1α,25-Dihydroxyvitamin D₃ Down-Regulates Estrogen Receptor Abundance and Suppresses Estrogen Actions in MCF-7 Human Breast Cancer Cells¹

Breast cancer is the most commonly diagnosed cancer and the second leading cause of cancer-related deaths among women in the United States (1). Because breast cancer is generally characterized by estrogen-dependent growth, the abundance of ERs³in these cells assumes critical importance (2). E₂ acts via the ER, a member of the steroid/thyroid/retinoid receptor superfamily (3). The factors and mechanisms that control the level of ER expression are important in determining the amplitude of E₂-mediated actions on the breast cancer cells (2).

1,25(OH)₂D₃, the biologically active form of vitamin D, is a major regulator of calcium and phosphate homeostasis in the body (4, 5). The regulatory effects of 1,25(OH)₂D₃ are mediated via the VDR, which is also a member of the steroid/thyroid/retinoid receptor superfamily (4, 5, 6). In addition to its effects on calcium and phosphate homeostasis,1,25(OH)₂D₃ is an important modulator of cellular proliferation and differentiation in a number of normal and malignant cells (4, 7, 8, 9). In breast cancer cells, 1,25(OH)₂D₃ has potent growth-inhibitory actions (10, 11, 12, 13). Although the growth-inhibitory effects of 1,25(OH)₂D₃ on breast cancer cells have been well established, the effects of 1,25(OH)₂D₃ on ER expression are less well documented. Studies on ERα in human breast cancer cell lines have reported minor decreases (14) or no change (15) in ER expression with 1,25(OH)₂D₃ treatment. A more recent study reported significant decreases in ER protein levels in the MCF-7 cells treated with EB-1089, a potent analogue of 1,25(OH)₂D₃(13). Studies have also been conducted to indicate that the E₂-mediated bioresponses are attenuated by 1,25(OH)₂D₃ treatment (16, 17). Although the above-mentioned reports have suggested potential cross-talk between 1,25(OH)₂D₃ and estrogen signaling pathways, the extent of the interaction and the mechanism by which 1,25(OH)₂D₃causes down-regulation of ER are still not clarified.

The purpose of the present investigation is to study the effects of 1,25(OH)₂D₃ on ER and E₂-mediated effects in MCF-7 breast cancer cells. As discussed below, the MCF-7 cells used in this study do not express ERβ, as measured by reverse transcription-PCR. Therefore, in these studies the effect is limited to ERα. For simplicity, we refer to the ERα in these cells as ER. To achieve our goal of investigating the effect of 1,25(OH)₂D₃on breast cancer cells, we have studied the relationship between changes in MCF-7 growth rate, levels of ER protein, steady-state ER mRNA, and gene transcription in cells treated with 1,25(OH)₂D₃ or its analogues. We have also assessed the correlation between the effects of 1,25(OH)₂D₃ and its analogues KH-1060, EB-1089, Ro 27-0574, and Ro 23-7553 on the growth of MCF-7 cells and the changes elicited in ER levels. Furthermore, we have investigated how changes in ER abundance induced by 1,25(OH)₂D₃ and its analogues alter the functional responses to E₂ in MCF-7 cells. We have established that the 1,25(OH)₂D₃-mediated effect on ER gene expression is at the transcriptional level.

[³H]Estradiol-17β-d-glucuronide(specific activity, 40 Ci/mmol) and[³H]progesterone (specific activity, 54.1 Ci/mmol) were purchased from DuPont NEN (Wilmington, DE). Radioinert steroids were obtained from Steraloids, Inc. (Wilton, NH). Nonradioactive 1,25(OH)₂D₃and its analogues 1,25-dihydroxy-16-ene-23-yne-cholecalciferol (Ro 23-7553) and 1,25-dihydroxy-23-yne-26,27-hexafluoro-20-cyclopropyl-19-nor-cholecalciferol(Ro 27-0574) were generous gifts from Dr. M. Uskokovic (Hoffmann La-Roche Co., Nutley, NJ). 1α,25-Dihydroxy-22,24-diene-24,26,27-trihomovitamin D₃ (EB-1089) and 1α,25-dihydroxy-20-epi-22-ene-24,26,27-trihomovitamin D₃ (KH-1060) were generous gifts from Dr. L. Binderup (Leo Pharmaceuticals, Ballerup, Denmark). MCF-7 cells were obtained from American Type Culture Collection (Rockville, MD). Culture media and other supplements were purchased from Mediatech (Herndon,VA). All other chemicals and reagents, including anti-actin antibody,were obtained from Sigma Chemical Co. (St. Louis, MO).

