Richard Poulin, Denis Baker and Fernand Labrie

MRC Group in Molecular Endocrinology, Laval University Medical Center, Quebec G1V 4G2, Canada.

F. Labrie, MRC Group in Molecular Endocrinology, Laval University Medical Center, Quebec G1V 4G2, Canada;

Summary

This study describes the inhibitory effect of 5α-dihydrotestosterone (5α-DHT) and its precursors testosterone (T) and androst-4-ene-3,17-dione (Δ⁴-DIONE) on the growth of the estrogen-sensitive human breast cancer cell line ZR-75-1. In the absence of estrogens, cell proliferation measured after a 12-day incubation period was 50-60% inhibited by maximal concentrations of 5α-DHT, T, or Δ⁴-DIONE with half-maximal effects (IC₅₀ values) observed at 0.10, 0.15 and 15 nM, respectively. This growth inhibition by androgens was due to an increase in generation time and a lowering of the saturation density of cell cultures. The antiestrogen LY156758 (300 nM) induced 25-30% inhibition of basal cell growth, its effect being additive to that of 5α-DFIT. The mitogenic effect of InM estradiol (E₂) was completely inhibited by increasing concentrations of 5α-DFIT with a potency (IC₅₀ = 0.10 nM) similar to that measured when the androgen was used alone. Ez had a more rapid effect on cell proliferation than 5α-DHT, the latter requiring at least 5 to 6 days to exert significant growth inhibition. As found in the absence of estrogens, maximal inhibition of cell proliferation in the presence of E₂ was achieved by the combination of the antiestrogen and 5α-DHT. Supraphysiological concentrations of E₂ (up to 1μM) were needed to completely reverse the growth inhibitory effect of a submaximal concentration of 5α-DHT (1 nM). The antiproliferative effect of androgens was competitively reversed by the antiandrogen hydroxyflutamide, thus indicating an androgen receptor- mediated mechanism. The present data suggest the potential benefits of an androgen-antiestrogen combination therapy in the endocrine management of breast cancer.

Key words: androgens, antiestrogen, antiandrogen, androgen receptor, estrogen receptor, breast cancer.

---

Introduction

Androgens such as testosterone propionate [1-3], fluoxymesterone [4, 5], and calusterone [6] have long been used in the adjuvant therapy of breast cancer with an efficacy comparable to that achieved with other types of endocrine manipulations [3, 7-9]. Since tumor regression induced by treatment with androgens is not restricted to premenopausal patients [7], it is likely that the therapeutic activity of androgens is not limited to an inhibitory effect on gonadotropin secretion.

While the presence of androgen receptors has been documented in normal [10,11] and neoplastic [11] breast tissue, as well as in several established breast cancer cell lines [12], very little is known about their functional significance. In the widely used in vitro model of estrogen-responsive human breast cancer, namely the MCF-7 cell line, phar-macological concentrations (0.1-1μM) of 5α-di-hydrotestosterone (5α-DHT) are mitogenic [13] and induce the secretion of a 52,000 Da glycoprotein, a protein known to be specifically regulated by estrogens [14]. This effect of 5α-DHT has been shown to be mediated by the low-affinity binding of high concentrations of androgens to the estrogen receptor [15,16]. On the other hand, physiological (0.1-10 nM) concentrations of androgens can coun-teract induction of the progesterone receptor by 17βestradiol (E₂) in MCF-7 cells through an an-drogen receptor-mediated mechanism [17, 18]. Moreover, in the T47-D human breast cancer cell line, 5α-DHT specifically induces the secretion of several proteins, an effect which is reversed by the antiandrogen flutamide [19, 20]. Androgens have also been found to modulate the number of specific prolactin binding sites in another human breast cancer cell line [21].

Although the above-mentioned studies indicate that breast cancer cells possess functional androgen receptors mediating various biochemical responses, no specific effect of androgens on cell proliferation has yet been reported. In the present study, we present evidence that androgens strongly inhibit proliferation of the well-characterized, es-trogen-sensitive human breast cancer cell line ZR-75-1 [22]. Moreover, we have used the pure antiandrogen hydroxyflutamide (OH-FLU) [23, 24] to assess the specificity of the growth-inhibitory action of androgens. Finally, we have compared the ability of androgens and of the antiestrogen [6-hydroxy-2-(4-hydroxyphenyl)benzo[b]thien-3- yl], (4-[2-(10piperidinyl) ethoxy]-phenyl) metha- none hydrochloride (LY156758) [25,26] to influen-ce basal cell growth and to counteract the mitogenic effect of E₂.

Materials and methods

Chemicals

All media and supplements for cell culture were purchased from Sigma, except for fetal bovine se-rum which was obtained from Hyclone (Logan, UT). The antiandrogen hydroxyflutamide (OH- FLU, SCH 16423) was kindly supplied by Drs J. Nagabhushan and R. Neri (Schering Corporation, Kenilworth, NJ). The antiestrogen LY156758 was generously provided by Dr. J.A. Clemens (Lilly Research Laboratories, Indianapolis, IN). 17β [2,4,6,7-₃H]estradiol (specific activity, 110 Ci/ mmol) was purchased from Amersham (Arlington Heights, IL). [ 17α.-methylO³H]methvltrienolone (R1881; 17β-hydroxy, 17α-methyl-4,9,ll-estra- trien, 3-one; specific activity, 87 Ci/mmol) and un-labeled R1881 were obtained from New England Nuclear (Lachine, Quebec, Canada). Steroids were obtained from Steraloids (Pawling, NY).

