Int. J. Cancer: 52, 778-784 (1992) 0 1992 Wiley-Liss, Inc.
'
Publication of the InternationalUnion Against Cancer Publication de I'Unlon InternationaleContre le Cancer
REGULATION OF ANDROGEN RECEPTOR GENE EXPRESSION BY STEROIDS AND RETINOIC ACID IN HUMAN BREAST-CANCER CELLS Rosemary E. Wayne D. T I L L E YMichael ~, J. M C P H A U Land ~ Robert L. SUTHERLANDl 'Cancer Biology Division, Garvan Institute of Medical Research, St. VincentS Hospital. Sydney, NS W 2010, Australia; ?Depar?mentof Internal Medicine, The Universiq of Texas Southwestern Medical Center, Dallas, T X 752358857, USA; and ' D e p a ~ m e nof t Surgery, Flinders Medical Center, Adelaide, SA S025, Australia. Although the androgen receptor (AR) has been detected by ligand-bindingassays, there is little known about the expression and regulation of the AR gene in human breast-cancer cells. AR mRNA, measured by Northern analysis in 18 cell lines, was found t o be expressed predominantly in oestrogen- and progesterone-receptor-positive (ER+, PR+) lines as a single species of approximately 10.5 kb but was also comparatively abundant in I ER- and PR-negative cell line, MDA-MB-453. Dexamethasone (Dex), Organon 2058 (Org 2058), dihydrotestosterone (DHT), and all-trans-retinoic acid (RA) down-regulated AR mRNA levels in T-47D (ER+, PR+) cells 6 hr after treatment, whereas oestradiol (E2) had no effect. In MDA-MB-453 (ER-, PR-) cells, regulation of AR mRNA by RA differed from the other cell lines: RA increased the level of AR mRNA. DHT-binding assays indicated a corresponding increase in AR protein. Transfection of the androgen-responsive mouse mammary tumour virus long-terminal repeat (MMTV LTR) linked t o a chloramphenicol acetyltransferase(CAT) reporter gene was used to examine the effect of altered AR levels on androgen action. The increased level of AR following RA pre-treatment in MDA-MB-453 cells resulted in enhanced induction of CAT activity by DHT and, conversely, a decrease in the level of AR following RA pretreatment in T-47D cells resulted in reduced induction of CAT activity by DHT. These data demonstrate that AR is expressed predominantly in ER+ and PR+ cell lines and i t s expression is regulated by ligands also known to regulate ER or PR. including progestins and retinoids. Androgen responsiveness measured by a transfected reporter gene was altered according t o the extent of up- or down-regulationof AR expression, demonstrating that responsivenessis dependent on receptor concentration. 8 1992 Wiley-Liss.lnc.
The AR is often co-expressed with the E R and PR in human breast tumours (Allegra et al., 1979; Teulings et al., 1980; Miller et al., 1985; Brentani etal.. 1986; Valyani et al., 1987; Lea et al., 1989). Compared with anti-oestrogens and progestins, however, ligands acting through the A R have received less attention as therapeutic agents in treating breast cancer, partly due to the lesser incidence and lower amounts of AR measured in breast biopsies (Trams and Maass, 1977; Allegra et al., 1979; Mercer et al., 1983; Miller et al., 1985; Lea et al., 1989). Only in patients with advanced disease has A R been targeted. Administration of pharmacological doses of androgens have shown anti-tumour activity (Manni et al., 1981). No objective response was observed in preliminary investigations with androgen antagonists, although disease stabilization in a few patients suggested that further evaluation is justified in women with less advanced disease (Millward et al., 1991), and anti-androgen treatment achieved a response in 7 out of 10 male breast-cancer patients (Lopez, 1985). The combined use of an androgen and anti-oestrogen was found to have a therapeutic advantage over anti-oestrogen treatment alone (Tormey et al., 1983; Ingle etal., 1991). There are also only limited data on the effect of androgens on the in vitro growth of breast-cancer cells and both stimulatory and inhibitory effects have been observed in different cell lines. Physiological and pharmacological concentrations of androgens stimulated growth of MCF-7 and EFM-19 cells (Hackenberg ef al., 1988), growth was inhibited with physiological conccntrations in ZR-75-1 (Poulin et ul., 1988) and MFM-223 cells (Hackenberg et al., 1991), and growth was
inhibited only by pharmacological concentrations of androgens in T-47D cells (Sutherland et al., 1988). To understand the role of A R in breast cancer, including the effects of androgens on cell proliferation, it was first necessary to characterize A R gene expression and regulation in breastcancer cells. We compared A R expression in breast cells with a prostate carcinoma cell line, because considerably more is known about AR expression and regulation in prostate cancer. In addition to the endogenous ligand for AR, DHT, A R regulation by agents which are known to regulate the other sex-steroid-hormone receptors, E R and PR, was also examined. Recent studies in male androgen-responsive tissues (for example prostate, epididymis, seminal vesicle), have demonstrated that androgens down-regulate A R mRNA (Quarmby et al., 1990; Shan et al., 1990; Krongrad et al., 1991). However, in MCF-7 breast-cancer cells 6 h of androgen treatment had no effect on the quantity of A R protein (Shapiro and Lippman, 1985). The level of A R protein in MCF-7 cells has been shown to be reduced by oestradiol (Stover et al., 1987) and by progestins (Hackenberg et al., 1990). Effects on A R mRNA have not been examined. We have previously demonstrated in T-47D cells that E R mRNA and P R mRNA are downregulated by progestins (Alexander et al., 1989, 1990) and P R is down-regulated by retinoic acid (Clarke et al., 1990, 1991). The aims of this study were to characterize the expression of A R in human breast-cancer cell lines, to examine regulation of AR, particularly by agents which are known to regulate the other sex-steroid-hormone receptors, and to determine the functional consequences of this regulation in terms of cellular sensitivity to androgens. MATERIAL AND METHODS
Material Multicel RPMI 1640 and DMEM culture media were products of Cytosystems, Castle Hill, Australia. Reagents for Northern blot analysis were from the sources listed previously (Hall et al., 1990). Ez, DHT, Dex and R A were obtained from Sigma, St Louis, MO. Unlabeled Org 2058, Sa-dihydro-['HH]testosterone and [3H)-acetyl-co-enzyme A were purchased from Amersham, Sydney, Australia. The steroids were stored in ethanol and RA in dimethylsulphoxide. Plasmids, pMSGCAT, containing MMTV LTR and pCHl10, containing bacterial P-galactosidase sequences, were obtained from Pharmacia
4Towhom correspondence and reprint requests should be sent. Fax:
61 2 332 4876.
