| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
and a Tumor-Specific Splice Variant1
Neuroendocrine Unit, Departments of Medicine and Neurosurgery (B.S.), Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: J. M. Alexander, Ph.D., Beth Israel Deaconess Medical Center, Harvard Institutes of Medicine, Room 944, 330 Brookline Avenue, Boston, Massachusetts 02215.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). Expression of ER and several of its
messenger ribonucleic acid (RNA) alternate splice variants has been
shown to be restricted to prolactinomas and gonadotroph tumors.
However, little is known about gene expression patterns of the novel
nuclear hormone receptor ERß in the neoplastic pituitary. ERß has
high homology to ER in the DNA- and ligand-binding domains, but
encodes a distinct transcriptional activating function-1 (AF-1) domain.
Using RT-PCR analysis of total RNA from 38 human pituitary adenomas, we
found that ERß messenger RNA was coexpressed with ER and its
splice variants in 60% of prolactinomas, 100% of mixed GH/PRL tumors,
and 29% of gonadotroph tumors. ERß gene expression was not limited
to ER-positive tumor subtypes, however, and was also found in 100%
of null cell tumors, 80% of somatotroph tumors, and 60% of
corticotroph tumors. Because ERß is coexpressed with ER and its
splice variants in prolactinomas and gonadotroph tumors, we
functionally characterized the potential interactions between ERß and
ER. We also examined the potential cooperative effects on
ERß-mediated gene expression of a tumor-specific truncated 
5ER
splice variant that has been shown to be coexpressed in the majority of
ER-positive tumors. This exon 5 splice variant encodes the AF-1
domain as well as regions critical for DNA binding and nuclear
localization, but lacks the ligand-binding and AF-2 domains. Mammalian
expression vectors encoding ER, 5ER, and/or ERß
complementary DNAs were transiently transfected along with an
E2 response element promoter-luciferase (ERELuc) reporter
into human ER/ERß-negative osteosarcoma U2-OS cells. ERß was
less potent than ER in activating E2-stimulated ERELuc
activity (4- vs. 14-fold relative to basal control
levels). However, when 5ER was coexpressed with ERß or ER,
E2-stimulated ERELuc activity was markedly increased to 8-
and 57-fold, respectively, relative to basal control levels when each
full-length isoform was expressed alone. Finally, coexpression of ERß
with ER did not significantly alter the E2-stimulated
ERELuc activity induced by ER alone. Cotreatment with tamoxifen
markedly inhibited all E2-stimulated ERELuc responses to
baseline levels. Together, these data suggest that ERß has a minor
role in mediating E2 responses in ER-positive tumors,
but may be the main mediator of E2-stimulated gene
expression when expressed alone in somatotroph, corticotroph, and null
cell tumors. This low, but significant, level of ERß
trans-activation potential may be enhanced by
coexpression of 5ER in neoplastic pituitary. Therefore,
E2-mediated gene expression in normal and neoplastic
pituitary appears to be highly dependent on the expression of ER and
ERß isoforms, which have varying transcriptional activities. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
), a
ligand-activated transcriptional factor of the steroid receptor
superfamily. RT-PCR and immunocytochemical studies using human
pituitary adenoma specimens have demonstrated that ER expression is
restricted to lactotroph and gonadotroph cells and is only detected in
other tumor subtypes when PRL biosynthesis is detected and a mixed
tumor phenotype is observed (7, 8, 9). However, other experimental
pituitary animal model systems have offered suggestive evidence that
E2 may have a direct, although modest, influence
on hormone biosynthesis and secretion in somatotroph and corticotroph
cells (10, 11, 12, 13, 14, 15). These data have been difficult to interpret in the
context of several RT-PCR and immunocytochemical studies that have
failed to demonstrate that these pituitary cell types express
detectable levels of ER (7, 8, 9).
The recent cloning of the novel ERß subtype affords a new opportunity
to examine ER isoform expression in normal and neoplastic
adenohypophysial cells (16, 17, 18). In particular, experiments using
primary animal pituitary cell cultures have documented GH release in
response to E2 from ER-negative somatotroph
cells (10, 12, 14). This suggests a potential role for ERß in
mediating E2-stimulated hormone release and gene
transcription. ERß has been shown to be expressed in numerous
tissues, including developing spermatids and ovarian granulosa cells
(16, 18). It has also been detected in several human neoplasms, such as
prostate and breast tumors (16, 19). It shares a 57% sequence homology
to human ER, and its genomic intron/exon structure is strikingly
similar to that of ER (16).