Maintenance of stock cultures

The ZR-75-1 human breast cancer cell line (83rd passage) was obtained from the American Type Culture Collection (Rockville, MD). The cells were routinely cultured in phenol red-free [27, 28] RPMI1640 medium supplemented with 10 nM E₂, 15 mM Hepes, 2mM L-glutamine, ImM sodium pyruvate, 100μg penicillin per ml, 100/rg strepto-mycin sulfate per ml, and 10% (v/v) fetal bovine serum (FBS), in a water-saturated atmosphere of 95% air: 5% C0₂ at 37°C. The cell cultures used for the experiments herein described were between passages 90 and 102.

Cell growth experiments

Stock cultures in their logarithmic growth phase were harvested with 0.05% trypsin/0.02% EDTA (w/v) in Hanks’ balanced salts solution and re-suspended in E₂- and phenol red-free RPMI 1640 medium containing 5% (v/v) dextran-coated char-coal (DCC)-treated FBS and 500 ng of bovine in-sulin per ml [29], but otherwise supplemented as described above for maintenance of stock cultures. In some experiments, cells were cultured for 4 weeks in this medium (hereafter referred to as SD medium) prior to harvesting, in order to maximize estrogenic stimulation of cell growth, as explained in the ‘Results’ section. Cells were plated in 24-well Linbro culture plates (Flow Laboratories) at a final density of 0.5-4.0 x 10⁴ cells/well.

Fourty-eight hours after plating, fresh SD medi-um containing the indicated concentrations of ste-roids or steroid antagonists was added. The final concentration of ethanol used for the addition of test substances did not exceed 0.12% (v/v), and had no significant effect on cell growth and morphology. The incubation media were replaced every other day and cells were harvested by trypsiniza- tion after 12 days of treatment, unless otherwise indicated. Cell number was determined with a Coulter Counter (model ZM).

Specific uptake of [₃H]E₂ and [₃H]R1881 by intact ZR-75-1 cell monolayers

The relative binding affinity (RBA) of steroids for androgen and estrogen specific binding sites was assessed by measuring the effect of increasing con-centrations of competitor on the uptake of [₃H]R1881 and [₃H]E₂, respectively, by ZR-75-1 cells in monolayer culture [18,29,30]. Briefly, cells from stock cultures were trypsinized, resuspended in SD medium at a density of 2-5 x 10⁴ cells/ml, and distributed to 24-well culture plates. Cultures were grown to a final yield of about 3-5 x 10⁵ cells/well in SD medium (±1 nM E₂, in some experiments). Growth medium was then replaced with 0.5 ml of phenol red-free RPMI1640 medium (supplemented with 2mM L-glutamine, ImM sodium pyruvate, 15 mM Hepes, and 0.1% (w/v) fatty acid acid-free bovine serum albumin) containing the indicated concentration of competitor plus either 2.5 nM [₃H]E₂ or 3.0nM [₃H]R1881 (for the determination of estrogen and androgen specific binding sites, respectively). Triamcinolone aceto- nide (4.5μM) was added for the measurement of [₃H]R1881 uptake in order to block binding to the progesterone receptor. Incubations with [₃H]R1881 and [₃H]E₂ were stopped after lh and uptake of radioligand measured as described [18, 30]. The apparent dissociation constant (K_D) and the number of androgen-specific binding sites per cell (Bmax) were estimated using Scatchard analysis [31] by adding increasing concentrations of [₃H]R1881 (0.1-4 nM) to monolayer cultures, plus or minus a 200-fold excess of unlabeled R1881 to account for non-specific uptake.

Calculations and statistical analyses

Apparent IC₅₀ values were calculated using an iter-ative least squares regression [32], while apparent inhibition constants (Ki values) were calculated according to Cheng and Prusoff [33]. Mean gener-ation times were calculated according to a least- squares regression program on triplicate log-trans-formed cell number values measured on at least 6 time intervals. Statistical significance was calculated according to the multiple-range test of Duncan- Kramer [34]. Values are presented as means ± SEM of measurements obtained from triplicate incubations. When no bar is shown, the SEM is smaller than the symbol used.

Results

As illustrated in Fig. 1, a 12-day incubation with increasing concentrations of 5α-DHT had a biphasic effect on the proliferation of ZR-75-1 cells incubated in the absence of estrogens in phenol redfree medium. Concentrations of 5α-DHT between 0.01 and 10 nM inhibited basal cell growth up to a maximum of 50% in a dose-dependent manner, a half-maximal effect (IC₅₀) being observed at about O.10nM 5α-DFlT. When concentration of the an-drogen was increased from 10 to 2000 nM, the am-plitude of the inhibitory effect gradually decreased, although cell number remained below control. A progressive decrease in cell growth was again ob-served at concentrations of 5α-DHT exceeding 2μM.

Figure 1. Inhibitory effect of increasing concentrations of 5α-DHT on the proliferation of ZR-75-1 human breast cancer cells in culture and its prevention by the antiandrogen OH-FLU. Forty- eight hours after plating (initial cell density = 1.0 x 10⁴ cells/ well), 5α-DHT was added to cell cultures at the indicated concentrations in the presence (□, ■) or absence (O, •) of 1 nM E₂. Cells also received 3 piM OH-FLU (#, ■) or the vehicle alone (O, □). The data are presented as means ± SEM of triplicate determinations from a representative experiment.

