| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
3) in Breast Cancer and the Consequences of Its Reexpression: Interference with Estrogen-Stimulated Properties of Malignant Transformation
Department of Cell Biology (I.E., B.S., L.O.) Rochelle Belfer Chemotherapy Foundation Laboratory (I.E., R.M.L., L.O) Department of Medicine Department of Obstetrics and Gynecology (B.S.) Mount Sinai School of Medicine New York, New York 10029
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
) isoform, missing exon
3 (ER
3), to the full-length
ER, in normal breast epithelium to that in
primary breast cancers and breast cancer cell lines revealed a 30-fold
reduction of this ratio in cancer cells (P < 0.0001).
To test what functions may have been affected by the loss of
ER3, stable clones of MCF-7 cells expressing ectopic
ER3 protein, at the range of
physiological ER, were generated. In
vector-transfected controls the ER3-mRNA
and protein were less than 10% while in the
ER3-expressing clones,
ER3-mRNA and protein ranged from 36–76%
of the total ER. Estrogen (E2)
stimulated the expression of pS2-mRNA in pMV7 vector control cells, but
the stimulation was reduced by up to 93% in
ER3-expressing clones. In addition,
several properties associated with the transformed phenotype were also
strongly affected when ER3 protein was
reexpressed. Compared with vector-transfected control cells, the
saturation density of the ER3-expressing
clones was reduced by 50–68%, while their exponential growth rate was
only slightly (14.5 ± 5%) lower. The in vivo
invasiveness of the ER3-expressing cells
was significantly reduced (P = 0.007) by up to 79%.
E2 stimulated anchorage-independent growth of
the pMV7 vector control cells, but reduced it to below baseline levels
in ER3 clones. The reduction of the pS2
response to E2 in the
ER3-expressing clones and the
E2 block of anchorage-independent growth to
below baseline were more pronounced than expected from the dominant
negative function of ER3. These
observations suggest that E2 may activate an
additional ER3-dependent inhibitory
pathway. The drastic reduction of ER3 to
ER ratio in breast cancer, and the fact that
when present in breast cancer cells this isoform leads to a
suppression, rather than enhancement, of the transformed phenotype by
E2 suggests that the regulation of
ER-mRNA splicing may need to be altered for
the breast carcinogenesis to proceed. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
) is the major mediator of estrogen action, the
precise mechanism by which ER contributes to altered
estrogen response in cancer remains unclear.
ER is one of many transcriptional regulatory proteins of the steroid
receptor family that act principally as ligand-activated DNA-binding
dimers (3). ER has distinct functional domains, including two
transcriptional activating regions (the N-amino-terminal AF1 and the
C-terminal, ligand-dependent, AF2), vs. internal zinc
fingers, DNA-binding domain, dimerization regions, and several nuclear
localization sequences (4) (Fig. 1A
).
Like other ligand-activated transcriptional regulators (5), ER is not a
single protein, but rather a set of proteins coded by two genes giving
rise to ER and ERß (6) as well as isoforms
generated by alternative splicing (exon skipping) of the single
ERpre-mRNA. Since alternatively spliced
ER-mRNAs were first noted in breast tumors and tumor
cell lines, before their normal counterparts were thoroughly examined,
it was proposed that overexpression of aberrant ER
isoforms is characteristic of breast cancer (7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24). One
ER isoform (ER3), missing exon 3,
which encodes the second zinc finger of the DNA-binding domain, was
shown to be an in-frame deletion that in an in vitro
translation reaction yielded a protein of 61.2 kDa (20). This protein
was unable to form specific complexes with estrogen response elements
(EREs) or transactivate an ERE-reporter plasmid in transient
transfection assays (20). At an ER3 to
ER ratio of 1:1, ER binding to an ERE and
transactivation of an ERE-chloramphenicol acetyltransferase (CAT)
reporter were inhibited by approximately 30%, suggesting that
ER3 can function as a dominant negative receptor
(20). The importance of dominant negative receptors in controlling
cellular responses to agonists and antagonists has been underscored by
several recent studies of the steroid receptor family (25, 26, 27, 28). Since
ER3 is a naturally occurring form of
ER, if expressed at high relative levels to full-length
ER, it may have a profound effect on several
estrogen-dependent functions. For example, ER3
expression in normal breast tissue may provide a means of regulating
the magnitude of estrogen responses, and a relative loss of
ER3 expression in breast tumor tissue may lead to
unchecked estrogen stimulation. Alternatively, a rise in
ER3 expression during breast carcinogenesis may
facilitate the disabling of the normal differentiation-inducing
function of estrogen. Finally, the isoform may represent such a minor
component that it would not influence estrogen-mediated pathways in
either normal or malignant breast tissue.
