Identification and Androgen Regulation of Two Proteins Released by T47D Human Breast Cancer Cells1

[CANCER RESEARCH 47, 2787-2792, June 1, 1987] Identification and Androgen Regulation of Two Proteins Released by T47D Human Breast Cancer Cells1

D. Chalbos, D. Haagensen, T. Parish, and H. Rochefort2 1NSERM U148, UnitÃ©d'Endocrinologie Cellulaire et MolÃ©culaire,34100 Montpellier, France [D. C., H. R.]; Department of Surgery, The Jewish Hospital of St. Louis, Washington University Medical Center, St. Louis, Missouri 63110 [D. H.J; and Cambridge Research Laboratory, Cambridge, Massachusetts 02139 [T. P.]

ABSTRACT scribed several glycoproteins isolated from breast gross cystic Three [35S|methionine-labeled polypeptides released by T47D human disease fluids, three of which have been identified and charac terized. They include a progesterone-binding protein (M, breast cancer cells have been identified as corresponding to two proteins 24,000), the Zn-a2-glycoprotein (M, 44,000), and a glycoprotein previously described in breast gross cystic disease fluid. A M, 43,000 protein was immunoprecipitated by polyclonal antibodies to the 7,n-Â«2- with a molecular weight of approximately 15,000, termed GCDFP-153 (8). These proteins have been purified to homo glycoprotein. A M, 18,000 and a M, 13,000 polypeptide were both immunoprecipitated by four monoclonal antibodies directed against four geneity, and antibodies (8-10) and cDNA (11, 12) probes for separate epitopes of M, 15,000 gross cystic disease fluid protein, a major GCDFP-15 have been prepared. protein that is characteristic of apocrine gland secretions. In sodium Specific immunoprecipitation was used here to identify two dodecyl sulfate-polyacrylamide gel electrophoresis, purified M, 15,000 proteins secreted by T47D cells as the Zn-a2-glycoprotein and gross cystic disease fluid protein migrated at the level of the M, 18,000 the GCDFP-15 protein of breast gross cystic disease fluid. The protein of T47Dcells. androgen-induced increase in the accumulation of GCDFP-15 The proteins were regulated by androgens and progestins. In addition to a general stimulation of protein secretion, 5a-dihydrotestosterone was confirmed by immunoprecipitation of cellular proteins. (DHT) specifically increased, by 2- to 20-fold, the release of the M, The regulation of these two proteins by progestin and androgen is probably mediated by the androgen receptor, based on dose- 43,000 and M, 18,000 proteins into the medium. DHT also increased the cellular level of the M, 18,000 protein, as shown by immunoprecipitation response experiments and inhibition by flutamide. with a M, 15,000 gross cystic disease fluid protein antibody, which suggests a stimulation of protein synthesis. The progestin 17,21-dimethyl-19-nor-4,9-pregnadiene-3,20-dione also increased the release in MATERIALS AND METHODS the medium of the M, 43,000 and M, 18,000 proteins, but higher molar Cell Culture. Clone 11 (passage 57) is a subline of the T47D cell line concentrations were required than in the case of DHT. The induction of (13), which was obtained by the agar method (14) in the laboratory of these proteins by DHT and 17,21-dimethyl-19nor,4,9-pregnadiene-3,20- Dr. I. Keydar (15). Single cells were grown in monolayers and recloned dione was specifically inhibited by the antiandrogen flutamide which has by the same method. Several single isogenic cell sublines were obtained, no effect on other progestin-regulated proteins. This suggests an effect including clone 11, which retains the marker chromosome of the mediated by the androgen receptor. parental line and contains high levels of progesterone receptors, as well This is the first report on the identification of two proteins induced by as estrogen, androgen, and glucocorticoid receptors (16).4 T47D clone androgens in a human breast cancer cell line. These proteins should be 