In the pituitary, wild-type ER messenger ribonucleic acid (mRNA) and
protein have been demonstrated in normal and adenomatous lactotrophs
and gonadotrophs (8, 9). In addition, tumor-specific splice variants
have been characterized in prolactinomas and gonadotroph tumors and are
not present in other pituitary tumor subtypes. However, the expression
of ERß in normal and neoplastic pituitary cells has never been
studied. Altered ERß gene expression in
E2-sensitive pituitary adenomas may modulate
normal ER function, affecting both neoplastic cell proliferation and
hormone secretion. Therefore, we investigated the expression of ERß
in 38 human pituitary adenomas of different phenotypes and normal
pituitary tissues using RT-PCR with ERß-specific primers. Functional
studies on E2 response element
promoter-luciferase (ERELuc) reporter constructs were carried out using
cotransfected expression vectors that synthesize ER, ERß, and the
tumor-specific alternatively spliced 5ER. Together, these studies
examine pituitary tumor subtype expression of ERß and functionally
evaluate the potential consequences of ERß coexpression with ER
and its tumor-specific splice variant 5ER.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Pituitary adenomas were obtained from 38 patients who underwent
transsphenoidal surgery and whose ER isoform expression has been
previously described (7). Patients with macroprolactinomas (n =
11) ranged in age from 17–60 yr (median, 27 yr), and all had elevated
PRL levels (210 to >8000 µg/L) and immunostaining consistent with
the diagnosis. Patients with clinically nonfunctioning adenomas (n
= 10) ranged in age from 37–79 yr (median, 54 yr). All patients had
normal serum glycoprotein hormone (GPH) and free -subunit
levels with serum PRL of less than 100 µg/L. Immunocytochemical
staining with specific antibodies against GPH subunits (common ,
FSHß, LHß, and TSHß) was positive in 7 of 10 tumors
(i.e. gonadotroph tumors), whereas antibodies against GH,
PRL, and ACTH were negative. Three other tumors were negative for all
specific pituitary hormone antibodies (i.e. null cell
tumors). Patients (n = 8) with GH-secreting adenomas ranged in age
from 27–78 yr (median, 38 yr). All patients exhibited clinical and
biochemical evidence of acromegaly, with elevated serum levels of GH
(9–90 µg/L) and insulin-like growth factor I (699–1233 µg/L). All
somatotroph adenomas had strong immunostaining for GH, and 3 tumors
(no. 4, 6, and 7) also showed rare to scattered PRL staining (GH/PRL
tumors). Patients with corticotroph tumors (n = 10), 5
microadenomas (diameter, <1 cm) and 5 macroadenomas, ranged in age
from 20–59 yr (median, 48 yr). All patients with microadenomas had
dexamethasone suppression testing and petrosal catheterization results
consistent with Cushing’s disease. All corticotroph adenomas exhibited
positive immunostaining for ACTH, whereas only one tumor (no. 3) also
had scattered staining for PRL (ACTH/PRL tumor). To test whether
surgical specimens were contaminated with normal pituitary cells, all
PRL-, GH-, and ACTH-secreting as well as null cell tumors studied using
LHß primers and were negative for LHß mRNA expression in the RT-PCR
assay (data not shown). All GPH-producing tumors were tested for Pit-1
expression, a member of the POU domain of transcription factors that is
specifically found in GH-, PRL-, and TSH cells, by RT-PCR and were
negative. Control normal pituitary tissues used in these studies were
obtained within 3–6 h postmortem and snap-frozen in liquid nitrogen
(National Disease Research Interchange, Philadelphia, PA).