|
3 and ER
expression in primary breast cancers, and breast cancer cell lines to
that in luminal and basal epithelium and fibroblasts purified from
reduction mammoplasty specimens. This comparison, and the subsequent
analysis of breast cancer cells expressing ectopic
ER3, yielded strong support for the hypothesis that
ER3 causes a profound change in cell response to
E2 and that the relative loss of ER3 may
be important in carcinogenesis. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and
ER3 Expression in Cancer and Normal
Breast Cells and ER3 expression using a
semiquantitative RT-PCR assay capable of distinguishing between mRNA
encoding the full-length and the ER3 forms of the
receptor. The two forms of ER were detected by Southern
blot analysis of ER-cDNA amplified with primers within
exons 2 and 4, using internal probes hybridizing either with exon 4, to
detect both forms of the receptor, or within exon 3 to detect only the
full-length ER. The median ratio of
ER3 to ER expression in breast tumors
was 0.11 (range of 0.03–0.47) (Fig. 1B
1 and 1D, group 1). Analysis of
nine cancer cell lines, (seven of which were ER
positive) confirmed the low relative levels of ER3 in
pure populations of cancer cells (median ratio of 0.10, range
0.06–0.30, Fig. 1B2 and 1D, group 2). A similar test was performed on
epithelial cells isolated from 10 reduction mammoplasties, which
included seven pools of epithelial cells and three purified populations
of luminal cells. [An immortalized, myoepithelial mammary cell line
(Hs 578Bst) was also examined]. The median ratio of
ER3 to ER in all 10 preparations of
epithelial cells was 3.4 (range 0.36 to 9.80) or approximately 30-fold
greater than the ratio found in breast cancer (P <
0.0001), (Fig. 1, C and D, group 3). The three 90% pure luminal cell
populations (purity determined by the K8/K5 ratio) had
ER3 to ER ratios of 0.53, 0.36, and
0.42 (Fig. 1D, group 3, encircled crosses). Individually,
these were 3- to 4-fold greater than the median ratio of
ER3 to ER in breast cancers and 7- to
10-fold greater than the median ratio in breast cancers with low ER
level (see below). This comparison is important because, like luminal
cells, a large proportion of breast cancers expresses cytokeratin K8
(29, 30), suggesting the possibility that breast cancer may arise in
luminal cells. Until formally excluded, however, the possibility that
breast cancer originates in stem cells found in the basal layers of
breast acini must also be considered. We found that in pools of
epithelial cells, which due to culture conditions had a preponderance
of stem/basal cells (Fig. 1D, group 3) (31), the ER3
to ER ratio was even greater (median 4.00, range
0.55–9.80). We also tested six preparations of breast fibroblasts,
which under culture conditions became predominantly (
80%)
myofibroblasts, as determined by their content of smooth muscle
-actin. Regardless of their origin (reduction mammoplasty, n =
4; or breast cancer, n = 2) (Fig. 1D, group 4), they were also
found to have high ER3 to ER ratios
(median 2.4, range 1.5–4.5).
In an attempt to determine whether the loss of ER3
was associated with the carcinogenic event per se, or with
cancer progression, we analyzed the breast cancer patients data by
stratifying them according to their menopausal status (pre- and
postmenopausal) or tumor size (<1.5 cm or >1.5 cm) or presence or
absence of lymph node involvement. With one exception, these analyses
did not identify a subpopulation of patients with a significant
difference in the ER3 to ER ratios
among the tumors. The exception was a group of tumors (6/33) with
ER levels lower than 5 fmol/mg protein, deemed
ER negative by clinical standards and considered more
aggressive, in which the median ratio (0.05) was significantly
different (P < 0.001) from the median ratio (0.11) of
all tumors.
Cumulatively, these results, showing that the ratio of
ER3 to full-length ER is substantially
reduced in all breast cancer cell lines and in breast cancers, even
when tumors are smaller than 1.5 cm and have not spread to the lymph
nodes, suggest that a loss of the ER3 isoform may be
associated with an early event in carcinogenesis.
Transfection and Isolation of MCF-7 Cells Expressing
ER3; Characterization of the Native
and Transgenic Protein
The findings described above suggested that restoring
ER3 in cancer cells to normal relative levels may
result in attenuation of their transformed phenotype. To test this,
ER3-cDNA was subcloned into a pMV7 vector (32). The
ER3/pMV7 construct was first tested by transiently
transfecting COS cells, negative for ER, an experiment
that showed (Fig. 2) that transfected
cells express and properly localize ER3 to the
nucleus and that the ER3 protein reacts with a well
characterized rat anti-ER antibody (H226) (33).
|
3-coding
constructs, or the pMV7 vector alone for negative control, and analyzed
both for ER3-mRNA, by RT-PCR (Fig. 3A), and protein expression, by
immunoprecipitation and Western blotting (Fig. 3B). Several pMV7vector
control clones were analyzed individually for ER and
ER3-mRNA and found to have similar overall levels of
ER and invariably very low levels (<10% of total ER)
of the ER3-mRNA. Thus, for all further experiments, a
pool of approximately 50 pMV7 vector-transfected MCF-7 clones was used
as a control. Attempts to generate specific ER3
antibodies that recognize the exon 2/exon 4 splice junction were
unsuccessful. Therefore, the ER3 protein was
identified on the basis of its reactivity with two antibodies
recognizing different N-terminal epitopes of ER, its
faster mobility than ER on SDS-PAGE, and the correlation
of its expression with that of ER3-mRNA.
|
, lane 5), control pMV7 cells expressed
predominantly full-length ER-mRNA and a small amount of
ER3, similar to that observed in the parental MCF-7
cell line (Fig. 1B2, lane 8). In contrast, all four
ER3 clones had a predominance of
ER3-mRNA (Fig. 3A, lanes 1–4; ratios of
ER3 to ER ranged from 1.3–2.1 or 57-
67% of total ER was ER3), indicating
that the transgene mRNA was efficiently expressed in these cells.