11 is routinely cultured in RPMI 1640 supplemented with 10% fetal useful in studying the role of androgens in human mammary tumors and calf serum and 0.6 Â¿tg/rnlbovine insulin (Sigma Chemical Co., St. their mechanism of action in cell culture. Louis, MO). The protocol used to test the effects of steroids on protein secretion INTRODUCTION has been described previously (2, 4). Briefly, cells were cultured with 10% (3 days) and then 3% (2 days) FCS-DCC to remove steroids. They Human mammary cancer cell lines have been found to con were then plated out in 0.8-cm-diameter microwells (15,000 cells/well) tain sex steroid receptors and to respond to sex steroid hor in medium plus 3% FCS-DCC. Two days later, steroids were added in mones with respect to specific protein synthesis and the control ethanol solution (final ethanol concentration, 0.2%) and solvent was of cell proliferation. Proteins induced by steroids in these in added to control cells. Flutamide was kindly donated by Schering vitro systems provide new potential markers of hormone re Laboratories (Bloomfield, NJ) and hormones were given by Roussel Uclaf (Romainville, France). The compounds were 99% pure as shown sponsiveness. These systems can also be used to study the by thin layer chromatography. After hormonal stimulation, cells were mechanisms by which steroid hormones modulate gene expres labeled for 6 h with 200 /iCi/ml [35S]methionine or [35S]cysteine (Ra- sion and cancer cell proliferation. In MCF7 cells, several estro diochemical Center, Amcrsham. England; specific activity, 600 Ci/ gen-regulated proteins have been reported (1,2). More recently, mmol) in 50 u\ of medium containing one-tenth the normal concentra progestin-regulated proteins have been found within cells (T47D tion of methionine or cysteine, respectively. For analysis of cellular or MCF7 cells) (3) or secreted by cells (T47D cells) (4). These proteins, cells were labeled for 4 h with [35S]methionine and washed proteins are defined according to their molecular weight under twice with phosphate-buffered saline. They were then dissolved in 50 ul of lysis buffer A, described by O'Farrell (17), containing 0.5% (w/v) denaturing conditions. sodium dodecyl sulfate. Cell lysis was completed by freezing and Among the progestin-regulated proteins secreted by T47D thawing 3 times (-20Â°C). cells, the M, 43,000, 22,000, and 18,000 proteins were also Immunoprecipitations. Radioactivity incorporated into proteins re found to be regulated by androgens (4, 5) and appear to be leased into the medium was determined by TCA precipitation of ali- related to the synandrogenic effect of progestins (6). quots of the media (5 ul); 60,000 cpm of TCA-precipitable material or Working separately, Haagensen and Mazoujian (7) have de- 'The common names and abbreviations used are: GCDFP-15, gross cystic Received1/3/86; revised10/20/86, 3/5/87; accepted3/10/87. disease fluid protein (M, 15,000); R5020 or promegestone, 17,21-dimethyl-19- The costs of publicationof this article were defrayedin part by the payment nor-4,9-pregnadiene-3,20-dione; DHT, Sa-dihydrotestosterone, 17j3-hydroxy-5o- of page charges. This article must therefore be hereby marked advertisementin androstane-3-one; RU486, 17/3-hydroxy-11Â¡3-(4-dimethylaminophenyl-1)-17a- accordancewith 18U.S.C. Section 1734solelyto indicatethis fact. (prop-l-ynil)estra-4,9-dien-3-one; flutamide, a,a,a-trifluoro-2-methyl-4'-nitro-m- 1This work was funded by the Institut National de la SantÃ©etde la Recherche propionotoluidide; FCS-DCC, fetal calf serum treated with dextran-coated char MÃ©dicale,theUniversity of Montpellier I, and the Association pour la Recherche coal; SDS, sodium dodecyl sulfate; TCA, trichloroacetic acid; cDNA, comple sur le Cancer. mentary DNA. 2To whom requests for reprints should be addressed. 4 D. Chalbos and F. Vignon, unpublished experiments. 2787

