Prevention of Brca1-Mediated Mammary Tumorigenesis in Mice by a Progesterone Antagonist

See allHide authors and affiliations

Science  01 Dec 2006:
Vol. 314, Issue 5804, pp. 1467-1470
DOI: 10.1126/science.1130471


Women with mutations in the breast cancer susceptibility gene BRCA1 are predisposed to breast and ovarian cancers. Why the BRCA1 protein suppresses tumor development specifically in ovarian hormone–sensitive tissues remains unclear. We demonstrate that mammary glands of nulliparous Brca1/p53-deficient mice accumulate lateral branches and undergo extensive alveologenesis, a phenotype that occurs only during pregnancy in wild-type mice. Progesterone receptors, but not estrogen receptors, are overexpressed in the mutant mammary epithelial cells because of a defect in their degradation by the proteasome pathway. Treatment of Brca1/p53-deficient mice with the progesterone antagonist mifepristone (RU 486) prevented mammary tumorigenesis. These findings reveal a tissue-specific function for the BRCA1 protein and raise the possibility that antiprogesterone treatment may be useful for breast cancerpreventioninindividuals with BRCA1 mutations.

Mutations in the breast cancer susceptibility gene BRCA1 are associated with an increased risk of breast and ovarian cancers (1). Reduced BRCA1 expression due to promoter methylation may contribute to breast cancer progression (2). The BRCA1 protein has been implicated in DNA damage repair, cell cycle checkpoint control, and transcriptional regulation [reviewed in (3, 4)]. The specific suppression of breast and ovarian carcinogenesis by the pleiotropic BRCA1 tumor suppressor has been attributed to its regulation of estrogen receptor α (ERα) and two progesterone receptor isoforms (PRs) (58), which play important roles in breast development (915). BRCA1 interacts with ER and PRs directly and modulates ligand-dependent and -independent transcription activities of ERα and PR, as well as nongenomic functions of ERα (58). However, the mechanisms by which the ER and PRs contribute to BRCA1-mediated carcinogenesis remain unclear.

Hormone replacement therapy with progesterone and estrogen, but not estrogen alone, has been associated with an elevation in breast cancer risk in postmenopausal women (1618). In mice, the long isoform of PR, PR-B, is required for full development of mammary gland (15, 19), and overexpression of the short isoform, PR-A, leads to abnormal mammary gland development and ductal hyperplasia (20). These results are consistent with the hypothesis that PRs play a role in breast carcinogenesis.

To address the specific roles of ER and PRs in Brca1-mediated tumorigenesis, we studied p53f5&6/f5&6Crec and Brca1f11/f11p53f5&6/f5&6Crec mice (fig. S1A) (21, 22). Inactivation of both Brca1 and p53 genes in the mouse mammary gland mimics the majority of human BRCA1-associated tumors, which also harbor p53 mutations (3, 4). Brca1 p53f5&6/f5&6Crec mammary glands from nulliparous mice at 2.5 months of age showed about 4.5-fold more branching points compared with wild-type or p53f5&6/f5&6Crec glands (Fig. 1A and fig. S1B). By 4 months of age, the nulliparous Brca1f11/f11p53f5&6/f5&6Crec mammary gland showed further accumulation of side branches and extensive alveolar formation (Fig. 1B). The mammary gland morphology of mature, nulliparous Brca1f11/f11p53f5&6/f5&6Crec was similar to that of wild-type pregnant mice, suggesting that proliferation of mammary epithelial cells (MECs) was altered. Proliferation of MECs is regulated by ovarian hormones (23). In the estrous phase, MEC proliferation as measured by 5-bromo-2-deoxyuridine (BrdU) incorporation was about five times higher in the Brca1f11/f11p53 f5&6/f5&6Crec mice than it was in wild-type or p53f5&6/f5&6Crec mice (Fig. 1C and fig. S1C). Increased MEC proliferation in Brca1f11/f11p53f5&6/f5&6Crec mice was also seen in the diestrous phase (Fig. 1C and fig. S1C). Previous studies have shown that progesterone exerts its functional effects through paracrine action (24). Indeed, BrdU-positive MECs were found adjacent to PR-positive cells; there were also BrdU and PR double-positive MECs in the hyperplastic Brca1f11/f11p53f5&6/f5&6Crec mammary gland (fig. S2), indicating that the paracrine action of PR was maintained, at least in most cases.

Fig. 1.