MCF-7 cells were routinely cultured in T-75 flasks at 37°C under an atmosphere of 5% CO₂. They were maintained in RPMI 1640 supplemented with 10% calf serum, 100 units/ml penicillin,and 100 μg/ml streptomycin. At confluence, the flasks contained approximately 1.5–2 × 10⁷ cells and were routinely subcultured every 7–10 days. For experiments, the growth medium was replaced with phenol red-free RPMI 1640 supplemented with 10% calf serum (CSS), twice stripped of endogenous hormones using charcoal and Dextran T-70. We used medium containing CSS in most of our studies to minimize the effects of E₂, which is known to down-regulate its own receptor (18). The medium containing CSS (treatment medium) and hormones, as indicated, was introduced to the cells 24 h after subculture, and fresh medium and hormones were replenished every 2 days. Stock solutions of steroid hormones were made in 100% ethanol and added to the treatment medium. All controls received ethanol vehicle at a concentration equal to that in the hormone-treated cells (0.1% v/v).

DNA content or attained cell mass was used as a measure of cell proliferation. MCF-7 cells were seeded in six-well tissue culture plates (Becton Dickinson, Lincoln Park, NJ) at a density of 50,000 cells/well in 3 ml of RPMI 1640 containing 10% calf serum. Twenty-four h later, fresh medium was added. Cells were grown in RPMI 1640 medium with 10% calf serum or CSS and were treated with various doses of 1,25(OH)₂D₃ or its analogues in the presence or absence of 10 nmE₂. Fresh medium and hormones were added every other day. At the end of 6 days, cell monolayers were processed as described earlier (19), and DNA contents were determined by the method of Burton (20).

MCF-7 cells growing in CSS-containing medium were treated with either E₂ (10 nm) or 1,25(OH)₂D₃ or analogues(1, 10, or 100 nm) for 2 days. Cells were then harvested,and high salt cell extracts were made as described previously (21). The protein concentration of the extracts was measured by the method of Bradford (22). Aliquots of the extracts were incubated overnight at 4°C with either 10 nm [³H]E₂ or 10 nm [³H]progesterone for ER and PR measurements, respectively. Two hundred-fold excess of nonradioactive hormone was used to correct for nonspecific binding. Bound and free hormones were separated using hydroxylapatite, and specific binding was calculated as described earlier (21).

Aliquots of cell extracts prepared as described above were mixed with 3× SDS sample buffer, boiled for 5 min, and subjected to 10%SDS-PAGE. After transfer to nitrocellulose membranes, immunoblotting with either the antimouse monoclonal antibody to human ER (H222; 1:500 dilution in 1% Carnation nonfat milk; a gift from Abbott laboratories)or antimouse monoclonal antibody to the human BRCA1 protein (C-20; 2μg/ml in 1% Carnation nonfat milk; from Santa Cruz Biotechnology)was carried out as described previously (19). The blots were then probed with a horseradish peroxidase-conjugated antimouse secondary antibody, and the immunoreactive bands were detected using an enhanced chemiluminescence (ECL) kit obtained from Amersham (Arlington Heights, IL). High molecular weight markers from Life Technologies,Inc. (Grand Island, NY) were used to estimate the sizes of the immunoreactive bands.

Total RNA was isolated from cells treated with either ethanol vehicle or 1,25(OH)₂D₃ as described previously (23, 24). Changes in ER mRNA levels attributable to 1,25(OH)₂D₃treatment were detected by Northern blot analysis. A 2.1-kb EcoRI fragment of the human ER cDNA was subjected to random prime labeling using [³²P]dCTP and the Rediprime labeling kit (Amersham), and the labeled fragment was used to probe the blots. To control for differences in RNA sample loading and transfer, the blots were also hybridized with a ³²P-labeled, 0.9-kb EcoRI fragment of the gene encoding the human L7 ribosomal protein. The membranes were exposed to X-ray films (Hyperfilm MP) for about 17 h at −80°C. Autoradiograms were scanned using a Molecular Dynamics Computing densitometer (model 300A; Molecular Devices, Menlo Park, CA), and the ER mRNA levels were indexed to the corresponding L7 mRNA levels.

