Differences in oestrogen receptors in malignant and normal breast tissue as identified by the binding of a new synthetic progestogen

Oestrogen receptor protein (ER) was detected in 9 of 11 samples of malignant breast tissue and 8 of 9 samples of normal breast tissue. Levels of cytosolic ER (ERJ) in malignant breast were 21-1102 fmol mgsoluble protein (Kd1.8x 10-9-3.1 x 10-8moll-1) and those of nucleosolic ER (ERj), 13-526fmolmg-' soluble protein (Kd 2.1 x 10-9-1.4 x 10 8mol '1). In normal breast tissue ERC levels were 33-640fmol mgsoluble protein (Kd 1.3 x 10-L-3.2 x 0mol -'), ER. was detected in only 2 samples, 8 and 87 fmolmg' soluble protein with Kd 3.2 x 109 and 1.4 x 1091 mol -' respectively. 17a-ethinyl13,-ethyl-17,-hydroxy-4,15gonadiene-3-one (gestodene), a new synthetic progestogen displaced 3H-oestradiol (3H-E2) from both ERc and ERn in malignant tissue but not in normal breast, or these receptors from endometrial tissue. In

gonadiene-3-one (gestodene), a new synthetic progestogen displaced 3H-oestradiol (3H-E2) from both ERc and ERn in malignant tissue but not in normal breast, or these receptors from endometrial tissue. In competition studies gestodehe was 3 times more effective in displacing 3H-E2 from ERC and ERn in malignant breast tissue than the natural ligand. Quantitation of ER by gestodene were ERr, 12-1134 fmol gestodene bound mgsoluble protein (Kd 1 x 10-9-8.1 x 10-9 mol l-1); ER., 17-531 fmol gestodene bound mg ' soluble protein (Kd 1.6 x O1-9.1 x 10 8moll-'). L-13-ethyl-17a-ethinyl, 17,B-hydroxy-gonen-3-one (levonorgestrel) showed no binding to ER in malignant breast, normal breast or endometrial tissue. In circulation both gestodene and levonorgestrel displaced E2 from sex hormone binding globulin more than any of the androgens tested. These results suggest that gestodene is a progestogen with oestrogenic and/or antioestrogenic properties and provide strong evidence for differences in ER from malignant and normal breast tissue.
Oestrogen receptor (ER) in human carcinoma of the breast is the most widely studied steroid receptor. ER has been extensively purified and characterised (Jensen et al., 1982) and perhaps together with the glucocorticoid receptor more is known about its physicochemical forms and characteristics than any other receptor (McGuire et al., 1978;Grody et al., 1982). However, many points of contention still remain (King & Green, 1984;Welshons et al., 1984;Szego & Pietra, 1985).
Recently, Jasper et al. (1985) reported different physicochemical forms of ER in rat uterus and pituitary gland based on the hypothesis of a monomer-dimer relationship, and Brown et al. (1984) found that the E2 dependent pS2 gene was expressed in the MCF-7 cell line and malignant breast samples but not in normal breast or ER negative cell lines. Similarly, there have been many reports on multiple receptor forms in tissues from animals of different ages and endocrine status (Jasper et al., 1985), on the sedimentation behaviour of molybdate stabilised, non-activated ER, on ER bound to E2 or to antioestrogens and salt or heat-activated receptor (Katzenellenbogen et al., 1978(Katzenellenbogen et al., , 1981McGuire et al., 1978;Grody et al., 1982 andKeen et al., 1984). To date, however, in no organ in any species have differences in binding of a particular steroid metabolite or analogue by ER been reported in a malignant tissue as compared to the normal tissue of the same organ.
Here we report the significant binding of a synthetic progestogen to ER in human malignant breast tissue and its total lack of binding to ER in normal breast tissue or to ER in endometrium. This is all the more surpising because the 'down regulation' of ER by progestogens has been reported (Katzenellenbogen, 1980) but the binding of this class of hormones to ER has not. As this new progestogen may form part of an oral contraceptive preparation, its binding in circulation and to specific proteins in tissues requires investigation and is reported here.
