| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Prince Henry’s Institute of Medical Research and the Monash University Department of Obstetrics and Gynecology, Monash Medical Center, Clayton, Victoria 3168, Australia
Address all correspondence and requests for reprints to: Dr. Peter J. Fuller, Prince Henry’s Institute of Medical Research, P.O. Box 5152, Clayton, Victoria 3168, Australia. E-mail: peter.fuller{at}med.monash.edu.au
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ERß gene expression in a
panel of ovarian tumors consisting of GCT and serous and mucinous
cystadenocarcinomas as well as in normal ovary. Expression was
determined by RT-PCR using gene- and isoform-specific primers and
probes combined with Southern blot analysis of the PCR products.
Widespread expression of ER was observed in all tumor types, but at
relatively low levels. ERß is expressed predominantly in GCT, with
lower levels in mucinous tumors and very low levels in serous tumors.
The ERß2 splice variant previously reported in rodents was not
observed. Only very low levels of the exon 5, exon 6, and exon 5/6
deletion variants were detected. The C-terminal truncation variant
ERßcx, however, exhibited widespread expression across
all the tumor types. As ERßcx has been shown to be a
ligand-independent antagonist of ER action, the relative ratios of
ERßcx, ER, and ERß may influence the response of a
tumor to antiestrogen therapy. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ERß) that are
encoded by separate genes (1). These two receptors share both
structural and functional homology. They are members of the
steroid/thyroid/retinoid superfamily of ligand-dependent
transcription factors (2, 3).
The granulosa cells of the developing follicle are the major site of
estrogen biosynthesis in the premenopausal woman. Both estrogen actions
and ERs have been identified in the normal ovary of various animal
models, although some controversy exists with respect to the presence
of ERs in human ovary (4). ERs and/or responses have also been reported
in some ovarian tumors (5). The recent identification of the second ER
(6, 7) may provide an explanation for some of the controversy
surrounding ER expression in the human ovary. The major site of ERß
gene expression in the female rat (8) and human (1) is the ovary, where
its expression clearly predominates over that of ER (1, 8). Within
the ovary of both species ERß is localized to the granulosa cells (1, 6).
Several splice variants of ERß have been identified. A 54-nucleotide insertion has been reported in the rat ERß transcript (9), which encodes a functional receptor (10) containing 18 amino acids inserted in the ligand-binding domain (9, 10). This isoform has also been reported in mice (11) and has been termed ERß2 (10). Low levels of ERß expression are observed in breast tumors (12), and although the above insertion was not observed (11), shortened transcripts with deletion of exon 5, exon 6, or both have been reported (11, 13). Alternate splicing at the 3'-end of the coding region has also been reported, which results in an additional isoform, ERßcx (14, 15).
The majority of malignant ovarian tumors are epithelial in origin, but approximately 10% are classified as ovarian sex cord tumors, of which most are granulosa cell tumors (16). In this study we sought to establish which of the two ER genes is expressed in granulosa cell tumors of the ovary and to compare these with epithelial tumors. In addition, we sought to determine which of the ERß isoforms are expressed in these tumors.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ovarian granulosa cell tumors (GCT; n = 4), mucinous
cystadenocarcinomas (n = 8), and serous cystadenocarcinomas
(n = 4) were obtained from a study of serum inhibin levels in
ovarian tumors (17). The tumors were consecutive tumors for which
adequate tissue was available for RNA extraction. Normal ovarian tissue
was obtained from three patients who had undergone elective
hysterectomy with oophrectomy for menorrhagia associated with benign
uterine lesions at the ages of 39, 46, and 48 yr. The clinical details
of the tumors (Table 1
) have been
presented previously (18). The RNA was extracted using the guanidine
thiocyanate/cesium chloride method as described previously (18).
|
One microgram of total RNA was reverse transcribed for 90 min at
42 C in a total volume of 20 µL using AMV reverse transcriptase
(Roche Molecular Biochemicals, Mannheim, Germany). First
strand synthesis was performed using 30 pmol of the specific antisense
primer (Table 2) alone or with the
ß2-microglobulin antisense primer. The
oligonucleotide primers for the
ß2-microglobulin gene have previously been
described (18). The ER primers were designed from the published
sequences (7, 19, 20) with OLIGO Primer Analysis software version 5.0
(Natural Biosciences, North Plymouth, MN). Primer pair 1 is for
ERß-specific amplification. The second primer pair consists of
universal primers, which will amplify both ER and ERß. Primer pair
3 spans the exon 5-exon 6 splice junction for the detection of ERß2.