Abbreviations: AR, androgen receptor; CAT, chloramphenicol acetylIransferase; Dex, dexamethasone; DHT, dihydrotestosterone; Ez, 17P-oestradiol; ER, oestrogen receptor; FCS, foetal calf serum; MMTV LTR, mouse mammary tumour virus long-terminal repeat; Org 2058, Organon 2058,16a-ethyl-21-hydroxy-19-nor-4-pregnene-3,20dione; PR, progesterone receptor; RA, all-trans-retinoicacid.
Received: April 13, 1992 and in revised form July 10,1992.
779
ANDROGEN RECEPTOR IN BREAST CANCER
LKB. Sydney. Australia. Micro BCA protein assay reagents were purchased from Pierce, Rockford, IL. Cell lines Eighteen human breast-cancer cell lines including a line derived from breast milk, HBL-100, were examined in this study (Kcddel et a/., 1985, and references therein). MCF-7 cells were obtained from Dr. C.M. McGrath, Meyer L. Prentis Cancer Center, MI, and from E.G. and G . Mason Research Institute. Worcester, MA, the latter being designated MCF-7M in this study. T-47D, MDA-MB-231. HBL-100 and Hs0578T werc supplied by E.G. and G. Mason Research Institute; MDA-MB-361, BT 474, MDA-MB-134, MDA-MB-330 and BT 20 werc obtained from both E.G. and G. Mason Research Institute and the ATCC, Rockville, MD; and ZR-75-1, Du 4475, SK BR 3. MDA-MB-436, MDA-MB-453, MDA-MB468, BT 483 and BT 549 were supplied by the ATCC. A human prostate carcinoma line, LNCaP, obtained from Dr. J.S. Horoszewicz, Roswell Park Memorial Institute, Buffalo, NY (Horoszcwicz et a/., 1983), was used for comparison of AR concentration with the breast lines in this study. Stock cell cultures were maintained in RPMI 1640 medium supplemented as described previously with glutamine, insulin, buffers, antibiotic and 10% FCS (Reddel et al., 1985). In regulation and transfection experiments, phenol-red-free RPMI 1640 or DMEM respectively was supplemented as for stock culture, except that the FCS was charcoal-stripped and its concentration reduced from 10 to 5%.
A R niRNA expression Northern blot analysis of 20 k g total RNA from breast cell lines cultured in stock medium was as previously described (Hall et a/., 1990). The human AR cDNA was a 0.7 kb probe that encompasses most of the DNA-binding domain and approximately half the hormone-binding domain of the receptor (Tilley et al., 1989). A 30-base oligonucleotide complementary to rat 18s rRNA (Chan et a/., 1984) was used as a control probe for RNA loading. Densitometric data from autoradiograms of the Northern blots from separate experiments were combined by standardizing the area of hybridization of the MCF-7 cell line to 10 ODmm. AR and PR mRNA regulation T-47D, MCF-7M, MDA-MB-134 and MDA-MB-453 cells were cultured for at least 2 days in phenol-red-free medium supplemented with 5% charcoal-stripped FCS before the addition of ethanol or DMSO vehicle (control cells) or receptor-saturating concentrations of El, Org 2058, D H T M), with 0.1% final M), Dex M) or R A vehicle concentration. Total R N A was isolated from cells at various times after treatment and A R mRNA levels determined by Northern blot and hybridization with the 3'P-labeled AR cDNA. T-47D cells were cultured in phenol-red-free medium containing either 5 % FCS or 5% charcoal-stripped FCS, to test the effects of serum components on A R and PR mRNA levels. A 1.2-kb human PR cDNA encompassing most of the DNAbinding domain and part of the A / B domain of the receptor (Misrahi ef a f . , 1987) was used to re-probe the Northern blots of T-47D cells both treated with E2 and cultured in medium containing either 5% FCS or 5% charcoal-stripped FCS. ['HIDHT-binding assays
AR protein levels were assayed in cytosol fractions derived from T-47D, MDA-MB-134, MDA-MB-453 and LNCaP cells, cultured in phenol-red-free medium containing 5% charcoalstripped FCS and either grown overnight in serum-free medium or washed in serum-free medium before harvesting. No difference was found in the level of AR measured following these 2 preparations. Cytosol fractions were prepared as described previously (Alexander et a/., 1989). Binding was