60 ni of culture medium, as specified in the figure legends, were methionine-labeled proteins secreted by T47D cells (clone 11) incubated with antibodies for l h at 20"< in 100 HIMTris (pH 8), 100 has been reported previously (4). Cells treated for 5 days with IHMNaCl, 0.5% Nonidet P-40, 2 mM phenylmethylsulfonyl fluoride, DHT secreted more proteins than control cells, as seen by and 0.5% (w/v) bovine serum albumin (final volume, 100 (Â¿1).The analyzing the same volume of medium on SDS-polyacrylamide antibodies were 10 n\ of rabbit antiserum to Zn-Â«2-glycoprotein or nonimmune rabbit serum; 2 Â¿ilofmouse monoclonal antibodies A5, gel (Fig. 1, a and b). Three polypeptides, Mr 43,000, 18,000, D6, B1, or B15, directed against GCDFP-15 (10); or 4 /Â¿Iofnonimmune and 13,000, accumulated more specifically, according to anal mouse IgG (1 mg/ml). Protein A-Sepharose (100 pi; 6% w/v in the ysis of the same amounts of TCA-precipitable material (Fig. 1, same buffer) was then added. Tubes were agitated for 2 h at 4Â°Cand c and d) which have been shown to be induced by R5020 and then centrifuged at 10,000 x g for 2 min. The protein A-Sepharose gel DHT (4). The different molecular weights of the two smaller was washed 4 times with 1.2 ml of Tris-LiCl buffer (100 mM Tris-HCl, peptides (M, 18,000 and 13,000) were obtained using more pH 8.6-500 mM LiCl-0.5% Nonidet P-40) and then once with 50 mM suitable 15% or 20% polyacrylamide gels. Tris-HCl, pH 6.8. The antigen-immunoglobulin complexes were dis In an attempt to identify these androgen-regulated polypep sociated from the protein A-Sepharose gel by heating for 10 min at 90Â°Cin 60 pi of electrophoretic sample buffer. Tubes were centrifuged tides, several antibodies to proteins of similar molecular weight at 10,000 x g for 2 min and the supernatant was loaded onto SDS- were tested. Polyclonal antibodies to the Zn-Â«2-glycoprotein polyacrylamide gel. present in human breast gross cystic disease proteins immuno- For immunoprecipitation of cellular extracts (6 x 10' cpm TCA- precipitated the M, 43,000 protein when the cells were first precipitable material), the protocol was the same except for an addi treated by DHT or R5020. As shown in Fig. 2, the immuno tional protein A-Sepharose step before the addition of antibodies. precipitation was specific for this protein, and nonimmune Protein extracts were preincubated with 100 pi of protein A-Sepharose rabbit antiserum was inactive. Approximately 65% of the 35S (6% w/v) for 4 h at 4Â°Cwith agitation. The supernatants (10,000 x g labeled Mr 43,000 protein was precipitated at a serum dilution for 10 min) were then processed for immunoprecipitation as indicated above. Antiserum to human Zn-Â«2-glycoprotein was obtained from Nordic Immunological Laboratories (Tilburg, The Netherlands). Monoclonal antibodies to GCDFP-15 are described elsewhere (10). ab cd Protein Analysis. SDS-polyacrylamide gel electrophoresis were per formed as described previously (2). Gels were processed for fluorography and exposed for 8 days at -80Â°Cto Kodak-S film (Eastman Kodak, Rochester, NY) with preflashing to an absorbance of 0.2 to linearize the response of the film to radioactivity (18). The fluorograms were scanned using a Vernon densitometer (Paris, France) and the amounts of M, 43,000, 18,000, and 13,000 proteins were estimated from the traces. Results are expressed as percentages of total released protein or as the total amount. Total amounts were estimated by measuring the absorbance of the film in the M, 43,000, 18,000, and 13,000 regions and correcting it for the volume of medium analyzed. Results are expressed in arbitrary units of absorbance. These units are not the same in each experiment (duration of exposure, densitometer setting) and the absolute values obtained in different experiments cannot be com pared. Percentages were calculated taking the absorbance of total 43K released proteins in each track as 100%. Gels were stained for l h in Coomassie Brilliant Blue R (0.2% in 45% methanol-10% acetic acid) and then destained for 6 h using several changes of 45% methanol and 7% acetic acid. The molecular weights - 31 - of proteins were evaluated by their mobilities relative to standard proteins of known molecular weight from Bio-Rad (Touzart et Matig non, Vitry sur Seine, France). Binding Studies. Androgen-binding sites were measured on T47D clone 11 cells maintained for 8 days in 10% FCS-DCC. The cells were washed twice with phosphate-buffered saline and scraped from the -21.5- flasks using a rubber policeman. The suspension was centrifuged at 800 x g and the cell pellet was resuspended in 10 mM Tris (pH 7.4), l mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 20 mM sodium m molybdate, and 10% glycerol (w/v). The cells were homogenized in a Dounce homogenizer and a cytosol was obtained from the homogenate by centrifugation at 120,000 x g for 45 min at 2Â°C.Cytosol (500 pi) was incubated overnight at 0Â°Cwith increasing concentrations of [3H]- DHT (specific activity, =40 Ci/mmole) with or without a 100-fold excess of unlabeled DHT. C DHT Specific binding was determined by dextran-coated charcoal assays Fig. 1. Effect of DHT on the proteins released into medium. T47D cells were (1% charcoal-0.1% dextran) for 30 min at 0Â°C.The radioactivity of the cultured in the presence of 10% FCS-DCC for 3 days and then 3% FCS-DCC for 2 days. They were plated in 96-microwell dishes ( 15,000 cells/well) in medium supernatant was counted in a scintillation mixture consisting of 1 ml plus 3% FCS-DCC. After 2 days, they were incubated for 5 days without hormone of ethanol, 5 ml of 0.3% PPO, and 0.01% POPOP in toluene. The (Q or with 1 nM DHT. Cells were then labeled for 6 h with [35S]methionine and saturable binding was obtained from the difference between total and the incorporation of radioactivity into proteins was determined by trichloroacetic acid precipitation of an aliquot, identical volumes of culture medium (a, b) or the nonspecific binding. same amounts of TCA-precipitable 35S-labeled material (c, d) were analyzed in SOS 12', polyacrylamide gel, as described in "Materials and Methods." The migrations of the M, 43,000,18,000, and 13,000 androgen-regulated polypeptides RESULTS (arrows) are shown along with those of the following standard proteins (in thousands) from Bio-Rad: phosphorylase b (92.5), bovine serum albumin (66), Immunoprecipitation of the M, 43,000 Protein by Antibodies inal Ininiin (45), carbonic anhydrase (31), and soybean trypsin inhibitor (21.5). to the Zn-a2-Glycoprotein. The effect of steroids on the [35S]- MW, molecular weight; K, thousands. 2788

