- Email: [email protected]
Steroid modulation of aromatase activity in human cultured breast carcinoma cells
J. steroid B&hem. Vol. 29, No. 4, pp. 393-3%1988 Printed in Great Britain.All rightsrecurred
STEROID ACTIVITY
Copyright0
0022-4731/88 $3.00 + 0.00 1988 Pergamon Press plc
MODULATION OF AROMATASE IN HUMAN CULTURED BREAST CARCINOMA CELLS
M. BLACKSTEIN and D. W. KILLINGER* Department of Medicine, The Wellesley Hospital, University of Toronto, Toronto, Ontario, Canada M5S lA8 E.
F%REL,
D. DANIILESCU,L. -LIP,
(Receiued 1 April 1987)
Stmtmary-Cortisol and steroids with progestational or androgenic activity were studied to determine the effects of these steroids on the conversion of androstenedione (A) to estrone (E,) in human cultured breast carcinoma cells. Cortisol (10e6 M) stimulated aromatase activity in two estrogen unresponsive cell lines (MD, DM) and in an estrogen responsive cell line (MCF,) with the maximum stimulation occurring during confluence. Cortisol inhibited the replication of MCF, cells but not MD and DM. Dihydrotestosterone, androsterone and Sa-androstanedione (10m6M) inhibited the conversion of A to Et by greater than 90% under basal and cortisol stimulated conditions. Progesterone (10e6M) had no effect on aromatase activity while the progestational agent R5020 (10e6M) produced a 30% inhibition. The anabolic steroids 19-nortestosterone and 19norandrostenedione which also have progestational activity inhibited the conversion of A to E, in a dose dependent manner with 90% inhibition at lOA M. Danazol ( 10m6M) a drug with both androgenic and progestational activity inhibited El formation by 30%. Under the same conditions, the known inhibitor of aromatase, 4-hydroxyandrostenedione ( 10m6M) decreased E, formation by more than 90% and aminoglutethimide (10e6 M) caused only 25% inhibition. These studies demonstrate that endogenous and exogenous steroids may have significant effects in modulating the local formation of estrogens from androgen precursors in cultured breast carcinoma cells. This effect on estrogen formation may be a factor in the biological response of breast tissue.
INTRODUCTION
gestational agents and endogenous aromatase activity.
The ability of breast carcinoma cells and breast adipose tissue to form estrogens from androgen precursors suggests that this local estrogen formation may be important in both physiological and pathological breast development. The control of peripheral aromatase is not well understood, but specific factors which can modify estrogen formation have been identified. The locally formed Sa-reduced metabolites of androstenedione and testosterone can inhibit the aromatization of these substrates in both breast carcinoma cells [1] and breast adipose stromal cells [2]. Glucocorticoids, in the presence of fetal calf serum, stimulate aromatase activity in adipose tissue stromal cells [3] and this stimulation can be inhibited by Sa-reduced androgens [2]. Drugs which inhibit peripheral estrogen formation [4-81 have been shown to be effective in the treatment of breast cancer. The present study was designed to determine the effects of cortisol on growth and aromatase activity in estrogen responsive and estrogen unresponsive breast carcinoma cell lines and to examine the effects of specific drugs, pro-
on
EXPERIMENTAL Materials
*Address for correspondence: Room 7366, Medical Sci-
ences Building, University of Toronto, Ontario, Canada MSS lA8.
androgens
Reference steroids obtained from Makor Chemicals (Jerusalem, Israel) and the Sigma Chemical Co. (St Louis, MO) were recrystallized prior to use.. Radioactive steroids [l,2,6,7-3H]androstenedione (A) (sp. act. 85 Ci/mmol), [4-‘4C]estrone (E,) (sp. act. 52 mCi/mmol) [4-14C]estradiol (Er) (sp. act. 50 mCi/ mmol) were obtained from New England Nuclear Corp. (Boston, MA), and were purified by paper chromatography prior to use in the systems hexane-benzene-methanol-water (67: 80:20, by vol) for E, and E, and hexane-methanol-water (10:9: 1, by vol) for A. 19-Nortestosterone (norT) and 19-norandrostenedione (norA) were purchased from Research Plus Steroid Laboratories in Denville, NJ, 17a-pregn-4-en-20-yno-(2,3-p) isoxanol-17-01 (Danazol) from Sterling-Winthrop Research Institute; Rensselaer NY and 17,21 -dimethyl- 19-nor-4,9pregnadiene-3,20dione(R5020) from New England Nuclear. Aminoglutethimide was a gift from Ciba Pharmaceutical and 4-hydroxyandrostenedione (4-OH-A) was synthesized in our laboratory by the method of Brodie[9].