To determine the half-life of ER mRNA, MCF-7 cells were grown as described earlier and treated with ethanol (control) or 100 nm 1,25(OH)₂D₃for 24 h. At the end of 24 h, transcription was terminated by the addition of 4 μm actinomycin D. Because the reported half-life of ER mRNA is ∼4 h, total RNA was extracted at regular time intervals up to 6 h after actinomycin treatment. Twenty μg of total RNA were used for Northern blot analysis as described earlier.

Nuclei were isolated from MCF-7 cells treated with 100 nm1,25(OH)₂D₃ for various time intervals according to the procedure described by Stott (25). Briefly, MCF-7 cells treated with 1,25(OH)₂D₃ or ethanol vehicle were harvested at various time intervals and resuspended in ice-cold nuclei isolation buffer [10 mm Tris-HCl (pH 7.4),10 mm NaCl, 5 mm MgCl₂,and 1 mm DTT] containing 0.5% NP40. The intact nuclei were then pelleted by centrifugation at 2000 rpm for 5 min. The pellets were resuspended in nuclei freezing buffer [50 nm Tris-HCl(pH 8.5), 50% w/v glycerol, 5 mmMgCl₂, and 0.1 mm EDTA] in aliquots of 10⁸ nuclei/ml and frozen as aliquots of 210μl at −70°C until needed. RNA elongation was carried out as described earlier (25). Frozen aliquots of nuclei were thawed on ice and incubated with [³²P]UTP and unlabeled ATP, CTP, and GTP at 30°C for 45 min. The radiolabeled nascent RNA transcripts were isolated using TRIzol reagent, followed by chloroform extraction and ethanol precipitation. Isolated RNA was then hybridized to nitrocellulose filters containing cDNAs for ERand L7 for 72 h at 65°C. Filters were then washed and exposed to X-ray film. Autoradiographs were scanned, and results were normalized by comparison to the transcriptional level of L7.

Data are presented as mean ± SD of three to four individual measurements. Statistical analysis was done by Student’s ttest or ANOVA using the Statview 4.5 software (Abacus Concepts,Berkeley, CA). P <0.05 is considered significant.

Fig. 1 demonstrates the effect of 1,25(OH)₂D₃ and its analogues on the growth of MCF-7 cells cultured in medium containing 10% calf serum. 1,25(OH)₂D₃ and each of the analogues tested caused a dose-dependent decrease in the cell growth. The analogues were more potent inhibitors of growth than 1,25(OH)₂D₃, with the order of potency being Ro 27-0574 > KH-1060 > EB-1089 > Ro 23-7553 >1,25(OH)₂D₃. The calculated IC₅₀ values are shown in Table 1.

Fig. 2 shows the effect of 1,25(OH)₂D₃ and its analogues on growth of MCF-7 cells cultured in phenol red-free RPMI containing 10% CSS in the absence or presence of 10 nmE₂. The total DNA content in these controls was slightly less than that observed with the controls grown in medium containing 10% calf serum (31 ± 6.6 versus 43 ±0.42 μg DNA/well). 1,25(OH)₂D₃ and its analogues demonstrated a higher degree of growth inhibition in medium containing serum (Fig. 1) than medium containing CSS (Fig. 2). From Fig. 2, E₂ treatment caused a 2-fold increase in growth (P <0.001) when compared with controls. Significant decreases in DNA content were seen after cotreatment with E₂ and 1,25(OH)₂D₃ or its analogues when compared with E₂ treatment alone. 1,25(OH)₂D₃ and its analogues were able to partially or completely counteract the E₂-mediated increase in growth, with the analogues being more potent than 1,25(OH)₂D₃.

