Antiestrogen Action of 2-Hydroxyestrone on MCF-7 Human Breast Cancer Cells*

The estrogen responsive human breast cancer MCF-7 cell culture was examined for its response to 2- hydroxyestrone a principal metabolite of estradiol. Addition of 2-hydroxyestrone to the cell cultures in concentration of 10-9-10-6 M had no effect on cell growth and proliferation because of rapid O-methylation of the catechol estrogen by catechol O-methyltransferase which is highly active in these cells. In the presence of quinalizarin, a potent catechol O-methyltransferase inhibitor which reduces the O-methylation of the steroid, lo-’ M and lo-’ M 2-hydroxyestrone markedly suppresses the growth and proliferation of the cells. The tumor cell growth-inhibitory action of the catechol estrogen was neutralized by the presence of lo-’ M estradiol. The catechol estrogen inhibition of cell growth is not observed in the estrogen receptor-nega-tive human breast cancer cell lines MDA-MB-231 and MDA-MB-330 providing evidence that the inhibition is specific and is estrogen receptor-mediated. In contrast, the 16a-hydroxylated metabolites of estradiol, estriol and 16a-hydroxyestrone, are effective stimu-lators of MCF-7 cell proliferation with the latter ex- hibiting potency in excess of that expected from its estrogen receptor affinity. The present results repre- sent the first observation of a specific receptor-me-diated antiestrogenic action of 2-hydroxyestrone and suggest that the physiological regulation of the agonist activity of the primary estrogen may involve in situ generation of

the primary estrogen for the receptor. The catechol estrogen may also act via their interaction with the catecholaminergic system. They have been reported to bind specifically to dopamine and al-adrenergic receptors in the brain and pituitary (9, 10). They inhibit tyrosine hydroxylase, the rate-limiting enzyme of catecholamine biosynthesis (11,12) and can thus decrease catecholamine content. Conversely they are highly effective competitive inhibitors of the metabolism of catecholamines by catechol O-methyltransferase and would thus decrease the turnover time of these biogenic amines (13,14). Indeed, it is now apparent that catechol estrogen administered in uiuo are rapidly metabolized by the action of tissue and erythrocyte catechol O-methyltransferase resulting in very high metabolic clearance rates for these compounds (15)(16)(17). This feature of their biology makes it difficult to define their intrinsic biological properties by in uiuo administration.
In the present study we have used the hormone-dependent human breast cancer cell line MCF-7 in tissue culture as a system for the study of the biological effects of 2-hydroxyestrone, under conditions where its further metabolism could be controlled, and its activity in peripheral estrogen target sites be evaluated. Our results indicate that the catechol estrogen behaves as an antiestrogen in suppressing the tumor cell proliferation both under control and estradiol stimulatory conditions. This activity of the catechol estrogen becomes apparent in the presence of the inhibitors of catechol 0methyltransferase an enzyme whose activity is greatly increased in this tumor tissue.

EXPERIMENTAL PROCEDURES
Materials-Human breast cancer cell lines MCF-7, MDA-MB-231, and MDA-MB-330 were obtained from the Breast Cancer Human Cell Culture Bank of the National Cancer Institute (maintained by EG & G Mason Research Institute, Worcester, MA). 2- [6,7-3H] hydroxyestrone (specific activity, 40-60 Ci/mmol), purchased from New England Nuclear, was purified by thick layer chromatography as described (18) to ensure the absence of any contamination by other steroids, particularly estradiol. Estradiol and estriol were obtained from Steraloids, Inc. (Wilton, NH), l6a-hydroxyestrone from Sigma, and quinalizarin from Aldrich Chemical Co. All steroids were checked for purity by chromatographic and spectral criteria, and were purified to homogeneity as required. RPMI 1640 medium, fetal bovine serum, penicillin-streptomycin suspension (10,000 units and 10,000 pg, respectively/ml), mycostatin suspension (10,000 units/ml), trypsin-EDTA (lX), Eagle's balanced salt solution, and Eagle's balanced salt solution without Mg+ and Ca2+ were obtained from Gibco Labs (Grands Island, NY). Sterilization filters (0.45 p and 250 ml or 500 ml capacity) were purchased from Nalge Co. (Rochester, NY).