RNA extraction and RT-PCR analysis of mRNA
Pituitary adenomas were obtained in phosphate-buffered saline after transsphenoidal surgery and frozen in liquid nitrogen. Total RNA was extracted using a single step acid guanidinium isothiocyanate phenol/chloroform technique (20), followed by enzymatic digestion of genomic DNA with 1 U RQ1 deoxyribonuclease/µg total nucleic acid at 37 C for 1 h (Promega, Madison, WI) and were quantitated by UV spectrophotometry. To prepare complementary DNA (cDNA), 1 µg total RNA was reversed transcribed in 50 mmol/L Tris-HCl (pH 8.3), 5 mmol/L KCl, 5 mmol/L MgCl2, 5 mmol/L dithiothreitol, 0.25 mmol/L spermidine, 200 µmol/L deoxy (d)-NTPs, and 12 U AMV reverse transcriptase (Promega), with random hexamers (1 µg) as first strand cDNA primers. RT reactions were carried out at 25 C for 10 min, followed by a 10-min elongation step at 42 C, and heat inactivation at 99 C for 5 min.
Oligonucleotide primers were designed using Oligo software (National Biosciences, Minneapolis, MN) and compared to GenBank sequence libraries to assure specific amplification of human ERß cDNA. Primer synthesis and analysis was based on human ERß NCBI accession X99101. Primer sequences were (5'-3'): ERß1 (exon 1, nucleotide 129), ggt cca tcg cca gtt atc ac; ERß2 (exon 5/6, nucleotide 943), ttc ccc tca tcc ctg tcc ag; ERß3 (exon 4, nucleotide 788), tca gct ggg cca aga aga ttc; and ERß4 (exon 8, nucleotide 1293), acc aca ttt ttg cac ttc atg. Results for ERß RT-PCR were confirmed using both sets of primers; however, only data using ERß3 and -4 are shown. Reactions without AMV reverse transcriptase were also carried out with each RNA sample to exclude genomic DNA contamination as a source of amplified signal. All tumors were negative for receptor signal in reactions without RT (data not shown). To control for potential nonspecific RNA degradation in pituitary tumor RNA preparations, samples were tested for the presence of GAPDH mRNA by PCR, and all were positive. Oligonucleotide primers for control PCR of human glyceraldehyde phosphate dehydrogenase (GAPDH), Pit-1, and LHß were (5'-3'): GAPDH-U, gag cca gat cgc tga gac; GAPDH-L, ttc tcc atg gtg gtg aag; Pit-1U, cat tta ctt cgg ctg ata; Pit-1L, agg ttg atg gct ggt ttc; LHß-U, gct cca ggg gct gct gct; and LHß-L, cga cag ctg aga gcc aca ggg.
All PCR amplifications used 40 ng first strand cDNA from a single RT
reaction. For ERß, all pituitary tumor samples were amplified
simultaneously using a common PCR reaction mixture to ensure that any
differences in receptor amplification between samples were not due to
variability in PCR reaction conditions. PCR was carried out in 50
mmol/L KCl, 10 mmol/L Tris-HCl (pH 9.0), 3.5 mmol/L MgCl2,
0.1% Triton X-100, 40 µmol/L dNTPs, and 0.125 U TaqI
polymerase (Promega) in a final volume of 25 µL. To control for
extraneous contaminating genomic DNA or cDNA in experiment reagents, a
tube containing the PCR reaction mixture with no template was included
in each ERß amplification and was negative for all experiments. PCR
products were visualized by incorporation of
[-32P]dCTP (100 nCi/reaction) in the PCR reactions.
PCR primers (12.5 pmol) were used for each reaction, and
amplifications were carried out in an MJ thermocycler (MJ Research,
Watertown, MA). All reactions were amplified for 35 cycles (1 min at 94
C, 1 min at the optimal annealing temperature, and 1 min and 15 s
at 72 C), except for GAPDH, which was amplified for 30 cycles. All
amplified products were fractionated by 6% nondenaturing
tris-borate-EDTA/polyacrylamide gel electrophoresis (Protogel,
National Diagnostics, Atlanta, GA), and exposed to Kodak XO-Mat film
for 2–48 h.
