Some good reading here....
(CANCER RESEARCH 49, 5312-5316, October 1. I989|
Up-Regulation of Estrogen Receptors by Nonsteroidal Antiestrogens in Human
Breast Cancer1
David T. Kiang,2 Rahn E. Kollander, Thresia Thomas, and B. J. Kennedy
Division of Medical Oncology, Department of Medicine, University of Minnesota Medical School, Minneapolis, Minnesota 55455
ABSTRACT
Development of resistance to hormonal therapy in breast cancer is
frequently associated with a decline or loss of cellular estrogen receptors.
Agents which up-regulate the receptor may reduce the incidence of
hormonal resistance. Antiestrogens at concentrations ranging from 0.1 to
l »IMproduced a 2- to 4-fold increase of estrogen receptors in MCF-7
and T-47D breast cancer cells. This increase, which occurred as early as
3 h and was sustained throughout the 4 days of continuous exposure to
tamoxifen, was primarily due to an enhancement in receptor synthesis.
INTRODUCTION
Hormonal therapy plays a significant role in breast cancer
management. Objective tumor regression can be achieved from
hormonal therapy in patients whose tumors contain a signifi
cant amount of ER' (l, 2). Nevertheless, hormonal resistance
eventually develops when a hormonally dependent, ER-rich
tumor evolves to one that is nonresponsive and ER poor (3, 4).
Preventing the development of hormonal resistance may im
prove patient survival. With this notion, up-regulation of tumor
ER may be one of the means to slow down or to reverse the
process of hormonal resistance.
In most instances, the dominant factors regulating the recep
tors are their own ligands. The effect of estrogens on cellular
ER varies depending on the developmental stage of the target
organ. In immature uteri, estrogens induce ER (5); while in
mature uteri and breast cancer, they down-regulate ER (5, 6),
probably via a feedback mechanism to avoid excessive estrogenie
function. We therefore hypothesize that tamoxifen acting
as an antiestrogen may deprive the cell of this estrogenic
feedback loop, thus up-regulating the ER. Furthermore, antiestrogens
have been shown to induce progesterone receptors in
uteri and breast cancers (7, 8), and it is conceivable that they
may also induce ER synthesis.
The mechanism of action of antiestrogens is that it competes
with estrogens at the ER-binding sites. Before the current
knowledge that the nuclear location of ER was available, it was
once thought that the ER-antiestrogen complexes remained in
the nuclei for a prolonged duration, rendering the cells insen
sitive to estrogenic stimulation (9). Such nuclear ER-antiestro
gen "retention" has been thought to be due to a reduction in
receptor degradation (processing) (9-12). In this study, we
attempt to investigate whether the mechanism of ER increase
could be based on another reason, i.e., an increase of ER
synthesis. The ER up-regulation, if proven to exist, could be a
useful tool in modulating the behavior of the cancer cell and in
preventing it from becoming hormonally resistant.
There are technical difficulties in studying the modulation of
ER in the presence of tamoxifen, because antiestrogens interfere
with the conventional ER assay in which ['HJestradiol is used
to measure the amount of ER. Earlier studies on ER modulation
by estrogens or antiestrogens have adopted a ligand exchange
method which, however, is not completely reliable. Because of
their extremely heat-labile properties, some receptor bindingactivities
are lost during ligand exchange at temperatures rang
ing from 25 to 34°C.Recent advances in anti-ER monoclonal
antibodies have made it possible to accurately determine the
receptor level in the presence of antiestrogens. The antigenic
sites of ER are much more physically stable than the biological
steroid-binding sites. In this study, we assess the effect of
tamoxifen on the rate of ER synthesis in human breast cancer
cells using both the anti-ER monoclonal antibodies and the
density-shift technique (13).
RESULTS
The effect of 1 MMtamoxifen on ER content in MCF-7 cells
is shown in Table 1. In this experiment, cells were exposed to
tamoxifen for 6, 24, and 48 h. There was a 3- to 4-fold increase
of ER content in MCF-7 cells. Fig. 1 shows the effects of
continuous exposure of cells to 1 or 0.1 MMtamoxifen for up
to 5 days. There was a 325% increase of ER with 1 JUM
tamoxifen at 48 h, but the receptor level decreased to that of
the control in 84 hr, presumably due to the cytotoxic effect
from this high concentration of tamoxifen. Although there was
a smaller increase in ER (220%) at the lower concentration of
tamoxifen (0.1 MM), this stimulatory effect was sustained
throughout the 5-day experimental period.