While the addition of 1 nM E₂ induced a 2-fold increase in cell number, the effect of the estrogen was completely prevented when 5α-DFIT was added up to 10 nM with an IC₅₀ value again seen at O.lOnM. Inhibition of the mitogenic effect of E₂ then remained on a plateau between 10 and 2000 nM 5α-DHT. The estrogen, however, prevented the growth inhibition induced by 5α-DHT alone up to 2 /U.M 5α-DHT, while higher concentrations of the androgen decreased cell number to a level similar to that observed in the absence of estrogen.

It can also be seen in Fig. 1 that the inhibitory effect of 5α-DHT was competitively neutralized by coincubating ZR-75-1 cells with the antiandrogen OFI-FLU (3 μM) in control as well as in E₂-treated cells, the IC₅₀ value of DHT action being increased from 0.10 to about 5nM. On the other hand, the sole addition of OH-FLU had no significant effect on the growth of ZR-75-1 cells.

The effect of two other physiologically important androgens, namely testosterone (T) and its precursor androst-4-ene-3,17-dione (Δ⁴-DIONE), on the proliferation of ZR-75-1 cells was next examined. Increasing concentrations of T (up to 10 nM) maximally decreased cell number by about 50% with a half-maximal effect measured at about 0.15 nM (Fig. 2). A modest but significant decrease in the inhibitory effect of T was apparent at concentrations of the androgen ranging between 0.2 and 2μM. Inhibition of cell proliferation was also in-duced by Δ⁴-DIONE, although the potency of this steroid was much lower (IC₅₀ ~ 15 nM) than that observed with 5α-DHT or T. It can also be seen in Figure 2 that the maximal inhibition achieved with Δ⁴-DIONE is approximately 30% lower than that achieved by 5α-DHT or T. Both T- and Δ⁴- DIONE-induced inhibitory effects were efficiently counteracted by the antiandrogen OH-FLU, thus indicating that their action is also mediated by the androgen receptor. From the above-described ex-periments, the apparent inhibition constant (Ki) of OH-FLU action was estimated [33] at 107 ± 31 nM (mean ± SEM from 3 independent experiments), a value in close agreement with that reported for the interaction of the antiandrogen with the androgen receptor in several tissues [24].

Figure 2. Effect of increasing concentrations of T or Δ⁴-DIONE on the proliferation of ZR-75-1 breast cancer cells in culture and its reversal by the antiandrogen OH-FLU. Forty-eight hours after plating (1.0 X 10⁴ cells/well), T (O, •), or Δ⁴-DIONE (□, ■) was added at the indicated concentrations in the presence (•, ■) or absence (control, O, □) of 3μM OH-FLU.

We next studied the time course of 5α-DHT action on ZR-75-1 cell growth in the presence or absence of E₂. When cells were initially plated at a density similar to that used in experiments shown in Figs. 1 and 2, the inhibitory effect of 5α-DHT (10 nM) on the growth rate became significant following a 5- to 6-day period in both the presence and absence of InM E₂ (Fig. 3A). While E₂ decreased the mean generation time from 71 ± 4 to 57 ± 2 h and increased saturation density by about 2.2-fold, a 47% lower plateau was observed in DHT-treated cells, which had a generation time of 89 ± 4h. In the presence of E₂, 5α-DHT-treated cells had an initial growth rate similar to that of cells incubated with E₂ alone, their proliferation rate progressively decreasing after 6 days to reach cell number similar to control values after 12 days in culture. Thus, 5α-DHT decreased not only the growth rate, but also the degree of confluency reached by both control and Entreated ZR-75-1 cell cultures. It is also apparent that the time required by androgens to inhibit cell growth is longer than that needed by E₂ to exert its mitogenic effect.

Figure 3. Time course of the effect of 5α-DHT and/or E₂ on the proliferation of ZR-75-1 cells. A) Cells were plated at 1 x 10⁴ cells/2.0-cm² well and 48h later (zero time), 1 nM E₂ (•), 10 nM 5α-DHT (□), or both steroids (■) were added and cell number determined at the indicated times. Control cells received the ethanol vehicle only. B) Same as in A, except that the initial density was 5.0 x 10³ cells/2.0-cm² well.

Even more striking antiproliferative effects of 5α-DFIT were observed when ZR-75-1 cells were initially plated at lower densities. As illustrated in Fig. 3B, addition of the androgen virtually stopped net cell growth after approximately one population doubling when cells were initially plated at 2.5 x 10³/cm². Under these conditions, control and E₂- treated cells had not yet reached a plateau at the end of a 25-day incubation period. The presence of 5α-DHT abolished the net mitogenic effect of the estrogen at all time intervals examined.

We then studied in detail the influence of cell density on the effects of E₂ and DHT. As shown in Fig. 4, increasing the initial inoculum allowed a higher initial growth rate and a much shorter lag period, thus resulting in a smaller relative decrease in cell number and saturation density at a fixed harvesting time following incubation with the an-drogen. This effect was especially marked in the case of E₂-treated cells, where the effect of 5α- DHT measured at 12 days became non-significant at initial cell densities exceeding 1.5-2 x 10⁴ cells/ cm². These results are consistent with the observation that the action of estrogens on cell growth is exerted faster than that of androgens, and that Entreated cultures were reaching confluency within 6 to 7 days at the highest cell densities studied (data not shown).

Figure 4. Influence of seeding density on the effect of androgens and estrogens on the proliferation of ZR-75-1 cells measured after a 12-day incubation period. Cells were plated at the in-dicated density (per 2.0-cm² well) and treated for 12 days with no steroid (O), InM E₂ (•), 10 nM 5α-DHT (□), or both steroids (■). Cell number was determined at the end of the incubation period.