measured with a single saturating concentration of ['HIDHT (20 nM) 2 100-fold molar excess of unlabeled DHT, at 4°C for 16-20 hr, and unbound ligand separated by dextran:charcoal adsorption as previously described (Reddel eta/., 1983). Exchangeable DHT-binding was assayed in whole cells by modifying the assay previously described for PR (Sutherland et a/., 1988) and using an incubation time similar to that used previously to assay A R (Grino ct a/., 1990). MDA-MB-453 and T-47D cells, cultured in phenol-red-free medium containing 5% charcoal-stripped FCS, were treated with vehicle or lo-'' M RA and assayed at 24,48,72 and 96 hr after treatment. The cells were washed with binding buffer (ie., serum-, insulin- and phenol-red-free medium containing 0.1% BSA), then incubated in binding buffer containing increasing concentrations of ['HIDHT (0.1-5 nM) ? 100-fold molar excess of unlabeled D H T at 37"C, for 1 hr. Specifically bound radioactivity was measured after removal of free label with 5% BSA and the data analyzed according to Scatchard (1949). Transient transfection of pMSG-CAT into MDA-MB-453 and T-47D cells M MDA-MB-453 and T-47D cells were treated with RA or vehicle for 24 hr before they were co-transfected with pMSG-CAT and pCHl10 by calcium-phosphate precipitation (Gorman, 1985). The cells were subjected to a I-min osmotic shock with 15% glycerol, 4 hr after the D N A precipitate was added, and the precipitate remained on the cells overnight. After transfection, 0.1-10 nM D H T was added and the RA or vehicle treatments also maintained, until the cells were harvested 48 hr later. Cells were lysed by freeze-thawing ( 3 ~and ) aliquots of the supernatant used to measure protein concentration with Micro BCA reagents (Pierce) P-galactosidase activity (Gorman, 1985) and chloramphenicol acetyltransferase (CAT) activity (Sleigh, 1986). RESULTS
A R expression The relative levels of A R mRNA in 18 human breast-cancer cell lines were measured by Northern analysis of 20 pg total RNA and A R + lines, expressing a single 10.5-kb mRNA, are represented in Figure 1. AR mRNA was detectable within 10 days of Northern blot exposure in cell lines which have previously been characterized as ER+ by ligand-binding assay (MCF-7, MCF-7M, T-47D, ZR-75-1, MDA-MB-361, BT474 and MDA-MB-134). A R mRNA, barely detectable in some E R - lines (HBL-100, MDA-MB-231, MDA-MB-330, BT 20 and Hs0578T) when Northern blots were exposed for 3 weeks with an intensifying screen, were classified as AR- (data not shown). A R mRNA was undetectable in the remaining E R cell lines: D u 4475, SK BR 3, MDA-MB-436, MDA-MB-468, B T 483 and B T 549, after 5 weeks exposure of the autoradiograms with an intensifyingscreen. One E R - cell line MDA-MB453, however, expressed comparatively abundant AR mRNA (Fig. lb). MDA-MB-453 cells assayed by exchangeable ligandbinding assays were confirmed E R - and PR- (data not shown). The level of cytosolic A R measured by ligand-binding assays was also greater in MDA-MB-453 cells compared with 2 of the other A R + breast cell lines, and was of a similar magnitude to that in the prostate carcinoma cell line LNCaP (Fig. 2). AR regulation Initially, regulation of AR mRNA in T-47D and MCF-7, both ER+, PR+ cell lines, was examined 6 hr after treatment with receptor-saturating concentrations of various agents known to regulate the other sex-steroid-hormone receptors (Fig. 3). Ez had no effect on AR mRNA expression in either cell line; Dex, Org 2058, D H T and R A reduced AR mRNA levels in T-47D cells; and Dex and D H T reduced AR mRNA levels in
780
HALL ETAL.
300 A
m
-E 200 %
Eo
.c Y
K
4
100
0
“ P
a
FIGURE 2 - Androgen receptor protein in human breast-cancer cells and LNCaP human prostate-carcinoma cells. AR was measured by ligand-binding in cytosols from 3 of the breast cell lines and LNCaP. Data represent the mean ( 2 range) of AR measured in 2 separate cytosol extracts, except for T-47D.