Mutation in Brca1/p53 leads to increased mammary ductal branching, alveologenesis, and proliferation. (A) Number of branching points in mammary glands of 2.5-month-old wild-type (wt), p53f5&6/f5&6Crec, and Brca1f11/f11p53f5&6/f5&6Crec mice was determined. The data represent averages of branch points in five randomly selected areas ± SD. (*P ≤ 0.05) (B) Alveolar development in 4-monthold p53f5&6/f5&6Crec and Brca1f11/f11p53f5&6/f5&6Crec mice. Arrows indicate alveoli. (C) Proliferation of mammary epithelial cells was determined at different estrous phases in wt, p53f5&6/f5&6Crec, and Brca1f11/f11p53f5&6/f5&6Crec mice by BrdU incorporation. Histogram shows the average number of BrdU-labeled cells per duct ± SD (*P ≤ 0.05). At least 15 mammary ducts per animal were evaluated (a minimum of three mice per genotype).

To assess the contribution of circulating estrogen and progesterone on MEC proliferation, we treated ovariectomized mice with progesterone daily for 3 days, followed by BrdU pulse-labeling 2 weeks after surgery. In vehicle-treated mice, no BrdU-positive MECs were found in wild-type and p53f5&6/f5&6Crec mammary glands, but ducts containing one or two BrdU-positive cells were detected in Brca1f11/f11p53f5&6/f5&6Crec mammary glands (Fig. 2A and fig. S3A). Importantly, the number of BrdU-positive MECs in Brca1f11/f11p53f5&6/f5&6Crec mammary glands increased significantly upon exposure either to estradiol or progesterone alone or to a combination of both hormones (Fig. 2A and fig. S3A). The strong mitogenic effect of estradiol and progesterone on Brca1f11/f11p53f5&6/f5&6Crec MECs prompted us to examine the expression of ER and PRs by immunostaining. No difference in ER expression was detected during diestrous or estrous phase.

Fig. 2.

Mitogenic effect of progesterone on Brca1f11/f11p53f5&6/f5&6Crec mammary gland and stabilization of progesterone receptor in Brca1f11/f11p53f5&6/f5&6Crec mammary epithelial cells. (A) Ovariectomized mice (14 to 20 weeks old) were treated with vehicle and 1 μg of E2, 1 mg of P4, or E2 and P4 (E2+P4) for 3 days, and BrdU was injected 2 hours before sacrifice. BrdU-positive MECs were detected by immunostaining and quantified in 15 mammary ducts. Average number of BrdU-positive cells in ovariectomized, E2-, P4-, and E2+P4-treated mice is shown in the histogram (error bars indicate SE, *P ≤ 0.05). (B and C) Expression of PR protein in wt, p53f5&6/f5&6Crec, and Brca1f11/f11p53f5&6/f5&6Crec mice. Mammary gland tissues harvested at the diestrous or estrous phase were subjected to immunohistochemical staining by using PR antibody (anti-PR). Histogram represents the average percentage of PR-expressing cells per duct (error bars indicate SD). A minimum of five ducts per animal was evaluated (B). Whole-cell extracts from the mammary gland tissues as in (B) were used for immunoprecipitation by using PR antibody followed by Western blotting analyses (C). Arrow indicates PR-A; arrowhead indicates PR-B. (D) Effect of proteasome inhibitor, MG132, on PR protein quantities in p53f5&6/f5&6Crec and Brca1f11/f11p53f5&6/f5&6Crec MECs. Cells were starved for 4 hours and then treated with or without MG132 (10 μM) and R5020 (10 nM) for 6 hours as indicated. Western blotting for PR was performed. β-actin served as a loading control. (E) Half-life of PR. MECs were treated with 100 μg/ml cyclohexamide and 10 nM R5020. Cells were harvested at the indicated time points. p53f5&6/f5&6Crec MECs, open circles; Brca1f11/f11p53f5&6/f5&6Crec MECs, solid circles.