Extracts of the individual clones shown in Fig. 3B were subjected to
immunoprecipitation with a polyclonal rabbit anti-ER
antibody followed by Western blotting with the H226 antibody,
recognizing the amino terminus of ER. In addition to the
65-kDa band, representing the full-length ER protein,
all ER3 clones also contained a prominent 61-kDa
band, which corresponded to the predicted molecular mass of the
ER3 protein. Expression of the ER3
protein in these clones ranged from 36–76% of total ER
(a relative ratio of ER3 to ER of
0.6–3.3), comparable to that observed in the normal mammary
epithelium. Although the relative ratios of the RNAs and the proteins
did not match perfectly, the two clones (clones 3 and 4, Fig. 3, A and
B) with the highest proportion of ER3-mRNA (64% and
67%) also had the highest proportion of ER3-protein
(76% and 60%). In parental MCF-7 extracts, a faint band (5% of
the total ER), comigrating with the
ER3 form, could be detected only when excess protein
was loaded onto the gel (Fig. 4, lane 2).
This and the correspondence between the low intensity of the
ER3-mRNA band and the 61-kDa protein band (Fig. 1B, lane 8; Fig. 3A, lane 5, and 3B, lane 5; Fig. 4, lane 2, respectively)
suggest that both pMV7-carrying and the parental MCF-7 cells produce
small amounts of the ER3 mRNA and protein. The
identity of the 61-kDa band as ER3, and not as an
underphosphorylated form of full-length ER, was
confirmed in a dephosphorylation experiment. Total immunoprecipitated
ER from pMV7 or ER3 cells was mixed
with protease inhibitors; one aliquot of each was dephosphorylated by
incubation with calf intestinal phosphatase (CIP), and the other was
incubated under identical conditions but without CIP. Analysis of
products by Western blotting showed that without CIP, ER
from both pMV7 pool vector control and ER3 cells
produced a comigrating doublet of bands, the upper corresponding to
full-length ER and the lower to ER3
protein (compare lanes 2 and 3 of Fig. 4). CIP treatment shifted the
migration coefficient of both bands in the vector control cells as well
as the ER3 clone to new positions, once more as a
comigrating doublet (Fig. 4, lanes 1 and 4). No lower bands or smear
was detected, indicating that proteolysis during CIP incubation was
effectively blocked by the protease-inhibitors cocktail. Comigration of
the lower molecular mass protein from pMV7 pool vector control cells
with that of the ER3 protein in the transfected
clone, both before and after dephosphorylation, strongly suggests the
presence of endogenously produced ER3 protein.
Similar results were obtained using the ER-positive Ishikawa cells, an
endometrial carcinoma cell line (data not shown).
|
3 Suppresses
Estrogen-Stimulated Gene Expression3 has been shown to interfere with
ER binding to its specific DNA response element in
vitro, as well as with E2-induced transcription of an
ERE-CAT reporter in transient transfection of COS cells in
vivo (20). These studies suggested that ER3
functions as a dominant negative receptor to inhibit ER
regulation of gene expression through its cognate DNA response element.
To determine whether the ER3 expressed in MCF-7 cells
can interfere with estrogen induction of an endogenous gene, the
expression of pS2, a gene with several imperfect EREs in its promoter,
was assessed. pMV7 pool vector control cells and three individual
ER3 clones were incubated either with the pure
antiestrogen, ICI 164,384 (1 x 10-7 M),
to establish the baseline of pS2 expression, or with E2
(1 x 10-8 M). Total RNA was prepared and
analyzed by Northern blot to determine pS2 expression. (Ribo-somal
RNA, Fig. 5, bottom, which
served as loading control, shows that similar amounts of RNA were used
in each lane). As shown in the blot in Fig. 5, in which the number of
scanned units is inserted under each lane, the ICI 164,384
treatment reduced pS2 expression to below detection in all but one
(ER3–4) clone. The maximal induction of pS2 by
E2 (5.8 scanned units) was observed in the pMV7 pool vector
control cells. In clone ER3–3, with the highest
ER3 protein level, E2 induced only 0.4
scanning units, a 93% reduction compared with pMV7 vector control. In
the two additional clones the effect was intermediate (21 and 60%
reduction). A longer exposure of the Northern blot allowed the
visualization of the pS2-band in all the ICI 164,384-treated cells and
the calculation of fold stimulation of pS2 by E2 in each
individual clone, compared with pMV7 pool vector control. This
comparison showed that the presence of ER3 blocked
the E2-induced stimulation by 83%, 91%, and 95% for
clone ER3–4, 2, and 3, respectively (results not
shown). Therefore, at least in the two clones (ER3–2
and 3) in which base line pS2 expression was similar to that of pMV7
pool vector control, ER3 prevented E2
stimulation of pS2 more effectively than predicted from its dominant
negative function. (The third clone, ER3–4, fell
into the same category if comparison of fold induction of pS2 by
E2 was considered). Overall these findings confirm that the
ER3 interferes with the ER-regulated
gene expression in vivo and suggests that more than a single
mechanism may be responsible for this effect.
