The Chemopreventive Effect of Mifepristone on Mammary Tumorigenesis Is Associated with an Anti-invasive and Anti-inflammatory Gene Signature

Progesterone is a major physiologic regulator of reproduction through the mammary gland, uterus, ovary, and hypothalamic-pituitary axis (1), and its interaction with the progesterone receptor (PR) is essential for lobuloalveolar development of the mammary gland (2). Although its physiologic function is complex, the PR is believed to drive proliferation of mammary epithelial cells in 2 waves, the first requiring PR-positive cells, and the second wave comprising mostly PR-negative cells by a paracrine mechanism involving RANK ligand (RANKL; ref. 3). PR-mediated signaling is also important in the development of mammary tumorigenesis as shown by a marked reduction in 7,12-dimethylbenz(a)anthracene (DMBA)-initiated mammary carcinogenesis in PR-null mice (4). Progesterone and estrogen alone or in combination enhances mammary tumorigenesis in p53-null mice (5), and treatment of Balb/c, but not C57BL/6, mice with medroxyprogesterone induces invasive ER⁺/PR⁺ ductal mammary carcinomas (6). In ovariectomized ACI rats, both estrogen and progesterone induce a 100% incidence of mammary tumors over 5 to 9 months (7). More importantly, PR regulation is also relevant to human breast cancer (8). PR and not ER is a marker for early-stage breast cancer (9), and PR expression increases in tumors with BRCA1 mutations (10). Progestins, but not estrogens, reactivate a subset of estrogen receptor (ER)⁻/PR⁻/cytokeratin (CK)5⁺ cells with stem-like properties (11) that have the capacity to generate ER⁺/PR⁺/CK5⁻/CK18⁺ cells (12). This feature may be especially important with regard to the association between estrogen/progestin hormone replacement therapy (HRT) and increased risk of invasive lobular breast disease (13, 14), where the use of progestin oral contraceptives (15) and HRT (16) poses a greater breast cancer risk than estrogen alone.

It was appreciated early in the development of the PR antagonist, mifepristone/RU486, that it could serve as a contraceptive by interruption of the progesterone-dependent luteal phase of the menstrual cycle (17). Mifepristone, and second-generation PR antagonists with greater PR selectivity, such as onapristone, have produced a high percentage of stable disease in postmenopausal women with metastatic breast cancer (18). Experimental studies have shown that mifepristone can mitigate relapse to tamoxifen in MCF-7 breast cancer xenografts (19), inhibit the growth of progestin-stimulated T47-D and BT-474 xenografts (20), and prevent tumorigenesis resulting from inactivation of BRCA1 and p53 (21). In primary mammary tumor models, mifepristone inhibited DMBA- and N-methyl-nitrosourea (NMU)-induced tumorigenesis in rats (22, 23), which in one instance was not associated with classical PR-mediated effects (22). Importantly, mifepristone has high affinity for both the PR and glucocorticoid receptor (GR; ref. 24), which may contribute to its antitumor activity in androgen-dependent and -independent prostate cancer (25).

In the present study, we examined whether mifepristone was equally efficacious as a chemopreventive agent in reducing DMBA mammary carcinogenesis in wild-type (WT) and mouse mammary tumor virus (MMTV)-Pax8PPARγ (Pax8) transgenic mice. The latter mouse model is unique in that the mammary gland expresses an increased percentage of ER⁺/PR⁺/CK5⁺/CK19⁺ and double CK14⁺/CK18⁺ progenitor cells and presents with ER⁺/PR⁺ mammary carcinomas following progestin/DMBA-mediated carcinogenesis, in contrast to ER⁻/PR⁻ tumors of mixed lineage in WT mice (26). Thus, it was hypothesized that if mifepristone acted by inhibiting PR function, tumor formation in Pax8 mice would be more sensitive to mifepristone in comparison with WT mice. We report that in contrast to Pax8 animals, WT mice were more sensitive to a low dose of mifepristone and that their response correlated with a unique gene expression signature. These results suggest that mifepristone acts through other signaling pathways apart from PR and that further testing of mifepristone and other PR antagonists may be warranted in patients with breast cancer and other malignancies.