Plasmid constructs
Full-length normal human ER and 5ER were cloned from
human pituitary tumor cDNA into the eukaryotic pBK-cytomegalovirus
(CMV) expression vector (Stratagene, La Jolla, CA). Briefly, pooled
pituitary first strand cDNA obtained from human lactotroph and
gonadotroph tumors by RT served as template for PCR amplification using
a high fidelity pfu DNA polymerase (Stratagene, La Jolla,
CA) as previously described (7). Primer pairs SpeI-ER
(5'-gga cta gtc cat gac cat gac cct CCA-3') and ER4L (5'-ttc gcc cag
ttg atc atg tg-3') were used to amplify the N-terminal portion of ER
(nucleotides 231-1298, GenBank accession no. X03635), whereas ER4U
(5'-gcc ccc cat act cta ttc-3') and ER-ClaI (5'-gga tcg
atg cag cag gga tta tct ga-3') were used to amplify the C-terminal part
of ER (nucleotides 1201–2063). PCR products of the appropriate
sizes for normal and variant ER fragments were purified from 1%
agarose gel using the GlasPac/GS purification kit (National Scientific,
San Rafael, CA). The N-terminal portion of the ER fragment was
digested with SpeI and HindIII, whereas the
C-terminal portions of ER and 5ER fragments were digested with
HindIII and ClaI before cloning into the pBKCMV
plasmid vector. Expression plasmids containing the complete
protein-coding region of ER or 5ER were constructed by
ligating the common N-terminal portion of ER to the isoform-specific
C-terminal portion via an overlapping unique HindIII site.
Positive clones were identified and confirmed by dideoxy sequencing
using primers covering the entire translated region of ER.
Full-length human ERß cDNA was provided by Jan-Ake Gustafsson
(Karolinska Hospital, Stockholm, Sweden), and was shuttled to pBKCMV
mammalian expression vector. Reporter plasmid EREtk81Luc contains two
copies of consensus ERE palindromic sequence (aggtcacagtgacct) at
positions -126 to -138 and -142 to -156 upstream of 81 bp of the
herpes simplex virus minimal thymidine kinase (tk) promoter in the
pA3Luc plasmid (courtesy of R. Pestell, Albert Einstein College of
Medicine, Bronx, NY).
Transient transfection and luciferase assay
U2-OS cells were plated in six-well plates at a density of
2 x 105 cells/well and allowed to adhere overnight.
One hour before transfection, cells were washed and incubated in phenol
red-free reduced serum OptiMEM (Gibco BRL, Grand Island, NY). A
lipid transfection mixture was prepared using a 1:400 mixture of
dioleoyl--phosphatidylethanolamine to
demethyldioctadecylammonium bromide dissolved in 100% ethanol (both
lipids were purchased from Sigma Chemical Co., St. Louis, MO). The
DNA/lipid mix (1:5 ratio) containing the appropriate pBKCMV/ER,
5ER, or ERß expression plasmid along with the luciferase
reporter plasmid was prepared in phenol red-free reduced serum OptiMEM
(1 mL/well) for 30 min at room temperature before addition to
triplicate wells containing U2-OS cells. Whenever applicable, empty
pBKCMV plasmid was used for experiments with pBKCMV-ER, 5ER,
or ERß to ensure that stoichiometrically equal amounts of CMV
promoter, plasmid DNA, and lipid mix were applied to each well. Rous
sarcoma virus/ß-galactosidase was used to monitor the variability of
transfection efficiency. After 5-h incubation at 37 C, the transfection
medium was replaced with serum-free, phenol red-free DMEM and incubated
for additional 24 h in the presence or absence of 10 nmol/L
E2 and/or tamoxifen. Transfected cells were lysed
with 300 µL lysis buffer containing 1% Triton X-100, 10% glycerol,
2 mmol/L ethylenediamine tetraacetate, 2 mmol/L dithiothreitol, and 25
mmol/L Tris-phosphate (pH 7.8), and the cellular debris was removed by
centrifugation. One hundred microliters of cell lysate were assayed for
luciferase activity by measuring light emission with a Luminometer
(EG&G Berthold, Gaithersburg, MD) in the presence of luciferin and
ATP.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Normal ER and 5ER full-length cDNA were cloned by RT-PCR
from pooled human gonadotroph and lactotroph tumor mRNA. These
pituitary tumor subtypes have been previously shown to express the
5ER variant in a tumor-specific manner along with full-length
ER (7). Figure 1
shows the exon/intron
structure and functional domains of normal ER and ERß as well as
the tumor-specific alternate splice variant, 5ER. All ER isoform
cDNAs were cloned into pBKCMV mammalian expression vector (Stratagene,
La Jolla, CA). The positions for two sets of RT-PCR primers for
amplification of ERß are shown at the bottom of Fig. 1.