To determine the effects of 1,25(OH)₂D₃ and its analogues on ER abundance, we used[³H]E₂ ligand binding assays in cells cultured in medium containing CSS treated with graded concentrations of 1,25(OH)₂D₃ or its analogues for 2 days. As shown in Fig. 3, %1,25(OH)₂D₃ induced a dose-dependent decrease in ER levels, which was modest at 1 and 10 nm and higher (∼50%) at 100 nm(P < 0.001). As in the proliferation experiments, the analogues were more potent than 1,25(OH)₂D₃, with the order of potency being Ro 27-0574 > KH-1060 > EB-1089 > Ro 23-7553 >1,25(OH)₂D₃. IC₅₀ values for ER down-regulation are shown in Table 1. Data in Table 1 demonstrate that the antiproliferative effect of 1,25(OH)₂D₃ or a given analogue correlated very well (R² =0.98) with ER suppression.

In the next set of experiments, MCF-7 cells grown in medium containing 10% CSS were treated with 100 nm1,25(OH)₂D₃, and the time course of the 1,25(OH)₂D₃effect on ER levels was investigated. Although decreases in ER levels could be seen as early as 6 h after treatment with 1,25(OH)₂D₃, changes were significant only at the end of 2 days (Fig. 4 A). At this time point, the ER levels declined by ∼50%(from 479 ± 41 to 213 ± 37 fmol/mg protein). This decrease was transient, because by day 4, the decrease in ER was less pronounced, and by day 6, ER levels were back to that seen in control cells. Scatchard analysis (data not shown) of[³H]E₂ binding revealed that 1,25(OH)₂D₃ treatment for 2 days caused a decrease in ER abundance (Nmax) from 492 fmol/mg protein to 251 fmol/mg protein, with no change in the affinity for E₂ (K_d =0.70 nm in controls versus 0.62 nm in 1,25(OH)₂D₃-treated cells).

Changes in ER protein levels were also assessed by Western blot analysis. Cell extracts from MCF-7 cells, grown in medium containing CSS, were treated with 1,25(OH)₂D₃ for various time intervals and probed with the H222 anti-ER monoclonal antibody. Western blot analysis (Fig. 4,B) revealed a pattern of changes similar to that observed with ligand binding studies (Fig. 4 A). A 66-kDa band representing the immunoreactive ER protein could be detected in both treated and untreated groups. A significant decrease (∼50%) in the ER protein levels was seen at the end of 2 days in the 1,25(OH)₂D₃-treated cells. The ER protein levels remained suppressed (∼30%) at day 4 and gradually returned to control levels by the end of 6 days. The experiment was repeated three times to confirm the pattern of changes and to rule out possible differences attributable to loading.

Because the growth-inhibitory effects of 1,25(OH)₂D₃ and its analogues were more evident in medium containing serum (Fig. 1), it was important to establish the down-regulation of ER in serum containing medium. Fig. 4,C demonstrates the effect of 1,25(OH)₂D₃ and its analogues on ER levels in the presence of medium containing 10% calf serum. A significant fall in ER levels could be seen with 1,25(OH)₂D₃ and all of the analogues tested. Expression of actin, which was used as a control, did not change. Thus, ER down-regulation by 1,25(OH)₂D₃ or its analogues could be seen in medium containing CSS as well as regular serum. These results correlate with the observations from ligand binding studies (Fig. 4,A) and growth assays (Fig. 2),confirming that the analogues were more potent than 1,25(OH)₂D₃ in both assays(Table 1).

To determine whether the changes in ER attributable to 1,25(OH)₂D₃ treatment occurred at the mRNA level, we examined the steady-state levels of ER mRNA by Northern blot analysis. Fig. 5,A is a representative Northern blot and Fig. 5 Bits corresponding densitometric scan. All experiments(n = 4) were conducted in the presence of CSS to minimize the effects of E₂ present in the medium. 1,25(OH)₂D₃ decreased ER mRNA (∼60%; P < 0.001) only at 24 h. The decreased levels of ER mRNA were transient, with ER mRNA returning to near normal levels by 48 h. Data from Northern blot analysis correlate with the observations from Western blot analysis and ligand binding studies, which showed an ∼50% drop in the ER protein levels after 48 h of treatment.