100
50
FIGURE 1 - Androgen receptor mRNA in human breast-cancer cell lines. (a) A representative Northern blot of 20 kg total RNA from 8 human breast-cancer cell lines hybridized with a 32Plabeled AR cDNA probe and with the 18s rRNA control probe. (b) Histogram of the mean (+ SE) AR mRNA levels in 20 pg total RNA from the cell lines presented above. RNA was extracted from the cell lines on 3 to 5 occasions. Autoradiograms were scanned with a densitometer and, after correction for loading, the data from separate blots were combined by standardizing the MCF-7 cell line to 10 ODmm. MCF-7 cells. MCF-7 cells were less sensitive to the progestin, Org 2058, possibly due to the fact that they have far fewer PRs than T-47D cells (Sutherland et al., 1988). MCF-7 cells were insensitive to RA treatment, and while we have investigated R A receptor expression in MCF-7 cells (Roman et al., 1992), the reason for this lack of sensitivity is not yet clear. The effect of E2 treatment on A R mRNA was examined at later times in T-47D and MCF-7 cells (Fig. 4), because the level of PR mRNA increases progressively with time after E2 treatment to 280% of control levels in T-47D cells by day 3 (Fig. 4). A R mRNA levels were unchanged up to 3 days after E2 treatment in both cell lines (Fig. 4).Following a change in culture medium from 5% FCS to 5% charcoal-stripped FCS, the level of AR mRNA increased, accompanied by a decrease in P R mRNA (Fig. 4, inset). This increase in AR mRNA was
0
E2
Dex
Org
DHT
RA
6 h treatment
FIGURE3 - Ligand specificity of androgen-receptor mRNA regulation in T-47D and MCF-7M cells. T-47D cells (open bars) and MCF-7M cells (hatched bars) were treated for 6 hr with M E2, lo,-’ M Dex, M Org 2058, M DHT, or M RA, as described in “Material and Methods”. AR mRNA levels in treated cells are expressed as ercentages of AR mRNA levels in vehicle-treated controls. “-478 data represent the mean ( 2 range or SE) of 3 experiments for all ligands except Dex (n = 2 ) . MCF-7M data represent the mean ( 5 range) of 2 experiments for all ligands except Ez (n = 3). also seen in MDA-MB-134 (234% control on day 3) and MDA-MB-453 cells (132% control on day 3, data not shown). In T-47D cells, RA (Fig. 5a), DHT (Fig. 5b) and Org 2058 (Fig. 5c) reduced A R mRNA levels compared with vehicletreated controls, extending the time course of the effects of these agents shown in Figure 3. A R mRNA was reduced
100
150: 50
-e
n
781
ANDROGEN RECEPTOR IN BREAST CANCER
A
A
c
E
0
0
8
Y
04
a z
0
K
1
i
24
48
B
E
K
DHT
n U E
T
0
K
a
1
2
3
Time (days) FIGURE4-Effect of oestrogen treatment on androgen and progesterone rece tor mRNA in T-47D and MCF-7M cells. AR mRNA in T-47D g p e n bars) and MCF-7M (closed bars) and PR mRNA levels in T-47D cells (hatched bars), treated with lo-@M E2are expressed as percentages of the respective mRNA levels in vehicle-treated controls. After 1 day of E2treatment in 2 separate experiments, the mean ( 2 range) AR mRNA level in each cell line was not significantly different from controls. (Inset) AR and PR mRNA levels in T-47D cells changed to medium containing 5% charcoal-stripped FCS were expressed as percentages of the respective mRNA levels in cells maintained in medium containing 5% FCS. The mean ( 2 SE) levels of AR and PR on days 1 to 7 and 4 to 7, respectively, after changing medium are shown.
A
a
E
K
U
O ! 0
I
1
24
48
I
i
0
24
48
C
maximally to approximately 40% control 12 hr after R A treatment (Fig. Sa), to between SO% and 70% control 24 hr after D H T treatment (Fig. 56) and to 30% control 12 hr after Org 2058 treatment (Fig. 5c), in T-47D cells. AR mRNA remained low 48 hr after these treatments. In MDA-MB-134 cells, R A and D H T also reduced A R mRNA levels (Fig. Sa and 6). However, in MDA-MB-453 cells, while D H T reduced
FIGURE5 -Effect of retinoic acid, dihydrotestosterone and Organon 2058 treatments on androgen receptor mRNA in 3 breast-cancer cell lines. T-47D (m), MDA-MB-134 (0) and cells were treated with (a) MDA-MB-453 (0) M RA, (6) M DHT or ( c ) or M Org 2058. AR mRNA levels in treated cells are expressed as percentages of AR mRNA levels in vehicle-treated controls. T-47D cells were examined in 2 to 5 experiments. No difference was observed between the effects of and M Org 2058, and data for both concentrations have been pooled. Data for MDA-MB-453 cells are from 2 to 4 experiments and for MDA-MB-134 from 2 experiments at 24 hr. (The range from the mean was smaller than the size of the symbol at 24 hr after RA treatment of MDA-MB-134 cells.)
Tlme (h)
782
HALL E T A L .
A 1.0
-
0.8
-
I
0 0
. 8
0.6
U
50
W
0.4
-
0.2
-
m
0
Y
0.0
I
0.0
0.1
I
0.2
I
0.3
1
0.4
Bound (nM)
O I
I
0.5
0
0
1
10
DHT concentration (nM) FIGURE7 - Effect of retinoic acid pretreatment of MDA-MB453 and T-47D cells on androgen induction of reporter gene activity. MDA-MB-453 (closed symbols) and T-47D cells open symbols), were treated with vehicle (0,0)or 10-6 M RA (&? 0) for 24 hr before being transfected with pMSG-CAT as described in “Material and Methods”. DHT (0-10-8 M) was added and CAT activity measured 48 hr later. The data are shown as fold induction above basal CAT activity.