|
RT-PCR analysis of ERß mRNAs in 38 human pituitary tumors is
summarized in Table 1. Full-length ER
mRNA was detected in 60% of lactotroph and 29% of gonadotroph
adenomas. ERß mRNA was also found in 89% of somatotroph adenomas,
60% of corticotroph tumors, as well as all three null cell tumors that
had no detectable hormone synthetic or secretory phenotype. Figure 2 shows representative RT-PCR data for
ERß mRNA expression in lactotroph and gonadotroph tumors as well as
normal pituitary specimens. Normal pituitary specimen 3 exhibited faint
ERß signal using primer set ERß1 and -2 (data not shown). Results
for ERß RT-PCR were confirmed using both sets of ERß-specific
primers. A summary of previously published data (7) for RT-PCR results
with 5ER and normal ER mRNA in this series of characterized
tumors is included in Fig. 2. Control RT-PCR of GAPDH mRNA was positive
in all tumors and normal tissue.
|
|
shows RT-PCR data for ERß
expression in corticotroph, somatotroph, and null cell tumors. These
pituitary tumor subtypes failed to express any ER isoforms, except
when PRL immunostaining was positive, as it was for corticotroph tumor
specimen 3, which had mixed ACTH/PRL immunoreactivity, and somatotroph
tumor specimens 4, 6, and 7, which had mixed GH/PRL immunostaining.
ER isoforms were never detected in any of the pure corticotroph or
somatotroph tumors that exhibited positive immunostaining for only GH
or ACTH, respectively (7). In contrast to ER, 60% of corticotroph
tumors and 88% of somatotroph adenomas exhibited a positive RT-PCR
signal for ERß. In addition, all three null cell tumors were positive
for ERß despite having no evidence of detectable levels of steady
state ER mRNA. Control RT-PCR of GAPDH mRNA was positive in all
tumors and is shown at the bottom of Fig. 3.
|
and ERß on E2-stimulated
ERE-directed gene expression
U2-OS cells have been reported to be unresponsive to
E2 and fail to express normal ER or ERß as
measured by RT-PCR. Figure 4 shows
transient transfection studies with a heterologous EREtk81Luc reporter
construct, and the effects of ER and ERß isoforms on modulating
E2-mediated gene expression. Figure 4
demonstrates that U2-OS cells in which no ER or ERß is expressed
are unresponsive to E2, with no significant
increases in ERE-driven luciferase activity above control wells.
However, when transiently transfected with pBKCMV/ER, U2-OS cells
exhibited significant (P < 0.001) up-regulation of
EREtk81Luc activity with E2 treatment. ERß
transient transfection also caused EREtk81luc activity to increase in
response to 10 nmol/L E2, but gene activation was
significantly lower than that observed when equivalent amounts of ER
were transfected (4- vs. 14-fold relative to basal control
levels; P < 0.001). There was no cooperative effect
observed when ER and ERß were coexpressed in U2-OS cells.
Tamoxifen treatment had no significant effect on basal EREtk81luc gene
expression in the presence of any ER isoform studied. Coadministration
of E2 and tamoxifen significantly
(P < 0.001) reduced
E2-stimulated EREtk81Luc reporter activity to
basal control levels.
|
5ER variant isoform modulates the
trans-activation by both ERß and ER on ERE-directed gene
expression
To examine the effects of normal ERß and 5ER on
E2-responsive gene transcription, U2-OS cells
were transiently transfected with ERß and/or 5ER along with
EREtk81Luc, followed by treatment with E2 and/or
tamoxifen. The results of these cotransfection studies are shown in
Fig. 5. 5ER had no constitutive or
E2-stimulated effect on EREtk81Luc transcription
when expressed alone in U2-OS cells. However, coexpression of 5ER
with normal ERß resulted in significant up-regulation of
E2-stimulated (P < 0.001)
EREtk81Luc activity 6-fold over basal levels in the control OS cells or
approximately 2-fold compared to E2-stimulated
cells transfected with ERß alone. Basal levels of ERß
trans-activation of EREtk81luc were not increased. Tamoxifen
significantly suppressed (P < 0.05)
E2-stimulated levels of EREtk81luc activity in
cells transfected with ERß alone or together with 5ER.
|
shows the effects of ER and
5ER coexpression on EREtk81luc gene activation in U2-OS cells.
5ER had no constitutive or E2-stimulated
effect on EREtk81Luc transcription when expressed alone in U2-OS cells.