Pax8 transgenic Friends virus type B (FVB) mice were generated as previously described (26). Age-matched WT FVB mice were obtained from Taconic Farms. All animal studies were conducted under protocols approved by the Georgetown University (Washington, DC) Animal Care and Use Committee.

Five-week-old WT and Pax8 mice were treated with medroxyprogesterone acetate and DMBA as previously described (27, 28). Briefly, mice were injected s.c. with a single dose of 15 mg medroxyprogesterone acetate suspension (Depo-Provera), and after 7 days were administered 4 weekly doses of 1 mg DMBA/0.1 mL cottonseed oil by gavage. Mifepristone was provided by the Chemoprevention Branch, National Cancer Institute and formulated as 2.5 and 7.5 mg 60-day controlled release pellets and similarly formulated placebos by Innovative Research of America, Inc. Pellets were implanted by trocar s.c. into the scapular region 1 day after the last dose of DMBA and again after 60 days. The doses of mifepristone used in this study did not produce toxicity.

The source of antibodies, their dilution, and use were as follows: rabbit anti-ERα [sc-542, Santa Cruz Biotechnology, 1:200 for immunohistochemistry (IHC), 1:1,000 for Western blotting]; rabbit anti-PgR (sc-538, Santa Cruz Biotechnology, 1:200 for IHC, 1:1,000 for Western blotting); rabbit anti-GRα (sc-1004, Santa Cruz Biotechnology, 1:200 for IHC, 1:1,000 for Western blotting); rabbit anti-DKK2 (06-1087, Millipore Corp., 1:200 for IHC); and rabbit anti-Neu/ErbB2 (sc-33684, Santa Cruz Biotechnology, 1:200 for IHC).

Immunohistochemical analysis was carried out as previously described (26–28).

Western blotting was carried out as previously described (26). Briefly, tissue was frozen in liquid nitrogen and pulverized in a mortar and pestle and mixed with lysis buffer containing: 0.1% SDS, 0.5% NP-40, phenylmethylsulfonyl fluoride, 1 mmol/L sodium vanadate, 50 mmol/L sodium fluoride, 10 mmol/L β-glycerophosphate, 5 mmol/L sodium pyrophosphate, and protease inhibitor cocktail (Roche Diagnostics). Following incubation on ice for 30 minutes, lysates were cleared by centrifugation for 15 minutes at 13,000 × g at 4°C. Protein concentrations were determined by the Coomassie Plus Protein Assay (Pierce), and 50 μg of lysate was separated in a 4% to 12% NuPAGE Bis-Tris gel (Invitrogen). After wet transfer, membranes were blocked for 1 hour at room temperature in TBS (pH 7.4) containing 5% non-fat dry milk and 0.1% Tween-20. Primary antibody was incubated overnight at 4°C, and secondary antibody was incubated for 1 hour at room temperature. Proteins were visualized with either SuperSignal West Pico or SuperSignal West Dura (Pierce).

RNA was obtained from 2 to 3 individuals per group when animals were 5 to 6 months of age; only adenocarcinomas were analyzed and not tumors of other or mixed lineage. Total RNA was extracted using an RNeasy mini kit (Qiagen) following the manufacturer's protocol as previously described (26, 29), and the quality of RNA was confirmed by 28S/18S rRNA profiling using a microfluidic chip (Agilent). Equal amounts of RNA from each group were pooled for Affymetrix GeneChip analysis. cRNA was synthesized using the Affymetrix protocol with minor modifications as described (27). Biotin-labeled cRNA was fragmented for 35 minutes at 94°C and hybridized overnight to an Affymetrix mouse 430A 2.0 GeneChip representing approximately 22,000 annotated mouse genes. Hybridization signals were detected with an Agilent Gene Array scanner, and grid alignment and raw data generation conducted with Affymetrix GeneChip Operating software 1.1. Common genes expressed in tumors from WT and Pax8 mice treated with 2.5 mg mifepristone with a signal ≥300 (log₂ ≥8.1) and ≥3-fold change (28–30) were analyzed by Pathway Studio 7.1 (Ariadne). Array data were deposited in the GEO database under accession no. GSE33753.