Breast tissue was obtained at operation; malignant breast samples were confirmed histologically and normal breast samples were obtained either from surrounding tissue (samples 1-6, Table I) which was histologically confirmed as normal or from operative breast reduction (samples 7-9, Table II) and were stored in liquid nitrogen until assayed. Estimation of ER Cytosol and nucleosol fractions were prepared as previously described (Greenway et al., 1981;Iqbal et al., 1983) and the original tissue weight:volume buffer was 1:20 in the incubates. Tissue samples were manipulated below 4°C and were homogenised in TED buffer (10mMTris, 1.5mMEDTA and 1.5mM dithiothreitol, pH 7.4) using an Ultra-Turrax homogeniser before centrifugAtion at 160,000g for 1h, the resulting supernatant was used as cytosol.
The remaining pellet was washed with TES buffer (10 mm Tris, 1 mm EDTA and 250 mm sucrose, pH 7.4), centrifuged at 800g for 10min and the supernatant discarded. The pellet was then homogenised in TSMK buffer (10mmTris, 250mM sucrose, 5mM MgCl2 and 25 mm KCI, pH 7.5), centrifuged at 800g for 10 min and the supernatant discarded. The pellet was washed twice with TSMK bufffer and finally suspended in TKED buffer (TED buffer containing 0.5 M KCl). The suspension was kept at 4°C for 1 h to extract nuclear receptor and then centrifuged at 15,000g for 30 min, the supernatant being retained as nucleosol.
ER was measured using the two-tier column microassay employing Cibacron Blue 3GA-Sepharose 6B for the affinity immobilisation of the receptor and the steroid bound to it (Iqbal et al., 1985). Aliquots (0.4ml) of cytosol or nucleosol were incubated with a constant amount (6,000c.p.m.) of 3H-E2 and increasing amounts of radioinert E2 (0-18.4 pmol). Cytosols were assayed after 2 h of incubation and nucleosols were assayed after 18h of incubation. In the assay 0.1 ml aliquots of these incubates were applied to the microassay columns in duplicate. The column comprises a glass microcolumn fitted with a cellulose acetate plug, the upper layer consisting of 0.5ml of Cibacron Blue 3GA-Sepharose 6B and the lower layer 1 ml of Sephadex LH-20. The columns were eluted with either 2.7 ml cytosol assay buffer (10 mm Tris, 1.5 mm EDTA, pH 7.4) for the cytosolic ER or 2.7 ml nucleosol assay buffer (10 mm Tris, 250mM sucrose, 5mM MgCl2, 25mmKCL, pH7.4). After cutting the columns at the interface of the two gels, the radioactivity in the Blue gel fraction was determined. Eleven samples of malignant breast, 9 of normal breast and 3 of endometrium were assayed.
Competition studies for ER In samples of normal breast (n = 9), malignant breast (n=ll) and endometrium (n=3), cytosolic and nucleosolic preparations were made as above. Aliquots (0.4ml) of these were incubated (cytosols for 2 h and nucleosols for 18 h) with a constant amount of 3H-E2 (6,000 c.p.m.) and varying amounts (0-16 pmol) of either gestodene or levonorgestrel, and also using varying amounts (0-18.4pmol) of radioinert E2. Displacement of the 3H-E2 was studied with the microassay.
Determination of ER employing gestodene as the binding ligand The assay for ER, and ER. were carried out on all samples of malignant, normal and endometrial tissues exactly as the ER assay described above except that a constant amount of 3H-gestodene (6,500c.p.m.) and varying amounts (0-16pmol) of radioinert gestodene were employed as the binding ligands. To prevent artefactual measurement of other receptors, large excesses (100 x fold) of radioinert DHT, progesterone and cortisol were included in the incubates to saturate androgen receptor, progesterone receptor and glucocorticosteroid receptor respectively. Protein concentrations were measured using the BCA protein assay system obtained from Pierce UK Ltd. employing human serum albumin as standard.
