Physical Properties of Estrogen Receptor Complexes in MCF-7 Human Breast Cancer Cells DIFFERENCES WITH ANTI-ESTROGEN AND ESTROGEN*

We have examined the binding of two high affinity radiolabeled anti-estrogens, 2-[4-(2-pyrrolidin-l-yl-eth-oxy)phenyl]-2-(3-[3H]-4-hydroxyphenyl)-l-phenyl-l- nitroethene ([3H]C1628M) and 1-[4-(2-dimethylamino-ethoxy)phenyl]-l-(4-hydroxyphenyl)-2-(4-[3H]phenyl)- but-l(2)-ene ([3H]trans-hydroxytamoxifen) to the estrogen receptor from MCF-7 human breast cancer cells and have used hydrodynamic methods to determine the molecular properties of estrogen and anti-estrogen receptor complexes from these cells. Saturation binding analysis indicates that each compound binds predomi-nantly to a single class of high afflnity binding sites with K d of 1.3 X 10“’ M for [3H]estra-1,3,5(10)-triene-3, 17p-diol (estradiol (Ez)), 1.4 X 10”’ M for [3H]trans- hydroxytamoxifen, and 2.2 X 10”’ M for [3H]C1628M. Marked differences are seen in the sedimentation rate and chromatographic properties of the nuclear estrogen receptor when complexed with anti-estrogen as opposed to

The differences in physical properties of nuclear estrogen and anti-estrogen receptor complexes are obliterated in the presence of 3 M urea where estrogen and anti-estrogen complexes of 81,000 molecular weight (3.9 S, 4.94-nm Stokes radius) are obtained. These data are consistent with the association of the nuclear antiestrogen receptor complex with an additional protein This work was supported by National Institutes of Health Grant CA 18119 and the Bane Estate. A preliminary report of portions of this work was presented at the 63rd Annual Endocrine Society Meeting, June, 1981 (Proceedings of the Sixty-third Annual Endocrine Society Meeting, p. 125, Abstr. 170). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
* To whom correspondence should be addressed. of -55,000 molecular weight, an association that can be reversed by 3 M urea. This anti-estrogen-promoted change in receptor association with another cellular component may be an important aspect of the estrogenantagonist and growth-inhibiting properties of these compounds.
The classical scheme of estrogen action, wherein estrogen binds to the cytoplasmic receptor to form an activated complex which then translocates to the nucleus and interacts with chromatin sites to modulate gene expression, still appears to be a suitable model for the interaction of estrogen in target cells, including the mammary gland (1-3). Anti-estrogens, which are effective inhibitors of the growth of hormone-dependent breast cancers (4-6), compete with estrogen for binding to cytoplasmic receptor sites and translocate the receptor complex to the nucleus (5,7). The nuclear anti-estrogen receptor complex, however, appears to be only partially active in promoting specific biological responses and is effective in blocking the action of estrogens (7)(8)(9)(10). This suggests that the receptor anti-estrogen complex might be physically different from the receptor estrogen complex; receptor conformation or size could be altered, or its interaction with other cellular components might be modified.
Most efforts to characterize the interaction of anti-estrogens with receptor have utilized indirect exchange or competition assays (5, 7, 11). Moreover, previous efforts to characterize directly anti-estrogen receptor complexes using radiolabeled anti-estrogens have been hampered by the low affinity of these compounds for estrogen receptor, which results in the dissociation of the complex during characterization (7). The recent availability of radiolabeled anti-estrogens, having affinities for estrogen receptor comparable to that of the natural estrogen estradiol,' now make possible the direct characterization of these complexes (12, 13).
The MCF-7 cell line, derived from a human metastatic breast carcinoma, is a model system for study of the regulation of breast cancer cell growth by estrogens and anti-estrogens. These cells contain high levels of estrogen receptor and estrogen-responsive progestin binding activity, and growth of these cells is inhibited by anti-estrogens (8)(9)(10)(11). In this paper, we compare the receptor interactions of two radiolabeled high affinity nonsteroidal anti-estrogens, ["H]CI628M and ["HI trans-hydroxytamoxifen, with those of the potent natural estrogen ['Hlestradiol and examine the physicochemical properties of receptor following anti-estrogen or estrogen binding. Our findings demonstrate a substantial alteration in sedimentation rate and chromatographic properties of the nuclear estrogen receptor from MCF-7 cells when complexed with either of the high affinity anti-estrogens, CI628M or transhydroxytamoxifen, as opposed to the estrogen estradiol. These changes in receptor properties may be an important aspect underlying the estrogen-antagonist antitumor properties of these compounds.