B
Androgen responsiveness Transfection of a plasmid (pMSG-CAT) containing the androgen-responsive MMTV LTR linked to a CAT reporter gene was used to examine the effect of altered A R levels on androgen action. There was enhanced induction of CAT activity 48 hr after D H T addition in MDA-MB-453 cells pre-treated with R A and reduced induction of CAT activity by 0.0 .01 .02 .03 .W D H T in T-47D cells pre-treated with R A (Fig. 7). The concentration of D H T needed to achieve half-maximum-fold Bound (nM) induction of CAT activitv was not markedlv different in control and in RA-treated cells’in both cell lines i0.4 nM and 0.3 nM FIGURE6 - Ligand-binding analysis of androgen receptor in MDA-MB-453 and T-47D cells treated with retinoic acid. ( a ) respectively in MDA-MB-453 cells, and 0.5 nM and 0.25 nM MDA-MB-453 and (b) T-47D cells were treated with vehicle (0, respectively in T-47D cells), suggesting that the sensitivity of 0)or 10V M RA 0) for 24 hr, then changed to binding buffer CAT induction by D H T was similar. This finding is consistent containing 0-5 nM [)H]DHT 5 100-foldmolar excess of unlabeled with the similar DHT-binding affinities in both cell lines (Fig. DHT, and incubated at 37°C for 1 hr. Specifically bound radioac- 6). We observed no differences in growth or transfection tivity was measured after removal of free label with 5% BSA, and efficiency following these short-term treatments with RA and the data analyzed according to the method of Scatchard. D H T (data not shown).
(m,
the level of AR mRNA (Fig. 5b), R A produced an increase in AR mRNA (Fig. 5a). Org 2058 treatment was not examined in MDA-MB-134 or MDA-MB-453 cells because they are PR-. A low level of PR, which was not regulated by EZ,has been reported in MDA-MB-134 cells (Reiner and Katzenellenbogen, 1986). but we have not detected PR mRNA in these cells (Hall el al., 1990). A R sites assayed by exchangeable ligand-binding were increased in MDA-MB-453 cells and decreased in T-47D cells following 24 hr treatment with RA (Fig. 6), indicating an effect on A R protein corresponding to that observed on A R mRNA (Fig. 5). The effects of RA treatment seen at 24 hr were maintained to 96 hr (data not shown). In control MDA-MB453 cells there were 74,500 sites/cell, compared with 120,000 sites/cell following R A treatment, and the dissociation constant (K, 0.4 and 0.55 nM respectively) was unchanged. In control T-47D cells there were 16,000 sites/cell, compared with 10,200sitesicell following R A treatment, and the dissociation constant (K, 0.3 and 0.4 nM respectively) was unchanged.
DISCUSSION
Unlike E R and PR, relatively little information is available with respect to the regulation of A R expression in breastcancer cells. We have demonstrated that A R mRNA is expressed predominantly in E R + human breast-cancer cell lines. Six E R + cell lines expressed AR mRNA, while only 1 of 12 ER- cell lines expressed A R mRNA. We have shown previously that the incidence and levels of A R and P R mRNA were also positively correlated in breast-cancer cell lines (Hall et al., 1990). The co-expression of the sex-steroid-hormone receptors is, however, not absolute: for example, MDA-MB453 cells were found to be A R + ,ER- and PR-; and MDA-MB134 cells are A R + , E R + , and PR-. Among the A R + breast lines, MDA-MB-453 cells expressed the greatest amount of A R mRNA, and the concentration of cytosolic AR protein was of a similar order to that of LNCaP prostate carcinoma cells. The high level of A R in MDA-MB-453 cells could not be explained by amplification of the A R gene (data not shown). Regulation of A R expression also differed in MDA-MB-453 cells, in that R A up-regulated AR. This difference from the
ANDROGEN RECEPTOR IN BREAST CANCER
other 3 cell lines examined may be related to the absence of E R in MDA-MB-453 cells, and this is currently being examined. It has been shown that the A R gene is transcribed from the same start site in both LNCaP prostate-cancer cells and MCF-7 and T-47D breast-cancer cells, implying that the A R promoter is the same in both tissues (Tilley et a/., 1990). Regulation of A R in breast-cancer cells in comparison with findings in prostate and other tissues, however, suggests that there are some tissue-specific differences. Also, comparison of A R regulation with that of the other sex-steroid-hormone receptors, E R and PR, in breast-cancer cells indicates coordinate regulation by some agents, e g . , progestin; and selective difference from other regulatory agents, eg., AR, unlike PR, was not induced by oestrogen. A R mRNA expression was unchanged following treatment with oestrogen and was reduced following treatment with glucocorticoid, progestin, androgen and retinoid in T-47D cells. Although the level of A R mRNA was unchanged following oestrogen treatment, both a reduction in ARbinding sites in MCF-7 cells (Stover et al., 1987), and increases in AR-binding sites in prostate (Moore et al., 1979) and chick oviduct (Tokarz et al., 1979) following oestrogen treatment have been reported. The reasons for these differences in regulation of A R by oestrogen require further examination. The data on progestin regulation of A R mRNA adds to the finding that progestin treatment reduced AR-binding sites in MCF-7 and EFM-19 breast-cancer cells (Hackenberg et al., 1990). The effect of glucocorticoid on A R mRNA was potentially interesting in view of our previous finding that glucocorticoid receptor expression was inversely correlated with the expression of AR, E R and P R in breast-cancer cell lines (Hall et al., 1990), but the interrelationships of their regulation have not been examined further. The down-regulation of A R mRNA by its own ligand in breast-cancer cells was similar to the autologous down-regulation observed in male androgenresponsive tissues, including the prostate-carcinoma cell line LNCaP (Quarmby et a/., 1990; Shan et a/., 1990; Krongrad et aL, 1991). The failure by Shapiro and Lippman (1985) to see an effect on A R binding sites, 6 hr after androgen treatment in MCF-7 cells, may relate to the half-life of A R protein, or differential effects of androgens on A R mRNA and protein concentrations (Krongrad et al., 1991). The down-regulation of AR expression by retinoic acid is of interest and potential mechanistic significance, since the mechanism by which retinoids inhibit carcinogenesis in breast, prostate and other tissues is not well defined (Moon and Itri, 1984). Retinoids also down-regulate PR mRNA (Clarke et a/., 1990, 1991), but do not affect E R mRNA (data not shown) or their own a and y receptors in breast-cancer cells (Roman et al., 1992). In changing from medium containing 5% FCS to medium containing 5% charcoal-stripped FCS, we found that the level of AR mRNA increased and at the same time PR mRNA decreased. These changes may be coincidental or the outcome of more than one regulatory factor. In addition to steroids, many factors in serum could be involved in this regulation.