In order to better understand the opposite influ-ences of androgens and estrogens on the proliferation of ZR-75-1 cells, we next examined the ability of increasing concentrations of E₂ to prevent growth inhibition induced by 5α-DFlT and/or the antiestrogen LY156758. In preliminary experiments, we observed that following prolonged deprivation (4 weeks) of ZR-75-1 cells from estrogenic influence in SD medium, the maximal amplitude of the E₂-induced mitogenic effect was increased. Using such a protocol, which allows an optimal stimulatory effect of E₂, increasing concentrations of E₂ induced a maximal 4-fold stimulation of cell proliferation, a half-maximal increase in cell number being observed at approximately 8μM E₂ (Fig. 5α). Addition of a submaximal concentration of 5α-DHT alone (InM) inhibited basal cell growth by 54%. The androgen effect was then reversed by E₂ in a biphasic manner, an intermediary plateau of the stimulatory estrogenic effect being observed between 0.3 and 10 nM E₂. This plateau corresponds to an approximately 75% inhibition by 5α- DHT of the maximal estrogenic stimulation of cell growth. At E₂ concentrations >10 nM, there was a second rise in growth of androgen-treated cells which reached values similar to those observed in the absence of 5α-DHT at 1 μM E₂. The presence of OH-FLU (1 μM). while having no effect by itself on the mitogenic effect of E₂, completely reversed the androgen-induced inhibition of cell growth.

Figure 5. Effect of increasing concentrations of E₂ on the proliferation of ZR-75-1 cells, and its inhibition by 5α-DHT and/or the antiestrogen LY156758. Prior to the experiment, cells were cultured for 4 weeks in SD medium. Cells were then plated at 8.0 x 10³ cells/well in fresh medium of identical composition. A) 48 hours after plating, E₂ was added at the indicated concentrations alone (control, O), or concomitantly with InM 5α-DHT (•), 1 μM OH-FLU (□), or both compounds (■). B) Same as in A, except that 200nM LY156758 was also added alone (A) or in the presence of 1 nM 5α-DHT (•), 1 μM OH-FLU (□), or both compounds (■).

LY156758 (200 nM) competitively blocked stimulation of cell proliferation by increasing concentrations of E₂ (Ki of LY156758 = 0.14 nM) and caused a maximal 25% decrease in basal cell number (Fig. 5B). In the absence of E₂, the simultaneous presence of both LY156758 and 5α-DHT slightly enhanced the inhibition induced by the androgen alone (54 and 63% for 5α-DHT and 5α- DF1T + LY156758, respectively). Moreover, the antiestrogen completely counteracted the partial reversal by E₂ (up to about 10 nM) of the growth- inhibitory effect of 5α-DHT in a competitive man-ner. Again, OH-FLU (1 μM) selectively and completely abolished the effect of low 5α-DHT on E₂- sensitive cell proliferation.

Additivity of the inhibitory effects of 5α-DHT and LY156758 is even more clearly illustrated in Fig. 6. In the absence of E₂, there was a dose- dependent inhibition (up to 30%) of cell proliferation by increasing concentrations of LY156758, with a half-maximal effect observed at about 0.3 nM (Fig. 6A). Again, incubation with 5α-DFIT (InM) induced a 55% inhibition of basal cell growth, while the addition of LY156758 further decreased cell number to a maximal 70% inhibition. In good agreement with the data presented in Figs. 1 and 5, the addition of InM 5α-DHT alone inhibited the mitogenic effect of E₂ by 80% (Fig. 6B). In the presence of increasing concentrations of LY156758, 5α-DHT further decreased cell number below the value reached with maximally effective concentrations of LY156758 alone to approximately 30% of basal cell growth (Fig. 6B). In every instance, OH-FLU (1 μM) completely reversed the effect of 5α-DHT without affecting the pattern of inhibition achieved with the antiestrogen.

Figure 6. Effect of increasing concentrations of the antiestrogen LY156758 on the proliferation of ZR-75-1 cells, and the influence of coincubation with 5α-DHT and/or E₂. Prior to the experiment, cells were cultured for 4 weeks in SD medium. Cells were then plated at 8.0 x 10³ cells/well in fresh medium of identical composition. A) 48 h after plating, LY156758 was added at the indicated concentrations, alone (control, O) or in the presence of 1 nM 5α-DHT (•), 1 μM OH-FLU (□), or both compounds (■). B) Same as in A, except that 1 nM E₂ was also present (A) or in association with 1 nM 5α-DHT (•), 1 μM OH-FLU (□), or both compounds (■).

Since high concentrations of 5α-DHT and T are known to bind to the estrogen receptor and induce nuclear retention of the resulting complexes with subsequent estrogenic responses [35-39], the binding affinity of ligands used in this study for the androgen and estrogen receptors present in ZR-75-1 cells was determined. As shown in Fig. 7, intact ZR-75-1 cells grown under steroid-deprived conditions contain 2.53 X 10⁴ androgen specific binding sites/cell with an apparent dissociation constant (K_D) or 0.69 nM, in agreement with data already published for the ZR-75-1 cell line [12, 22]. Incubation of ZR-75-1 cells for up to 7 days in the presence of InM E₂ had no significant effect on either the number (2.69 x 10⁴ sites/cell) nor K_D (0.79 nM) of the androgen specific binding sites, while increasing cell number by 58% (data not shown) during the same period.