Four days prior to use, cells were transferred to harvest medium, which was the same as growth medium except that calf serum was replaced by charcoal-dextran stripped calf serum prepared by incubating 100 ml of serum with 5 ml of charcoal-dextran solution (0.50% Norit A, 0.05% dextran in 0.14 M NaCI) for 30 min a t 55 "C. This procedure removed 96-98% of a trace amount of ['H]estradiol. The charcoal was removed by centrifugation-filtration, and the serum was filter-sterilized and stored at -20 "C. Harvest medium was also supplemented with 10 nM hydrocortisone. Cells were fluid-changed with fresh harvest medium on 4 consecutive days prior to each experiment to enable the removal of endogenous steroids.
[2,4,6,7-"H]Estradiol (108 Ci/mmol) was obtained from Amersham. sodium phosphate, pH 7.4 at 4 "C, 10 mM thioglycerol, and 10% glycerol (PTG buffer), and homogenized in a Dounce homogenizer using the B-pestle. The homogenate was centrifuged (10 min, 800 X g), and the supernatant was collected. The crude nuclear pellet was washed twice more at 0-4 "C with buffer and the supernatants combined. The supernatant fraction (cytosol) was centrifuged at 180,000 X g (30 min) and then diluted to 12 ml with PTG buffer. Aliquots of cytosol ( These times gave maximal localization of anti-estrogen and estradiol receptor complexes in the nucleus (10). The cells were harvested, washed as described above, and homogenized in 1 ml of PTG buffer. The homogenate was centrifuged at 800 X g for 10 min, and the resulting nuclear pellet was washed twice with 1 ml of iced PTG buffer. The washed pellet was resuspended in 0.125 ml of PTG buffer to which 0.375 ml of 10 mM Tris-HCI, pH 8.5, a t 4 "C, 1.5 mM EDTA, 10 mM thioglycerol, 10%) glycerol, 0.8 M KC1 was added and incubated a t 4 "C for 1 h with resuspension every 15 min. This procedure extracted reproducibly over 90% of the ["H]E. or ["Hlanti-estrogen receptor complexes as determined by ethanol extraction. The suspension was centrifuged for 30 min at 180,000 X g, and the supernatant was treated for 8 min at 0-4 "C with charcoal-dextran slurry (1 part of slurry, 9 parts of extract). A 300-pI aliquot was layered on linear 5-20% sucrose gradients (3.6 ml) formed in PTG buffer containing 0.4 M KCI. The gradients were centrifuged at 357,000 X g in an SW 60 rotor for 15 h, and two-drop fractions were collected for counting. Cells were also treated with radioactive ligand plus 100-fold excess of radioinert estradiol to assess the level of nonspecific binding. Hecovery on all gradients was greater than 90%. Sedimentation coefficients were determined according to Martin and Ames (16) relative to the internal markers ['4C]ovalbumin (3.5 S) and ["C]y globulin (6.6 S) which were included in each gradient (13).
Binding to cytoplasmic sites was assessed in the cell-free cytosol (supernatant (180,000 x g X 30 m i d ) prepared from control previously untreated cells or following incubation of cells with tritiated estradiol or anti-estrogen for 1.5 h at 0-4 "C. In the latter case, the cells were then harvested, washed, and homogenized in 0.5 ml of PTG buffer, as above. The homogenate was centrifuged at 800 X g for 10 min, and the resulting nuclear pellet was washed twice with 0.25 ml of iced PTG buffer. The supernatants were combined, centrifuged at 180,000 X g for 30 min, treated for 8 min with charcoal-dextran slurry, and then sedimented for 15 h on 5-20% sucrose gradients formed in PTC buffer containing 0.4 M KCI. Sucrose gradients (5-20% sucrose) containing 3.0 M urea and 0.4 M KC1 were prepared in PTG buffer. Centrifugation was for 17 h a t 357,000 X g a t 4 " C .