783
Insulin, for example, is known to increase PR levels in MCF-7 cells (Katzenellenbogen and Norman, 1990). Since insulin (10 pg/ml) was always present as a supplement in our culture medium, however, it is unlikely to be the mediating factor. Previous reports of increased AR expression after changing the medium to charcoal-stripped FCS (Stover et a/., 1987) or serum-free medium (Quarmby et al., 1990) have attributed the difference to removal of oestrogen and androgen respectively. Our demonstration that treatment with androgen decreased A R mRNA levels supports the latter conclusion, but additionally, glucocorticoids, progestins and retinoids in FCS could be contributing to this effect on AR. The changes in fold induction of CAT activity measured after the level of A R had been modulated by R A pretreatment demonstrate a clear link between androgenreceptor concentration and androgen responsiveness. Despite the competing regulatory actions of R A (up-regulation of AR) and D H T (down-regulation of AR) in MDA-MB-453 cells, the effect of R A on A R mRNA levels predominates in this combined treatment regimen. In separate experiments (data not shown) with the same regimen as used in the transfection experiments of 24 hr (1 FM) RA pre-treatment followed by 48 hr combined R A and 10 nM D H T treatments, A R mRNA levels were 178 ? 13% control ( D H T alone) levels (n = 2). Thus, when MDA-MB-453 cells were pre-treated for 24 hr with RA, we observed an increased magnitude of CAT induction by DHT. In control T-47D cells which possess a much lower concentration of A R than MDA-MB-453 cells, the fold induction of CAT activity was less, and pre-treatment with RA, dampened this induction. It has previously been recognized in stable transfectants expressing different levels of glucocorticoid receptor that steroid-responsiveness is proportional to receptor concentration (Vanderbilt et al., 1987), and in other studies from this laboratory it has been demonstrated that down-regulation of E R by Org 2058 (Alexander et al., 1990) and P R by R A (Clarke et al., 1991) in T-47D cells reduced oestrogen and progestin-responsiveness respectively. To summarize, we have found A R to be expressed predominantly in E R + human breast-cancer cells, although one of 12 E R - cell lines, MDA-MB-453, was A R + . The steroids and retinoid examined in this study reduced A R expression, with the exception that RA up-regulated AR expression in MDAMB-453 cells. Androgen responsiveness measured by a transfected reporter gene was dependent on A R concentration. Endogenous genes may be similarly sensitive to androgenreceptor concentration, as is suggested by the observation that some cases of androgen-insensitivity syndrome are attributable to reduced A R levels (Grino et al., 1990; Marcelli et al., 1991). ACKNOWLEDGEMENTS
We are grateful to Dr. P. Chambon, Strasbourg, France, for providing the hPR cDNA probe. This research was supported by the National Health and Medical Research Council of Australia. R.E.H. is a recipient of a NHMRC Biomedical Postgraduate Scholarship.
REFERENCES
I.E., CLARKE,C.L., SHINE,J. and SUTHERLAND, R.L., ALEXANDER, Progestin inhibition of progesterone receptor gene expression in human breast-cancer cells. Mol. Endocrinol.. 3,1377-1386 (1989).
androgen and glucocorticoid receptors in human breast cancer. Cancer Res., 39,1447-1454 (1979). BRENTANI, M.M., FRANCO,E.L., OSHIMA, C.T.F. and PACHECO, M.M., ALEXANDER, I.E., SHINE,J. and SUTHERLAND, R.L., Progestin regula- Androgen, estrogen and progesterone receptor levels in malignant and tion of estrogen receptor messenger RNA in human breast-cancer benign breast tumors: a multivariate analysis approach. Int. J. Cancer, 38,637-642 (1986). cells. Mol. Endocn'nol., 4,821-828 (1990). CHAN,Y.-L., GUTELL,R., NOLLER.H.F. and WOOL, I.G., The ALLEGRA, J.C., LIPPMAN, M.E., THOMPSON, E.B., SIMON, R., BARLOCK. nucleotide sequence of a rat 18 S ribosomal ribonucleic acid gene and A., GREEN,L., HUFF.K.K., DO, H.M.T. and AITKEN,S.C., Distribu- a proposal for the secondary structure of 18 S ribosomal ribonucleic tion, frequency, and quantitative analysis of estrogen, progesterone, acid. J. biol. Chem.. 259,224-230 (1984).
784
HALL E T A L .