We next evaluated PR protein expression. In diestrous phase, PR was detected in the nuclei of a scattered subset of epithelial cells in mice of all genotypes (Fig. 2B and fig. S3B). In estrous phase, 86.8 ± 4% of Brca1f11/f11p53f5&6/f5&6Crec MECs were PR positive, compared with 27.8 ± 3.4% and 28.2 ± 2.4% of MECs in wild-type and p53f5&6/f5&6Crec mammary glands, respectively. Elevated expression of PR-A in Brca1f11/f11p53f5&6/f5&6Crec mammary gland in the estrous phase was confirmed by Western blotting (Fig. 2C). Consistent with a previous report, only a low amount of PR-B was detected in nulliparous mice (Fig. 2C) (25). The staining pattern of PR in Brca1+/– MECs was similar to that of Brca1f11/f11p53f5&6/f5&6Crec MECs (fig. S3B), indicating that Brca1 deficiency correlates with PR accumulation. No difference in PR transcript amounts was found between different genotypes (fig. S3C) or in T47D cells with or without BRCA1 suppression (fig. S4A). The elevated PR protein quantity in Brca1-deficient MECs was accompanied by overexpression of the PR target gene Bcl-xl (fig. S4B). Interestingly, PR expression is increased in normal MECs of breast cancer patients with a germline mutation of the BRCA1 gene (26). Thus, BRCA1 may regulate PR at the posttranscriptional level.

To explore this possibility, we established MEC cultures from the mammary glands of Brca1f11/f11p53f5&6/f5&6Crec and p53f5&6/f5&6Crec mice at 2 months of age. Ligand treatment induced pronounced degradation of PR-A in p53f5&6/f5&6Crec MECs (Fig. 2D, lanes 1 and 2) compared with Brca1 f11/f11p53 f5&6/f5&6Crec MECs (Fig. 2D, lanes 5 and 6). PR becomes polyubiquitinated upon exposure to progesterone and is subsequently degraded by the proteasome (27). Thus, we found that treatment with the proteasome inhibitor, MG132, led to accumulation of PR-A in p53f5&6/f5&6Crec MECs (Fig. 2D, lanes 2 and 4) but not in Brca1f11/f11p53f5&6/f5&6Crec MECs (Fig. 2D, lanes 6 and 8). In the presence of cyclohexamide, 90% of PR-A remained in Brca1f11/f11p53f5&6/f5&6Crec MECs, whereas only 40% of PR-A was detected in p53f5&6/f5&6Crec MECs 6 hours after the addition of the synthetic progesterone R5020 (Fig. 2E). These data suggest that Brca1 regulates PR stability.

To confirm that BRCA1 exerts a similar regulatory role in human breast cancer cells, we depleted BRCA1 with small interfering RNA (siRNA) in T47D cells. This treatment led to a 3.5-fold increase in PR-A protein as well as an increase in PR-B, but it did not affect the ERα protein concentration (Fig. 3A, lanes 1 and 2). Treatment with MG132 stabilized both PR isoforms (PR-A and PR-B) in control cells but not in siBRCA1-treated cells (Fig. 3A, lanes 1 and 3 versus 2 and 4). In the absence of ligand, PR-A and PR-B were stable, and only a slight increase was detected in siBRCA1-treated cells (Fig. 3A, lanes 5 and 6). Ligand-induced phosphorylation of PR-B and PR-A, as demonstrated by the mobility shift, was not affected by BRCA1 suppression (Fig. 3A). Previous studies have shown that phosphorylation of Ser294 of PR-B is required for ligand-dependent proteasome degradation (27). However, BRCA1 suppression did not decrease PR-B phosphorylation at Ser294 (Fig. 3A).

Fig. 3.

Effect of BRCA1 on PR stability, phosphorylation, and polyubiquitination in human breast epithelial cells. (A) Effect of BRCA1 inhibition on ERα and PR quantities and PR phosphorylation. Human breast cancer T47D cells were infected with SiLuc or SiBRCA1 adenovirus. After being starved overnight, cells were treated with cycloheximide (80 μg/ml), R5020 (10 nM), and/or MG132 (10 μM) for 6 hours as indicated. Quantities of ERα and PR protein were compared by using Western blotting. Phosphorylation of PR-A and PR-B upon R5020 treatment was determined by mobility shift and change of protein mobility upon phosphatase treatment. PR phosphorylation at Ser294 was evaluated with Ser294-phospho-specific antibody. α-tubulin served as a loading control. (B) BRCA1 regulates PR ubiquitination. MCF10A cells were transiently co-transfected with hemagluttinin (HA)-tagged ubiquitin, PR-A, and BRCA1 constructs (wt or mutants) as indicated, followed by infection with adenovirus expressing siRNA targeting BRCA1 or luciferase. Before harvest, cells were treated with 10 μM MG132 for 2 hours, followed by incubation with 10 nM R5020 for 2 hours. Anti-PR immunoprecipitates (IP) were analyzed by immunoblotting (IB) by using HA (top) and PR (middle) antibodies. Anti-BRCA1 immunoblotting shows efficiency of siBRCA1 (bottom). PP1, protein phosphatase 1.