The fourth sense primer hybridizes to exon 4, and the antisense primer
hybridizes to exon 8 to detect ERß
5, ERß6, and ERß5/6.
The primers for ERßcx are the same as those
used by Moore et al. (15). Two microliters of each RT
reaction were amplified in a single stage PCR for 30 cycles with 10
pmol gene-specific primers and 2.5 U Taq polymerase
(Roche Molecular Biochemicals), in a total volume of 50
µL. The thermal cycling profile for the receptor consisted of a
denaturing step at 95 C for 5 min and subsequently for 30 s,
annealing at 60 C for 30 s, and extension at 72 C for 45 s,
with a final 72 C incubation for 5 min. The products were visualized on
a 1.8% agarose gel, stained with ethidium bromide, and photographed
under UV transillumination. Controls for the RT-PCR were the reaction
mixtures described above but with reverse transcriptase omitted. The
identity of the amplicons was confirmed by direct sequencing (21).
|
For Southern blot analysis using gene-specific
32P-labeled probes (Table 1
), the PCR products
described above were transferred to Hybond N+
membranes (Amersham Pharmacia Biotech, Aylesbury, UK) as
described previously (18, 22).
PAGE
PCR products amplified with primer pair 4 were electrophoresed through an 8% nondenaturing polyacrylamide gel and transferred to Hybond N+ membranes (Amersham Pharmacia Biotech) using the Mini Trans-Blot Cell Apparatus (Bio-Rad Laboratories, Inc., Regents Park, Australia). The membranes were used for Southern blot analysis as described above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The pattern of relative expression
for ERß is GCT>>mucinous tumors>>serous tumors. There is, however,
significant variation between tumors within each tumor type for the
mucinous tumors and one of the serous tumors. This variability with the
mucinous tumors does not obviously correlate with stage, age of
patient, or serum hormone levels (Table 1). The relatively low levels
in the normal ovary presumably reflect the relatively small number of
granulosa cells in these ovaries, which are from older premenopausal
women.
|
and ERß in
these tumors, universal primers were designed to amplify both receptors
(Fig. 2). The RT-PCR also included
primers for ß2-microglobulin (18) to provide an
internal control. The results of the initial amplification for 30
cycles are shown in the upper panel of Fig. 2. To ensure
that the amplification remained in the linear phase, the samples were
also amplified for 25 cycles before Southern blotting with
gene-specific 32P-labeled internal
oligonucleotide probes (Fig. 2). ER shows widespread low levels of
expression, which are similar to or less than those seen in the
endometrium. The pattern of ERß expression is described in Fig. 1.
Although a direct quantitative comparison between the levels of ER
and ERß expression is difficult, the levels of ERß expression can
be inferred from the ethidium bromide-stained amplicons in tissues that
predominantly express ERß compared with those that express only
ER. ERß expression, at least in the GCT, is severalfold higher
than that of ER.
|
). The ERß2
isoform was not observed in either normal ovary or ERß-expressing
ovarian tumors. The previously published results in rats (9) are also
shown in Fig. 3 for comparison.
|
and 2) was amplified using primers spanning
both exons 5 and 6. Although the wild-type transcript was inevitably
seen, the ERß5 and ERß6 isoforms were extremely difficult to
detect. Where faint bands were seen it was very difficult to determine
which of the two isoforms they represented, given that they differed by
only six nucleotides. A faint band corresponding to the ERß5/6
isoform was observed. Southern blot analysis of the PCR products with a
32P-labeled internal oligonucleotide that would
detect all isoforms is shown (Fig. 4);
this emphasizes that these isoforms are an extremely minor component
relative to the wild-type transcript.
|
). Widespread expression of the
ERßcx isoform was observed in all tumor types,
with relative levels that showed much less variation than that observed
for ERß (Figs. 1 and 2). Also observed were two minor splice
variants, ERß4 and ERß5, described by Moore et al. (15).