Cell Culture-The MCF-7 cells were subcultured in T-75 flasks by daily changes of the "passage" medium (50 ml) consisting of RPMI 1640 supplemented with fetal bovine serum (lo%, v/v). which were maintained in a humidified chamber at 37 "C with 5% COZ and 95% air. The cells were harvested at confluency by trypsinization and suspended at a cell density of 200,000 to 300,000 cells/ml in a small volume of the growth medium to be used.
In a typical experiment in which the effects of hormones on the cell growth were to be examined, aliquots of the cell suspension containing about 20,000 to 30,000 cells were replicately plated in dishes (60 mm in diameter) containing 3 ml of the growth medium consisting of RPMI 1640 supplemented with charcoal-stripped1 fetal bovine serum (5%, v/v) and of penicillin-streptomycin (10 units and 10 pg, respectively/ml of medium) and mycostatin (2.5 units/ml) and incubated at 37 "C in the humidified chamber. On day 1 when the cells adhered firmly to the bottom surface of the dishes, the medium was aspirated off and the cells then challenged by hormones; groups of replicate dishes received 3 ml of fresh growth medium containing a hormone at the indicated concentrations. Another group of replicate dishes which received only the vehicle served as the nonbormonetreated controls. All dishes, unless otherwise stated, also received quinalizarin (dissolved in dimethyl sulfoxide), a potent inhibitor of catechol 0-methyltransferase (19), at a concentration of M, a concentration which does not exert any effect on cell growth but is effective in inhibiting the enzyme. Each group received a change of the fresh growth medium (3 ml) every 24 h unless otherwise noted, at which time the hormone-treated group received the hormone at the indicated levels and the control group the vehicle. At various time intervals, i.e. 24 h, triplicate dishes were withdrawn for cell counting; after removal of the medium by aspiration, the adhered cells were washed twice with 1 ml of Eagle's balanced salt solution without Ca2+ and M e and dislodged from the dish by trypsinization involving the incubation with 0.5 ml of trypsin-EDTA solution (1X) for 5 min at 37 "C, after which the protease reaction was stopped by adding 2.5 ml of a "stop" medium consisting of RPMI 1640 and 5% fetal bovine serum. The cells were then collected by centrifugation with subsequent suspension in the medium. An aliquot of the suspension was subjected to cell counting using a hematocytometer. In the present study we focused on cell proliferation as an index of estrogen response since other endpoints such as [3H]thymidine incorporation are subject to complex variables (20,21).

MCF-7 Cancer Cells
The estrogen-independent human breast cancer cell lines MDA-MB-231 and MDA-MB-330 were cultured in a manner identical with that described above for MCF-7.
Metabolism of 2-Hydroxyestrone and Identification of Its Metabolic Products-150 pmol of 2-[6,7-3H]hydroxyestrone (containing about ' The calf bovine serum was stripped of endogenous estrogens by treatment with charcoal (1 g/lOO ml of serum) with stirring in a dark cold room for 24 h, after which the serum was recovered by filtering under vacuum on acid-washed Celite paper, followed by sterilization filter. Complete removal of the unconjugated steroids was verified by a test in which an exogenously added labeled steroid was completely removed by the identical treatment. It is known however (38) that this treatment does not result in complete removal of estrogen sulfates which could serve as sources of estrogen stimulus upon deconjugation by the tissue. 1.32 million dpm) were added to groups of replicate MCF-7 cells on day 6 when they were growing exponentially (see Fig. l ) , and then exposed for the various time intervals indicated, after which all of the medium (3 ml) was collected for analysis of the metabolites. Then 25 pg of 2-hydroxyestrone (dissolved in AMA solution consisting of 20% ascorbic acid, 78% methanol, and 2% acetic acid) was added as a carrier. The medium was extracted with 3 ml of ethyl acetate by mixing vigorously in a vortex for 1 min, and collecting the upper organic layer after centrifugation at 2500 X g for 10 min and evaporating to dryness. The residue was then taken up in 100 pl of methanol and received 10 pg of 2-methoxyestrone as a standard, after which an aliquot of the methanolic solution was subjected to thin layer chromatography on Silica Gel G in a solvent system consisting of ethanol/ chloroform = 5/95 (v/v). 2-Hydroxyestrone and 2-methoxyestrone had RF values of 0.51 and 0.76, respectively, in this solvent system.