Figure 7. Specific uptake of increasing concentrations of [₃H]R1881 by intact ZR-75-1 cell monolayers. Cells were grown in 24-well culture plates either in the absence (O) or in the presence (•) of 1 nM E₂, for a 7-day incubation period. The specific uptake of [₃H]R1881 was then determined as described in ‘Materials and Methods’. Inset: Scatchard analysis of the same data.

y 58% (data not shown) during the same period. The RBA of various ligands for androgen and estrogen specific binding sites in ZR-75-1 cell monolayers was next studied (Fig. 8). Table 1 summarizes the apparent Ki values of the steroids used, as calculated [33] from the K_D values of R1881 (Fig. 7) and E₂ [29] of their respective receptors. The competition studies show that E₂, Δ⁴-DIONE, and OH-FLU were much weaker ligands of the androgen receptor than 5α-DHT and T, which had affinities close to that of R1881 (Fig. 8A). The apparent Ki values of 5α-DHT and T were somewhat higher than the IC₅₀ values measured for the same androgens on the proliferation of ZR-75-1 cells (0.10 and 0.15 nM, respectively; Figs. 1 and 2). Similar discrepancies between the measured affinity of E₂ for the estrogen receptor and the half-maximal effective concentration of the estrogen needed to induce a mitogenic effect have been noted [29, 40, 41]. While the apparent Ki value found for the antiandrogen OH-FLU agrees well with the calculated Ki for its reversal of the androgen effect on cell growth (110nM), the affinity of Δ⁴-DIONE for androgen specific binding sites was about 13-fold lower than that predicted from its potency to inhibit cell proliferation (IC₅₀== 15 nM; cf. Fig. 2). In-terestingly, the calculated Ki value of E₂ for the specific uptake of [₃H]R1881 was about 50 nM.

Figure 8. Competition for the uptake of [₃H]R1881 (A) or [₃H]E₂ (B) by increasing concentrations of various unlabeled ligands. Cells were grown in SD medium and the uptake of radiolabeled ligand determined, as described in ‘Materials and Methods’, in the presence of the indicated concentrations of competitor.

Table 1. Apparent dissociation constants (Ki) of various ligands towards the androgen (AR) and estrogen receptors (ER) pre-sent in ZR-75-1 cells. Ki values were calculated according to Cheng and Prusoff [33], using the RBA values determined from Fig. 8. The K_D values of R1881 and E₂ were estimated by Scatchard analysis under identical experimental conditions (Fig. 7.) [29].

Except for E₂ itself, only 5α-DHT and T were found to efficiently compete with the high affinity uptake of [₃H]E₂ in intact ZR-75-1 cells, although only at high concentrations (Ki = 441 and 135 nM, respectively) (Fig. 8B and Table 1). Likewise, OH- FLU had no significant affinity for estrogen specific binding sites as reported in other tissues [42].

Discussion

The present study provides the first demonstration that naturally occuring androgens exert a potent inhibitory action on the proliferation of an estab-lished, estrogen-sensitive human breast cancer cell line. Thus, in ZR-75-1 cells, concentrations of 5α- DHT similar to the plasma levels found in normal women [35, 43, 44] and breast cancer patients [45] (0.3-0.7 nM) are potent inhibitors of the mitogenic effect of E₂ and inhibit growth in the absence of estrogens. Furthermore, T, at concentrations within the physiological range of concentrations observed in adult women (1-3 nM) [35,43-45], is also a potent inhibitor of cell growth. Δ⁴-DIONE also led to significant growth inhibition in ZR-75-1 cells, although the active concentrations (IC₅₀~ 15 nM) are in the upper range of the plasma concentrations (1-10 nM) found in women [35, 43-15].

Several lines of evidence show that the potent growth-inhibitory effect of androgens observed in ZR-75-1 cells is mediated through their specific interaction with the androgen receptor. Firstly, the potency of 5α-DHT and T to induce antiproliferative effects (IC₅₀~ 0.10 and 0.15 nM, respectively) is in agreement with their relative binding affinity for androgen specific binding sites in intact ZR-75-1 cells as well as in other human breast cancer cells [12, 46]. It also compares well with the potency of 5α-DHT to specifically stimulate the secretion of the Zn-a2-glycoprotein [19] and the GCDFP-15 glycoprotein [19, 20] in T47-D human breast cancer cells. The ability of Δ⁴-DIONE to induce an antiproliferative effect (IC₅₀ ~ 15 nM) is more likely to result from its metabolic transofrma- tion into T and 5α-DHT [47-49] than from its direct interaction with the androgen receptor (K_D~ 200nM). Secondly, the antiandrogen OH-FLU competitively reversed the effect of 5α-DHT, and Δ⁴-DIONE with an apparent dissociation constant (Ki~ 110 nM) consistent with its known affinity for the androgen receptor [24]. This compound is devoid of any significant affinity for steroid recep-tors other than the androgen receptor [42].

It should be mentioned that growth inhibition by 5α-DHT lagged behind the mitogenic effect of E₂, either in the presence or absence of the estrogen (Fig. 3), suggesting different pathways for the action of estrogens and androgens on cell proliferation. Moreover, the growth-inhibitory effect of 5α-DHT is clearly additive to that induced by maximally effective concentrations of the antiestrogen LY156758, thus indicating an action mediated by a mechanism different from interaction with the es-trogen receptor. Thus, the present evidence leaves little doubt that the antiproliferative effect of an-drogens does not result from competition for binding to the estrogen receptor, but rather is mediated by an androgen receptor-mediated mechanism which is additive to blockade of the estrogen receptor by LY156758.