Sephadex G-200 Column Chromatography-MCF-7 cells from two near-confluent T-150 flasks were incubated with each compound as described above. The extracts were also prepared as described above, except that the final volume of the nuclear and cytosol extracts was 1.1 ml. After charcoal-dextran t,reatment (9 parts of extract, 1 part of slurry) and removal of an aliquot for determination of bound radioactivity, 1.0 ml of extract was applied to a Sephadex G-200 column (1.5 X 24 cm) equilibrated at 4 "C with PTG buffer containing 0.4 M KCI. Samples were eluted at a flow-rate of 5.4 ml/h, and 1.5-ml aliquots were collected. For columns containing urea, the Sephadex column pack was equilibrated a t 4 "C with PTG buffer containing 0.4 M KC1 and 3.0 M urea prior to pouring columns (1.5 X 24 cm).

Analysis of the Binding of Anti-estrogens to the MCF-7
Cytosol Receptor-Saturation binding analyses for estradiol and the anti-estrogens are shown in Fig. 1. These data indicate the presence of one class of high affinity cytosol binding sites with equilibrium dissociation constants ( K d ) of 1.3 X 10"" M for E?, 2.2 X 10"" M for CI628M, and 1.4 X 10"" M for transhydroxytamoxifen binding to the estrogen receptor. Scatchard binding analysis (17), therefore, indicates that the affinity of these anti-estrogens for receptor is very close to that of estradiol and that each compound binds to a similar number of sites. Moreover, competitive binding analysis (10) and the full displacement of ['Hlanti-estrogen binding by unlabeled estradiol (see below) indicate that these compounds bind in a mutually competitive manner to the cytosol estrogen receptor. Sedimentation Behauior of Tritiated Anti-estrogen a n d Estrogen Receptor Complexes-To determine whether the physicochemical properties of the receptor anti-estrogen complex differ from those of the receptor estrogen complex, MCF-7 cells were incubated for 0.5-1 h at 37 'C with a receptorsaturating concentration of radiolabeled E2, CI628M, or transhydroxytamoxifen. The nuclear receptor sites were then extracted with 0.6 M salt and analyzed on sucrose gradients. (This extraction procedure solubilized greater than 9090 of the tritiated estradiol or anti-estrogen receptor complexes.) Fig. 2 shows that the receptor estradiol complex sediments at 4.1 S while CI628M and trans-hydroxytamoxifen complexes sediment more rapidly, a t 5.5 S. By contrast, the cytoplasmic receptor, when complexed with any of these estrogen or antiestrogen ligands, following in vitro cytosol or whole cell labeling a t 0-4 "C, showed similar sedimentation rates (4.1 s) on high salt. sucrose gradients (Fig. 3).
Gel Chromatography of Nuclear and Cytoplasmic Antiestrogen a n d Estrogen Receptor Complexes-Nuclear and cytoplasmic receptor complexes with radiolabeled anti-estrogen or estradiol were chromatographed on a Sephadex G-200 column that had been standardized previously with a series of proteins of known Stokes radii (18)(19)(20). The elution profiles of the receptor complexes are shown in Fig. 4 (top), and the elution positions of the complexes (IC,") are indicated on the column calibration line (Fig. 4, bottom). The nuclear CI628M and trans-hydroxytamoxifen complexes eluted ahead of the nuclear estradiol complex, indicating that the nuclear antiestrogen receptor complexes have a larger Stokes radius (5.

regardless of the ligand, eluted at an identical position and somewhat later than the nuclear form (4.39-nm Stokes radius).
Utilizing the Stokes radii estimated from Sephadex G-200 column chromatography and the sedimentation coefficients determined from sucrose gradient analyses, the molecular weights of the complexes were calculated by the equation of Siege1 and Monty (21) and are summarized in Table I. The cytoplasmic receptor, whether complexed with estradiol or anti-estrogen, has a calculated molecular weight of 75,600.

Anti-estrogen-Estrogen Receptor Complexes
The nuclear estradiol receptor complex has a slightly greater most likely represents an increase in mass as a result of molecular weight (83,400) due to a larger Stokes radius. This association with another component (see "Discussion"). small change may simply reflect a change in conformation.
Characterization ofEstradiol and Anti-estrogen Receptor The nuclear anti-estrogen receptor complexes, by contrast, Complexes in the Presence of Urea-Urea is known to dishave a far greater molecular weight (137,000). Such a large sociate protein molecules held together by weak interactions increase in calculated molecular weight is unlikely to be (22, 23). Since the binding of estradiol and anti-estrogen to ascribable to conformational change or to a change in Ls and receptor was found to be very stable in concentrations of urea ov I    (Fig. 5). Each nuclear receptor complex sedimented as a single peak a t 3.9 S; this is in contrast to the profiles observed in the absence of urea (Fig.   2). Cytoplasmic receptor sites also sedimented a t 3.9 S, regardless of ligand (Fig. 5 ) . Analysis of estradiol and antiestrogen receptor complexes on Sephadex G-200 columns equilibrated with 3 M urea and 0.4 M KC1 and calibrated with a series of protein standards showed the cytoplasmic or nuclear receptors complexed with E2 or anti-estrogen to elute similarly, corresponding to a Stokes radius of 4.94 nm (Fig. 6). The molecular weights calculated from the Stokes radii and sedimentation coefficients are 81,000 for all the receptor species.