The isomeric 3-methyl ether of 2-hydroxyestrone another possible product of 0-methylation was coincident with 2-methoxyestrone. The resulting thin layer chromatographic plate was scanned by Packard Model 385 Radiochromatogram Scanner.

Rapid Metabolism of 2-Hydroxyestrone by 0-Methylation-
Our initial attempt to demonstrate the effect of 2-hydroxyestrone on the tumor cell growth and proliferation of a hormonedependent human breast cancer cell line, MCF-7, in cell culture, met with a lack of response. Analysis of the metabolites of the catechol estrogen, however, revealed that the failure to observe any effect was due to rapid metabolism of the steroid via 0-methylation by catechol 0-methyltransferase which was highly active in these cells. As shown in Fig. 1, 2-hydroxyestrone, when presented at a concentration of M to the 600,000 cells growing exponentially in RPMI 1640 medium supplemented with 5% calf bovine serum stripped of endogenous estrogen, was rapidly metabolized to 0-methylated products as analyzed by thin layer chromatography. The metabolism is essentially complete in 1-2 h with 30 to 40% of the total dose being 0-methylated and remaining unchanged subsequently for up to 48 h. The remainder of the dose is retained intracellularly; the nature of the cell-retained material was not investigated. It should be noted that 2hydroxyestrone appears to undergo a two-stage metabolism in this system; an early phase conversion to unidentified polar metabolites (Fig. 1) and a subsequent biotransformation to 2methoxyestrone. In any event no intact 2-hydroxyestrone is Day   FIG. 2 (left). Tumor cell growth inhibition of a hormone-dependent human breast cancer cell line MCF-7 by the catechol estrogen, 2-hydroxyestrone. The cells were plated on day 0 as described under "Experimental Procedures" and exposed to the hormones on day 1 (as indicated by an arrow) by changing the medium to the growth medium containing each hormone at the indicated concentrations. Quinalizarin was present in the medium for all hormone-treated cell groups at a concentration of M to inhibit catechol @methyltransferase (19). The hormone-untreated control and the catechol 0-methyltransferase inhibitor control cell groups received the medium containing the vehicle alone and the inhibitor alone, respectively. Media changes were made every other day (on days 3 and 5).  (Fig. 3), suggesting that the 2-hydroxyestrone and the primary estrogen compete with each other for the cellular site(s) of action which is presumably the estrogen receptor. This result is consistent with the notion that the tumor growth-inhibitory action of 2-hydroxyestrone is mediated by the estrogen receptor.
Inability of Estradiol to Rescue the 2-Hydroxyestrone-inhibited MCF-7 Cells-In contrast to the result of the experiment in which the simultaneous challenge on day 1 of the primary estrogen and the 2-hydroxyestrone results in the reversal of the catechol estrogen inhibition, the addition of the former (at a concentration of lo-" or lo-' M) on day 4 under identical condition failed to rescue the catechol estrogen-inhibited cells (Fig. 4), suggesting that in the absence of estradiol during the early days of cell growth the catechol estrogen effects an irreversible alteration of the estrogen-responsiveness of the MCF-7 cells. The irreversible action of the catechol estrogen may also involve estrogen receptor-linked events.

Effects of 2-Hydroxyestrone on the Tumor Cell Growth of Estrogen Receptor-negatiue Human Breast Cancer
Cells-An additional line of evidence supporting the notion that the tumor cell growth-inhibitory action of 2-hydroxyestrone is brought about by estrogen receptor-mediated events stems from the results of the experiments (Fig. 5 ) indicating that estrogen receptor-negative human breast cancer cell lines, such as MDA-MB-231 and MDA-MB-330, grown under conditions similar to those for the MCF-7 cells, do not exhibit any growth-inhibitory response to 2-hydroxyestrone at a concentration of M (Fig. 5b) in the presence of catechol 0methyltransferase inhibitors. Under identical conditions the estrogen receptor-positive MCF-7 cell exhibited the growthinhibitory response to 2-hydroxyestrone as well as the growthstimulatory response to estradiol (Fig. 5a). Both of the estrogen receptor-negative cancer cell lines did not respond to estradiol even at as high as lo-' M (Fig. 56), confirming their estrogen-unresponsiveness of both cell lines. These results, taken together, suggest that the action of 2-hydroxyestrone on the MCF-7 cells is mediated by the estrogen receptorrelated events.