393
E. F'ERELet al.
394
Precoated TLC plates, SIL-G U.V.254 were purchased from Brinkman Instruments (Westbury, New York). Whatmann No. 1 paper for chromatography was purchased from Whatmann (U.K.). Preparation
of cell lines of breast carcinoma
The MD and DM cell lines utilized in these studies were isolated from primary breast carcinomas from 2 patients. The tumour cells were grown out in soft agar as described by Salmon et a[.[101 and Salmon[ 1I]. Individual colonies were picked, grown out in monolayer and subcultured. The cell lines have been maintained in monolayer culture in u-MEM supplemented with 15% fetal calf serum, 0.3 U/ml insulin and 50 pg/ml amikacin. The MCF, cells used in this study were obtained from Dr R. Buick through the courtesy of Dr K. Osbourne and the ZR,, cells were obtained from the American Tissue Type Culture collection. Studies of cell growth and metabolism
Metabolic activity was studied in each cell line at different phases of growth. A total of 5 x lo5 cells were plated into 75cm’ flasks with 25 ml of a-medium containing 15% fetal calf serum, 0.3 p/ml insulin and 50 pg/ml amikacin with or without cortisol (1O-6 M). The medium was changed every third day. To study metabolic activity at regular intervals throughout the culture period, medium was replaced with 10 ml of medium containing 20 x 106dpm [‘HIA. After 8 h the medium was removed, extracted and E, , and E,, were isolated as described below. The cells were counted in a hemocytometer after trypsinization. All studies were carried out in duplicate and the studies shown are representative of at least 2 sets of experiments. Flasks containing medium to which no cells were added were used as blanks. To determine the effect of drugs and Sa-reduced androgens on aromatase activity, MD cells were grown in 75 cm* flasks as described above. When the cells were approaching confluence (days 6-8), the medium was changed and new medium containing the drug to be tested was added and preincubated for 24 h prior to the aromatase assay. The compounds studied were norT (10e6, lo-‘, lo-* M), norA (10m6, lo-‘, lo-* M), progesterone [P] ( 10m6M), R-5020 ( 10e6 M), aminogldtethimide ( 10m6M), Danazol ( 10e6 M), 4-OH-A ( 10V6M), 5a-androstanedione [Sa-A-dione] (10m6M), androsterone (AND) ( low6 M), dihydrotestosterone [DHT] ( 10m6M), epiandrosterone [EPI] ( 10m6M), Sa-androstane 3cr, 178 diol, [3a-diol] (10m6M), 5a-androstane-3p, 17/I diol [3/I-diol] (10e6 M) and etiocholanolone [ETIO] ( 10e6 M). The steroids were extracted from the medium with ethyl acetate after addition of [“Cl E,, E2, (5000 dpm) to correct for losses and 100 pg of unlabelled steroid to facilitate identification. Following extraction, each residue was subjected to phenolic partition. The phenolic fraction containing E, and E,
was purified by TLC after acetylation as previously described (Perel and Killinger, 1979). The radioactivity isolated from each fraction from the flasks containing no cells was subtracted prior to calculating the conversion to each product. Recovery of E, and E2 during the isolation procedure ranged from 20 to 30%. Radioactive
counting
Isotopes were counted with a dual label program in a Philips model PW 4700 liquid scintillation analyser using Permaflour (Packard Instruments) as a scintillation mixture. Quench correction and correction for spillover of j4C into the ‘H channel was carried out using an internal standard, quenched controls and an internal computer program. Under the conditions used, ‘H was counted with an efficiency of 28% and 14Cwith an efficiency of 57%. RESULTS
The relationship between cell growth and the conversion of A to E, , and E,, in the MD, DM, and MCF, cell lines in the presence and absence of cortisol is shown in Figs 1-3. Under basal conditions, the MD and DM cells (Figs 1 and 2) in which neither estrogen or progesterone receptors could be demonstrated, had the highest levels of aromatase activity. This activity was maximum as the cells reached confluence and decreased rapidly thereafter in spite of a constant cell number. The presence of cortisol ( 10m6M) in the culture medium during cell growth produced an increase in the conversion of A to E, and E, and peak stimulation of aromatase activity occurred after the cells had reached confluence. There was no effect of cortisol on the growth of cells in either the MD or DM cell lines. In the MCF, cell line which contains both the estrogen and progesterone receptors the formation of E, from A was approx 10% of that seen in the MD and DM lines under basal conditions (Fig. 3). With cortisol (1O-6 M) in the medium during cell growth there was again an increase in aromatase activity and the highest levels were observed after the cells had reached confluence. The growth of these cells was inhibited by 50% in the presence of cortisol. In the ZR,, cell lines (which contains both estrogen and progesterone receptor) no aromatase activity could be demonstrated either in the presence or absence of cortisol. There was, however, greater than 60% inhibition of the growth of these cells in the presence of cortisol. To determine the effect of varying concentrations of cortisol on aromatase activity, studies were carried out in the MD cell line in the preconfluent phase of growth (Fig. 4). Stimulation of aromatase activity was noted at lo-* M with maximum stimulation at 10e6 M. 5a-A-dione, AND and DHT had previously been shown to inhibit aromatase activity in the MD cell line and Table 1 shows the effect of these compounds on aromatase activity under basal and
Steroid modulation of aromatase
395
*O-IX I
n +iic
lo-
F X
z3
5 8
(4
l.O-
CL
O.lo
2
4
II Iii
(b)
I 0
-77
it! 8
0.1 1.0 0.01 :-6 ‘fi! 6
B
16
Fig. 1. The effect of cortisol (HC) on growth and aromatase activity of MD breast carcinoma cells in culture: (a) conversion of A to E,; (b) conversion of A to E,.
cortisol stimulated conditions. In all cases, there was greater than 90% inhibition of El formation by Sa-A-dione, AND and DHT at a concentration of 10e6 M. Similar studies with EPI, 3a-A-diol,3/I-Adial and ET10 at a concentration of 10m6M resulted in inhibition of E, formation by 25, 45, 45 and 20% respectively under basal conditions. It had previously been shown by Schwarzel ef a1.[4] that NorT could inhibit aromatase activity in placental microsomes. Studies were therefore carried out with MD cells with norT and norA at concentrations of lo-’ to 10e6 M to determine the effects of these compounds on aromatase activity. As can be seen on
Table 1, both compounds inhibited aromatase by 40% at lo-’ M, and 80% inhibition was obtained at concentration of 10m6M. Additional studies were performed on the MD cells to determine the effects of a variety of compounds used in the treatment of breast carcinoma or benign breast disease on aromatase activity. These results are also shown in Table 1. Progesterone at a concentration of 10T6M had no effect on aromatase activity while the progestational agent R5020 inhibited aromatase by over 60% at a similar concentration. Aminoglutethimide, a drug which is active in the treatment of breast carcinoma showed relatively modest inhibition (25%) of aroma-
E.
PEREL et al. *II
-HC
.