The degree of growth inhibition exerted by 5α- DHT, especially in the presence of E₂, was critically dependent on cell density at the time of steroid addition, mainly because of the relatively long lag needed for androgen action on cell growth. On the other hand, estrogens and androgens can antagonize each other at earlier time intervals on a number of parameters or regulated gene products, as shown in the rat uterus [50, 51] and mammary tumors [50], in the Syrian hamster kidney [52], and in human hepatoma cells [53]. In the MCF-7 [17, 18] and the ZR-75-1 breast cancer cells (R.P. and F.L., manuscript in preparation), low androgen concentrations inhibit the E₂-dependent induction of progesterone receptor via an androgen-recep-tor-mediated mechanism. We were unable to detect any estrogen receptor-mediated ‘down-regulation’ of androgen specific binding sites in ZR-75-1 cells, such as that recently reported for MCF-7 cells [54]. Reasons for this discrepancy are not clear, but might reflect divergent control of cellular functions by androgens and estrogens in different human breast cancer cell lines. For instance, while T47-D and ZR-75-1 mammary tumor cells express the GCDFP-15 mRNA and glycoprotein under androgen stimulation, MCF-7 cells do not [19, 20], possibly indicating an apocrine tissue character for the former cell lines [19].

According to the antagonistic relationship be-tween estrogen and androgen receptor-mediated events, only when the antiestrogen LY156758 was added together with E₂ did the antiproliferative effect of 5α-DHT reach the degree observed in estrogen-deprived, androgen-treated ZR-75-1 cells (Figs. 5,6). On the other hand, high concentrations of E₂ can interact with the androgen receptor (Fig. 8) [35], a fact which may account for the progressive reversal of androgen-induced inhibition of cell proliferation observed at concentrations of E₂ ex-ceeding 10nM (Fig. 5α).

Antiestrogens such as LY156758 inhibit both E₂- induced and basal cell proliferation through an es-trogen receptor-mediated mechanism. Although the physiological basis for growth inhibition of es-trogen-deprived cells by antiestrogens is still debated [55, 56], it cannot be solely attributed to the estrogenic effect of phenol red [27, 28], which was absent from the media used in the present experi-ments. Furthermore, the possible prolonged reten-tion of estrogens following steroid withdrawal [57] was unlikely to contribute significantly to the effect of LY156758 on basal cell growth, since a 4-week period of estrogen deprivation did not attenuate this phenomenon. Whatever the underlying mechanism for antiestrogen-mediated inhibition of growth in the apparent absence of estrogens, the present data clearly show that complete suppression of estrogen action by an antiestrogen is a prerequisite in order to observe the full inhibitory effect of androgens on breast cancer cell growth.

Interpretation of the role of androgens in mam-mary cancer has been complicated by the known dual affinity of 5α-DHT and T for both the androgen (K_D ~ 0.01-1 nM) and the estrogen (K_D-~ 100-1000 nM) receptors in various tissues (Fig. 8) [15,16, 35, 36]. Moreover, binding of androgens to the estrogen receptor results in the induction of several characteristic estrogenic responses in the rat uterus [36, 37], in 7,12-dimethylbenz(a)anthra-cene-induced rat mammary tumors [38, 39], and in human breast cancer cells [13-16]. In ZR-75-1 cells, high (100-2000 nM) concentrations of 5α-DHT and to a lesser degree T, are less inhibitory than those in the 0.1-10 nM range (Figs. 1,2), this biphasic effect probably resulting from increasingly significant in-teraction of DHT and T with the estrogen receptor at higher concentrations. Similar observations have been made in vivo concerning carcinogen- induced rat mammary tumors [58, 59] and in the immature mouse uterus [38], where intermediary doses of various androgens were the most effective in counteracting estrogen-induced tissue weight in-creases.

In conclusion, we have presented unequivocal evidence that physiologically relevant concentrations of 5α-DHT, T, and Δ⁴-DIONE can potently counteract the mitogenic effect of E₂ as well as markedly accentuate growth inhibition induced by an antiestrogen in an established human breast cancer cell line. These effects are mediated viaspecific interaction of these steroids with the an-drogen receptor. Whether similar growth-inhibitory action of androgens can apply to human breast tumor cells in the clinical situation remains to be established. However, the overwhelming clinical evidence for tumor regression observed in 20 to 50% of pre- and post-menopausal breast cancer patients treated with various androgens [7] favors the view that naturally occuring androgens might constitute an as yet overlooked, direct hormonal control of mammary cancer cell growth. It is thus reasonable to suggest that an imbalance between androgenic and estrogenic influences could modify the overall growth rate of breast tumors in much the same way as that suggested for progestins in estrogen target tissues [60]. Interestingly, the indication that an increased response rate might be obtained by combining androgens and an antiestrogen therapy in breast cancer patients [5] is in agreement with our observation that the mechanisms of inhibition by both types of agents are different, and their effects, at least in part, additive.

Acknowledgements

This work was supported in part by grants from the Medical Research Council of Canada and the Terry Fox Cancer Foundation.

---

References

Fels E: Treatment of breast cancer with testosterone pro-pionate. A preliminary report. J Clin Endocrinol 4: 121— 125,1944 

Segaloff A, Gordon D, Horwitt BN, Schlosser JV, Murison PJ: Hormonal therapy in cancer of the breast. 1. The effect of testosterone propionate therapy on clinical course and hormonal excretion. Cancer 4: 319-323, 1951 

Cooperative Breast Cancer Group: Testosterone propionate therapy of breast cancer. J Amer Med Assoc 188:10691072, 1964 . 

