Breast Cancer Research and Treatment 23: 77-86, 1992. © 1992 KluwerAcademic Publishers. Printed in the Netherlands.
Report
Secretion of breast gross cystic disease fluid proteins by T47D breast cancer cells in culture - modulation by steroid hormones
Darrow E. Haagensen, 1Peter Stewart, William G. Dilley and Samuel A. Wells
1Department of Surgery, Methodist Hospital, 7500 Timberlake Way, Sacramento, California, and Department of Surgery, Washington University, St. Louis, Missouri, 63110, USA Key words': hormone modulation, T47D cells, GCDFP-15, GCDFP-24, GCDFP-44 Summary The effect of steroid hormones on modulating the secretion rates of three human breast gross cystic disease fluid proteins (GCDFP-15, GCDFP-24, and GCDFP-44) by T47D breast carcinoma cells in tissue culture was evaluated. Androgens (dihydrotestosterone or fluoxymesterone) were capable of stimulating the secretion rates for all three GCDFIrs while showing a minimal trend toward slowing the growth rate of T47D cells. This is the first study which shows that androgens can specifically stimulate all three of the major breast GCDFP's concomitantly. Progesterone, and three synthetic progestins, all showed inhibition of the growth rate of T47D cells while causing enhancement of the secretion of GCDFP-15 and GCDFP-44, and only minimal effect on the secretion rate of GCDFP-24. Estradiol was essentially neutral to the growth rate of the T47D cells in our test system. Estradiol did cause a mild enhancement of GCDFP-44 secretion rate, with no appreciable effect on GCDFP-15 or GCDFP-24 secretion rates. These findings suggest that an androgenic stimulus may be involved in the secretion of GCDFFs associated with breast gross cystic disease.
Introduction Human breast gross cystic disease is the most common benign breast disease [1]. It is predominately a premenopausal disease with a peak incidence occurring between ages 40 and 50 years. The cause of breast cystic disease remains unidentified. Controversy exists with regard to its relationship to breast carcinoma [2]. Several studies have shown a statistical relationship between the occurrence of breast gross cystic disease and the development of breast carcinoma [1-7]. The morphological feature common to breast cyst development is the occurrence of metaplastic apocrine breast epithelial cells lining the wall of developing breast cysts [8-9]. Bio-
chemical analysis of the fluid secreted into breast cysts has revealed unique steroid [10-13] and protein [14-21] profiles. Three major proteins are found in breast gross cystic disease fluid and they account for approximately 70% of the total protein content of the fluid [15]. These three proteins have been isolated, characterized, and termed GCDFP-15, GCDFP-24, and GCDFP-44 for gross cystic disease fluid proteins of 15,000, 24,000, and 44,000 monomer molecular size on SDS acrylamide gel electrophoresis [15, 16, 20]. They are present in breast cyst fluid at very high concentrations averaging 3mg/ml for GCDFP-15, 30 mg/ml for GCDFP-24, and 1mg/ml for GCDFP-44 [16, 22, 23].
Address for offprints: Darrow E. Haagensen, The Methodist Hospital, 8120 Timberlake Ave., Suite 112, Sacramento, California 95823, USA
78
DE Haagensen et al.
The amino acid sequence for GCDFP-15 has been determined [16] and the c-DNA has been cloned [24-25]. The genomic gene is located on chromosome 7 [25]. This gene sequence has also been termed by Myal et al. as the prolactin inducible protein [25]. The GCDFP-24 has recently been shown to have an aminoacid sequence identical to apolipoprotein D [19]. Antiserum which recognize GCDFP-24 show immunological identity for recognition of apolipoprotein D [26]. The GCDFP-44 protein has been shown to be immunologically identical with human serum Zn-alpha-2-glycoprotein [15, 16]. Immunohistochemical staining for GCDFP-15 has shown that it is restricted in normal tissues to apocrine glandular systems and to tissues which express apocrine glandular features [16, 27]. The GCDFP-24 has been identified in a number of normal tissues with steroid metabolizing capability including adrenal cortex and ovarian corpus luteum [28]. The GCDFP-44 is present in all normal apocrine glandular systems as well as in benign and malignant prostatic tissues [29-31]. The presence of GCDFP-15, GCDFP-24 and GCDFP-44 has also been evaluated in breast carcinoma tissues by immunohistochemical techniques [27, 28, 30, 32, 33]. The proteins are each expressed by a proportion of breast carcinomas; however, the concomitant presence of all three proteins in breast carcinoma tissues occurs in only approximately 25% of breast carcinomas [28]. The secretion of breast gross cystic disease fluid proteins by breast carcinoma cells in tissue culture gives a tool to investigate the stimuli which induce or inhibit the secretion of these proteins. This paper reports on our findings of the influence of various classes of steroid hormones on the concomitant secretion of GCDFP-15, GCDFP-24 and GCDFP-44 by T47D cells.