The abundance of these forms closely parallels that of ERßcx. The
identity of ERßcx was confirmed by direct
sequencing of the amplicon. The ERßcx variant
appears relatively abundant; however, it is difficult to make a direct
comparison with wild-type ERß; the strategy used in Fig. 2 to compare
ER and ERß is not applicable. As the ERßcx
transcript will be a component of the transcript detected by the RT-PCR
used in Fig. 1 but not in Fig. 2, it would seem likely that at least in
the GCT they may still make a relatively small contribution to the
total pool of ERß transcript. This may not be the case in the
mucinous tumors.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although studies of the ER in normal ovaries have focused on granulosa cells, studies of ovarian tumors have tended to focus on epithelial tumors, which are thought to arise from the simple cuboidal surface epithelium of the ovary. Hillier et al. (28) reported the expression of both ER genes in cultured human surface epithelial cells; whether these receptors are also expressed in vivo in the epithelium has yet to be determined.
The major novel findings of this study are the demonstration of
abundant expression of ERß isoforms in granulosa cell tumors of the
ovary and the comparison of these expression profiles with the two
subtypes of ovarian epithelial tumor. The relatively abundant
expression of ERß in the granulosa cell tumors parallels the patterns
of expression for ER and ERß in normal granulosa cells (1).
Relatively lower expression of ERß was observed in the mucinous
tumors examined, and even lower levels were observed for the serous
tumors. Brandenberger et al. (29) reported expression of
both ER and ERß in serous cystadenocarcinomas of the ovary, with
relatively higher expression of ER than ERß, at least compared to
that in normal ovary. Pujol et al. (30) explored the
relative expression of ERß and ER in ovarian tumor cell lines and
a range of benign and malignant epithelial tumors. As in the study by
Brandenberger et al. (29), they relate increased expression
of ER to the neoplastic process. They examined eight serous
carcinomas, only two mucinous carcinomas, and no GCT. Our results are
consistent with both of these studies with regard to expression in
serous tumors. Our larger sample size suggests, however, considerable
variation in ERß expression by the mucinous tumors, at least half of
which predominantly express ERß.
Numerous splicing variants of the human ER gene have been reported
(31) in both normal tissue and, more particularly, in human breast
tumors. More recently, several splice variants of the ERß gene have
also been described, and their expression was also sought in the
present study. The ERß2 variant, a 54-nucleotide insertion at the
exon 5–6 junction, has been reported by several groups in rat (9, 10, 32, 33) and murine (11) tissue, where the levels of expression of the
ERß2 isoforms appear similar to those of ERß1 (wild-type
transcript). ERß2 is able to bind ligand and mediate expression
through an estrogen response element (10, 32), but it requires
1000-fold greater concentrations of estradiol than ERß1. ERß2 is
able to suppress ERß1- and ER-mediated transcription (32, 33),
such that it would appear to be a negative regulator of estrogen action
(33). Although its presence has been reported in a human ovarian cancer
cell line (32), no evidence for its expression was found in normal
ovary or ovarian tumors. Lu et al. (11) also failed to
detect the ERß2 isoform in human tissue. At this stage, therefore,
the importance of ERß2 appears to be restricted to rodents.
Two groups have recently reported the existence of ERß messenger RNA
variants in breast tumors (11, 13) and normal tissues, including ovary
(11). Vladusic et al. (13) reported coexpression of
wild-type ERß and a variant in which 139 bp corresponding to exon 5
are deleted in an ER-negative breast cancer cell line and in three
of four breast tumors. Lu et al. (11) reported a similar
finding; they examined four ERß-positive human breast tumors and
observed bands corresponding to an exon 5 deletion, an exon 6 deletion,
and an exon 5–6 deletion. All are predicted to encode truncated
proteins, making ligand binding unlikely (11). They reported similar
transcripts in normal breast, ovary, and uterine tissue. At this stage
no evidence has been presented that these variants are translated (11, 13). Our studies clearly identified coexpression of the exon 5, 6, and
5/6 deletion variants of ERß in ovarian tumors, as reported by Lu
et al. (11). We also found them to be of much lesser
abundance compared to the wild-type transcript. The ER exon 5
deletion splicing variant has been reported in normal ovary, uterine
endometrium, and cervix and in tumors arising from these tissues (34).