Kennedy BJ: Fluoxymesterone therapy in treatment of ad-vanced breast cancer. N Engl J Med 259: 673-675, 1958 

Tormey DC, Lippman ME, Cassidy JG: Evaluation of Tamoxifen doses with and without fluoxymesterone in ad-vanced breast cancer. Ann Intern Med 98: 139-143, 1983 

Gordan GS, Halden A, Horn Y, Fuery JJ, Parsons RJ, Walter RM: Calusterone (7p,17α-dimethyltestosterone) as primary and secondary therapy of advanced breast cancer. Oncology 28: 138-146, 1973  

Gordan GS: In: Kochakian CD (ed) Anabolic-androgenic steroids. Handbook of Experimental Pharmacology, vol Springer-Verlag, New York, 1976, pp 499-513 

Segaloff A: The use of androgens in the treatment of neoplastic disease. Pharmac Ther C2: 33-37, 1977 

McGuire WL, Carbone PP, Sears ME, Escher GC: Estrogen receptors in human breast cancer: an overview. In: McGuire WL, Carbone PP, Vollmer EP (eds) Estrogen Receptors in Breast Cancer, Raven Press, New York, 1975, PP 1-7 

Wittliff JL: Steroid binding proteins in normal and neoplastic mammary cells. In: Busch H (ed) Methods in Cancer Research, vol 11 Academic Press, New York, 1975, pp 298-304 

Allegra JC, Lippman ME, Thompson EG, Simon R, Barlock A, Green L, Huff KK, Do HMT, Aitken SC: Distribution, frequency, and quantitative analysis of estrogen, progesterone, androgen, and glucocorticoid receptors in human breast cancer. Cancer Res 39: 1447-1454,1979 

Horwitz KB, Zava DT, Thilager AK, Jensen ET, McGuire WL: Steroid receptor analyses of nine human breast cancer cell lines. Cancer Res 38: 2434-2439, 1978 

Lippman M, BolanG,Huff K: The effects of androgens and antiandrogens on hormone-responsive human breast cancer in long-term tissue culture. Cancer Res 36: 4610-4618, 1976 

Westley B, Rochefort H: A secreted glycoprotein induced by estrogen in human breast cancer cell lines. Cell 20: 353-362,1980 

Zava DT, McGuire WL: Human breast cancer: androgen action mediated by estrogen receptor. Science 199: 787788,1978 

Zava DT, McGuire WL: Androgen action through estrogen receptor in a human breast cancer cell line. Endocrinology 10³: 624-631,1978 

Mclndoe JH, Etre LA: An antiestrogenic action of androgens in human breast cancer cells. J Clin Endocrinol Metab 53: 836-842,1981 

Shapiro E, Lippman ME: Onset of androgen action in MCF-7 human breast cancer cells is not accompanied by receptor depletion. J Steroid Biochem 22: 15-20, 1985 

Chalbos D, Haagensen D, Parish T, Rochefort H: Identification and androgen regulation of two proteins released by T47D human breast cancer cells. Cancer Res 47:2787-2792, 1987 

Murphy LC, Tsuyuki D, Myal Y, Shiu RPC: Isolation and sequencing of a cDNA clone for a prolactin-inducible protein (PIP). J Biol Chem 262: 15236-15241,1987 

Simon WE, Palinke VG, Hôlzel F: In vitro modulation of prolactin binding to human mammary carcinoma cells by steroid hormones and prolactin. J Clin Endocrinol Metab 60:1243-1249,1985 

Engel LW, Young NA, Tralka TS, Lippman ME, O’Brien SJ, Joyce MJ: Establishment and characterization of three new continuous cell lines derived from human breast carcinomas. Cancer Res 38: 3352-3364, 1978 

Neri R, Peets E, Watnick A: Antiandrogenicity of flutamide and its metabolite Schl6423. Biochem Soc Trans 7: 565-569,1979 

Simard J, Luthy I, Guay J, Bélanger A, Labrie F: Characteristics of interaction of the antiandrogen Flutamide with the androgen receptor in various target tissues. Mol Cell Endocrinol 44: 261-270,1986 

Clemens JA, Bennett DR, Black LJ, Jones CD: Effects of a new antiestrogen, keoxifene (LY156758), on growth of car-cinogen-induced mammary tumors and on LH and prolactin levels. Life Sci 32: 2869-2875, 1983 

20. Simard J, Labrie F: Keoxifene shows pure antiestrogenic activity in pituitary gonadotrophs. Mol Cell Endocrinol 39: 141-144, 1985 

Berthois Y, Katzenellenbogen JA, Katzenellenbogen BS: Phenol red in tissue culture media is a weak estrogen: implications concerning the study of estrogen-responsive cells in culture. Proc Natl Acad Sci USA 83: 2496-2500, 1986 

Hubert JF, Vincent A, Labrie F: Estrogenic activity of phenol red in rat anterior pituitary cells in culture. Biochem Biophys Res Commun 141: 885-891,1986 

Poulin R, Labrie F: Stimulation of cell proliferation and estrogenic response by adrenal C19-A5-steroids in the ZR-75-1 human breast cancer cell line. Cancer Res 46: 4933-4937, 1986 

Taylor CM, Blanchard B, Zava DT: A simple method to determine whole cell uptake of radiolabeled oestrogens and progesterone and their subcellular localization in breast cancer cell lines in monolayer cultures. J Steroid Biochem 20:1083-1088,1984 

Scatchard G: The attraction of proteins for small molecules and ions. Ann NY Acad Sci 51: 660-672, 1959 

Rodbard D: Apparent positive cooperative effect in cyclic AMP and corticosterone production by related adrenal cells in response to ACTH analogs. Endocrinology 94: 1427-1437,1974 