However, coexpression of 5ER with normal ER resulted in
significant up-regulation of both basal (P < 0.01) and
E2-stimulated (P < 0.001)
EREtk81Luc activity (6- and 57-fold, respectively) over basal levels in
the control OS cells, or approximately 4-fold compared to the
E2-stimulated level in cells transfected with
ER alone. When all three ER isoform vectors were cotransfected along
with EREtk81luc, E2-stimulated reporter activity
was markedly up-regulated. Tamoxifen significantly suppressed
(P < 0.05) both basal and
E2-stimulated levels of EREtk81luc activity in
cells transfected with ER alone or cotransfected with ER and
5ER.
|
-positive pituitary adenomas also synthesized
ERß and the tumor-specific splice variant 5ER. To examine the
effects of coexpression of these three ER isoforms on
E2-responsive gene transcription, U2-OS cells
were transiently transfected with ER, ERß, and 5ER along
with EREtk81Luc, followed by treatment with E2
and/or tamoxifen. The results of these cotransfection studies are shown
in Fig. 7. This combination of ER
isoforms further induced E2-stimulated
trans-activation of the EREtk81Luc reporter, which was
significantly greater (P < 0.05 to P
< 0.001) compared to E2 responses by all other
combinations of ER isoforms, except for coexpression of ER and
5ER, in which E2-stimulated EREtk81Luc
expression was not significantly different from cotransfection with all
three isoforms. Figure 7B shows the inhibitory effects of tamoxifen on
both basal and E2-stimulated EREtk81Luc activity
by ER/ERß/5ER coexpression. Tamoxifen significantly
inhibited (P < 0.001) both basal and
E2-stimulated reporter activity down to the
baseline levels seen in parental ER-negative U2-OS cells.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
and its
splice variants in 60% of prolactinomas, 100% of mixed GH/PRL tumors,
and 29% of gonadotroph tumors. Unlike ER, which is expressed only
in prolactinomas, gonadotroph tumors, and mixed phenotype tumors with
detectable PRL immunostaining, ERß mRNA was also found in 100% of
null cell tumors, 80% of somatotroph tumors, and 60% of corticotroph
tumors. Based on our series, ERß is expressed in the majority of
human pituitary tumors (63%) across all tumor phenotypes. ERß was
found to be coexpressed with ER and its tumor-specific splice
variants in 47% of estrogen-responsive adenomas, i.e.
prolactinomas and gonadotroph tumors.
Although multiple isoforms of ER variants have been shown to be
coexpressed with wild-type ER in human pituitary tumors of lactotroph
and gonadotroph origins (7), the functional consequences of this
observed coexpression on ERE-regulated gene activity have not been
tested in vitro. We therefore designed transient
transfection studies that would functionally characterize the potential
interactions between ERß and ER and determine whether coexpression
of both isoforms altered E2 responsiveness of ERE
promoter/reporter constructs in an ER-negative U2-OS cell line. We also
examined the potential cooperative effects on ERß-mediated gene
expression of a tumor-specific truncated 5ER splice variant that
has been shown to be coexpressed in the majority of ER-positive
tumors. Therefore, to reliably manipulate the relative levels of ER
and 5ER variant in a mammalian cell expression system and test
this hypothesis, we required an ER-negative human cell line that had
lost the ability to express ERß, ER, and ER variants during
cellular transformation and/or clonal selection. U2-OS cells were
chosen because 1) they failed to express any ER isoform; and 2) they
were derived from a normally E2-responsive human
osteoblast cell type. Although studies using such an immortalized cell
line come with the caveat that they cannot recreate the pituitary
cellular phenotype to reflect true physiological conditions in each of
the human pituitary tumor subtypes, U2-OS cells were chosen as the best
available model system to assure that our functional studies were
performed in an ER-negative cellular context as well as in a human
cell line.
These functional data on ERß trans-activation of
ERE-regulated luciferase reporter constructs are consistent with at
least one other report examining ERß activity in vitro
(21). ERß was less potent than ER in activating
E2-stimulated ERELuc activity (4- vs.
14-fold relative to controls). However, when 5ER was coexpressed
with ERß or ER, E2-stimulated ERELuc
activity was markedly increased to 8- and 57-fold, respectively,
relative to basal control levels when each full-length isoform was
expressed alone. Conversely, coexpression of ERß with ER did not
significantly alter E2-stimulated ERELuc activity
induced by ER alone. Cotreatment with the E2
antagonist tamoxifen markedly inhibited all
E2-stimulated ERELuc responses to baseline
levels.