Fig. 2 shows the dose-dependent effect of tamoxifen and its
derivative, 4-hydroxytamoxifen, on ER content. Stimulation of
ER by tamoxifen or 4-hydroxytamoxifen was seen in both
MCF-7 and T-47D breast cancer cells and was dose dependent.
The elevation of ER by 4-hydroxytamoxifen was observed at a
concentration as low as 0.1 nM and reached a maximum at 10
UM.On the other hand, the stimulatory effect of tamoxifen was
not seen until the concentrations exceeded 10 nM. A concentra
tion of l MMtamoxifen was required to attain the equivalent
stimulatory effect of 10 HM 4-hydroxytamoxifen. Hence, 4-
hydroxytamoxifen was at least 100 times more potent than
tamoxifen in elevating ER. When 0.1 MMconcentrations of
various antiestrogens were added to MCF-7 cells, their effec
tiveness was of the following order: tamoxifen = keoxifene <
trioxifene = LY117018 < 4-hydroxytamoxifen (Table 2).
One aspect which needs to be addressed is that the increase
in ER detected by monoclonal antibody techniques might be
attributed to an increase of available antigenic sites following
the tamoxifen binding. This phenomenon has recently been
described by Martin et al. (17) in a different buffer system. We
therefore performed experiments to study the effect of anties
trogens on ER content when they were added to the cell lysate
(Table 3). Tamoxifen up to 2 MMconcentration or 4-hydroxy
tamoxifen up to 1MMadded into the cell lysate did not increase
the ER measurement, while an increase of ER occurred in cells
exposed to antiestrogens in the culture medium. This increase
was observed in cells grown in regular media (Table 3), as well
as in estrogen-deprived media (data not shown).
Increase of the ER content by antiestrogens could be attrib
uted to either a stimulation in receptor synthesis or a decrease
in ER processing. The protein synthesis inhibitor, cycloheximide,
was added into the culture medium to examine its effect on
receptor synthesis (Table 4). After exposure to 0.01 ¿IM4-
hydroxytamoxifen for 3 and 6 h, the ER content of MCF-7
cells increased to 126.4 and 148.6% of the control, respectively.
Cycloheximide (20 Mg/ml) completely blocked the ER stimu
lation from 4-hydroxytamoxifen in 6 h.
Using the dense amino acid labeling technique, we measured
the amount of newly synthesized ERD, and the remaining ERL
in MCF-7 cells with or without exposure to 0.5 MMtamoxifen.
Whole cell ER was extracted by 0.4 M KC1 and layered on top
of 5-20% sucrose gradients containing 0.4 MKC1-TEMG buffer
in deuterium oxide. After ultracentrifugation, 4-drop fractions
were collected and the ER content in each fraction was meas
ured by the ER-EIA method. The ERL was sedimented at
approximately 3.8S, while the sedimentation of ERD shifted to
the 5.4S region (Fig. 3). It is interesting to note that tamoxifen
exposure not only increased the magnitude of the ERD peak
but also caused a further shift of ERD density from 5.4S to 6.6S
(Fig. 3, A and B). The shift of ERD-tamoxifen complexes to a
heavier density region made it technically easier for us to
identify the newly synthesized ERD. This ER density shift from
tamoxifen, as described by other investigators, can be prevented
by the addition of urea (18,19). In a separate set of experiments,
3 Murea was added to the sucrose gradients. Indeed, 'the ERD-
tamoxifen peak at 6.6S was converted back to 5.4S in theUsing the protein density-shift technique as shown in Fig. 3,
the amounts of ERD and ERL were calculated, and the control
and tamoxifen-treated groups were compared using the control
values as 100%. The average total ER in control cells was 1.85
pmol/mg DNA and the ERD was 1.20 pmol/mg DNA or 65%
of total ER in 3 h. Following tamoxifen (0.5 n\t) treatment for
3 and 6 h, there were 1.5-1.6 times as much total ER (P <
0.01). These increases were totally derived from the newly
synthesized ERD (Table 5). The newly synthesized ERD in
tamoxifen-treated MCF-7 cells was 1.97 times higher than the
control at 3 h and 1.84 times higher at 6 h (P < 0.01). On the
other hand, the tamoxifen did not appear to affect the ER
processing judging by a rather stable l-'.R, level at 3 and 6 h in
comparison with that of the control group.