+HC
(4
*a-HC x
./’
+lic
7
Fig. 2. The effect of cortisol (HC) on growth and aromatase activity of MD breast carcinoma cells in culture: (a) conversion of A to E, ; (b) conversion of A to E2.
tase activity at 10e6M while 4-OH-A at the same concentration produced greater than 90% inhibition. Danazol which is used in the treatment of benign breast disease [12] showed 30% inhibition of E, formation at 10m6M. DISCUSSION
These studies confirm the observations of Lippman that glucocorticoids inhibit the growth of MCF, cells in culture [13, 141. It was their conclusion that this inhibition of growth required the presence of glucocorticoid receptors since incorporation of 13H]thymidine was not inhibited in cell lines lacking et al.
these receptors. No relationship was noted in that study between glucocorticoid inhibition of [3H]thymidine uptake and the presence of estrogen receptors. In the present study, only the MCF, and ZR,, cell lines which contain estrogen, progesterone and glucocorticoid receptors were inhibited in their growth by glucocorticoids [15, 161. The MD and DM cell lines used in these studies do not contain estrogen or progesterone receptors and their glucocorticoid receptors status is not known. It is possible that the lack of growth inhibition of these cells by glucocorticoids is due to a lack of glucocorticoid receptors.
Steroid modulation of aromatase
0.1, 0
‘
,
2
(
,
4
I
6
,
8
1
I
10
397
I
I
12
0.01
Fig. 3. The effect of cortisol (HC) on growth and aromatase activity of MCF-7 breast carcinoma al]s in culture.
3
II
2,$38
Fig. 4. The effectof different ~n~tratio~ of cortisol (HC) on aromatase activity in confluent h4D breast carcinoma cells in culture. The MD and DM lines which are lacking an estrogen receptor have the highest level of aromatase activity. These cells are also unresponsive to estrogen with respect to growth stimulation. There was an
increase in aromatase activity in the presence of cortisol while the cells were in the exponential phase of growth but cortisol had no effect on cell replication. In the MCF, cells aromatase activity was relatively low under basal conditions and was stimulated by cortisol while cell replication was inhibited. In all three cell lines, the stimulation of aromatase activity was greatest during confluence. In the conditions used in these studies, there was 15% fetal calf serum in the culture medium giving optimum cell growth. The amount of A measured in the fetal calf serum was 1.1 nM so the final concentration of A in the culture medium was 2OOpM. The estrogen formed through aromatase activity would have been insufficient to promote cell growth in the estrogen responsive cell lines. The con~ntration of cortisol required to maximally stimulate aromatase activity was low6 M and concentrations as low as lo-* M caused a consistent increase in E, formation. This indicates that changes in cortisol concentrations within the physiological range found in plasma can significantly alter aromatase activity. It had been previously reported [2] that
Table 1. The eiTc& of SK-redwed androgcns and compounds used in tbe treatment of breast disease on aromatasc activity in the MD line of breast carcinoma cells Compound (IO-*M)
% Inhibition of aromatasc
Sa-A-dione DHT
:;
AND
95
4-OH-A NorT (IO-’ M) (IO-” M) NorA (10-I M)
3a-A-dial 3/I-A-dial EFI ET10
45 45 2s 20
DanazylO- M, ~ino~utcthi~de RSO20 Progcstcrone
Compound (K+M)
% Inhibition of aromatasc z : 80 :‘: 65 0
398
E. PERELel al.
Su-A-dione, AND and DHT were able to inhibit the conversion of A to E, in breast carcinoma cells. The present study demonstrates that these Sa-reduced androgens can also inhibit the cortisol stimulated increase in estrogen formation in these cells and may provide a further mechanism for local modulation of aromatase activity. Other Sa-reduced androgens including EPI, 3a-A-dial and 3/l-A-dial were less effective as aromatase inhibitors and the 5b-reduced androgen ET10 had minimal effect. The relatively weak inhibition of aromatase by aminoglutethimide noted in this system is in contrast to the potent inhibition reported by Thompson and Siiteri[ 171 using a placental microsome preparation and Santen et al. in in vivo studies [18]. The concentration used in our incubations (10m6M) may have been lower than the plasma levels achieved by Santen et a/.[181 in their in vivo studies. The inhibition noted in our report is comparable to that found by Tilson-Mallett et al. in their in vitro studies using homogenates of breast carcinoma tissue [l9]. NorT is an anabolic steroid used clinically in the form of a phenylpropionate or decanoate ester [20,21]. In this study norT and norA were found to be excellent inhibitors of aromatase activity at relatively low concentrations (lo-’ M). These compounds can be aromatized [22] by placental aromatase and may inhibit E, formation by competing with A for binding to the aromatase enzyme. In studies in non-pregnant subjects in vivo, however, the conversion of norT to estrogens was found to be minima1 [23]. Anabolic steroids have been used in the treatment of breast cancer [24] and their effectiveness could be related to the inhibition of aromatase, or to their androgenic [21] or progestationa1[22] activity. These observations suggest that a re-evaluation of the use of these compounds in the treatment of breast cancer is indicated. Danazol a steroid which has progestational as well as androgenic activity (251 produced 30% inhibition of aromatase activity. The norT derivatives and Danazol bind to SHBG [26,27] and this binding may be required for the biological activity of these compounds as has been suggested by Avvakumov et al.[28] and Siiteri and Simbert[29]. These observations demonstrate that a variety of compounds including endogenous steroids and drugs used in the treatment of benign and malignant breast disorders can modulate the formation of estrogen. Aromatase inhibition may be one mechanism by which these compounds are active in the treatment of breast disease.