To determine whether the effect of 1,25(OH)₂D₃ on ER required de novo protein synthesis, we assessed the ability of 1,25(OH)₂D₃ to regulate ER mRNA in the presence and absence of cycloheximide, a potent inhibitor of protein synthesis. MCF-7 cells were treated with 100 nm1,25(OH)₂D₃ in the presence of cycloheximide (10 μg/ml of culture medium) for 24 h. Controls received either ethanol vehicle or cycloheximide alone. At the end of 24 h, total RNA was extracted from each treatment group and used for Northern blot analysis. As can be seen from Fig. 6, inhibition of protein synthesis by cycloheximide did not prevent the 1,25(OH)₂D₃-mediated decrease in steady-state levels of ER mRNA compared with either the ethanol-treated cells or the cells treated with only cycloheximide. An∼75% decrease in ER mRNA was seen 24 h after treatment with 1,25(OH)₂D₃ (Fig. 6, Lanes 2 and 4), suggesting that ongoing protein synthesis is not necessary for the down-regulation of ER mediated by 1,25(OH)₂D₃.

To measure the half-life of ER mRNA, MCF-7 cells were pretreated with 100 nm1,25(OH)₂D₃ or vehicle, and transcription was terminated at the end of 24 h by the addition of 4 μm actinomycin D. RNA was isolated at various times after the addition of actinomycin D and subjected to Northern blot analysis of ER and L7. ER/L7 mRNA at the time of actinomycin D addition (zero time) was represented as 100% for both control and 1,25(OH)₂D₃treated cells. It is clear from Fig. 7 that the half-life of ER mRNA did not decrease with 1,25(OH)₂D₃ treatment;rather, a small increase was seen. The half-life was ∼4.5 h in the controls versus ∼6 h in the 1,25(OH)₂D₃-treated cells,a difference that was not statistically significant, suggesting that 1,25(OH)₂D₃ had no effect on ER mRNA turnover.

To determine whether the 1,25(OH)₂D₃ induced decrease in ER mRNA was a transcriptional event, ER gene transcription run-on assays were performed using nuclei isolated from MCF-7 cells treated with either 1,25(OH)₂D₃ or ethanol vehicle for various time intervals. L7 gene transcription was used as an internal control. ER:L7 mRNA ratio was calculated for vehicle-treated controls for each time point and was represented as 100%. ER:L7 mRNA for 1,25(OH)₂D₃-treated cells were calculated and represented as a percentage of its corresponding vehicle-treated control. Fig. 8 demonstrates a decrease in the level of ER gene transcription with 1,25(OH)₂D₃ treatment. ER transcription decreased at 6 h (50% inhibition) and reached a nadir at 17 h (80% inhibition) before rising back to control levels at 48 h. The decrease was specific for ERtranscription because L7 did not change with 1,25(OH)₂D₃ treatment. These data suggest that 1,25(OH)₂D₃ down-regulates ER gene expression at the transcriptional level.

This set of experiments evaluated the impact of ER regulation on estrogen-dependent responses, i.e., the induction of PR and expression of BRCA1, which is stimulated by E₂.

Fig. 9 A shows the changes in PR induction by E₂ in response to 1,25(OH)₂D₃ treatment. Cell extracts were made at the end of 2 days of treatment with E₂,1,25(OH)₂D₃, or a combination of both; and PR levels were determined. 1,25(OH)₂D₃ by itself did not have any effect on PR levels. A 6-fold increase in PR levels was seen at the end of 2 days of E₂ treatment when compared with controls. E₂ failed to induce a measurable increase in PR in the presence of 1,25(OH)₂D₃, suggesting that this functional response of ER was suppressed in the presence 1,25(OH)₂D₃.

BRCA1 expression was assessed by Western blot analysis in MCF-7 cells treated with 10 nm E₂ and/or 100 nm 1,25(OH)₂D₃. The immunoreactive protein representing BRCA1 could be seen as a single M_r 210,000 species using the C-20 anti-BRCA1 monoclonal antibody (Fig. 9 B). Detectable increases in BRCA1 levels were seen at the end of 2 days of E₂ treatment (Lane 3). 1,25(OH)₂D₃, on its own,did not have any effect on the levels of BRCA1 protein (Lane 2). However,1,25(OH)₂D₃ suppressed the estrogen-mediated rise in BRCA1 levels (Lane 4). No changes could be seen in the levels of actin used to control for loading differences.