Although the experiment shown in Fig. 1 indicated that the
increased cellular ER level could be maintained over the entire
5-day experimental period, it was critical to know whether the
elevated ER synthesis by tamoxifen could also be maintained
through this period. To study this aspect, MCF-7 cells grown
in estrogen-deprived media were exposed to a relatively nontoxic
concentration (20) of 0.2 ^M tamoxifen for periods rang
ing from 3 to 96 h. The cells were then grown in dense amino
acid media containing zero or 0.5 UMtamoxifen for 6 h before
being harvested. The amounts of newly synthesized ERD meas
ured by the protein density shift technique were 1.6 to 2.2 times
higher in cells that were preexposed to tamoxifen for 3 to 96 h
than in control cells with no exposure to tamoxifen (Fig. 4).
Also, there was no difference in DNA content between control
and experimental groups. Thus, cell growth was not suppressed
by 0.2 UMtamoxifen, and a change in DNA content was not
the cause of the observed ER increase. This study indicated
that the ER increase was mainly caused by a stimulation of ER
synthesis which persisted even through the 4-day tamoxifen
preexposure.
DISCUSSION
In this study, we have shown that the up-regulation of ER by
by the addition of urea (18,19). In a separate set of experiments, tamoxifen and its analogues is reproducible in two breast cancer
3 Murea was added to the sucrose gradients. Indeed, 'the ERD- cell lines, is dose-dependent, and is relatively long-term (at least
tamoxifen peak at 6.6S was converted back to 5.4S in the 5 days). It was not due to an increase in receptor antigenic sites caused by tamoxifen. Using hydroxytamoxifen (20 UM)in the
MCF-7 culture system, other investigators also demonstrated a
2-fold increase of nuclear hydroxytamoxifen-ER complexes
measured by [:>H]-4-hydroxytamoxifen binding, a measurement
which was unrelated to the epitope expression of ER antigen
(12). These results support our findings that there is a genuine
ER up-regulation from tamoxifen and its analogues. Our study
further illustrated that the ER up-regulation was due to an
increased rate of ER synthesis.
In our experiments, MCF-7 cells were from a range of
subculture numbers (passages 36 through 55). In spite of the
fact that the control ER contents in MCF-7 cells varied from
62 to 218 fmol/mg, the up-regulation of ER by antiestrogens
was persistently observed. There is a trend that the lower the
initial ER level, the greater is the ER up-regulation in response
to tamoxifen. This phenomenon indirectly supports the concept
that the ER response is a biological event rather than a physicochemical
alteration of receptors. The latter, if it occurs, would
provide a fixed percentage increase of antigenic sites at each
given concentration of tamoxifen.
It is interesting to note that ER up-regulation (3-fold in
crease) was also observed in MCF-7 cells growing in an estro
gen-deprived medium for 5-6 months (21). We therefore pos
tulate that the effects of estrogen depletion and tamoxifen
exposure may involve a common mechanism, i.e., a lack of an
estrogen-induced feedback loop. Further elucidation on the
mechanism of ER modulation by tamoxifen at the molecular
level is needed.
The mechanism of the up-regulation of ER by tamoxifen
may take place at the transcriptional or translational level. It is
interesting to note that the antiestrogens, hydroxytamoxifen
and LY117018, had no effect on ER mRNA levels (22, 23).
There is a possibility that tamoxifen may regulate the ER at
the translational level via a mechanism similar to the human
ferritin regulation. Ferritin biosynthesis is altered without a
corresponding change in the level of total ferritin mRNA (24);
instead, an iron-responsive element in the 5'-untranslated por
tion of the ferritin mRNA mediates iron-dependent control of
its translation (25). The phenomenon that rapid ER synthesis
was observed within 3 h of tamoxifen exposure certainly is
compatible with the hypothesis of translational regulation.
In clinical settings, successful tamoxifen therapy is frequently
associated with an ER reduction in the residual tumor (26, 27).
Such ER reduction is attributed to clonal selection by eliminat
ing an ER-rich hormonally sensitive population of tumor cells.
The effects of a low-dose tamoxifen on the tumor ER content
is still unknown. It is interesting to note that in athymic nude
mice bearing MCF-7 cells, tamoxifen treatment for 4 months
led to an increase in tumor ER (28). Such ER enhancement
may be involved in perpetuating the antiestrogenic and antitumor
effects of tamoxifen, a mechanism that may contribute
to the excellent tumor regression frequently observed with
tamoxifen therapy. There is also ample evidence which implies
a dissociation of proliferative function from biological activity
of antiestrogens (29, 30). As demonstrated in this study, ta
moxifen and other antiestrogens at their optimal sublethal
concentrations may alter the biological function via ER modu
lation without interfering with cell proliferation. In other words,
they may potentially be used as "maturation" agents in pre
venting the tumor progression from a hormonally sensitive to
resistant status.