Ig promoters via protein-protein interactions with both Oct-1 and
Oct-2 proteins. The existence of similar glial- and/or
neuronal-specific coactivators of Oct-2-mediated transcriptional
activation remains to be demonstrated.
In summary, the present results provide an initial insight into the
molecular mechanisms underlying the contribution of hypothalamic
astroglial cells to the neuroendocrine control of female sexual
development. Regulation of TGF gene expression by Oct-2 gene
isoforms probably represents only one component of the complex
regulatory system controlling glial and neuronal function at the advent
of reproductive maturation. Recent studies have shown that the POU
domain gene Oct-1 is required for the transcriptional activity of the
neuron-specific LHRH gene enhancer (37), and that SCIP/Tst-1, another
POU-domain gene, is a potent repressor of LHRH gene activity (34). It
is thus plausible that POU domain proteins are intrinsic components of
the developmental program that coordinates the interactive functions of
those neuronal and astrocytic subsets involved in the neuroendocrine
control of mammalian puberty.
|
Acknowledgments |
|---|
promoter.
We also thank Dr. Peter Gruss (Max Planck Institute, Gottingen,
Germany) and Dr. Walter Schaffner (Institute for Molecular Biology,
University of Zurich, Zurich, Switzerland) for providing us with the
pSCT-Oct-2c and pSCT-Oct-2a expression vectors, respectively. |
Footnotes |
|---|