Cheng Y,PrusoffWH: Relationship between the inhibition constant (Ki) and the concentration of inhibitor which causes 50% inhibition (IC₅₀) of an enzymatic reaction. Biochem Pharmacol 22: 3099-3108,1973 

Kramer CY: Extension of multiple-range test to group means with unique numbers of replications. Biometrics 12: 307-310,1956 

Rochefort H, Garcia M: The estrogenic and antiestrogenic activities of androgens in female target tissues. Pharmac Ther 23: 193-216, 1983 

Garcia M, Rochefort H: Androgen on the oestrogen receptor: II. Correlation between nuclear translocation and uterine protein synthesis. Steroids 29: 111-126,1977 

Huggins C, Jensen EV, Cleveland AS: Chemical structure of steroids in relation to promotion of growth of the vagina and uterus of the hypophysectomized rat. J Exp Med 100: 225-240,1954 

Hilf R: Anabolic-androgenic steroids and experimental mammary tumors. In: Kochakian CD (ed) Anabolic-androgenic steroids. Handbook of Experimental Pharmacology, vol 43, pp 191-210,1976 

Garcia M, Rochefort H: Androgen effects mediated by estrogen receptor in 7,12-dimethylbenz(a)anthracene-in- duced rat mammary tumors. Cancer Res 38: 3922-3929, 1978 

Lippman M, Bolan G, Huff K: The effects of estrogens and antiestrogens on hormone-responsive human breast cancer in long-term tissue culture. Cancer Res 36:4595-4601,1976 

Reiner GCA, Katzenellenbogen BS, Bindal RD, Katzenel-lenbogen JA: Biological activity and receptor binding of a strongly interacting estrogen in human breast cancer cells. Cancer Res 44: 2302-2308, 1984 

Raynaud JP, Ojasoo T: Receptor binding as a tool in the development of selective new bioactive steroids and non-steroids. In: Harms AF (ed) Innovative Approaches in Drug Research. Elsevier, Amsterdam, 1986, pp 47-72 

Abraham GE: Ovarian and adrenal contribution to peripheral androgens during the menstrual cycle. J Clin Endocrinol Metab 39: 340-346,1974 

Vermeulen A, Verdonck L: Factors affecting sex hormone levels in postmenopausal women. J Steroid Biochem 11: 899-904,1979 

Mistry P, Griffiths K, Maynard PV: Endogenous C19-ste- roids and estradiol levels in human primary breast tumor tissues aifd their correlation with androgen and estrogen receptors. J Steroid Biochem 24: 1117-1125, 1986 

Mclndoe JH, Woods GR, Lee FJ: The specific binding of androgens and the subsequent distribution of androgen receptor complexes within MCF-7 human breast cancer cells. Steroids 38: 439-452, 1981 

Perel E, Daniilescu D, Kharlip L, Blackstein ME, Killinger DW: The relationship between growth and androstene- dione metabolism in four cell lines of human breast carcinoma cells in culture. Mol Cell Endocrinol 41: 197-203, 1985 

Griffiths K, Jones D, Cameron EHD, Gleave EN, Forrest APM: Transformation of steroids by mammary cancer tissue. In: Dao TL (ed) Oestrogen Target Tissues and Neoplasia. University of Chicago Press, Chicago, 1972, pp 125-136 

Perel E, Killinger DW: The metabolism of androstene- dione and testosterone to C19-metabolites in normal breast, breast carcinoma, and benign prostatic hypertrophy tissue. J Steroid Biochem 19: 1135-1139, 1983 

Ip M, Milholland RJ, Kim U, Rosen F: Androgen control of cytosol progesterone receptor levels in the MT-W9B transplantable mammary tumor in the rat. J Natl Cancer Inst 69: 673-691, 1982 

Jellinek PH, Newcombe AM: Androgen receptor-mediated inhibition of estrogen-induced uterine peroxidase. J Steroid Biochem 19: 1713-1717, 1983 

Li SA, Li JJ: Estrogen-induced progesterone receptor in the Syrian hamster kidney. I. Modulation by antiestrogens and androgens. Endocrinology 103: 2119-2128, 1978 

TamS-P, Archer TK, Deeley RG: Biphasic effects of estrogen on apolipoprotein synthesis in human hepatoma cells. Mechanism of antagonism by testosterone. Proc Natl Acad Sci USA 83: 3111-3115, 1986 

Stover EP, Krishman AV, Feldman D: Estrogen down- regulation of androgen receptors in cultured human mammary cancer cells (MCF-7). Endocrinology 120:2597-2603, 1987 

Taylor CM, Blanchard B, Zava DT: Estrogen-receptor mediated and cytotoxic effects of the antiestrogens tamoxifen and 4-hydroxytamoxifen. Cancer Res 44: 1409-1414, 1984 

Bardon S, Vignon F, Montcourrier P, Rochefort H: Steroid receptor-mediated cytotoxicity of an antiestrogen and an antiprogestin in breast cancer cells. Cancer Res 47: 1441— 1448, 1987 

Strobl JS, Lippman ME: Prolonged retention of estradiol by human breast cancer cells in tissue culture. Cancer Res 39: 3319-3327, 1979 

Huggins C, Briziarelli G, Sutton H Jr: Rapid induction of mammary carcinoma in the rat and the influence of hormone on the tumors. J Exp Med 109: 25-42, 1959 

Heise E, Gorlich M: Growth and therapy of mammary tumors induced by 7,12-dimethylbenz(a)anthracene in rats. Br J Cancer 20: 539-545, 1966 

Horwitz HB: The structure and function of progesterone receptors in breast cancer. J Steroid Biochem 27: 447-457, 1987