Control D HT R5020 Control DHT DHT 123456789 1 2 3 4 5 6 7 9 10 11 12 13 14 MW ,10-3

â€”Â«SÃ•.- 18K Â«** i

31- 18K-.4

21.5- -18K T Ign A5 T Ign AS Ã„5 D6 D6 B1 B15 T A5 Â«-13K GP15 GP15 T Ign Igi T Ign Igi T Ign Igi Fig. 3. Immunoprecipitation of the secreted M, 18,000 and 13,000 proteins by monoclonal antibodies directed against GCDFP-15. The protocol was the Fig. 2. Immunoprecipitation of the secreted M, 42,000 protein by an antiserum same as in Fig. 2. T47D cells were either untreated (Tracks 1-3) or treated with I directed against the Zn-a2-glycoprotein. The protocol was as in Fig. 1. Cells were nM 5a-dihydrotestosterone (Tracks 4~I3). The same amounts of TCA-precipitable incubated for 5 days without hormone (Control) or with 1 nin DHT or RS020 35S-labeled secreted proteins (60,000 cpm) were analyzed before (T) or following and labeled with [35S]methionine. TCA-precipitable material (60,000 cpm) was immunoprecipitated (as described in "Materials and Methods") with nonimmune precipitation with the monoclonal antibodies AS, D6, Bl, and B15 directed against GCDFP-15 or with nonimmune mouse IgG (Ign). In Tracks 7 and 9, 5 rabbit serum (Ign) or serum directed against the Zn-a2-glycoprotein (Igi). Im- Mgof purified GCDFP-15 were added during incubation with antibodies. Tracks munoprecipitates and 60 Â¿ilofmedium containing total proteins (T) were analyzed 1-12, 15% SDS-polyacrylamide gel; Tracks 13 and 14, 20% SDS-polyacrylamide in 15% SDS-polyacrylamide gel. The molecular weights (MW) in thousands (K) gel. The molecular weights (MW) in thousands (K) were determined from the were determined from the migration, shown on the left of standard proteins from Bio-Rad: phosphorylase h (92.5), bovine serum albumin (66), ovalbumin (45), migration of standard proteins. The M, 14,400 protein is lysozyme. carbonic anhydrase (31), and soybean trypsin inhibitor (21.5). staining of 1:10. No production of the M, 43,000 protein was detected 1 2 MW in the control cells, which confirmed its regulation by DHT and K10-3 R5020. Immunoprecipitation of the M, 18,000 and 13,000 Proteins by 45- Monoclonal Antibodies to GCDFP-15. The 4 monoclonal anti bodies that recognize 4 separate epitopes of GCDFP-15 (10), which is another protein found in mammary cyst fluid, specifi cally immunoprecipitated the M, \ 8,000 and 13,000 polypep- 30- tides (Fig. 3). Eighty-five to 95% of the M, 13,000 and 18,000 polypeptides were immunoprecipitated by the 4 antibodies at a dilution of 2:100. In the control sample, the A/r 18,000 protein 21.5- was secreted in very small amounts (Track 3). Following DHT treatment, the two polypeptides (M, 18,000 and 13,000) were 14.4- secreted at much higher concentrations and were specifically immunoprecipitated. When the antibodies were first incubated with the pure antigen GCDFP-15, the 35S-labeIed M, 18,000 and 13,000 polypeptides were not immunoprecipitated, dem GP15 A5 onstrating the specificity of this interaction. To estimate the Fig. 4. Comigration of the M, 18,000 protein and GCDFP-15 in SDS- polyacrylamide gel electrophoresis. T<7D cells were steroid withdrawn, plated, molecular weight of these 2 peptides more accurately, we also and cultured with 1 nM DHT as described in Fig. 1. They were then labeled for analyzed them in 20% SDS-polyacrylamide gel (Tracks 13-14). 6 h with ("Sjmethionine (Tracks 3 and 4) or [35S]cysteine (Tracks 5 and 6). Total proteins in the medium (/') or immunoprecipitates obtained as described in Fig. The immunoprecipitate of the M, 18,000 protein indicated 2 with monoclonal antibody A5 were analyzed in 15% SDS-polyacrylamide gel. some heterogeneity with a smear of higher molecular weight. Standard proteins and 5 /ig (Track 1) or 10 Mg(Track 2) of purified GCDFP-15 Whether this was due to microheterogeneity in the peptide or (GP15) were loaded onto the same gel and revealed by staining with Coomassie oligosaccharide composition has not been determined. Brilliant Blue R. MW, molecular weight; K, thousands. To compare the GCDFP-15 purified from human gross cystic fluid and the M, 18,000 and 13,000 proteins of T47D cells, we Androgen and Progestin Stimulation of GCDFP-15 Synthesis. then comigrated these proteins on polyacrylamide gel (Fig. 4). The increased accumulation in the culture medium of the M, The GCDFP-15 migrated at the same level as the A/r 18,000 18,000 and 13,000 proteins after cells had been incubated with protein of T47D cells. When the gel was overloaded with R5020 or DHT may have been due to increased protein synthe GCDFP-15, a band migrating at M, 13,000 was also detected sis, an effect on protein processing, and/or increased protein (data not shown). However, the proportion of this band was secretion or shedding. In an attempt to specify the mechanism, smaller than in the medium of T47D cells. When T47D cell we analyzed the cellular proteins. No variation in the pattern proteins were labeled by [35S]cysteine, the same M, 18,000 and of total [35S]methionine-labeled cellular proteins was observed 13,000 proteins were immunoprecipitated, which is in agree in one-dimensional gel electrophoresis 5 days following treat ment with the cysteine content of GCDFP-15 (11). ment of cells with DHT or R5020 (Fig. 5, Tracks 1, 4, and 7). 2789