We next evaluated ligand-induced polyubiquitination of PR in control and BRCA1-depleted normal human MCF10A MECs. PR ubiquitination was reduced in BRCA1-depleted cells. Conversely, overexpression of wild-type BRCA1 increased the amount of ubiquitinated PR (Fig. 3B). A protein complex consisting of BRCA1 and BARD1 displays E3 ubiquitin ligase activity (28). To test whether E3 ligase activity of BRCA1 is required for PR polyubiquitination, we introduced wild-type, ubiquitin ligase–defective ring-domain mutant BRCA1C61G (28), and ZBRK1 interaction–defective mutant BRCA1Q356R (29, 30) into MCF10A cells. PR-A ubiquitination increased in cells overexpressing wild type and BRCA1Q356R, but not BRCA1C61G (fig. S5), indicating that E3 ligase activity of BRCA1 is required for PR polyubiquitination. However, PR failed to be ubiquitinated by BRCA1 and its interacting protein, BARD1, as assessed by an in vitro assay (fig. S6).

Because progesterone is a potent mitogen for Brca1f11/f11p53f5&6/f5&6Crec MECs, we next tested whether blockade of PR activity by a progesterone antagonist could prevent or delay mammary carcinogenesis in Brca1f11/f11p53f5&6/f5&6Crec conditional knockout mice. Mice were treated with a placebo pellet or with a pellet containing the antiprogesterone mifepristone (RU 486). The pellet released the drug over a 60-day period, and the mice were monitored weekly for tumor formation. The median tumor latency of Brca1f11/f11p53f5&6/f5&6Crec mice was 6.6 months (n = 25) with complete penetrance (Fig. 4A). All the control untreated mice as well as the placebo-treated mice (n = 4) developed palpable tumors by 8.7 and 5.2 months of age, respectively. In contrast, no palpable tumors were detected in the mifepristone-treated mice (n = 14) at 12 months of age. Five weeks of mifepristone treatment substantially reduced branching and suppressed alveologenesis in the mammary glands of Brca1f11/f11p53f5&6/f5&6Crec mice (Fig. 4B). By using R26R reporter mice to monitor Cre activity (31), we found LacZ-positive normal Brca1f11/f11p53f5&6/f5&6 MECs as well as hyperplastic foci in the mifepristone-treated mice (Fig. 4C). These foci did not progress to tumors, however. These results suggest that PR function is critical for Brca1-mediated mammary carcinogenesis and that antiprogesterone treatment can prevent or delay mammary carcinogenesis in Brca1/p53 conditional knockout mice. In contrast, previous work has shown that treatment of Brca1f11/f11p53f5&6/f5&6 MMTV-Cre mice with the selective estrogen receptor modifier, tamoxifen, can increase mammary tumor incidence, an effect attributed to the estrogenic activities of tamoxifen in Brca1-deficient cells (32).

Fig. 4.

Antiprogesterone treatment inhibits mammary tumorigenesis by decreasing ductal branching and alveolar proliferation in Brca1f11/f11 p53f5&6/f5&6Crec mice. (A) Nulliparous adult female Brca1f11/f11 p53 f5&6/f5&6Crec mice, ages 3 to 4 months, were implanted with either a pellet containing 35 mg/60 day constant-release mifepristone (n = 14) or a placebo pellet (n = 4). Mice were monitored weekly for tumor formation. (B) Mammary gland branching in control pellet (left) or mifepristone-treated (right) Brca1f11/f11p53f5&6/f5&6Crec mice. Mammary glands were removed 5 weeks after pellet implantation. (C) Whole mounts of mammary glands from age-matched Brca1f11/f11p53f5&6/f5&6Crec mice without (a and c) or with (b and d) mifepristone pellet implantation. Boxed areas in a and b were enlarged (c and d, respectively). Mammary glands were removed 60 days after pellet implantation. Positive staining with X-galactosidase for LacZ expression marks the cells with an active Cre transgene.

In a recent study, exposure to a progesterone pellet was found to dramatically increase mammary gland volume in Brca1 conditional knockout mice but had little effect in wild-type mice (8). Our findings of deregulated PR turnover and mitogenic effect of progesterone in Brca1-deficient MECs are consistent with these results (8). Importantly, the mifepristone-mediated inhibition of mammary tumorigenesis in our Brca1/p53-deficient model provides a molecular framework for future clinical evaluation of antiprogesterones as a potential chemopreventive strategy in women who carry BRCA1 mutations.

Supporting Online Material

Materials and Methods

Figs. S1 to S6


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article