|
3 Expression Alters
Saturation Density of MCF-7 Cells3-transfected cells formed smaller colonies
than the parental cells or the vector-transfected clones, and fewer
number of cells were obtained from confluent cultures. To further
evaluate this difference, ER3 clones and pMV7 pool
vector controls were plated at 2 x 105 cells per
60-mm dish in medium with 10% FBS; the medium was changed every third
day. Cells were detached and counted over a period of 8 days. Results
in Fig. 6 (mean of three determinations)
show that there were only small differences in the rate of cell
division between the clones and the control cells during the
exponential phase of their growth. The control cells underwent 0.78
divisions per day, while the clones underwent from 0.61 to 0.7
divisions per day (a 14 ± 5% reduction). It was striking,
however, that no increase in cell number of the
ER3-expressing clones was observed once 2.1 to
3.4 x 106 cells accumulated per dish. This plateau
cell density level was only 32–50% of that found in the pMV7 pool
vector control cells (6.9 x 106) (Fig. 6)
(P < 0.0001), perhaps indicating that cells expressing
ER3 are more sensitive to signals of contact
inhibition. Similarly reduced saturation density of
ER3-expressing clones was obtained when these cells
were cultured in 10-8 M E2
supplement (results not shown). This was the first suggestion that the
expression of ER3 shifts the transformed phenotype of
breast cancer cells toward behavior expected of normal cells.
|
3 to
ER Protein Ratio3-expressing cells into estrogen
depleted, charcoal-stripped FBS (csFBS) changed the
ER3 to ER protein ratio. Western blot
detection of ERs immunoprecipitated from an equal amount of protein of
clone ER3–2 cells, grown either in the presence of
csFBS or E2-supplemented FBS, showed that there was more
overall receptor protein in cells grown in medium with csFBS (Fig. 7) and that the gain was predominantly in
the full-length receptor, thus increasing the ER to
ER3 protein ratio. (Similar results were obtained
with clone ER3–1; data not shown.) To shift the
ER3 to ER ratio in favor of the
transgenic protein, all further experiments were carried out on cells
grown in medium with FBS and estradiol (although for daily cell
maintenance the ER3-expressing clones were kept in
medium supplemented with csFBS).
|
3 Attenuates the Transformed
Phenotype of MCF-7 Cells3 expression on the anchorage-independent growth of
MCF-7 cells (Fig. 8). As shown by others,
estrogen stimulated the ability of parental MCF-7 cells (and of the
pMV7 pool vector control cells) to form colonies in soft agar. In
contrast, hormone treatment drastically reduced the ability of
ER3 expressing clones to form colonies in agar, even
below the baseline level (Fig. 8). Accordingly, these data suggest
that, in vivo, ER3 may reverse the
tumorigenic phenotype of breast cancer cells through an as yet
undetermined mechanism.
|
3 expression on the ability
of MCF-7 cells to invade host tissue, which is linked to protease
production known to be under the control of E2 in these
cells (34, 35, 36, 37, 38), we inoculated chick embryo chorioallantoic membrane
(CAMs), in vivo, with pMV7 pool vector control cells or
ER3-expressing clones 1, 2, 3, and 4 grown in the
presence of E2 and metabolically labeled with
[125I]UdR for 24 h. Invasion was measured 24 h
after inoculation by a previously described method (39). We determined
that, compared with the pMV7 pool vector control cells or the parental
MCF-7 cells (results not shown), the ability of
ER3-expressing clones to invade the CAM was reduced
by 52–79% (Fig. 9).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
isoform
(ER3) is expressed in normal breast epithelial cells
with a median ratio of ER3 to ER of
3.40, with a subset of purified luminal epithelial cells having a
median ratio of ER3 to ER of 0.40. In
contrast, the median ER3 to ER ratio
in breast tumors and tumor cell lines is only 0.10, indicating a
substantial underrepresentation of ER3 in cancer
cells (P < 0.0001).
It is curious that the same hormone, estrogen, exerts the tightly
controlled effects on growth and differentiation of normal breast cells
during puberty and on their cyclical proliferation in an adult
nonpregnant female, while also acting as a potent mitogen in breast
cancer during its uncontrolled growth and invasion (1, 2). This
dichotomy suggests that, during oncogenic transformation, mammary
epithelial cells may undergo signaling pathway changes leading to
aberrant or inappropriate estrogenic responses. The evidence presented
in the current study is the first demonstration that a selective loss
of ER3 may contribute to the phenotypic changes of
cancer. The observation showing that small tumors, or tumors that have
not spread to the lymph nodes, have ratios of ER3 to
ER as low as the more advanced tumors, suggests that the
loss of ER3 may be an early event in carcinogenesis.
(However, the finding of significantly lower ratios in tumors with
ER < 5 fmol/mg, considered more aggressive, hints that
a further drop of ER3 may be associated with disease
progression.) It is also worth noting that, in spite of the fact that
this latter group of tumors and the normal breast cells have similarly
low levels of the receptor mRNA, they have contrasting ratios of
ER3 to ER, indicating that the high
relative level of ER3 in normal cells is not the
consequence of their overall low receptor level.
The high relative expression of ER3 in normal breast
epithelium and fibroblasts may provide them with a mechanism to
regulate and limit the magnitude of responses to estrogen. It can be
argued then, that to attain maximum estrogen stimulation of growth and
invasive potential during carcinogenesis, breast cells need to be
released from the effects of ER3. Accordingly we
demonstrated that a selective loss of this receptor occurs in breast
tumors and breast cancer cell lines and that the reintroduction of
physiologically relevant levels of ER3 into breast
cancer cells attenuates the mitogenic action of estrogen and reverses
several features that distinguish transformed from normal cells.
Most studies of ER in normal human mammary tissue have
used relatively insensitive immunohistochemical or biochemical
techniques. Consequently, only a subset of luminal epithelial cells,
and no other cells in normal breast tissue, were considered receptor
positive (40). Our study, using RT-PCR, demonstrated ER
expression both in luminal and basal/myoepithelial cells of the normal
breast epithelium (Fig. 1C), as well as in stromal fibroblasts.