Control R 5020 =48 fmol/mg protein and a KÃ s 0.2 nM in agreement with a 1 2 3 789 DHT-androgen receptor interaction. To determine the nature of the receptor(s) responsible for the regulation of GCDFP-15 and the Zn-a2-glycoprotein in T47D cells, we have compared the effects of increasing DHT and R5020 concentrations and used specific antihormones. To quantify the specific hormone- regulated (Mr 43,000, 18,000, and 13,000) secreted proteins, the regulation of total secreted proteins by DHT and R5020 also had to be taken into account. Cells treated with DHT secreted more proteins than control cells (Fig. 1), whereas R5020 decreased this production, as reported previously (4). The general stimulation of secreted proteins was not due to an effect of steroids on cell proliferation since, under our experi mental conditions (cells at low confluency when steroids were added), none of the steroids significantly affected total DNA content (4). To account for this general effect and quantify the specifically affected M, 43,000,18,000, and 13,000 proteins we analyzed the same amount of TCA-precipitable material as in â€¢Â«-18K Fig. 1, c and d. Dose-response experiments with DHT and R5020 indicated that DHT was 20 to 100 times more effective 14.4 than R5020 in inducing the three polypeptides, whether ex pressed in absolute amounts (Fig. la) or in percentages of total T Ignlgi T Ignlgi T Ignlgi secreted proteins (Fig. Ib). At optimal concentrations, R5020 induced higher percentages of the M, 43,000, 18,000, and Fig. 5. Immunoprecipitation of M, 18,000 cellular protein by monoclonal antibodies directed against GCDFP-15. T47D cells were stripped of steroids and 13,000 proteins than DHT, because of its overall inhibitory incubated for 5 days with I nm R5020. 1 nM DHT, or no hormone. They were effect on most of the released proteins. However, DHT and then labeled with ["Sjmethionine for 4 h, the incorporation of ("Sjmethionine R5020 induced similar absolute amounts of the M, 43,000, into intracellular proteins being linear for the first 5 h. Cellular proteins were extracted as described in "Materials and Methods." The same amounts (6 x 10* 18,000, and 13,000 proteins. The M, 43,000 and 13,000 pro cpm) of TCA-precipitable cellular proteins were immunoprecipitated as described in "Materials and Methods" with the AS monoclonal antibody (Â¡gÃ¬)orwith teins were undetectable without hormonal stimulation, while the A/r 18,000 protein had a basal level of expression. Dexa- nonimmune mouse IgG (Ign). Immunoprecipitates and 120,000 cpm of TCA- precipitated total cellular proteins were analyzed in 15% SDS-polyacrylamide methasone had some effect at very high concentrations, whereas gel. T, total cellular proteins; M W, molecular weight: K, thousands. estradiol caused no induction of the three proteins. We then used hormone antagonists to determine whether R5020 acts via the androgen receptor or the progesterone receptor. RU486,

Â£ 100

2 4 Total â€¢ (nM) Fig. 6. Binding of ['I I|l >II I in cytosols of I <-I>clone 11 cells, a, saturation curve. T47D cytosol (protein concentration, 2.3 mg/ml) was incubated overnight at O'C with I'HIDHT (0.1 to 5 nM) in the presence (x) or absence (O) of unlabeled I /Â¿MDHT as described in "Materials and Methods." The specific binding (â€¢) was obtained by subtracting nonspecific binding (x) from total (O) binding, b, boundScatchard over plot unbound analysis ligand of specific (II l'i binding versus the representing specificity the bound ratio ligand of the (B). specifically

However, when a sufficient amount of TCA-precipitable mate rial (6 x 10*cpm/track) was immunoprecipitated with GCDFP- 15 monoclonal antibodies, the amounts of cellular M, 18,000 protein was markedly increased in the DHT- and R5020-treated samples compared to the control (Fig. 5). The microhetero- steroid concentration steroid concentration geneity of the M, 18,000 protein was again observed, the Fig. 7. Dose-response curve of the effect of steroids in the accumulation of percentage of higher molecular weight species being greater the M, 43,000, 18,000 and 13,000 proteins. T47D cells were steroid withdrawn than in culture medium. The M, 13,000 polypeptide was not and plated as described in Fig. 1. They were then stimulated for S days with the indicated concentrations of DHT (â€¢),RS020 (O), dexamethasone (Dex, A) or immunoprecipitated by GCDFP-15 antibodies in cell extracts. 170-estradiol (E2, A). After labeling with [35S)methionine, 5,000 cpm of TCA- Nature of the Receptors) Mediating the Effects of Androgens precipitable material from each medium were analyzed by 15% SDS-polyacryl and Progestins. T47D clone 11 cells contain 2500 fmol/mg amide gel electrophoresis. Gels were fluorographed and films were scanned using protein of progesterone receptors.4 We have verified that in our a Vernon densitometer. In a, the amounts of M, 43,000, 18,000, and 13,000 proteins were evaluated from absorbance in the M, 43,000, 18,000. and 13,000 conditions these cells also contained noticeable amount of regions corrected for the volume of medium analyzed and expressed in arbitrary units. In b, the percentages of the U, 43,000, 18,000, and 13,000 proteins were androgen receptors. Fig. 6 shows results of a saturation analysis calculated by taking the absorbance of total released proteins in each track as of ['H]DHT binding to cytosol indicating a concentration of 100%. 2790