Materials and methods Chemicals
Tissue culture media (RPMI-1640/I-MEM; a 50/
50 mixture) and a cellharvesting solution of 0.05% trypsin containing 0.02% ethylenediamine tetraacetic acid (EDTA) were obtained from the Tissue Culture Center at Washington University, St. Louis, MO. Bovine calf serum (BCS) was obtained from GIBCO Laboratories, Grand Island, Maine. The BCS was charcoal-extracted prior to use [34]. Estradiol (E2), dihydrotestosterone (DHT), fluoxymesterone (HAL), progesterone (PROG), medroxyprogesterone acetate (MPA), norethindrone (NOR), megestrol acetate (MEG), gentamicin, deoxyribonucleic acid (DNA), Type III, from salmon testis; Triton X-100; bovine albumin, crystallized; and sodium azide were obtained from Sigma Chemical Company, St. Louis, MO. Casodex (ICI-176334) was a gift from Dr. B.J.A. Furr, Imperial Chemical Industries, Cheshire, England. Anandron (RU23908) and Mifepristone (RU486) were a gift from Dr. D. Martini, Roussel Uclaf, Romainville, France. Tamoxifen (TAM) was purchased from Stuart Pharmaceuticals, Wilmington, DE. 3,5-Diaminobenzoic acid dihydrochloride (DABA) was purchased from Aldrich Chemical Company, Milwaukee, WI. Hydrochloric acid (HC1), ammonium hydroxide (NH4OH), glass assay tubes, and polypropylene microcentrifuge tubes and caps were from Fisher Scientific, Pittsburgh, PA. Kynar was from Pennwalt Company, Philadelphia, PA. Goat anti-rabbit immunoglobin was from Pel-Freez, Rogers, AR.
Maintenance o f cell culture
The T47D cell line [35] was obtained from the American Type Culture Collection in its 98th passage. It was routinely grown in RPMI-1640/IMEM media containing phenol red and supplemented with 3% BCS-charcoal treated, and with gentamicin (40mg/L). The T47D cells were subcultured weekly in Costar, Cambridge, MA, 75 cm 2 flasks containing 45 ml of media. The cell cultures were maintained at 37°C in a 5% CO2 incubator.
Modulation of GCDFP's in T47D cells Cell modulation experiments" Maintenance cultures of T47D cells were harvested by addition of 15 ml of trypsin~DTA solution. The cells were incubated until they loosened, and the reaction was stopped by addition of 10ml of BCS. The cell suspension was pelleted by centrifugation, the supernatant decanted, and the cells resuspended in maintenance culture medium. The cells were counted by hemocytometer; then approximately 4 x 105 cells per well were subcultured into 6 well dishes (Coming, Ithaca, N.Y., model 25810 culture dishes). The medium (3 ml per well) was changed after three days in culture and new medium added which contained the test steroids, each at a final concentration of 10nM. All steroids were diluted and added to the culture media in ethanol solution with the final ethanol concentration being 0.1%. The control culture medium also had added to it ethanol to a 0.1% concentration. Each experiment was set up with 5 replicate test dishes. In each dish two wells were utilized for control wells and four wells for steroid modulation. Each test dish had the medium removed from each well at two day intervals and saved for analysis of GCDFP-15, GCDFP-24, and GCDFP-44 content (see below). At each two day test interval one replicate dish was analyzed for the DNA cellular content in each test well. The experiment was completed at day ten when the final replicate dish was assayed for DNA content. Results were expressed as the ng quantity of the GCDFP secreted per two day interval per gg of cellular DNA. For the experiments where the five replicate test dish analysis was repeated more than once, a standard error of the mean (SEM) was calculated with the results (see Table 1).
Radioimmunoassays GCDFP-15, GCDFP-24, and GCDFP-44 have been isolated and characterized previously [15, 36]. Rabbit antiserum specific for GCDFP-15 and GCDFP-24 have been developed [15, 36]. Rabbit anti-serum specific for GCDFP-44 (Zn-alpha-2-
79
glycoprotein) was purchased from Behring Diagnostics, Sommerfield, N.J. Each protein was radiolabeled with ~25iodineusing Iodogen (Pierce, Rockfort, IL.) by the method of Fraker and Speck [37]. Specific activity for radiolabeled GCDFP-15 was approximately 20gCi/gg; for GCDFP-24 and for GCDFP-44 it was approximately 10gCi/~tg. An isotope dilution format assay was established for measurement of each protein [38]. The assay sample (100 gl of culture medium) was first mixed with a known quantity of trace labeled antigen (also in 100gt), followed by addition of a defined amount of antibody solution in 100gl. Antibody solution for the GCDFP-15 assay was used at 1/ 50,000 per assay tube; for the GCDFP-24 assay at 1/70,000; and for the GCDFP-44 assay at 1/20,000. The total assay volume was brought to 1ml by addition of assay buffer (0.02M sodium azide containing 1mg/ml of crystallized BSA). Assay tubes were incubated overnight at room temperature followed by termination of the assay with addition of 400 gl of a solid phase suspension of goat anti-rabbit antibody attached to Kynar [39]. Assay tubes were centrifuged, the supernatant decanted and the pellet counted in a Packard Prias gamma counter with a counting efficiency of approximately 60%. The percentage of added label bound by antibody (Bo) for each assay was set to be approximately 50% of added counts. The standard curve for each assay utilized 0, 5, 10, 15, 20 and 25 ng of the specific antigen. If the initial culture medium volume analyzed (100gl) was off scale then further dilution of the culture medium into control culture medium was made and the sample re-analyzed. All assay data points were determined in duplicate. Duplicates which varied by greater than + 2% were repeated. The coefficient of variation for each of these RIA's was approximately +_7%. The GCDFP-15, GCDFP-24, and GCDFP-44 are immunologically unrelated proteins and no cross reactions were evident on RIA.
DNA analysis T47D cell growth was determined at two day in-
DE Haagensen et al.
80
Results 140, 130.
I
120, 110,
100" 90' ,