Moreover, mammary fibroblasts, demonstrated to be estrogen responsive
(41, 42), were confirmed to express ER protein by more
sensitive immunofluorescence techniques using a strepavidin
amplification of anti-ER antibodies (our unpublished
results). Thus, several different cell types in the normal adult breast
may respond directly to E2. Also, as the high
ER3 ratio (median 2.4) is preserved in breast cancer
fibroblasts, it is likely that their presence in cancer tissue may
contribute to some degree to the difference in ER3 to
ER ratios found in tumors.
The current study has identified the ER3 protein in
cell lines expressing the ER3 transgene as well as in
parental MCF-7 cells and Ishikawa cells; in all cases the
ER3 to ER protein ratios were similar
to the ER3 to ER RNA ratios (Figs. 3 and 4). The very low abundance of ER in normal mammary
cells precludes such a direct analysis of ER protein in
these cells. However, the finding, that in cell lines
ER-RNA ratios reflect those of the corresponding
proteins, makes the likelihood of such correspondence in normal cells
highly plausible.
The relevance of our findings is further underscored by the
demonstration of an autoregulatory loop in the clones with reexpressed
ER3. In these cells exposure to estrogen can shift
the complement of estrogen receptors from mostly ER to
predominantly ER3 (Fig. 7). This is achieved by a
more pronounced down-modulation of ER than of
ER3 and, as shown previously (43, 44), may occur via
several mechanisms, including mRNA and protein stability. If a similar
mechanism of auto-regulation exists in endogenous tissue, it is likely
that during periods of peak estrogen availability, a rise in the
ER3 to ER protein ratio may protect
breast tissue from overstimulation. Thus, oncogenic transformation of
breast cancer cells, resulting in a selective reduction in
ER3 expression, would lead to a disruption of this
response, promote unchecked estrogen action, and establish permissive
conditions for further carcinogenic events.
The reestablishment of a less tumorigenic phenotype in the
ER3-transfected MCF-7 cells deserves further comment
because certain of the properties, such as reduced plateau density and
reduced invasion, may be the result of dominant negative inhibition by
ER3, while others, such as anchorage-independent
growth, may be mediated via additional pathways. Since, as noted, in
the ER3-transfected clones, the relative level of
this isoform is highest in the presence of E2, it is
interesting that E2 treatment of these clones causes a
slight reduction of growth and, more importantly, a much lower
saturation density, as is characteristic of a normal phenotype. These
effects are specific to the ER3 isoform, since a
similar transfection of full-length ER into either MCF-7
or T47D cells (which are also ER-positive) did not
reduce their proliferative response to hormone (45). Although not yet
examined, a testable hypothesis is that a dominant negative receptor
interferes with E2 stimulation of genes critical for growth
regulation, such as cyclin D1, myc, and the fos/jun family of
transcription factors (46, 47). These gene products, in turn, may
reduce growth factor receptor expression, resulting in a lower
saturation plateau.
The reduced invasiveness of the ER3-expressing cells
may also be mediated via a dominant negative effect. It is known that
E2 is necessary for MCF-7 tumor growth and metastasis in
nude mice. E2 also stimulates the expression of several
proteolytic enzymes (such as plasminogen activators, collagenase IV,
cathepsin D), shown to be involved in cancer invasion (34, 35, 36, 37, 38). A
testable hypothesis is that the presence of ER3 will
effectively interfere with stimulation of these proteases by
E2 to result in reduced invasiveness.
In contrast to the above effects, E2 not only failed to
stimulate anchorage-independent growth in
ER3-expressing cells, but inhibited it to below
baseline levels. This observation cannot be explained purely on the
basis of a dominant negative effect, since even in the two
ER3-expressing clones, in which the isoform protein
was only 36–40% of the total ER, E2
reduced the growth in agar to below baseline level. Moreover, in a
clone with only 40% of ER3,
E2-stimulated pS2 expression was reduced by more than 60%,
and in a clone with 75% of ER3 it was almost
completely blocked (Fig. 5). Comparing these findings with the
published results, showing that a cotransfection of
ER3 and ER proteins, at ratios
comparable to those present in our clones, produced only a 30%
inhibition of estrogen-dependent transactivation (20), suggests the
existence of an additional pathway of ER3 action.
Since the total ER level in the clones is either equal
to or less than that in the parental MCF-7 cells, the observed effect
could not be due to the general overexpression of ER
protein, shown by some to lead to E2 inhibition of growth
(48, 49). Although we have not yet investigated the mechanism of the
suppressive signal transduction pathway of ER3, it is
likely that this receptor isoform, in addition to its dominant negative
action, participates in the nonclassic regulation of gene expression
via protein-protein interaction with other transcription factors, which
have been shown recently to be 1) independent of ER
binding to DNA (50), and importantly, 2) independent of the
ER-DNA-binding domain (51).