The significance of these variant transcripts is obscure. If they are
translated into stable receptor molecules, it would be unlikely that
they bind ligand; it is possible that they may act either as
ligand-independent constitutively active receptors, as the ER exon 5
splice variant does (34) or, alternatively, as ligand-independent
dominant negative inhibitors of estrogen action.
Two groups have recently reported a further isoform of human
ERß that involves alternative splicing of the C-terminal region
(14, 15). The predominant variant ERßcx is
truncated compared with that of ERß, but has an extra 26 amino acids
at the C-terminus and a novel 3'-untranslated region as a result of
alternative splicing. This pattern of splicing is also seen for
isoforms of other members of the steroid hormone receptor superfamily
(3). Several other isoforms resulting from alternate splicing of this
region have been described, but both their abundance and tissue
distribution appear relatively limited compared with those of
ERßcx (15). ERßcx is
unable to bind ligand (14). ERßcx
preferentially forms a heterodimer with ER rather than ERß and
elicits a dominant negative effect on ER-mediated
trans-activation (14). Expression of
ERßcx has been reported in normal ovary (14, 15), as observed in this study, and in breast tumors (35). Ogawa
et al. (14) were able to detect
ERßcx protein in several cell lines using
Western blot analysis. Most striking is the widespread expression of
ERßcx in all of the tumors examined at levels
above those observed in normal ovary. The levels of
ERßcx expression parallel those for ERß,
although the magnitude of the variation in levels among the tumors is
much less for ERßcx than for ERß. Given that
the ERß transcripts measured in Fig. 1 include transcripts with all
of the C-terminal variants, differences in the levels of expression
between tumors, particularly for the GCT, must represent a predominant
increase in wild-type ERß expression. The ERß4 and ERß5 isoforms
described by Moore et al. (15) were observed, albeit as a
minor product whose abundance clearly closely parallels that of
ERßcx.
Many of the tumors coexpress ER and ERß, although whether this
coexpression also occurs at a cellular level has not yet been
systematically examined. Pujol et al. (30) used in
situ hybridization to examine ER and ERß expression in a
serous tumor. They reported the expression of both receptors in the
epithelial cells. The same may not be true of the GCT; in the normal
rat ovary, ERß expression has been reported to be restricted to the
granulosa cells, whereas ER has been reported to be expressed at
very low levels (25) or to be absent (36) in these cells, but is
observed elsewhere in the ovary (36). Where coexpression does occur in
tumors, the observed ability of these receptors to heterodimerize
in vitro (14) may be relevant.
The recognition that most ovarian tumors express one or both ERs
suggests firstly that estrogens may have a role in pathogenesis, and
secondly that appropriate agonists or antagonists may have a
therapeutic role. Previous studies of the therapeutic effects of
antiestrogens, particularly tamoxifen, may have been confounded by the
relative heterogeneity of the tumors being examined, particularly with
regard to ER vs. ERß expression as well as expression
of the ERßcx isoform. Some studies suggest
subtle tissue- and promoter-specific differences in the way in which
ER and ERß respond to ligands, particularly antagonists (37, 38, 39).
Consideration of the types and isoforms of ER expressed in an ovarian
tumor in future clinical studies may enable the therapeutic agents to
be tailored to the specific tumor.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Recipient of a Principal Research Fellowship from the National
Health and Medical Research Council of Australia.
Received June 10, 1999.
Revised September 17, 1999.