Effects of 16a-Hydroxy Metabolites of Estradiol on the Cell Growth of MCF-7"We examined 16a-hydroxyestrone and estriol, the estradiol metabolites formed in the 16a-hydroxylative pathway, for their ability to affect the growth of the MCF-7 cells. The results, shown in Fig. 6, indicate that these metabolites, unlike the 2-hydroxylated derivatives, promote the cell proliferation of the breast cancer cells a t concentra- . " -. , . " -. ,  tions in the range of to lo-@ M in the potency order of estradiol > 16a-hydroxyestrone > estriol. The agonistic (estrogenic) potency order of these estradiol metabolites observed in the human breast cancer MCF-7 cells is not representative of their relative affinity for the estrogen receptor since the very much more potent 16a-hydroxyestrone is a less effective ligand for the receptor than estriol (22).

DISCUSSION
The data presented indicate that 2-hydroxyestrone, the most ubiquitous catechol estrogen, possesses a growth-inhibitory activity on an estrogen-dependent human breast cancer cell line MCF-7. The evidence that this effect of the catechol estrogen is mediated through the estrogen-receptor is 2-fold. Firstly, the effect is antagonized by physiological levels of the primary estrogen estradiol in the range of lo-" to lo-' M (Fig.   3). Secondly, the growth-inhibitory effect of the catechol estrogen is not observed in the estrogen receptor-negative human breast cancer cell lines MDA-MB-231 and MDA-MB-330 (Fig. 5b) under conditions where the estrogen receptorpositive cell line MCF-7 responds to both the catechol estrogen and the primary estrogen (Fig. 5a). These results, together with the data on the in situ formation of the catechol estrogen from estradiol in human breast cancer tissues (23, 24) and others (a), suggest that the catechol estrogen may function as an endogenous antiestrogen, subserving a physiological mechanism by which the catechol estrogen normally exerts a negative regulatory control to oppose the agonist action of the primary estrogen.
Our data on the metabolism of the catechol estrogen in the tumor tissue indicate that the steroid is rapidly metabolized by 0-methylation by the action of catechol O-methyltransferase in the tissue to form a metabolically stable but inactive derivative (Fig. 1). 2-Methoxyestrone is generally regarded as a physiologically inert metabolite of the catechol estrogen although it has been shown to possess a hypocholesteremic activity (25,26); it is, however, devoid of the uterotropic activity (18), inactive as an effector of gonadotrophin and prolactin secretion (18), and has no affinity for the estrogen receptor ( 5 ) . It seems therefore that the initial failure to observe the effect of the catechol estrogen on the cell growth and proliferation of the MCF-7 cells can be ascribed to its rapid metabolism by the enzyme. When the tissue catechol 0-methyltransferase activity was diminished by the inclusion of a catechol 0-methyltransferase inhibitor, quinalizarin (19), the antiestrogenic activity of the catechol estrogen could be expressed. It has been reported that in human breast cancer tissues catechol 0-methyltransferase was elevated 3-to 20fold relative to the normal human mammary gland (23, 27). In line with these data is our present result indicating that this enzyme activity is also greatly elevated in the cultured human breast cancer MCF-7 cells. It seems therefore likely that catechol 0-methyltransferase plays a role in the overall expression of the "estrogenic" action of estradiol in many if not all estrogen target tissues where the primary estrogen is converted to the catechol estrogen or which are accessible to catechol estrogen produced elsewhere. The recent data by Hersey et al. (28) which indicates that the inhibition of uterine catechol 0-methyltransferase by an inhibitor, U-0521, potentiates the action of catechol estrogen on the synthesis of induced protein is another case in point albeit in this instance the agonist activity of the catechol estrogen was apparently enhanced. Although catechol 0-methyltransferase is known to be an ubiquitous enzyme its exact cellular and subcellular localization and control, especially in estrogen target tissues, has not been determined. Information on its localization and activity in relation to estrogen 2-hydroxylase could therefore add to our understanding of the role of the aromatic hydroxylation and the subsequent 0-methylation reactions in the expression of estrogen action. Our present data indicating that the enzyme plays an important role in the expression of the catechol estrogen action in an estrogen-responsive cancer cell line MCF-7 is indicative of a possible role for this enzyme in the regulation of hormonal actions of estrogens in normal target tissues.