Thus we have demonstrated a novel function for a non-DNA-binding ER
isoform in breast biology. Relative high expression of this isoform in
normal mammary tissue may provide a mechanism for attenuating
estrogenic effects, and its reduction in breast cancer may lead to
excessive, unregulated mitogenic action of this hormone. Our results
indicate that, as with tumor suppressor WT1 (52), carcinogenic events
in breast can lead to alteration of splice choice pathways, but unlike
suggested for other ER isoforms (8, 9, 10, 11, 12, 14, 15, 17, 18, 19, 21), rather than being elevated in cancer, the relative ratio of this
isoform is diminished. Further studies of the mechanisms through which
ER3 exerts its effect will clarify its role in
controlling E2 responsiveness in mammary cells. Identifying
ways to redirect the pathway toward enhanced expression of
ER3, or finding alternative means of increasing its
relative ratio, may provide a novel avenue for future breast cancer
therapy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
ER RT-PCR
Total RNA was extracted using RNAzol B reagent (Biotecx
Laboratories, Houston, TX), and 1 µg was reverse-transcribed using
Superscript reverse transcriptase (GIBCO BRL, Gaithersburg, MD) and a
primer specific to exon 4 (5'-GGAGACATGAGAGCTGCCAAC-3') of
ER. This exon 4 primer and a primer specific to exon 2
(5'-CCGCAAATGCTACGAAGTGG-3') were used to amplify
ER-cDNA in a 25-cycle reaction of 1 min each at 95 C, 60
C, and 72 C. PCR products were fractionated on a 2% agarose gel,
Southern blotted onto Hybond nylon membrane (Amersham, Arlington
Heights, IL), and probed using either a 32P end-labeled
internal exon 4 probe (5'-GAATGTTGAAACACACAAGCGCC-3'), detecting
full-length ER and ER3, or an exon
3-specific probe (5'-CCGCAAATGCTACGAAGTGG-3'), detecting full-length
ER only. Quantification was performed using the
PhosphorImager ImageQuant program (Molecular Dyanamics, Sunnyvale,
CA).
Preparation of ER3
(ER3/pMV7) Expression Vector
A partial ER-cDNA fragment, containing exons 1,
2, and 4, but missing exon 3 (21), (a gift of Dr. R. Miksicek, State
University of New York, Stony Brook, NY) was used to replace exon 1–4
in a similarly digested HEGO vector (a gift of Dr. P. Chambon,
Strasbourg, France). The resulting ER3 coding
sequence was purified and ligated into a retroviral expression vector
pMV7 (32) under the MuLV promoter. This vector also contains the
neomycin resistance gene. The "empty" pMV7 plasmid served as a
vector control. Both vectors were used to transform DH5 bacteria and
the DNAs purified using Wizard Maxi-Prep kit (Promega, Madison, WI). To
prepare ER3-coding retrovirus for infection,
ER3/pMV7 DNA was transfected into an amphotropic
packaging cell line 
-CRIP and selected with G418, and the virus was
collected, pooled, and used for infection.
Maintenance of Control and
ER3 Clonal Cells
MCF-7 cells were maintained in RPMI-1640 medium supplemented
with insulin (5 µg/ml), penicillin (50 U/ml), streptomycin (50
µg/ml), and 10% FBS (JRH, Lenexa, KS) (final concentration of
E2 from serum, 3 x 10-12
M). All transfected cells were maintained in selection
medium with 500 µg/ml G418. For growth of ER3
clones, 10% csFBS was used, unless otherwise noted.
Separation of Epithelial and Stromal Cells
Normal reduction mammoplasty specimens or breast cancer samples
were obtained from the Pathology Department, Mount Sinai Medical
Center. Epithelial organoids were separated from stroma by mincing
normal breast tissue and incubating it overnight in
hyaluronidase/collagenase mixture as described previously (53). The
organoids were collected, by filtering the digest through a 400-mesh
sieve, and trypsinized into single-cell suspensions. The filtered
single cells were plated and enriched for fibroblasts by differential
trypsinization. Tumor tissue was minced and incubated in collagenase
for 2 h at 37 C. Tumor digest was plated without filtration, and
the cultures were enriched for fibroblasts by differential
trypsinization. Fibroblasts passaged on plastic differentiated into
myofibroblasts so that after two to three passages 80% stained
positively with antibody to smooth muscle -actin.
Separation of Basal and Luminal Epithelium
Basal epithelial cells were positively selected from the
filtered digest using a monoclonal antibody to the common acute
lymphoblastic leukemia (CALLA) antigen (DAKO, Carpinteria, CA) and
Dynabeads (Dynal, Norway) coated with goat anti-mouse IgG (a 10:1 bead
to cell ratio) essentially as described elsewhere (31). These basal
epithelia-enriched cells were cultured in mammary epithelial cell
growth medium (MEGM; Clonetics, San Diego CA) with 5 µg/ml
transferrin and 10 µM isoproterenol. The CALLA-negative
fraction, containing the luminal cells, was densely seeded onto
collagen I-coated dishes in MEGM. After a week in culture, RNAs were
extracted and cell purity determined by Northern blot analysis of
cytokeratin expression (K8-luminal and K5-basal). Cell preparations
with K8 to K5 ratios of 10:1 or 1:10, were defined as luminal or basal
cells, respectively. Some epithelial cell preparations, not purified
further, were used and designated "unselected." RNAs were extracted
from these cell types and used in RT-PCR assay for
ER3 and ER analysis.
Generation of MCF-7 Cells Expressing
ER3 (Transfection/Infection)
Three micrograms of ER3/pMV7 or pMV7 DNA were
transfected with Lipofectin into MCF-7 cell, as per manufacturer’s
recommendation. For retroviral infection (54) 2 ml of growth medium
containing the virus and 8 µg/ml of polybrene were added to
semiconfluent MCF-7 cells; the cells were rocked for 2 h at 37 C
and the inoculum was removed, after which cells were incubated in
medium with serum for 48 h and transferred to medium with G418
(selection medium). Infected and transfected cells were maintained in
G418-containing medium for 1–2 months before clone isolation.