The growth-stimulatory effects of E₂ on breast cancer cells have been well established (26, 27). MCF-7, the most studied and well-characterized breast cancer cells,have been shown to express high levels of ERα (28). Although there have been reports indicating that ERβ may also be expressed in MCF-7 cells (29, 30), we could not detect any ERβ mRNA by reverse transcription-PCR in our cultures of MCF-7 cells. Hence, we have interpreted our observations on the assumption that the MCF-7 cells used in this study express only ERα. Further investigations are needed to ascertain whether these observations can be extrapolated to other breast cancer cell lines. It will also be of great interest to study the effects of 1,25(OH)₂D₃ on ER regulation in cells that express both the α and β isoforms.

In our studies with MCF-7 cells,1,25(OH)₂D₃ and its structural analogues clearly demonstrate dose-dependent antiproliferative properties (Fig. 2). These results confirm earlier observations (10, 11, 13, 14, 31, 32, 33, 34, 35) and show that the analogues, in addition to being less calcemic (36, 37),are more potent than 1,25(OH)₂D₃ in their ability to inhibit growth of MCF-7 cells. The growth-inhibitory effects are more evident in serum containing medium than medium containing CSS. Furthermore, 1,25(OH)₂D₃and its analogues are antiproliferative, even in the presence of added E₂. Although 1,25(OH)₂D₃ and Ro 23-7553 could only partially counteract the E₂-mediated growth of MCF-7 cells, the other analogues are clearly more potent because they are able to completely abolish the stimulatory effects of E₂.

We next investigated whether the antiproliferative effects of 1,25(OH)₂D₃ on the MCF-7 cells could be attributable to its action to down-regulate ER levels. Although previous studies have shown that 1,25(OH)₂D₃ down-regulates ER protein, the results in different studies have been variable (13, 14). Using ligand binding studies and Western blot analysis, we have demonstrated significant down-regulation of ER levels in medium containing CSS as well as in the presence of complete serum. There is a high degree of correlation(R² = 0.98) between the ability of the different 1,25(OH)₂D₃analogues to inhibit cell growth and their ability to decrease ER levels. These results suggest that the actions of 1,25(OH)₂D₃ and its analogues to inhibit cell growth might be, in part, attributable to their ability to down-regulate ER levels. However, because 1,25(OH)₂D₃ has also been shown to inhibit the growth of ER-negative breast cancer cell lines (35), we emphasize that ER down-regulation is only one of several pathways through which 1,25(OH)₂D₃ and its analogues act to inhibit breast cancer cell growth.

In an attempt to elucidate the mechanism of ER regulation, we studied the effect of 1,25(OH)₂D₃on the steady-state levels of ER mRNA and ER gene transcription. The steady-state levels of ER mRNA are decreased 60–80% (Figs. 5 A and 6) by 1,25(OH)₂D₃, indicating that the regulation occurs at the mRNA level. 1,25(OH)₂D₃ does not significantly alter the half-life of ER mRNA, and the decrease in ER mRNA is not dependent on new protein synthesis, suggesting that the 1,25(OH)₂D₃ effects on ER are transcriptional rather than posttranscriptional. This hypothesis is confirmed by nuclear run-on assays, which demonstrate significant decreases in ER gene transcription after 1,25(OH)₂D₃ treatment. The magnitude of changes in transcription rate is more pronounced, and they seem to occur earlier than those observed with the changes in steady-state ER mRNA levels. The trends, however, are similar. Taken together, these data suggest that the predominant mechanism of 1,25(OH)₂D₃-mediated down-regulation of ER gene expression is at the transcriptional level.

Evidence from our study supports the hypothesis that the 1,25(OH)₂D₃-bound VDR interacts directly with a nVDRE in the ER gene promoter inhibiting its transcription. Further support for this hypothesis will require the identification of an nVDRE in the ER gene promoter. nVDRE sequences have been described in other genes (38, 39).