Both full-length ER and ERß encode several functional domains
important for hormone binding, DNA binding, and maximal transcription
activation (22, 23). Truncated receptors lacking specific domains due
to alternative mRNA splicing may create variant ER isoforms with
altered function. Altered ER isoform coexpression with either normal
ER or ERß may affect cellular phenotype by several mechanisms: 1)
by acting as a competitor of normal ER and ERß function by binding
with high affinity to E2, 2) by compromising or
disrupting stable ER and/or ERß homo/heterodimers after
E2 binding, or 3) by altering the
trans-activation of full-length ER isoforms at
E2-responsive gene promoters and modulating
full-length receptor activity in either a dominant positive or a
negative manner. Therefore, differential exon alternative splicing and
expression of ER can give rise to variant receptor isoforms that may
potentiate the diverse actions of E2 through a
single receptor gene. Despite the observed complexity of ER splice
variant expression in human pituitary adenomas, our RT-PCR analysis of
ERß gene expression in these tumors failed to consistently offer any
evidence of alternate splicing of any portion of the ERß mRNA
transcript.
Our RT-PCR data in pituitary tumors and functional studies in human
cell lines indicate that both ER and ERß isoforms may be mediating
the well documented mitogenic and hormonal regulatory effects of
E2 in prolactinomas and gonadotroph tumors. In
addition, the tumor-specific splice variant 5ER appears to act as
a dominant positive ER isoform by markedly enhancing the
E2 responsiveness of ERE promoter when
coexpressed with either full-length ER or ERß. Recent studies
investigating 5ER effects on the activating protein-1
class of E2-responsive genes demonstrate that
this positive dominant effect is specific for classical ERE promoters
(24). These data suggest an important role of this alternatively
spliced ER variant in promoting E2-regulated
tumor proliferation and hormonal biosynthesis and secretion.
These descriptive data documenting ERß expression in human
corticotroph, somatotroph, and null cell tumors may account for the
observed E2 responsiveness described in these
ER-negative pituitary tumor subtypes. For example, in
vitro studies examining the E2
responsiveness of cultured primary animal pituitary tissue indicates a
direct effect of E2 on ER-negative somatotroph
cells, with documented increases in both GH mRNA steady state levels
and immunoreactive GH in the culture medium (10, 12, 14, 15).
Therefore, our observed amplification of ERß mRNA from ER-negative
human somatotroph tumors suggests that ERß may be critical for the
observed modest responses of these cells to E2
administration.
For other ER-negative tumor types, evidence for
E2 effects on hormone biosynthesis and/or cell
proliferation is less well documented. E2 has
been shown to have modest effects on ACTH biosynthesis and secretion
in vivo (11), but the role of E2 in
the growth of corticotroph or null cell tumors is unknown. Furthermore,
potential E2 responses mediated via ERß may be
masked by the relatively weak trans-activation potential of
ERß in a specific subset of adenohypophysial cells, which may
represent a small percentage of the primary normal or neoplastic
pituitary cell culture. Further studies will be required to clarify the
potential effect of E2 via ERß on the growth
and hormone biosynthesis of ER-negative pituitary tumors.
Together, these data suggest that ERß has a minor role in mediating
E2 responses of ERE-regulated genes in
ER-positive tumors, but may be the main mediator of such
E2-stimulated gene expression when expressed
alone in somatotroph, corticotroph, and null cell tumors. This low, but
significant, level of ERß trans-activation potential may
be enhanced by coexpression of 5ER in neoplastic pituitary.
Therefore, E2-mediated gene expression in
neoplastic adenohypophysial cells appears to be highly dependent on the
expression of ER and ERß isoforms, which have varying
transcriptional activities. Coexpression and interaction of various ER
isoforms in pituitary tumors may be of pathophysiological relevance for
the regulation of pituitary neoplastic cell proliferation and hormone
biosynthesis in response to E2.
|
Footnotes |
|---|
Received April 20, 1998.