CLARKE,C.L.. GRAHAM, J., ROMAN,S.D. and SUTHERLAND, R.L., receptor deduced from cloned cDNA. Biochem. Biophys. Res. Comm., Direct transcriptional regulation of the progesterone receptor by 143,740-748 (1987). retinoic acid diminishes progestin responsiveness in the breast-cancer MOON, R.C. and ITRI,L.M., Retinoids and cancer. In: M.B. Sporn, cell line T-47D. J. bid. Chem., 266, 18969-18975 (1991). A.B. Roberts and D.S. Goodman (ed.), The Retinoids, pp. 327-371, CLARKE, C.L., ROMAN. S.D., GRAHAM, J., KOGA,M. and SUTHERLAND,Academic Press, London (1984). R.L., Progesterone receptor regulation by retinoic acid in the human J.D., Regulation of cytoplasbreast-cancer cell line T-47D. J. biol. Chern., 265,12694-12700 (1990). MOORE,R.J., GAZAK,J.M. and WILSON, mic dihydrotestosterone binding in dog prostate by l7b-estradiol. J. GORMAN, C., High efficiency gene transfer into mammalian cells. In: elin. Invest., 63,351-357 (1979). D.M. Glover (ed.), DNA Cloning VlI. A practical approach, pp. POULIN, R., BAKER,D. and LABRIE,F., Androgens inhibit basal and 143-165, IRL Press, Oxford (1985). estrogen-induced cell proliferation in the ZR-75-1 human breastGRINO,P.B., GRIFFIN, J.E. and WILSON,J.D., Androgen resistance due cancer cell line. Breast Cancer Res. Treat., 12,213-225 (1988). to decreased amounts of androgen receptor: a reinvestigation. J. QUARMBY, V.E., YARBROUGH, W.G., LUBAHN, D.B., FRENCH, F.S. and Steroid Biochem., 35,647-654 (1990). WILSON,E.M., Autologous down-regulation of androgen receptor HACKENBERG, R., HOFMANN, J., HOLZEL,F. and SCHULZ,K.-D., messenger ribonucleic acid. Mol. Endocrinol., 4,22-28 (1990). Stimulatory effects of androgen and antiandrogen on the in vitro REDDEL,R.R., MURPHY, L.C., HALL,R.E. and SUTHERLAND, R.L., proliferation of human mammary carcinoma cells. J. Cancer Res. clin. Differential sensitivity of human breast-cancer cell lines to the Oncol., 114,593-601 (1988). growth-inhibitory effects of tamoxifen. Cancer Rex, 45, 1525-1531 HACKENBERG, R., HOFMANN, J., WOLFF,G., HOLZEL,F. and SCHULZ, (1985). K.-D., Down-regulation of androgen receptor by progestins and L.C. and SUTHERLAND. R.L., Effects of interference with estrogenic or androgenic stimulation of mammary REDDEL,R.R., MURPHY, carcinoma cell growth. J. Cancer Res. elin. Oncol., 116,492498 (1990). biologically active metabolites of tamoxifen on the proliferation kinetics of MCF-7 human breast-cancer cells in vitro. Cancer Res., 43, J., KUNZMANN, R., 46184624 (1983). HACKENBERG, R., LUVCHENS,S., HOFMANN, HOLZEL,F. and SCHLILZ, K.-D., Androgen sensitivity of the new REINER, G.C.A. and KATZENELLENBOGEN, B.S., Characterization of human breast-cancer cell line MFM-223. Cancer Res., 51, 5722-5727 estrogen and progesterone receptors and the dissociated regulation of (1991). growth and progesterone receptor stimulation by estrogen in MDAI.E., SHINE,J., CLARKE, C.L. MB-134 human breast-cancer cells. Cancer Rex, 46,1124-1131 (1986). HALL,R.E., LEE, C.S.L., ALEXANDER, and SUTHERLAND, R.L.. Steroid-hormone-receptor-gene expression in ROMAN,S.D., CLARKE,C.L., HALL, R.E., ALEXANDER, I.E. and human breast-cancer cells: inverse relationship between oestrogen R.L., Expression and regulation of retinoic acid recepand glucocorticoid receptor messenger RNA levels. Int. J. Cancer, 46, SUTHERLAND, tors in human breast-cancer cells. Cancer Res., 52,2236-2242 (1992). 1081-1087 (1990). G., The attractions of proteins for small molecules and HOROSZEWICZ, J.S.,LEONG, S.S., KAWINSKI. E., KARR,J.P., ROSENTHAL, SCATCHARD, H., CHU,T.M., MIRAND, E.A. and MURPHY, G.P., LNCaP model of ions.Ann. N.Y. Acad. Sci., 51,660-672 (1949). human prostatic carcinoma. Cancer Rex, 43, 1809-1818 (1983). SHAN,L.