Accepted November 19, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ß. Endocrinology. 138:863–870. (ER-) and ß (ER-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 -ß messenger RNAs as a
potential marker of ovarian carcinogenesis. Cancer Res. 58:5367–5373. in the rat ovary. Endocrinology. 140:963–971. and ERß at
AP1 sites. Science. 277:1508–1510. and ß. Biochem Biophys Res Commun. 236:140–145.[CrossRef][Medline]
and
estrogen receptor ß to partial estrogen agonists/antagonists. Mol
Pharmacol. 54:105–112.This article has been cited by other articles:
 |
|
 |
|
  U. Ohnemus, M. Uenalan, J. Inzunza, J.-A. Gustafsson, and R. Paus The Hair Follicle as an Estrogen Target and Source Endocr. Rev., October 1, 2006; 27(6): 677 - 706. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Kalfa, A. Ecochard, C. Patte, P. Duvillard, F. Audran, C. Pienkowski, E. Thibaud, R. Brauner, C. Lecointre, D. Plantaz, et al. Activating Mutations of the Stimulatory G Protein in Juvenile Ovarian Granulosa Cell Tumors: A New Prognostic Factor? J. Clin. Endocrinol. Metab., May 1, 2006; 91(5): 1842 - 1847. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A L Ferry, D M Locasto, L B Meszaros, J C Bailey, M D Jonsen, K Brodsky, C J Hoon, A Gutierrez-Hartmann, and S E Diamond Pit-1{beta} reduces transcription and CREB-binding protein recruitment in a DNA context-dependent manner J. Endocrinol., April 1, 2005; 185(1): 173 - 185. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Chu, Y. Nishi, T. Yanase, H. Nawata, and P. J. Fuller Transrepression of Estrogen Receptor {beta} Signaling by Nuclear Factor-{kappa}B in Ovarian Granulosa Cells Mol. Endocrinol., August 1, 2004; 18(8): 1919 - 1928. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. E. Gargett, M. Zaitseva, K. Bucak, S. Chu, P. J. Fuller, and P. A. W. Rogers 17{beta}-Estradiol Up-Regulates Vascular Endothelial Growth Factor Receptor-2 Expression in Human Myometrial Microvascular Endothelial Cells: Role of Estrogen Receptor-{alpha} and -{beta} J. Clin. Endocrinol. Metab., September 1, 2002; 87(9): 4341 - 4349. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Gargett, K. Bucak, M. Zaitseva, S. Chu, N. Taylor, P. J. Fuller, and P. A.W. Rogers Estrogen receptor-{alpha} and -{beta} expression in microvascular endothelial cells and smooth muscle cells of myometrium and leiomyoma Mol. Hum. Reprod., August 1, 2002; 8(8): 770 - 775. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. T. K. Saunders, M. R. Millar, S. Macpherson, D. S. Irvine, N. P. Groome, L. R. Evans, R. M. Sharpe, and G. A. Scobie ER{beta}1 and the ER{beta}2 Splice Variant (ER{beta}cx/{beta}2) Are Expressed in Distinct Cell Populations in the Adult Human Testis J. Clin. Endocrinol. Metab., June 1, 2002; 87(6): 2706 - 2715. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Chu, S. Rushdi, E.T. Zumpe, P. Mamers, D.L. Healy, T. Jobling, H.G. Burger, and P.J. Fuller FSH-regulated gene expression profiles in ovarian tumours and normal ovaries Mol. Hum. Reprod., May 1, 2002; 8(5): 426 - 433. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Fuller, E. T. Zumpe, S. Chu, P. Mamers, and H. G. Burger Inhibin-Activin Receptor Subunit Gene Expression in Ovarian Tumors J. Clin. Endocrinol. Metab., March 1, 2002; 87(3): 1395 - 1401. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. S. Choong, P. J. Fuller, S. Chu, Y. Jeske, F. Bowling, R. Brown, P. Borzi, N. D. Balazs, R. Suppiah, A. M. Cotterill, et al. Sertoli-Leydig Cell Tumor of the Ovary, a Rare Cause of Precocious Puberty in a 12-Month-Old Infant J. Clin. Endocrinol. Metab., January 1, 2002; 87(1): 49 - 56. [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 |