DISCUSSION
In the past, studies of receptor anti-estrogen interaction have utilized indirect exchange assays to detect receptor (4,5,7). This was necessary because of the lack of radiolabeled anti-estrogens with high affinities for estrogen receptor. In the present study, two high affinity radiolabeled anti-estrogens were utilized (7,12,13) to study the physical properties of receptor anti-estrogen complexes in MCF-7 human breast carcinoma cells.
Our findings indicate that anti-estrogens promote a change in receptor form that may be the result of a change in its association with other cellular components such that the MCF-7 nuclear estrogen receptor migrates as a distinctly heavier form when complexed with anti-estrogen as compared with estrogen. The fact that the cytoplasmic receptor shows similar properties when complexed with estrogen or anti-estrogen ligands and that this heavier form is not seen upon incubation of the cytosol anti-estrogen receptor complex at 30 or 37 "C suggests that the shift in sedimentation rate and size of the receptor appears to be associated with exposure to the nuclear compartment.
The nuclear anti-estrogen complexes sedimented at 5.5 S and eluted with a Stokes radius of 5.93 nm, corresponding t,o a calculated molecular weight of 137,000. This represents an increase in molecular weight over that of the nuclear estradiol receptor complex too great to be explained by conformational change and suggests that the receptor is associating with other components (23). T o determine whether the anti-estrogen receptor complexes were associated with other cellular components, properties of the estradiol and anti-estrogen complexes were studied on sucrose gradients and on Sephadex columns equilibrated with 3 M urea. Under these conditions, cytoplasmic and nuclear receptors, whether complexed with estradiol or anti-estrogen, sedimented at 3.9 S , and they all eluted from Sephadex columns with a St,okes radius of 4.94 nm. These values correspond to a calculated molecular weight of 81,000.
Urea alters the hydrodynamic properties of the cytoplasmic receptor complexes and nuclear estradiol complex, with the Stokes radius increasing and the sedimentation coefficient decreasing, in a manner expected for moderate concentrations of urea. However, the molecular weights of the cytoplasmic receptor complexes and the nuclear estradiol receptor complex calculated from the Stokes radius and sedimentation coefficient remain unchanged. The presence of 3 M urea has a more profound effect on the nuclear anti-estrogen receptor complex. The changes in hydrodynamic properties observed in the presence of urea are consistent with dissociation of a component of approximately 55,000 molecular weight, resulting in a form of the nuclear anti-estrogen receptor complex that is now indistinguishable from that of the nuclear estradiol receptor complex.
In some estrogen target tissues, including the uterus and pituitary, activation of estrogen receptor (a process by which the cytoplasmic estradiol receptor complex is converted to a form with enhanced affinity for nuclear acceptor sites) is believed to involve a dimerization of receptor or its interaction with a dissimilar protein and is manifest by a change in sedimentation velocity (the 4 S to 5 S transformation) on high salt sucrose gradients (24)(25)(26). It is important to stress that the alteration in sedimentation rate and size of the antiestrogen nuclear receptor complexes probably does not represent activation of the receptor in the classical sense, as the estradiol receptor complex from MCF-7 cells is not changed in these properties upon activation. Both cytoplasmic and nuclear estradiol receptor complexes from MCF-7 cells sediment a t 4.1 S (as shown previously, Ref. 27) and have similar Stokes radii.
Our findings suggest that anti-estrogens may antagonize estrogen action by promoting a change in receptor that results in its association with another cellular component. It is tempting to speculate that this anti-estrogen-promoted interaction may be biologically nonproductive and would prevent or block receptor from associating with biologically important chromatin-binding or "acceptor" sites (28). Such a model could explain the deficient activation of estrogen-dependent responses by anti-estrogens and the gr0wt.h inhibition observed in MCF-7 cells treated with anti-estrogens (5,8,10). Further studies should allow a continuing examination of this hypothesis.