-X., RODRIGUEZ, M.C. and JANNE,O.A., Regulation of INGLE.J.N., TWITO,D.I., SCHAID, D.J., CULLINAN, S.A., KROOK,J.E., androgen-receptor protein and mRNA concentrations by androgens in MAILLIi\RD, J.A., TSCHETTER.L.K., LONG, H.J., GERSTNER, J.G., rat ventral prostate and seminal vesicles and in human hepatoma cells. WINDSCHITL, H.E., LEVIT, R. and PFEIFLE,D.M., Combination Mol. Endocrinol., 4,1636-1646 (1990). hormonal therapy with tamoxifen plus fluoxymesterone versus tamox- SHAPIRO. E. and LIPPMAN, M.E., Onset of androgen action in MCF-7 ifen alone in post-menopausal women with metastatic breast cancer. human breast-cancer cells is not accompanied by receptor depletion. J. An updated analysis. Cancer, 67,886-891 (1991). Steroid Biochem., 22, 15-20 (1985). KATZEKELLENBOGEN, B.S. and NORMAN, M.J., Multihormonal regula- SLEIGH,M.J., A non-chromatographic assay for expression of the tion of the progesterone receptor in MCF-7 human breast-cancer cells: chloramphenicol acetyltransferase gene in eucaryotic cells. Anal. interrelationships amon insuliniinsulin-like growth factor-1, serum Biochem., 156,251-256 (1986). and estrogen. Endocn'nokgy, 126,891-898 (1990). STOVER,E.P., KRISHNAN, A.V. and FELDMAN, D., Estrogen downKRONGRAD. A,, WILSON, C.M., WILSON,J.D., ALLMAN,D.R. and regulation of androgen receptors in cultured human mammary cancer MCPHAUL, M.J., Androgen increases androgen-rece tor protein while cells (MCF-7). Endocn'nology, 120,2597-2603 (1987). decreasing receptor mRNA in LNCaP cells. Mol. ce!. Endocrinol., 76, SUTHERLAND, R.L., HALL,R.E., PANG,G.Y.N., MUSGROVE, E.A. and 79-88 (1991). CLARKE, C.L.. Effect of medroxyprogesterone acetate on proliferation S. and THORSEN. T., Improved measurement and cell-cycle kinetics of human mammary carcinoma cells. Cancer LEA.O.A.. KVINNSLAND. of androgen receptors 'in human breast cancer. Cancer Res., 49, Res., 48,5084-5091 (1988). 7162-7167 (1989). TEULINFS, F.A.G., VAN GILSE,H.A., HENKELMAN, M.S., PORTENGEN, LOPEZ,M., Cyproterone acetate in the treatment of metastatic cancer H. and ALEXIEVA-FIGUSCH, J., Estrogen, androgen, glucocorticoid and of the male breast. Cancer, 55,2334-2336 (1985). progesterone receptors in progestin-induced regression of human MANNI,A,, ARAFAH,B.M. and PEARSON, O.H., Androgen-induced breast cancer. Cancer Rex, 40,2557-2561 (1980). remissions after antiestrogen and hypophysectomy in Stage-IV breast TILLEY, W.D., MARCELLI, M. and MCPHAUL, M.J., Expression of the cancer. Cancer, 48,2507-2509 (1981). human androgen receptor gene utilizes a common promoter in diverse W.D., ZOPPI, S., GRIFFIN, J.E., WILSON, J.D. human tissues and cell 1ines.J. bid. Chern., 265,13776-13781 (1990). MARCELLI. M., TILLEY, and MCPHAUL, M.J.. Androgen resistance associated with a mutation TILLEY. W.D., hfARCELL1, M., WILSON,J.D. and MCPHAUL,M.J., of the androgen receptor at amino acid 772 (Arg-Cys) results from a Characterization and expression of a cDNA encoding the human combination of decreased messeneer ribonucleic acid levels and androgen receptor. Proc. nat. Acad. Sci. (Wash.), 86,327-331 (1989). impairment of receptor function. -J. elin. endocn'nol Metab., 73, TOKARZ. R.R.. HARRISON. R.W. and SEAVER. S.S.. The mechanism of 318-325 (1991). androgen and estrogen synergism in the chick oviduct. J. hiol. Chern., MERCER, R.J., LIE,T.H., RENNIE, G.C., BENNETT, R.C. and MORGAN, 254,9178-9184 (1979). F.J., Hormone-receptor assays in breast cancer. A five-year experiTORMEY, D.C., LIPPMAN, M.E., EDWARDS, B.K. and CASSIDY, J.G., ence. Med. J. Aust., 1,365-369 (1983). Evaluation of tamoxifen doses with and without fluoxymesterone in MILLER, W.R.. TELFORD, J., DIXON, J.M. and HAWKINS, R.A., Andro- advanced breast cancer. Ann. Intern. Med., 98, 139-144 (1983). gen-receptor activity in human breast cancer and its relationship with G. and MAASS, H., Specific binding of estradiol and dihydrotesoestrogen and progestogen receptor activity. Europ. J. Cancer elin. TRAMS, tosterone in human mammary cancers. Cancer Res., 37, 258-261 Oncol., 21,539-542 (1985). (1977). J. MILLWARD, M.J., CANTWELL. B.M.J., DOWSETT,M., CARMICHAEL, VALYANI, S.H., ROBERTS, J.V., COLLE~TA, A.A. and IQBAL,M.J., and HARRIS, A.L., Phase-I1 clinical and endocrine study of Anandron (RU-23908) in advanced post-menopausal breast cancer. Brit. J. Sex-steroid receptors in human breast cancer. Incidence and interrelationships.Anricancer Res., 7, 1281-1284 (1987). Cancer, 63,763-764 (1991). J.N.. MIESFELD, R., MALER,B.A. and YAMAMOTO, K.R., MISRAHI, M., ATGER,M., D'AURIOL, L., LQOSFELT, H., MERIEL,C., VANDERBILT, FRIDLANSKY, F., GUIOCHON-MANTEL, A., GALIBERT, F. and MILGROM, Intracellular receptor concentration limits glucocorticoid-dependent enhancer activity. Mol. Endocrinol., 1,68-74 (1987). E., Complete amino acid sequence of the human progesterone