a progestin and glucocorticoid antagonist, inhibited the effects The A/r 13,000 polypeptide may be a proteolytic product of of both DHT and R5020 (not shown). Since RU486 is also an the Mr 18,000 protein formed following its secretion, since it androgen antagonist, this result could not be used to discrimi was not detected in the cellular extract. However, several pro nate between androgen and progesterone receptors. By contrast, tease inhibitors (leupeptin, plicnylmctlnIsullonyl fluoride and the androgen antagonist flutamide is considered to act specifi trasylol) were unable to prevent its appearance in the medium. cally on androgen receptor-mediated responses. In previous The small difference in the molecular weight of GCDFP-15, experiments with these cells, flutamide was shown totally inef successively reported to be 15,000 (8) and 18,000 (the present fective in preventing R5020 induction of the M, 48,000 and report), may be explained by variations in the technique used 250,000 proteins, which are two progestin-specific responses to determine molecular weight, such as different concentrations (3-5). At concentrations known to inhibit androgen action (19), of acrylamide gel. We verified that authentic purified breast flutamide prevented the induction of the M, 43,000, 18,000, cystic disease fluid GCDFP-15 comigrated with the M, 18,000 and 13,000 polypeptides by R5020 and DHT (4) (Fig. 8, a and protein of T47D cancer cells. A molecular weight of 18,000 is Â¿>)whereasit had no effect on these responses when used alone also consistent with the amino acid sequence of GCDFP-15 (Fig. Ib). In the same experiments, it had no effect on the determined by cDNA clone analysis.5 accumulation of the progestin-regulated M, 48,000 protein (Fig. GCDFP-15 and the Zn-o2-glycoprotein are secreted by apo- 6a). We therefore suggest that R5020 induces these proteins crine glands (20) and these glands, like human sweat glands via the androgen receptor, but we cannot totally exclude addi and rabbit salivary glands (21), are known to be stimulated by tional mediation by the progesterone receptor which may be androgens. The Zn-a2-glycoprotein is a major protein of apo- inhibited by flutamide in the case of this specific gene product. crine and sweat gland secretions (22). Its concentration in breast gross cystic disease fluid is only 10 times higher than in plasma, where its level is approximately 90 Mg/ml (23). By contrast, the DISCUSSION GCDFP-15 concentration in mammary cysts is much higher Separate reports have described two types of proteins secreted than its plasma level (20 to 100 ng/ml) which is elevated in some metastatic breast cancers (7). Moreover, the GCDFP-15 by human mammary cells. In vitro, the M, 43,000, 18,000, and 13,000 proteins secreted by the T47D breast cancer cells have plasma concentration increases in response to androgen treat been shown to be regulated by androgens and progestins (4, 5). ment (halotestin) in some patients with metastatic breast car In vivo, several glycoproteins from human breast gross cystic cinoma (24), which supports the present results. disease fluids have been isolated and characterized (7). Using The effects of androgens on the growth of human breast antibodies to two of these proteins, we now found that the M, cancer are complex, and their mechanisms are not understood. 43,000 protein is immunologically related to the Zn-a2-glyco- Although androgens have been used in treating metastatic protein and we identified the M, 18,000 and 13,000 proteins as breast cancer, it has been proposed that a relative excess of the GCDFP-15 of breast cystic disease. circulating androgens is correlated with an increased risk of breast carcinoma (25, 26). In vitro, androgens stimulate the growth of the Shionogi mouse mammary cancer cell line (27) a MW Â«1CT3 and inhibit (28) or stimulate (29) that of human mammary cancer cells. The complexity of androgens is partly due to the fact that they can act directly or after aromatization into estrogens and 3 that their action can also be mediated by the estrogen receptor (29). Progestins may also act via glucocorticoid and/or andro gen receptors (6). Part of our ignorance may be due to the absence of specific androgen-regulated responses in cultured human mammary cells. The two g'oss cyst fluid proteins re leased by T47D cells are the first examples of such androgen- regulated proteins. The mechanism by which secreted GCDFP- 31- 15 is increased by androgen and R5020 is due, at least in part, to an increase in protein synthesis, as shown by immunoprecipitation of cellular extracts. This increase appears to follow an 215- increased accumulation of GCDFP-15 mRNA, as shown by *18K hybridization with a cloned cDNA probe in dot blot experi ments and reported by Kelton and Monahan (12) and confirmed â€¢W.4- Â»13K by us.6-7 However, other mechanisms cannot be excluded. The mechanism by which the M, 43,000 Zn-Â«2-protein is regulated log flutamide concentration could not be specified, since no cDNA probe was available. Fig. 8. Effect of flutamide on the hormone-induced accumulation of the M, Two types of receptor can potentially mediate the effect of 43,000, 18,000, and 13,000 polypeptides. In a, TÂ«7Dcellswere steroid withdrawn and plated as described in Fig. 1. Two days later, they were incubated for 5 days androgens, i.e., the androgen and the progesterone receptors, with their vehicle (O, 0.5 mi RS020 alone (R). or along with 500 DMflutamide both of which are present in T4?D cells (clone 11) and absent (R + fluÃ¬);the media were changed every 2 days. Following ["Sjmethionine in BT20 cells, in which no androgen-regulated proteins could labeling, 20 u\ of culture medium were analyzed in 15% SDS-polyacrylamide gel MW, molecular weight; K, thousands. In It, steroid-withdrawn T47D cells were be detected (4). Two series of experiments strongly suggest that incubated for 5 days with the indicated concentrations of flutamide alone (O) or these proteins are regulated by the androgen receptor and not in the presence of 0.5 nM 5o-dihydrotestosterone (â€¢)or R5020 (A). Following [35S]methionine labeling, 5,000 cpm of TCA-precipitable material from each the progesterone receptor: (a) DHT is more effective than medium were analyzed on a 15% SDS-polyacrylamide gel. The total amount of the M, 43,000, 18,000, and 13,000 proteins were evaluated from absorbance in 5J. E. Monahan and C. A. Kelton, personal communication. ' D. <'bulbos, unpublished results. the M, 43,000, 18,000, and 13,000 regions and was corrected according to the volume of medium analyzed. 7 H. Haagensen and W. Dilley, unpublished results. 2791