Accepted June 11, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
and ß mRNAs in normal ovary, ovarian serous
cystadenocarcinoma and ovarian cancer cell lines: down-regulation of
ER-ß in neoplastic tissues. J Clin Endocrinol Metab. 83:1025–1028. and ER-ß
at AP1 sites. Science. 277:1508–1510.This article has been cited by other articles:
 |
|
 |
|
  M. Cosma, J. Bailey, J. M. Miles, C. Y. Bowers, and J. D. Veldhuis Pituitary and/or Peripheral Estrogen-Receptor {alpha} Regulates Follicle-Stimulating Hormone Secretion, Whereas Central Estrogenic Pathways Direct Growth Hormone and Prolactin Secretion in Postmenopausal Women J. Clin. Endocrinol. Metab., March 1, 2008; 93(3): 951 - 958. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. D. Curtis, V. S. Likhite, I. X. McLeod, J. R. Yates, and A. M. Nardulli Interaction of the Tumor Metastasis Suppressor Nonmetastatic Protein 23 Homologue H1 and Estrogen Receptor {alpha} Alters Estrogen-Responsive Gene Expression Cancer Res., November 1, 2007; 67(21): 10600 - 10607. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Veldhuis, D. M. Keenan, and C. Y. Bowers Peripheral estrogen receptor-{alpha} selectively modulates the waveform of GH secretory bursts in healthy women Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2007; 293(4): R1514 - R1521. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. P. Gillam, M. E. Molitch, G. Lombardi, and A. Colao Advances in the Treatment of Prolactinomas Endocr. Rev., August 1, 2006; 27(5): 485 - 534. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. Lee and J L. Jameson Gene therapy of pituitary diseases J. Endocrinol., June 1, 2005; 185(3): 353 - 362. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. P. Giraldi, M. Moro, and F. Cavagnini Gender-Related Differences in the Presentation and Course of Cushing's Disease J. Clin. Endocrinol. Metab., April 1, 2003; 88(4): 1554 - 1558. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Schreihofer, D. F. Rowe, E. F. Rissman, E. M. Scordalakes, J.-a. Gustafsson, and M. A. Shupnik Estrogen Receptor-{alpha} (ER{alpha}), But Not ER{beta}, Modulates Estrogen Stimulation of the ER{alpha}-Truncated Variant, TERP-1 Endocrinology, November 1, 2002; 143(11): 4196 - 4202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K.-i. Matsuda, I. Ochiai, M. Nishi, and M. Kawata Colocalization and Ligand-Dependent Discrete Distribution of the Estrogen Receptor (ER){alpha} and ER{beta} Mol. Endocrinol., October 1, 2002; 16(10): 2215 - 2230. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. J. Lee, W. R. Duan, M. Jakacka, B. D. Gehm, and J. L. Jameson Dominant Negative ER Induces Apoptosis in GH4 Pituitary Lactotrope Cells and Inhibits Tumor Growth in Nude Mice Endocrinology, September 1, 2001; 142(9): 3756 - 3763. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Schreihofer, E. M. Resnick, V. Y. Lin, and M. A. Shupnik Ligand-Independent Activation of Pituitary ER: Dependence on PKA-Stimulated Pathways Endocrinology, August 1, 2001; 142(8): 3361 - 3368. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. F. Palter, A. B. Tavares, A. Hourvitz, J. D. Veldhuis, and E. Y. Adashi Are Estrogens of Import to Primate/Human Ovarian Folliculogenesis? Endocr. Rev., June 1, 2001; 22(3): 389 - 424. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Nealen, K. V. Vijayan, E. Bolton, and P. F. Bray Human Platelets Contain a Glycosylated Estrogen Receptor {beta} Circ. Res., March 2, 2001; 88(4): 438 - 442. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Schreihofer, M. H. Stoler, and M. A. Shupnik Differential Expression and Regulation of Estrogen Receptors (ERs) in Rat Pituitary and Cell Lines: Estrogen Decreases ER{alpha} Protein and Estrogen Responsiveness Endocrinology, June 1, 2000; 141(6): 2174 - 2184. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. K. Hodges, L. Tung, X.-D. Yan, J. D. Graham, K. B. Horwitz, and L. D. Horwitz Estrogen Receptors {alpha} and {beta} : Prevalence of Estrogen Receptor {beta} mRNA in Human Vascular Smooth Muscle and Transcriptional Effects Circulation, April 18, 2000; 101(15): 1792 - 1798. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||
5ER