REFERENCES
Perel E., Stolee K. H., Kharlip L., Blackstein M. E. and Killinger D. W.: The intracellular control of aromatase activity by Sa-reduced androgens in human breast carcinoma cells in culture. J. clin. Endocr. Metab. 53
drogens in stromal cells from human breast adipose tissue. J. c/in. Endocr. Metab. 62 (1986) 314318. 3. Simpson E. R., Ackerman G. E., Smith M. E. and Mendelson C. R. Estrogen formation in stromal cells of adipose tissue of women: induction of glucocorticoids. Proc. natn. Acad. Sci. U.S.A. 78 (1981) 569&5694. 4. Schwarzel W. C., Kruggel W. and Brodie H. J.: Studies
of the mechanism of estrogen biosynthesis. VIII. The development of inhibitors-of the -enzyme system in human placenta. Endocrinoloav 92 (1973) 866880. 5. Santen R. J., Santner S., DaviyB., Vkldhuis J., Samojlik E. and Rubt E.: Aminoglutethimide inhibits extraglandular estrogen production in post-menopausal women with breast carcinoma. J. clin. Endocr. Mefab. 47 (1978) 1257-1265. 6. Santen R. J., Worgul T. J., Lipton A., Harvey H.,
Boucher A., Samojlik E. and Wells S.: Aminoglutethimide as treatment of postmenopausal women with advanced breast carcinoma: correlation of clinical and hormonal responses. Ann. intern. Med 96 (1982) 94-101. 7. Brodie A. M. H., Brodie H. J., Romanoff L., Williams J. G., Williams K. I. M. and Wu J. T.: Inhibition of estrogen biosynthesis and regression of mammary tumors by aromatase inhibitors. Harm. Cancer adu. exp. med. Biol. 138 (1982) 179-190. 8. Goss P. E., Powles J., Dowsett
M., Hutchison G., Brodie A. M. H., Gazet J. C. and Coombes R. C.: Treatment of advanced postmenopausal breast cancer with an aromatase inhibitor, 4-hydroxyandrostenedione: Phase II report Cancer Res. 46 (1986) 48234826. 9. Brodie A. M. H., Schwarzel W. C., Shaikh A. A. and Brodie H. J.: The effect of an aromatase inhibitor 4-hydroxy-4-androstene-3,17-dione, an estrogendependent process in reproduction and breast cancer. Endocrinology 100 (1977) 1684-1695. 10. Salmon S. E., Hamburger A. W., Soehnlen B. J., Durie B. G. M., Alberts D. S. and Moon T. C.: Quantitation of differential sensitivity of human-tumour stem cells to anticancer drugs. N. Eng. J. Med. 298 (1978) 131-137. 11. Salmon S. E. (Ed): Cloning of Human Tumor Stem Cells Alan Liss, New York (1980) p. 223. 12. Greenblatt B. and Ben-Nun I.: Danazol in the treatment of mammary dysplasia. Drugs (1980) 349-355. 13. Lippman M., Bolan G. and Huff K.: The effects of glucocorticoids and progesterone on hormone responsive human breast cancer in long-term tissue culthe. Cancer Rex 36 (1976) 4602-4609. 14. Osborne C. K.. Monaco M. E.. Kahn C. R.. Huff K.. Bronzert D. and Lippman M. E.: Direct inhibition of growth and antagonism of insulin action of glucocorticoids in human breast cancer cells in culture. Cancer Res. 39 (1979) 2422-2428. 15. Horowitz K. B., Zava D. T., Thilager A. K., Jensen E. M. and McGuire W. L.: Steroid receptor analysis of nine human breast cancer cell lines. Cancer Res. 38 (1978) 2434-2437. 16. Lippman M. E., Allegra J. C. and Strobl J. S.: Growth requirements of a human breast cancer cell line in serum free medium. In Hormones and Cell Cultures (Edited by G. Sato and R. Ross). Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, Vol. 6 (1979) 545-558. 17. Thompson E. A. Jr and Siiteri P. K.: The involvement of human placental microsomal cytochrome P-450 in aromatization. J. biot. Chem. 249 (1974) 5373-5378. 18.
Santen R. J., Santner S., Davis B., Veldhuis J.. Samojlik E. and Ruby E.: Aminoglutethimide inhibits extraglandular estrogen production in postmenopausal
(1984) 467-472.
women with breast cancer. J. clin. Endocr. Metab 47
Perel E., Daniilescu D., Kindler S., Kharlip L. and Killinger D. W.: The formation of Sa-reduced an-
(1978) 1257-1265. 19. Tilson-Mallett
N., Santner S. J., Feil P. D. and Santen
Steroid modulation of aromatase
20. 21. 22.
23.
24.
25.
R. J.: Biological significance of aromaactivity in human breast tumors. J. c&t. Endoer. Met& 57 (1983) 1125-1128. Overbeck G. A. and De Visser J.: Nor-androstenolonephenyl-propionate, a new potent anabolic ester. Acra endocr., Copenh. 24 (19S7) 209-219. De Visser J. and Overbeek G. A.: Pharmacological properties of nondrolone decanoate. Acra endocr., Copenh. 35 (1960) 405-412. Ryan K. J.: Estrogen formation by the human placenta: studies on the mechanism of steroid aromatization by mammalian tissue. Acfa endoer., Cope& 35, Suppl. 51 (1960) 697~98‘ Townsley J. D. and Brodie H. L.: Low conversion of 19nortestosterone to urinary estrogens. Lancer 2 (1970) 1039. Turner R.: The value of anabolic steroids for patients receiving chemotherapy. Clin Oncol. Sot. Aust. Ann Meeting APCS No. 9, Excerpta Medica, Amsterdam (1981). Jenkin G., Cookson C. I. and Thorbum G. D.: The
26.
27.
28.
29.
399
interaction of human endometrial and myometrial steroid receptors with Danaxol. Cfin. Endocr. 19 (1983) 377-388. Schwartz S., Tappeimer G. and Hinter H.: Hormone binding globulin levels in patients with hereditary angioedema during treatment with Danazol. Clin. Endocr. 14 (1981) 563-570. Hammond G. L., Lahteanmaki P. L. A., Lahteenmaki P. and Luukkainen T.: Distribution and percentage of non-protein bound contraceptive steroids in human serum. J. rferoid Bioehem. 17 (1982) 375-380. Awakumov G. V., Zhuk N. 1. and Stret’chyonok 0. A.: Subcellular dist~bution and seetectivity of the proteinbinding component of the recognition system for sexhormone-binding proteinestradiol complex in human d&dual endometrium. Biochem. biophys Acta 816 (1986) 490-498. Siiteri P. K. and Simberg N. H.: Changing concents of active androgens in blood. Clin. En&$ Net&. 15 (1986) 247-258.