Immunoprecipitation and Western Blotting
Total cell protein was prepared by four freeze/thaw cycles in a
high- salt lysis buffer (0.4 M NaCl, 10% glycerol, 1
mM dithiothreitol, 100 mM Tris, 10
mM EDTA, 50 µg/ml leupeptin, 50 µg/ml aprotinin, 10
µg/ml pepstatin). ER and ER3 were
immunoprecipitated with a rabbit anti-ER antibody
(Zymed, San Francisco, CA) and protein G agarose (Boerhinger Manheim,
Indianapolis, IN) from 400 µg total protein diluted with lysis buffer
without NaCl for a final NaCl concentration of 0.2 M.
Immunoprecipitated material was resuspended in 50 µl of loading
buffer and electrophoresed on an 11.5% SDS/PAGE gel for 8–10 h at 200
V. Protein was transferred onto nitrocellulose membrane (Amersham),
blocked overnight with 5% nonfat milk, washed in Tris-buffered saline
Tween-20/1% nonfat milk, and Western blotted with H226 (0.7 mg/ml) rat
anti-ER primary antibody (1:100 dilution) overnight at 4
C, and incubated with an horseradish peroxidase-conjugated goat
anti-rat secondary antibody (1:10,000 dilution) (Sigma, St. Louis, MO)
for 1 h at room temperature. Enhanced chemiluminescence (ECL Kit,
Amersham) detected bands were quantitated by densitometry.
Phosphatase Treatment of Protein Extract from pMV7 and
ER3 Clone
Total ER was immunoprecipitated from 2 mg of
protein from pMV7pool and ER3 clonal cells, using
rabbit polyclonal anti-ER antibody (Zymed).
Immunoprecipitated material was split into two equal aliquots,
resuspended in 25 µl 1x phosphatase buffer (Boerhinger Mannheim,
10x phosphatase buffer: 0.5 M Tris-HCl, pH 8.5, 1
mM EDTA), containing a 2x protease cocktail (100 µg/ml
leupeptin, 200 µg/ml bacitracin, 100 µg/ml aprotinin, 20 µg/ml
pepstatin). One aliquot each of pMV7pool and ER3–3
were treated with 3 U of CIP (Boehringer Mannheim) and, along with the
mock-treated aliquots, were incubated for 30 min at 30 C. The reaction
was terminated by the addition of 25 µl of 2x loading buffer and
heating to 95 C for 3 min. Western blot analysis was performed as
described above.
Expression of pS2
pMV7pool and ER3- 2, 3, and 4 clonal
cells (1 x 106) were grown for 3 days in 100-mm
tissue culture dishes in the presence of FBS and subsequently treated
with either ICI 164,384 (1 x 10-7 M) or
E2 (1 x 10-8 M) for 2 days.
pS2 expression was assessed by Northern blotting 20 µg of total RNA,
hybridized with random primed pS2 cDNA probe. Ethidium bromide-stained
ribosomal RNA served as loading control. pS2-mRNA level was determined
by densitometric scanning.
Saturation Density
pMV7pool and ER3 clonal cells (0.2 x
106) were plated in 60-mm tissue culture dishes (three
dishes per cell type per time point) in the presence of FBS. Medium was
changed every 3 days, and the cells were detached with trypsin and
counted on day 1, 2, 4, 6, and 8. The ER3-expressing
clones were deemed to be in saturation plateau when 2 days of growth
did not produce a further increase in cell number. The five groups of
cells (four clones and pMV7 pool vector control) were analyzed by ANOVA
statistics.
Growth in Soft Agar
A two-layer low melt agarose (Seaplaque; FMC
Corporation, Rockland, ME) system was used to assess
anchorage-independent growth of pMV7pool and ER3
clonal cells. A 1% lower layer and an 0.4% upper layer of agarose,
prepared in DMEM medium supplemented with insulin (5 µg/ml),
penicillin (50 U/ml), streptomycin (50 µg/ml), and 10% FBS
(±E2 1 x 10-8 M) inoculated
into 60-mm gridded plates. Cells (2 x 103 cells/ml),
distributed in the upper layer, were allowed to grow for 2 weeks and
colony formation in the two conditions was scored. The effect of
E2 was determined by comparison to cloning efficiency in
FBS.
Chorioallantoic Membrane Invasion
Invasion was assayed as previously described
(39). ER3 clones or pMV7pool cells were grown in the
presence of selection medium supplemented with 10% FBS and estradiol
(1 x 10-8 M) for 48 h. Cells were
trypsinized, counted, allowed to attach overnight in the same medium
(4 x 106 cells per 100-mm dish), and labeled with 0.2
µCi/ml of [125]UdR for 24 h (specific activity of
0.1–0.2 cpm/cell). An artificial air chamber above the CAM of a
10-day-old embryo was created, and the CAM was allowed to reseal for
22 h and the labeled cells (3 x 105 per CAM)
were inoculated onto the CAM. After a 24-h incubation, CAMs were washed
with PBS, excised, incubated for 20 min in trypsin-EDTA (0.05%
trypsin, 1 mM EDTA) to remove surface-attached tumor cells,
and rinsed with PBS. The radioactivity remaining in CAMs after the
trypsin-EDTA incubation and the PBS wash, expressed as percent of total
radioactivity (associated with CAMs before trypsinization and present
in washes), represent the proportion of cells that invaded. The ANOVA
test was used for the statistical analysis.
|
ACKNOWLEDGMENTS |
|---|
3 clone, S. Lehrer for tumor samples, P. Fedi for
breast cancer cell lines, and S. Waxman for helpful discussions and
support. |
FOOTNOTES |
|---|
This work was supported by a predoctoral fellowship from The Department of Defense, DAMD17–94-J-4140 (to I.E); NIH Grants HD-27557 and ES-90228 (to B.S); NIH Grant CA-54273 (to R.M.L.); NIH Grant CA-40758 (to L.O.); and funds from the Rochelle Belfer Chemotherapy Foundation and Samuel Waxman Cancer Research Foundation.