Coincident with decreases in ER levels, the functional responses to E₂ are also attenuated because of 1,25(OH)₂D₃ treatment (Fig. 9). It was demonstrated earlier (17) that EB-1089, a potent analogue of vitamin D, down-regulates ER expression in MCF-7 cells and limits E₂ responsiveness measured as the induction of PR protein and pS2 mRNA. In our study, the E₂ induction of PR protein is completely inhibited by 1,25(OH)₂D₃,although there is only a 50% decrease in ER levels. Thus, the attenuation of the ER functional response is greater than that of ER down-regulation. One possible explanation for this finding is that 1,25(OH)₂D₃ might act at multiple sites in the E₂-mediated pathway. Demirpence et al. (16) have demonstrated that 1,25(OH)₂D₃ decreases ER binding to an ERE element, suggesting that in addition to ER regulation, 1,25(OH)₂D₃ has other sites of action in the ER pathway to inhibit E₂-mediated transactivation of target genes.

We also studied the effect of 1,25(OH)₂D₃ on BRCA1 expression, which has been implicated in the development and/or progression of hereditary breast cancer (40). In addition to its role as a tumor suppression gene (41, 42), BRCA1 has been shown to be regulated by E₂ (43, 44) and also linked to cell cycle events (45). Marks et al.(46) have suggested that the increased expression of BRCA1 with E₂ treatment is attributable to an increase in the percentage of cells undergoing DNA synthesis. If this is indeed the case, a decrease in BRCA1 levels would be expected with 1,25(OH)₂D₃ treatment,because 1,25(OH)₂D₃inhibits MCF-7 cell proliferation (Fig. 1). However, Campbell et al. (47) have shown that vitamin D analogues increase the expression of BRCA1 in MCF-7 cells. In our experiments, BRCA1 expression remained unchanged with 1,25(OH)₂D₃ treatment alone. However, 1,25(OH)₂D₃caused a significant reduction of the E₂-mediated increase in BRCA1 protein expression, probably because of the down-regulation of ER.

The ability of 1,25(OH)₂D₃and its analogues to down-regulate ER levels and suppress E₂ actions may have important clinical ramifications. 1,25(OH)₂D₃and its analogues have received increasing attention as potential therapeutic agents in the treatment and/or prevention of cancers in a number of organs including breast, colon, and prostate. However,1,25(OH)₂D₃ given in pharmacological doses routinely induces hypercalcemia, thus limiting its use in a clinical setting (48). Because the analogues are more potent in addition to being less calcemic, it is hoped that they will exhibit improved efficacy with reduced side effects as anticancer agents.

In conclusion, our evaluation of the effects of 1,25(OH)₂D₃ on MCF-7 cells is based on two different actions: (a) its effect to down-regulate ER expression; and (b) its effect to inhibit growth. Our data show a high degree of correlation between the antiproliferative properties of 1,25(OH)₂D₃ and its analogues and their ability to decrease ER expression. On the basis of these observations, we speculate that some of the growth-inhibitory actions of 1,25(OH)₂D₃ and its analogues in ER(+) cells are linked to their ability to down-regulate ER levels. The major mechanism of this effect is a suppression of the ER gene transcription. It is likely that 1,25(OH)₂D₃ acts at several points along the estrogen response pathway, affecting the levels of ER as well as their ability to function as enhancers of transactivation. This study adds down-regulation of ER to the many postulated mechanisms by which 1,25(OH)₂D₃inhibits breast cancer cell growth.

A potential nVDRE has recently been identified in the ER promoter (Stoica et al., J. Cell Biochem., 75:640–651, 1999).

We thank Dr. M. Uskokovic (Hoffmann La-Roche Co., Nutley, NJ)for providing the 1,25(OH)₂D₃ Ro 27-0574 and Ro 23-7553; Dr. L. Binderup (Leo Pharmaceuticals, Ballerup, Denmark)for providing us with the EB-1089 and KH-1060; and Dr. P. Chambon(University of Louis Pasteur, Strasbourg, France) for providing the human ER cDNA. The authors thank Drs. Xiao Yan Zhao and Stephen Sarabia for assistance in the preparation of the manuscript.