STEROID ACTIVITY
Copyright0
0022-4731/88 $3.00 + 0.00 1988 Pergamon Press plc
MODULATION OF AROMATASE IN HUMAN CULTURED BREAST CARCINOMA CELLS
M. BLACKSTEIN and D. W. KILLINGER* Department of Medicine, The Wellesley Hospital, University of Toronto, Toronto, Ontario, Canada M5S lA8 E.
F%REL,
D. DANIILESCU,L. -LIP,
(Receiued 1 April 1987)
Stmtmary-Cortisol and steroids with progestational or androgenic activity were studied to determine the effects of these steroids on the conversion of androstenedione (A) to estrone (E,) in human cultured breast carcinoma cells. Cortisol (10e6 M) stimulated aromatase activity in two estrogen unresponsive cell lines (MD, DM) and in an estrogen responsive cell line (MCF,) with the maximum stimulation occurring during confluence. Cortisol inhibited the replication of MCF, cells but not MD and DM. Dihydrotestosterone, androsterone and Sa-androstanedione (10m6M) inhibited the conversion of A to Et by greater than 90% under basal and cortisol stimulated conditions. Progesterone (10e6M) had no effect on aromatase activity while the progestational agent R5020 (10e6M) produced a 30% inhibition. The anabolic steroids 19-nortestosterone and 19norandrostenedione which also have progestational activity inhibited the conversion of A to E, in a dose dependent manner with 90% inhibition at lOA M. Danazol ( 10m6M) a drug with both androgenic and progestational activity inhibited El formation by 30%. Under the same conditions, the known inhibitor of aromatase, 4-hydroxyandrostenedione ( 10m6M) decreased E, formation by more than 90% and aminoglutethimide (10e6 M) caused only 25% inhibition. These studies demonstrate that endogenous and exogenous steroids may have significant effects in modulating the local formation of estrogens from androgen precursors in cultured breast carcinoma cells. This effect on estrogen formation may be a factor in the biological response of breast tissue.
INTRODUCTION
gestational agents and endogenous aromatase activity.
The ability of breast carcinoma cells and breast adipose tissue to form estrogens from androgen precursors suggests that this local estrogen formation may be important in both physiological and pathological breast development. The control of peripheral aromatase is not well understood, but specific factors which can modify estrogen formation have been identified. The locally formed Sa-reduced metabolites of androstenedione and testosterone can inhibit the aromatization of these substrates in both breast carcinoma cells [1] and breast adipose stromal cells [2]. Glucocorticoids, in the presence of fetal calf serum, stimulate aromatase activity in adipose tissue stromal cells [3] and this stimulation can be inhibited by Sa-reduced androgens [2]. Drugs which inhibit peripheral estrogen formation [4-81 have been shown to be effective in the treatment of breast cancer. The present study was designed to determine the effects of cortisol on growth and aromatase activity in estrogen responsive and estrogen unresponsive breast carcinoma cell lines and to examine the effects of specific drugs, pro-
on
EXPERIMENTAL Materials
*Address for correspondence: Room 7366, Medical Sci-
ences Building, University of Toronto, Ontario, Canada MSS lA8.
androgens
Reference steroids obtained from Makor Chemicals (Jerusalem, Israel) and the Sigma Chemical Co. (St Louis, MO) were recrystallized prior to use.. Radioactive steroids [l,2,6,7-3H]androstenedione (A) (sp. act. 85 Ci/mmol), [4-‘4C]estrone (E,) (sp. act. 52 mCi/mmol) [4-14C]estradiol (Er) (sp. act. 50 mCi/ mmol) were obtained from New England Nuclear Corp. (Boston, MA), and were purified by paper chromatography prior to use in the systems hexane-benzene-methanol-water (67: 80:20, by vol) for E, and E, and hexane-methanol-water (10:9: 1, by vol) for A. 19-Nortestosterone (norT) and 19-norandrostenedione (norA) were purchased from Research Plus Steroid Laboratories in Denville, NJ, 17a-pregn-4-en-20-yno-(2,3-p) isoxanol-17-01 (Danazol) from Sterling-Winthrop Research Institute; Rensselaer NY and 17,21 -dimethyl- 19-nor-4,9pregnadiene-3,20dione(R5020) from New England Nuclear. Aminoglutethimide was a gift from Ciba Pharmaceutical and 4-hydroxyandrostenedione (4-OH-A) was synthesized in our laboratory by the method of Brodie[9].