(ii) Displacement from corticosteroid binding globulin (CBG) The two-tier columns were prepared as above except that the Sephadex LH-20 gel in the lower tier was replaced by 1 ml of Sephadex G-25. The rest of the experimental conditions were as in (i) above. Displacement of 3H-progesterone and 3H-cortisol was studied by using varying concentrations (0-180 pmol) of their respective radioinert ligands and in a parallel series of experiments .displacement of 3H-progesterone and 3H-cortisol was studied by using varying amounts of radioinert gestodene or levonorgestrel (0-165 pmol).

Results
Of the 11 samples of the carcinoma of the breast assayed 9 were ER, positive (1-6, 8, 10 and 11, Table I) and 8 were ERC and ER. positive (1-3, 5, 8, 10 and 11) employing E2 as the binding ligand (Table I, Figure 1). In normal breast obtained from the corresponding carcinoma of the breast tissue (1-6, Table I) ERC was positive in samples 1, 3-6 (Table II); ERC in samples 7-9 was also positive (Table II, Figure 2). ERn in normal breast was detected only in two samples (3 and 8, Table II).
Competition studies showed that gestodene displaced 3H-E2 from ER, and ER, in malignant breast tissue samples by a factor 3 times greater than the natural ligand (Figure 3), 50% displacement of E2 being caused by 7 x lO-4nmol E2 added as compared to 2.2 x 10-4 nmol gestodene added. No displacement of 3H-E2 was observed by gestodene in any sample of normal breast nor the 3 samples of endometrium either from ER, or ER,.
Levonorgestrel showed no displacement of 3H-E2 from ERC or ER, obtained from any sample of either malignant or normal breast (Figures 1 and 2) or endometrial tissue.
When the other receptors had been saturated with excess of their natural ligands ER measured by gestodene showed values comparable to those obtained when the natural ligand had been employed (Table I) with approximately similar Kd values. ERC or ERn observed to be negative using E2 as the ligand were also found to be negative using gestodene as the ligand (Table I) Fifty percent displacement of 3H-DHT, 3Htestosterone, 3H-E2, 3H-5a-androstane-3a, 17f-diol and 3H-5a-androstane-3p, 17f,-diol from SHBG in circulation by gestodene were achieved by concentrations 560%, 235%, 65%, 75% and 80% of their natural ligand respectively. Similar studies with levonorgestrel showed that gestodene was -20% more effective in these displacements than was the former analogue. No displacement of 3Hprogesterone or 3H-cortisol from CBG in circulation was caused by either gestodene or levonorgestrel.

Discussion
Gestodene is structurally closely related to levonorgestrel and its optical isomer d-norgestrel. The 0.10 latter compound has been shown to displace sexsteroids from SHBG in circulation (Victor et al., m 1u 1976). While it is surprising that a progestogen £ \t, should not displace progesterone from CBG in 0.05 n A x * \ X serum, the displacement of sex-steroids from SHBG may be related to the close structural similarities of hydroxy-gonen-3-one (levonorgestrel). D In relation to the binding of gestodene to ER in malignant breast tissue, the findings are much more unexpected and have far reaching implications. This study demonstrates for the first time that a steroidal compound exhibits binding to ER from malignant breast but not to that from normal breast. This indicates a structural difference between the two receptors. In competition studies the results show that gestodene can displace 3H-E2 by about 3-fold as compared to the natural ligand, however, when ER from malignant breast tissue is measured using this synthetic steroid there is little difference in the total steroid bound or in the dissociation constant suggesting that gestodene prevents binding of E2 not only by competing for the binding site on ER but perhaps also by interfering with the formation of E2-ER complex.
The high positive correlation between ERn measured with E2 and that measured with gestodene supports this hypothesis.
The evidence presented here suggests that the binding site and therefore the structure of ER extracted from malignant breast is different from that in either normal breast of endometrium and that gestodene may be of clinical value as an antioestrogen in the management of malignant breast disease.
The financial assistance of Schering Chemicals (UK.) Ltd. is gratefully acknowledged.