The actions of 2-hydroxyestrone in the presence of the catechol 0-methyltransferase inhibitor on the MCF-7 cell cultures are similar to those exerted by the synthetic triphenylethylene estrogen antagonists such as tamoxifen. These substances also inhibit cellular proliferation in estrogen receptor-positive cells (21), an action which can be reversed at an early stage by estradiol but which also becomes irreversible upon prolonged exposure to these agents (21,29). It has been proposed that these antiestrogens act via estrogen receptorrelated nuclear events at the level of chromatin or at the level of nuclear receptor "processing" (30). The 2-hydroxyestrogens also mimic the synthetic nonsteroidal antiestrogens in their in vivo mode of action; they behave as mixed estrogen agonists-antagonists and like the synthetics are capable of acting as estrogens under certain circumstances and as antiestrogens under others (2)(3)(4).
The irreversible nature of the catechol estrogen inhibition of MCF-7 cell growth after 4 days of exposure to catechol estrogen (Fig. 4) may shed some light on the mechanism by which the catechol estrogen exerts its effect on the tumor cell growth and proliferation. It is evident that the reversible action of the catechol estrogen is mediated through some estrogen receptor-related event since the primary estrogen can reverse the action of the catechol estrogen when present simultaneously on day 1 (Fig. 3). In the absence of the estrogen, however, it is considered that the catechol estrogen affects irreversibly a component involved in the estrogen receptor-mediated event, rendering it unresponsive to the subsequent stimulation of the estrogen. The Z-hydroxyestrogens readily undergo oxidation to form powerful semiquinonelike electrophiles which can bind covalently to the nucleophilic sites of biological macromolecules (31,32). It is, therefore, possible that the irreversible action of 2-hydroxyestrone on the cell growth of MCF-7 is the result of such covalent bonding to cellular macromolecules including presumably the estrogen receptor itself. The same mechanism for the irreversibility of action can be invoked in the case of the triphenylethylene estrogen antagonists. These substances are subject to oxidative metabolism (33) leading to phenolic derivatives with the potential for similar covalent binding either directly or via further transformed derivatives. In this context it is interesting that Reddel et al. (34) have reported the reversal of the inhibitory action of 4-hydroxytamoxifen on MCF-7 cell growth requires many fold greater concentrations of estradiol than in the case of tamoxifen. It is also possible that the action of the catechol estrogen may also be mediated through its binding to an antiestrogen binding site reportedly found in the cytosol of many estrogen-responsive cells (35) including the MCF-7 human breast cancer cells (36), which is distinct from the classical estrogen receptor (35,37). Until the mechanism of catechol estrogen on antiestrogen action on MCF-7 cells is more precisely known, it is not possible to define whether the noted similarities reflect common biological pathways or whether the catechol estrogen acts as the endogenous analogue of the triphenylethylene drugs.
The present data provide evidence that 2-hydroxyestrone, a natural and qualitatively significant estrogen, can inhibit the growth of the estrogen-dependent human breast tumor cell line MCF-7. This is accomplished at a concentration, which although not physiological, is orders of magnitude lower than any inhibitory concentration of the other estrogens. As noted previously the exposure to catechol O-methyltransferase and biological availability of endogenous intracellularly produced catechol estrogen may differ greatly in a quantitative sense from exogenously supplied 2-hydroxyestrone. The mechanism of this inhibition is unclear but strong evidence suggests the involvement of estrogen receptor. What is clear, however, is that the action of the catechol estrogen is inhibited by the extraordinarily high activity of catechol O-methyltransferase present in these tumor cells. It is tempting to speculate that this distortion in catechol 0-methyltransferase activity may have been one of the precipitatory events which led to the disruption of cellular control of estrogen expression and to eventual cell transformation. The clear estrogen agonist activity of the 16a-hydroxylated metabolites of estradiol which are unaffected by catechol 0-methyltransferase presence but which regulate its activity' emphasize further that the direction of estradiol metabolism either in situ or in the periphery can have a profound influence on the expression of its hormonal activity.