393
E. F'ERELet al.
394
Precoated TLC plates, SIL-G U.V.254 were purchased from Brinkman Instruments (Westbury, New York). Whatmann No. 1 paper for chromatography was purchased from Whatmann (U.K.). Preparation
of cell lines of breast carcinoma
The MD and DM cell lines utilized in these studies were isolated from primary breast carcinomas from 2 patients. The tumour cells were grown out in soft agar as described by Salmon et a[.[101 and Salmon[ 1I]. Individual colonies were picked, grown out in monolayer and subcultured. The cell lines have been maintained in monolayer culture in u-MEM supplemented with 15% fetal calf serum, 0.3 U/ml insulin and 50 pg/ml amikacin. The MCF, cells used in this study were obtained from Dr R. Buick through the courtesy of Dr K. Osbourne and the ZR,, cells were obtained from the American Tissue Type Culture collection. Studies of cell growth and metabolism
Metabolic activity was studied in each cell line at different phases of growth. A total of 5 x lo5 cells were plated into 75cm’ flasks with 25 ml of a-medium containing 15% fetal calf serum, 0.3 p/ml insulin and 50 pg/ml amikacin with or without cortisol (1O-6 M). The medium was changed every third day. To study metabolic activity at regular intervals throughout the culture period, medium was replaced with 10 ml of medium containing 20 x 106dpm [‘HIA. After 8 h the medium was removed, extracted and E, , and E,, were isolated as described below. The cells were counted in a hemocytometer after trypsinization. All studies were carried out in duplicate and the studies shown are representative of at least 2 sets of experiments. Flasks containing medium to which no cells were added were used as blanks. To determine the effect of drugs and Sa-reduced androgens on aromatase activity, MD cells were grown in 75 cm* flasks as described above. When the cells were approaching confluence (days 6-8), the medium was changed and new medium containing the drug to be tested was added and preincubated for 24 h prior to the aromatase assay. The compounds studied were norT (10e6, lo-‘, lo-* M), norA (10m6, lo-‘, lo-* M), progesterone [P] ( 10m6M), R-5020 ( 10e6 M), aminogldtethimide ( 10m6M), Danazol ( 10e6 M), 4-OH-A ( 10V6M), 5a-androstanedione [Sa-A-dione] (10m6M), androsterone (AND) ( low6 M), dihydrotestosterone [DHT] ( 10m6M), epiandrosterone [EPI] ( 10m6M), Sa-androstane 3cr, 178 diol, [3a-diol] (10m6M), 5a-androstane-3p, 17/I diol [3/I-diol] (10e6 M) and etiocholanolone [ETIO] ( 10e6 M). The steroids were extracted from the medium with ethyl acetate after addition of [“Cl E,, E2, (5000 dpm) to correct for losses and 100 pg of unlabelled steroid to facilitate identification. Following extraction, each residue was subjected to phenolic partition. The phenolic fraction containing E, and E,
was purified by TLC after acetylation as previously described (Perel and Killinger, 1979). The radioactivity isolated from each fraction from the flasks containing no cells was subtracted prior to calculating the conversion to each product. Recovery of E, and E2 during the isolation procedure ranged from 20 to 30%. Radioactive
counting
Isotopes were counted with a dual label program in a Philips model PW 4700 liquid scintillation analyser using Permaflour (Packard Instruments) as a scintillation mixture. Quench correction and correction for spillover of j4C into the ‘H channel was carried out using an internal standard, quenched controls and an internal computer program. Under the conditions used, ‘H was counted with an efficiency of 28% and 14Cwith an efficiency of 57%. RESULTS
The relationship between cell growth and the conversion of A to E, , and E,, in the MD, DM, and MCF, cell lines in the presence and absence of cortisol is shown in Figs 1-3. Under basal conditions, the MD and DM cells (Figs 1 and 2) in which neither estrogen or progesterone receptors could be demonstrated, had the highest levels of aromatase activity. This activity was maximum as the cells reached confluence and decreased rapidly thereafter in spite of a constant cell number. The presence of cortisol ( 10m6M) in the culture medium during cell growth produced an increase in the conversion of A to E, and E, and peak stimulation of aromatase activity occurred after the cells had reached confluence. There was no effect of cortisol on the growth of cells in either the MD or DM cell lines. In the MCF, cell line which contains both the estrogen and progesterone receptors the formation of E, from A was approx 10% of that seen in the MD and DM lines under basal conditions (Fig. 3). With cortisol (1O-6 M) in the medium during cell growth there was again an increase in aromatase activity and the highest levels were observed after the cells had reached confluence. The growth of these cells was inhibited by 50% in the presence of cortisol. In the ZR,, cell lines (which contains both estrogen and progesterone receptor) no aromatase activity could be demonstrated either in the presence or absence of cortisol. There was, however, greater than 60% inhibition of the growth of these cells in the presence of cortisol. To determine the effect of varying concentrations of cortisol on aromatase activity, studies were carried out in the MD cell line in the preconfluent phase of growth (Fig. 4). Stimulation of aromatase activity was noted at lo-* M with maximum stimulation at 10e6 M. 5a-A-dione, AND and DHT had previously been shown to inhibit aromatase activity in the MD cell line and Table 1 shows the effect of these compounds on aromatase activity under basal and
Steroid modulation of aromatase
395
*O-IX I
n +iic
lo-
F X
z3
5 8
(4
l.O-
CL
O.lo
2
4
II Iii
(b)
I 0
-77
it! 8
0.1 1.0 0.01 :-6 ‘fi! 6
B
16
Fig. 1. The effect of cortisol (HC) on growth and aromatase activity of MD breast carcinoma cells in culture: (a) conversion of A to E,; (b) conversion of A to E,.
cortisol stimulated conditions. In all cases, there was greater than 90% inhibition of El formation by Sa-A-dione, AND and DHT at a concentration of 10e6 M. Similar studies with EPI, 3a-A-diol,3/I-Adial and ET10 at a concentration of 10m6M resulted in inhibition of E, formation by 25, 45, 45 and 20% respectively under basal conditions. It had previously been shown by Schwarzel ef a1.[4] that NorT could inhibit aromatase activity in placental microsomes. Studies were therefore carried out with MD cells with norT and norA at concentrations of lo-’ to 10e6 M to determine the effects of these compounds on aromatase activity. As can be seen on
Table 1, both compounds inhibited aromatase by 40% at lo-’ M, and 80% inhibition was obtained at concentration of 10m6M. Additional studies were performed on the MD cells to determine the effects of a variety of compounds used in the treatment of breast carcinoma or benign breast disease on aromatase activity. These results are also shown in Table 1. Progesterone at a concentration of 10T6M had no effect on aromatase activity while the progestational agent R5020 inhibited aromatase by over 60% at a similar concentration. Aminoglutethimide, a drug which is active in the treatment of breast carcinoma showed relatively modest inhibition (25%) of aroma-
E.
PEREL et al. *II
-HC
.