Received for publication May 21, 1997. Revision received September 12, 1997. Accepted for publication September 15, 1997.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Wong and C. S. Weickert Transcriptional Interaction of an Estrogen Receptor Splice Variant and ErbB4 Suggests Convergence in Gene Susceptibility Pathways in Schizophrenia J. Biol. Chem., July 10, 2009; 284(28): 18824 - 18832. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. P. Venables, R. Klinck, A. Bramard, L. Inkel, G. Dufresne-Martin, C. Koh, J. Gervais-Bird, E. Lapointe, U. Froehlich, M. Durand, et al. Identification of Alternative Splicing Markers for Breast Cancer Cancer Res., November 15, 2008; 68(22): 9525 - 9531. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Maynadier, J.-M. Ramirez, A.-M. Cathiard, N. Platet, D. Gras, M. Gleizes, M. S. Sheikh, P. Nirde, and M. Garcia Unliganded estrogen receptor {alpha} inhibits breast cancer cell growth through interaction with a cyclin-dependent kinase inhibitor (p21WAF1) FASEB J, March 1, 2008; 22(3): 671 - 681. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. A. Ishunina, D. F. Swaab, and D. F. Fischer Estrogen Receptor-{alpha} Splice Variants in the Medial Mamillary Nucleus of Alzheimer's Disease Patients: Identification of a Novel MB1 Isoform J. Clin. Endocrinol. Metab., June 1, 2005; 90(6): 3757 - 3765. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. H. Herynk and S. A. W. Fuqua Estrogen Receptor Mutations in Human Disease Endocr. Rev., December 1, 2004; 25(6): 869 - 898. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Garcia Pedrero, P. Zuazua, C. Martinez-Campa, P. S. Lazo, and S. Ramos The Naturally Occurring Variant of Estrogen Receptor (ER) ER{Delta}E7 Suppresses Estrogen-Dependent Transcriptional Activation by Both Wild-Type ER{alpha} and ER{beta} Endocrinology, July 1, 2003; 144(7): 2967 - 2976. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Tiffoche, C. Vaillant, D. Schausi, and M.-L. Thieulant Novel Intronic Promoter in the Rat ER{alpha} Gene Responsible for the Transient Transcription of a Variant Receptor Endocrinology, September 1, 2001; 142(9): 4106 - 4119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Pasqualini, D. Guivarc'h, J.-V. Barnier, B. Guibert, J.-D. Vincent, and P. Vernier Differential Subcellular Distribution and Transcriptional Activity of {{Sigma}}E3, {{Sigma}}E4, and {{Sigma}}E3-4 Isoforms of the Rat Estrogen Receptor-{{alpha}} Mol. Endocrinol., June 1, 2001; 15(6): 894 - 908. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Platet, S. Cunat, D. Chalbos, H. Rochefort, and M. Garcia Unliganded and Liganded Estrogen Receptors Protect against Cancer Invasion via Different Mechanisms Mol. Endocrinol., July 1, 2000; 14(7): 999 - 1009. [Abstract] [Full Text]
|
|
|
|
|
|
A. Bollig and R. J. Miksicek An Estrogen Receptor-{alpha} Splicing Variant Mediates Both Positive and Negative Effects on Gene Transcription Mol. Endocrinol., May 1, 2000; 14(5): 634 - 649. [Abstract] [Full Text]
|
|
|
|
|
|
F. Pakdel, R. Metivier, G. Flouriot, and Y. Valotaire Two Estrogen Receptor (ER) Isoforms with Different Estrogen Dependencies Are Generated from the Trout ER Gene Endocrinology, February 1, 2000; 141(2): 571 - 580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. L. O’Brien, K. Park, Y. In, and O.-K. Park-Sarge Characterization of Estrogen Receptor-{beta} (ER{beta}) Messenger Ribonucleic Acid and Protein Expression in Rat Granulosa Cells Endocrinology, October 1, 1999; 140(10): 4530 - 4541. [Abstract] [Full Text]
|
|
|
|
|
|
C. Pasqualini, D. Guivarc’h, Y. v. Boxberg, F. Nothias, J.-D. Vincent, and P. Vernier Stage- and Region-Specific Expression of Estrogen Receptor {alpha} Isoforms during Ontogeny of the Pituitary Gland Endocrinology, June 1, 1999; 140(6): 2781 - 2789. [Abstract] [Full Text]
|
|
|
|
|
|
R. L. Balleine, S. M. N. Hunt, and C. L. Clarke Coexpression of Alternatively Spliced Estrogen and Progesterone Receptor Transcripts in Human Breast Cancer J. Clin. Endocrinol. Metab., April 1, 1999; 84(4): 1370 - 1377. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