radio immunoassay for a major component protein. J. Nati. Cancer Inst., R5020. The concentration of R5020 needed to induce the 62: 239-247, 1979. accumulation of these proteins was 100 times higher than that 9. Haagensen, D. E. Biochemical relationship between gross cystic disease and of DHT, and the affinity of R5020 for the androgen receptor breast carcinoma. In: C. D. Haagensen, C. Bodian, D. E. Haagensen, Jr. (eds.), Breast Carcinoma Risk and Detection, pp. 300-338. Philadelphia: W. is 100 times lower than that of DHT (30); (b) the accumulation B. Saunders Co., 1981. of these proteins by R5020 was inhibited by the antiandrogen 10. Parish, T., Wu, Y. S., Rittershans, C., Davis, S., Haagensen, D., and Ip, S. flutamide. However, an effect mediated by the progesterone H. C. Generation of monoclonal antibodies reactive with glycoprotein marker of apocrine epithelium and metastatic breast carcinoma. Protides Biol. receptor, which is highly concentrated in T47D cells, cannot be Fluids, Proc. Colloq. in press, 1987. totally excluded, since flutamide may also inhibit some proges- 11. Miinalimi. J. E., and Kelton, C. A. The cDNA cloning of the 15,000 daltons tin-specific responses (but not the M, 48,000 and 250,000 human breast gross cystic disease fluid protein. Fed. Proc., 43:1419, 1984. 12. Kelton, C. A., and Monahan, J. E. Androgen induction of CDP mRNA, a proteins), and since it has been shown that two different steroid human breast tumor marker. In: D. Putt (ed.), Advances in Gene Technol receptors can trigger the same specific gene (31). In MCF7 cells, ogyâ€”Molecular Biology of the Endocrine System, ICSU Short Reports, Vol. 4, pp. 300-301. Cambridge, England: Cambridge University Press, which are from another metastatic human breast cancer cell 1986. line, no androgen-regulated proteins could be detected and 13. Keydar, I., Chen, L., Karby, S., Weiss, F. R., Delarea, J., Radu, M., Chaitcik, antibodies to GCDFP-15 were unable to detect the correspond S., and Brenner, H. J. Establishment and characterization of a cell line of human breast carcinoma origin. Eur. J. Cancer, 15:659-670, 1979. ing protein.6 This may be due to a higher androgen responsive 14. Macpherson, I., and Montagnier, L. Agar suspension culture for the selective ness in T47D cells or may reflect a different type of breast assay of cells transformed by polyoma virus. Virology, 23: 291-294, 1964. 15. Keydar, I., Ohno, T., Nayak, R., Sweet, R., Simoni, F., Weiss, R. F., Karby, carcinoma. In contrast with the MCF7 and ZR75 cell lines, T47D S., Mesa-Tejada, R., and Spiegelman, S. Properties of retrovirus like panicles cells produce less M, 52,000 protein (32) and no pS2 mRNA produced by human breast carcinoma cell line: immunological relationship (33) but their production of two proteins (GCDFP-15 and Zn- with mouse mammary tumor virus particles. Proc. Nati. Acad. Sci. USA, 81: 4188-4192, 1984. a2-glycoprotein) also found in breast gross cystic fluid suggests 16. Horwitz, K. B., Zava, D. T., Thilager, A. K., Jensen, E. M., and McGuire, these cells have some apocrine characteristics (7). W. L. Steroid receptor analyses of nine human breast cancer cell lines. Cancer Res., 38:2434-2437, 1978. To conclude, we identified two proteins specifically induced 17. O'Farrell, P. H. High resolution on two dimensional electrophoresis of by androgens in the T47D human breast cancer cell line. The proteins. J. Biol. Chem., 250: 4007-4021, 1975. 18. Laskey, R. A., and Mills, A. D. Quantitative film detection of 3H and "C in system may be of considerable interest in studying the regula polyacrylamide gels by fluorography. Eur. J. Biochem., 56: 335-341, 1975. tion of gene expression by androgens in cultured human cells 19. Liao, S., Unwell. D. K., Chang, T. M. Action of a nonsteroidal antiandrogen, and should also prove to be useful in cellular pharmacology for flutamide, on the receptor binding and nuclear retention of 5a dihydrotestosterone in rat vetrai prostate. Endocrinology, 94:1205-1209, 1974. testing new drugs with potential androgenic or antiandrogenic 20. Mazoujian, G., Pinkus, G., Haagensen, D. E., Jr. Extramammary Paget's activity (34). diseaseâ€”evidence for an apocrine origin. Am. J. I'at hoi.. 8:43-50, 1982. 21. Wales, N. A. M., Ebling, F. J. The control of the apocrine glands of the rabbit by steroid hormones. J. Endocrinol., 51: 763-770, 1971. 22. Brugi, W. and Schmid, K. Preparation and properties of Zn-a2-glycoprotein ACKNOWLEDGMENTS of normal human plasma. J. Biol. Chem., 236: 1066-1074, 1961. 