+HC
(4
*a-HC x
./’
+lic
7
Fig. 2. The effect of cortisol (HC) on growth and aromatase activity of MD breast carcinoma cells in culture: (a) conversion of A to E, ; (b) conversion of A to E2.
tase activity at 10e6M while 4-OH-A at the same concentration produced greater than 90% inhibition. Danazol which is used in the treatment of benign breast disease [12] showed 30% inhibition of E, formation at 10m6M. DISCUSSION
These studies confirm the observations of Lippman that glucocorticoids inhibit the growth of MCF, cells in culture [13, 141. It was their conclusion that this inhibition of growth required the presence of glucocorticoid receptors since incorporation of 13H]thymidine was not inhibited in cell lines lacking et al.
these receptors. No relationship was noted in that study between glucocorticoid inhibition of [3H]thymidine uptake and the presence of estrogen receptors. In the present study, only the MCF, and ZR,, cell lines which contain estrogen, progesterone and glucocorticoid receptors were inhibited in their growth by glucocorticoids [15, 161. The MD and DM cell lines used in these studies do not contain estrogen or progesterone receptors and their glucocorticoid receptors status is not known. It is possible that the lack of growth inhibition of these cells by glucocorticoids is due to a lack of glucocorticoid receptors.
Steroid modulation of aromatase
0.1, 0
‘
,
2
(
,
4
I
6
,
8
1
I
10
397
I
I
12
0.01
Fig. 3. The effect of cortisol (HC) on growth and aromatase activity of MCF-7 breast carcinoma al]s in culture.
3
II
2,$38
Fig. 4. The effectof different ~n~tratio~ of cortisol (HC) on aromatase activity in confluent h4D breast carcinoma cells in culture. The MD and DM lines which are lacking an estrogen receptor have the highest level of aromatase activity. These cells are also unresponsive to estrogen with respect to growth stimulation. There was an
increase in aromatase activity in the presence of cortisol while the cells were in the exponential phase of growth but cortisol had no effect on cell replication. In the MCF, cells aromatase activity was relatively low under basal conditions and was stimulated by cortisol while cell replication was inhibited. In all three cell lines, the stimulation of aromatase activity was greatest during confluence. In the conditions used in these studies, there was 15% fetal calf serum in the culture medium giving optimum cell growth. The amount of A measured in the fetal calf serum was 1.1 nM so the final concentration of A in the culture medium was 2OOpM. The estrogen formed through aromatase activity would have been insufficient to promote cell growth in the estrogen responsive cell lines. The con~ntration of cortisol required to maximally stimulate aromatase activity was low6 M and concentrations as low as lo-* M caused a consistent increase in E, formation. This indicates that changes in cortisol concentrations within the physiological range found in plasma can significantly alter aromatase activity. It had been previously reported [2] that
Table 1. The eiTc& of SK-redwed androgcns and compounds used in tbe treatment of breast disease on aromatasc activity in the MD line of breast carcinoma cells Compound (IO-*M)
% Inhibition of aromatasc
Sa-A-dione DHT
:;
AND
95
4-OH-A NorT (IO-’ M) (IO-” M) NorA (10-I M)
3a-A-dial 3/I-A-dial EFI ET10
45 45 2s 20
DanazylO- M, ~ino~utcthi~de RSO20 Progcstcrone
Compound (K+M)
% Inhibition of aromatasc z : 80 :‘: 65 0
398
E. PERELel al.
Su-A-dione, AND and DHT were able to inhibit the conversion of A to E, in breast carcinoma cells. The present study demonstrates that these Sa-reduced androgens can also inhibit the cortisol stimulated increase in estrogen formation in these cells and may provide a further mechanism for local modulation of aromatase activity. Other Sa-reduced androgens including EPI, 3a-A-dial and 3/l-A-dial were less effective as aromatase inhibitors and the 5b-reduced androgen ET10 had minimal effect. The relatively weak inhibition of aromatase by aminoglutethimide noted in this system is in contrast to the potent inhibition reported by Thompson and Siiteri[ 171 using a placental microsome preparation and Santen et al. in in vivo studies [18]. The concentration used in our incubations (10m6M) may have been lower than the plasma levels achieved by Santen et a/.[181 in their in vivo studies. The inhibition noted in our report is comparable to that found by Tilson-Mallett et al. in their in vitro studies using homogenates of breast carcinoma tissue [l9]. NorT is an anabolic steroid used clinically in the form of a phenylpropionate or decanoate ester [20,21]. In this study norT and norA were found to be excellent inhibitors of aromatase activity at relatively low concentrations (lo-’ M). These compounds can be aromatized [22] by placental aromatase and may inhibit E, formation by competing with A for binding to the aromatase enzyme. In studies in non-pregnant subjects in vivo, however, the conversion of norT to estrogens was found to be minima1 [23]. Anabolic steroids have been used in the treatment of breast cancer [24] and their effectiveness could be related to the inhibition of aromatase, or to their androgenic [21] or progestationa1[22] activity. These observations suggest that a re-evaluation of the use of these compounds in the treatment of breast cancer is indicated. Danazol a steroid which has progestational as well as androgenic activity (251 produced 30% inhibition of aromatase activity. The norT derivatives and Danazol bind to SHBG [26,27] and this binding may be required for the biological activity of these compounds as has been suggested by Avvakumov et al.[28] and Siiteri and Simbert[29]. These observations demonstrate that a variety of compounds including endogenous steroids and drugs used in the treatment of benign and malignant breast disorders can modulate the formation of estrogen. Aromatase inhibition may be one mechanism by which these compounds are active in the treatment of breast disease.
REFERENCES
Perel E., Stolee K. H., Kharlip L., Blackstein M. E. and Killinger D. W.: The intracellular control of aromatase activity by Sa-reduced androgens in human breast carcinoma cells in culture. J. clin. Endocr. Metab. 53
drogens in stromal cells from human breast adipose tissue. J. c/in. Endocr. Metab. 62 (1986) 314318. 3. Simpson E. R., Ackerman G. E., Smith M. E. and Mendelson C. R. Estrogen formation in stromal cells of adipose tissue of women: induction of glucocorticoids. Proc. natn. Acad. Sci. U.S.A. 78 (1981) 569&5694. 4. Schwarzel W. C., Kruggel W. and Brodie H. J.: Studies
of the mechanism of estrogen biosynthesis. VIII. The development of inhibitors-of the -enzyme system in human placenta. Endocrinoloav 92 (1973) 866880. 5. Santen R. J., Santner S., DaviyB., Vkldhuis J., Samojlik E. and Rubt E.: Aminoglutethimide inhibits extraglandular estrogen production in post-menopausal women with breast carcinoma. J. clin. Endocr. Mefab. 47 (1978) 1257-1265. 6. Santen R. J., Worgul T. J., Lipton A., Harvey H.,
Boucher A., Samojlik E. and Wells S.: Aminoglutethimide as treatment of postmenopausal women with advanced breast carcinoma: correlation of clinical and hormonal responses. Ann. intern. Med 96 (1982) 94-101. 7. Brodie A. M. H., Brodie H. J., Romanoff L., Williams J. G., Williams K. I. M. and Wu J. T.: Inhibition of estrogen biosynthesis and regression of mammary tumors by aromatase inhibitors. Harm. Cancer adu. exp. med. Biol. 138 (1982) 179-190. 8. Goss P. E., Powles J., Dowsett
M., Hutchison G., Brodie A. M. H., Gazet J. C. and Coombes R. C.: Treatment of advanced postmenopausal breast cancer with an aromatase inhibitor, 4-hydroxyandrostenedione: Phase II report Cancer Res. 46 (1986) 48234826. 9. Brodie A. M. H., Schwarzel W. C., Shaikh A. A. and Brodie H. J.: The effect of an aromatase inhibitor 4-hydroxy-4-androstene-3,17-dione, an estrogendependent process in reproduction and breast cancer. Endocrinology 100 (1977) 1684-1695. 10. Salmon S. E., Hamburger A. W., Soehnlen B. J., Durie B. G. M., Alberts D. S. and Moon T. C.: Quantitation of differential sensitivity of human-tumour stem cells to anticancer drugs. N. Eng. J. Med. 298 (1978) 131-137. 11. Salmon S. E. (Ed): Cloning of Human Tumor Stem Cells Alan Liss, New York (1980) p. 223. 12. Greenblatt B. and Ben-Nun I.: Danazol in the treatment of mammary dysplasia. Drugs (1980) 349-355. 13. Lippman M., Bolan G. and Huff K.: The effects of glucocorticoids and progesterone on hormone responsive human breast cancer in long-term tissue culthe. Cancer Rex 36 (1976) 4602-4609. 14. Osborne C. K.. Monaco M. E.. Kahn C. R.. Huff K.. Bronzert D. and Lippman M. E.: Direct inhibition of growth and antagonism of insulin action of glucocorticoids in human breast cancer cells in culture. Cancer Res. 39 (1979) 2422-2428. 15. Horowitz K. B., Zava D. T., Thilager A. K., Jensen E. M. and McGuire W. L.: Steroid receptor analysis of nine human breast cancer cell lines. Cancer Res. 38 (1978) 2434-2437. 16. Lippman M. E., Allegra J. C. and Strobl J. S.: Growth requirements of a human breast cancer cell line in serum free medium. In Hormones and Cell Cultures (Edited by G. Sato and R. Ross). Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, Vol. 6 (1979) 545-558. 17. Thompson E. A. Jr and Siiteri P. K.: The involvement of human placental microsomal cytochrome P-450 in aromatization. J. biot. Chem. 249 (1974) 5373-5378. 18.
Santen R. J., Santner S., Davis B., Veldhuis J.. Samojlik E. and Ruby E.: Aminoglutethimide inhibits extraglandular estrogen production in postmenopausal
(1984) 467-472.
women with breast cancer. J. clin. Endocr. Metab 47
Perel E., Daniilescu D., Kindler S., Kharlip L. and Killinger D. W.: The formation of Sa-reduced an-
(1978) 1257-1265. 19. Tilson-Mallett
N., Santner S. J., Feil P. D. and Santen
Steroid modulation of aromatase
20. 21. 22.
23.
24.
25.
R. J.: Biological significance of aromaactivity in human breast tumors. J. c&t. Endoer. Met& 57 (1983) 1125-1128. Overbeck G. A. and De Visser J.: Nor-androstenolonephenyl-propionate, a new potent anabolic ester. Acra endocr., Copenh. 24 (19S7) 209-219. De Visser J. and Overbeek G. A.: Pharmacological properties of nondrolone decanoate. Acra endocr., Copenh. 35 (1960) 405-412. Ryan K. J.: Estrogen formation by the human placenta: studies on the mechanism of steroid aromatization by mammalian tissue. Acfa endoer., Cope& 35, Suppl. 51 (1960) 697~98‘ Townsley J. D. and Brodie H. L.: Low conversion of 19nortestosterone to urinary estrogens. Lancer 2 (1970) 1039. Turner R.: The value of anabolic steroids for patients receiving chemotherapy. Clin Oncol. Sot. Aust. Ann Meeting APCS No. 9, Excerpta Medica, Amsterdam (1981). Jenkin G., Cookson C. I. and Thorbum G. D.: The
26.
27.
28.
29.
399
interaction of human endometrial and myometrial steroid receptors with Danaxol. Cfin. Endocr. 19 (1983) 377-388. Schwartz S., Tappeimer G. and Hinter H.: Hormone binding globulin levels in patients with hereditary angioedema during treatment with Danazol. Clin. Endocr. 14 (1981) 563-570. Hammond G. L., Lahteanmaki P. L. A., Lahteenmaki P. and Luukkainen T.: Distribution and percentage of non-protein bound contraceptive steroids in human serum. J. rferoid Bioehem. 17 (1982) 375-380. Awakumov G. V., Zhuk N. 1. and Stret’chyonok 0. A.: Subcellular dist~bution and seetectivity of the proteinbinding component of the recognition system for sexhormone-binding proteinestradiol complex in human d&dual endometrium. Biochem. biophys Acta 816 (1986) 490-498. Siiteri P. K. and Simberg N. H.: Changing concents of active androgens in blood. Clin. En&$ Net&. 15 (1986) 247-258.