23. Poortmans, J. R., and Schmid, K. The level of Zn a2 glycoprotein in normal We are grateful to M. EgÃ¨aand E. Barrio for secretarial assistance human body fluids and kidney extract. J. Lab. Clin. Med., 71: 807-811, and to F. Depadova and C. Rougeot for their excellent technical 1968. assistance. We thank Dr. I. Keydar for providing T47D cells, Roussel 24. Dilley, W. G., Leight, G. S., Jr., Silva, J. S., Ammirata, S., Haagensen, D. E., and Wells, S. A., Jr. Androgen stimulation of gross cystic disease fluid Uclaf for steroids and KI 4X6. and Schering Laboratories for flutamide. protein and carcino-embryonic antigen in patients with metastatic breast carcinoma. J. Nati. Cancer Inst., 70:69-74, 1983. 25. Grattarola. R. Androgens in breast cancer I. Atypical endometrial hyperplasia REFERENCES and breast cancer in married premenopausal women. Am. J. Obstet. Gynecol., 116:423-428, 1973. 26. Adams, J. B., Archibald, L., Clarke, C. Adrenal dehydroepiandrosterone and 1. Edwards, D. P., Adams, D. J., Savage, N., and McGuire, W. L. Estrogen human mammary cancer. Cancer Res., 38:4036-4040, 1978. induced synthesis of specific proteins in human breast cancer cells. Biochem. 27. Sutherland, D. J. A., Robins, E. C, Meakin, J. W. Effects of androgens on Biophys. Res. Commun., 93: 804-812, 1980. Shionogi carcinoma 115 cells in vitro. J. Nati. Cancer Inst., 52.-37-48,1974. 2. Westley, B., and Rochefort, H. A secreted glycoprotein induced by estrogen 28. Maclndoe, J. H. and Etre, L. A. An antiestrogenic action of androgens in in human breast cancer cell lines. Cell, 20: 353-362, 1980. human breast cancer cells. J. Clin. Endocrinol. Metab., 53: 836-842, 1981. 3. Chalbos, D., Rochefort, H. A 250-kilodalton cellular protein is induced by progestins in two human breast cancer cell lines MCI-'- and T47D. Biochem. 29. Rochefort, II., Garcia, M. The estrogenic and antiestrogenic activities of androgens in female target tissues. Pharmacol. Ther., 23:193-216, 1984. Biophys. Res. Commun., 121:421-427, 1984. 30. Ojasoo, T., and Raynaud, J. P. Unique steroid congeners for receptor studies. 4. Chalbos, D., Rochefort, H. Dual effects of the progestin RS020 on proteins Cancer Res., 38:4186-4198, 1978. released by the T47D human breast cancer cells. J. Biol. Chem., 259: 1231- 31. Scheidereit, C, Krauter, P., Von der Ahe, D., Janich, S., Rabenau, O., Cato, 1238, 1984. A. C. B., Suske, G., Westphal, H. M., and Beato, M. Mechanism of gene 5. Rochefort, H., Chalbos, D. Prosgestin-specific markers in human cell lines: regulation by steroid hormones. J. Steroid Biochem., 24: 19-24, 1986. biological and pharmacological applications. Mol. Cell. Endocrino!., 36: 3- 32. Garcia, M., Contesso, G., Duplay, H., Cavailles, V., Derocq, D., Delante, J. 10, 1984. C., Krebs, B., Sancho-Gamier, H., Richer, G., Domergue, J., Namer, M., 6. Janne, O. A., Bardin, C. W. Steroid receptors and hormone action: physio and Rochefort, H. Immunohistochemical distribution of the 52K protein in logical and synthetic androgens and progestins can mediate inappropriate mammary tumors: a marker associated with cell proliferation rather than biological effects. Pharmacol. Rev., 36: 355-425, 1984. with hormone responsiveness. J. Steroid Biochem., in press, 1987. 7. Haagensen, D. E., Mazoujian, G. Relationship of glycoproteins in human 33. Brown, A., Jeltsch, J. M., Kobens, M., and ChambÃ³n, P. Activation of pS2 breast cystic disease fluid to breast carcinoma. In: A. Angeli, H. L. Bradlow, gene transcription is a primary response to estrogen in the human breast and L. Dogliotti (cils.). Endocrinology of Cystic Breast Disease, pp. 149- cancer cell line MCF7. Proc. Nati. Acad. Sci. USA, 81:6344-6348,1984. 169. New York: Raven Press, 1983. 34. Rochefort, H. Sex steroid regulated proteins in human breast cancer cell 8. Haagensen, D. E., Mazoujian, G., Diley, W. G., Pederson, C. E., Kister, S. lines: their interest for pharmacology. In: S. Garattini (ed.), New Tests for J., Wells, S. A., Jr. Breast gross cystic disease fluid analysis I. Isolation and New Drugs, pp. 183-196. Milan: Wichtig Editore, 1987.

2792

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1987 American Association for Cancer Research. Identification and Androgen Regulation of Two Proteins Released by T 47D Human Breast Cancer Cells

D. Chalbos, D. Haagensen, T. Parish, et al.

Cancer Res 1987;47:2787-2792.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/47/11/2787

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/47/11/2787. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.