Total RNA was extracted from 2 to 3 individual adenocarcinomas per group using an RNAeasy mini kit (Qiagen) according to the manufacturer's protocol as previously described (26, 29). The quality of RNA was confirmed by 28S/18S rRNA profiling using a microfluidic chip device, and equal amounts of RNA pooled from these samples for Affymetrix GeneChip analysis. Animals were 5 to 6 months of age when tumors were collected. One microgram of RNA was reverse-transcribed in a total volume of 20 μL using the Cloned AMV First-Strand cDNA Synthesis Kit (Invitrogen). PCR was carried out in triplicate in an ABI 7900 instrument (Applied Biosystems) using SYBR Green detection (Applied Biosystems) according to the manufacturer's protocol. Quantitative real-time PCR (qRT-PCR) primers were designed using the primer design tool at Integrated DNA Technology (31). Efficiencies of all primer sets (Supplementary Table S1) were validated using a standard curve of 5 serial cDNA dilutions in water in duplicate. Primers were acceptable if the deviation from the slope of the standard curve was less than 0.3 and if the melting curve showed only one product. The expression of each target gene was normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and the relative quantification method was applied using SDS2.3 software (Applied Biosystems). Primers are listed in Supplementary Table S1.

Survival curves were analyzed by the χ² log-rank test and tumor incidence by the 2-tailed paired Student t test at a significance level P ≤ 0.05.

To evaluate the chemopreventive effect of mifepristone on breast cancer development, mammary carcinogenesis was induced by progestin/DMBA in WT and Pax8 transgenic mice and animals were administered 2 doses of either placebo, 2.5 mg, or 7.5 mg mifepristone formulated as 60-day extended release pellets 60 days apart beginning 1 day after the last dose of DMBA (Fig. 1). Survival was increased in WT mice treated with either 2.5 or 7.5 mg mifepristone (Fig. 1A) and correlated with reduced tumor formation (Fig. 1B). In contrast, Pax8 mice were responsive only to the higher dose of mifepristone (Fig. 1C and D). Assessment of the proliferative marker Ki-67 indicated somewhat higher expression in Pax8 versus WT mice and a greater reduction in WT versus Pax8 mice following treatment with 2.5 mg mifepristone (Supplementary Fig. S1A). Consistent with these results was a greater degree of nucleosome cleavage as assessed by terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay in mifepristone-treated WT mice versus Pax8 mice (Supplementary Fig. S1B). Dissipation of the effects of mifepristone after 120 days (Fig. 1) resulted in an increase in tumorigenesis in both WT and Pax8 mice, suggesting that mifepristone must be continuously present to produce an antitumor effect.

Common genes differentially expressed in tumors from WT and Pax8 animals treated with either placebo or 2.5 mg mifepristone are presented in Table 1 and Fig. 2A and B. A complete list of ≥3-fold changes in gene expression is presented in Supplementary Tables S2 and S3. The response of WT and Pax8 mice treated with 2.5 mg mifepristone was defined by a subset of 79 genes common to both groups. Of this subset, 52 genes responded in an opposite fashion (highlighted in bold) and distinguished the antitumor response in WT mice to 2.5 mg mifepristone from the lack of response of Pax8 mice at this dose. Pathway linkage analysis of these changes in gene expression is depicted in Fig. 3. Most notable was the association between an anti-inflammatory and anti-invasive gene signature in tumors from WT mice treated with mifepristone, as reflected in the marked reduction in expression of Mmp13, Klk6, Klk7, S100a14, Saa1, and Saa2 (Table 1) and the gene ontology of these responses (Table 2). Interestingly, approximately 25% of the 52 genes are known to be regulated by PR and/or GR (Table 3), despite the lack of consistent changes in PR and GR expression by Western blotting (Supplementary Fig. S1C) and IHC (Supplementary Fig. S1D). In addition, there was no association between PR/GR-regulated genes and those regulated by ER (Table 3). Because expression of the Pax8PPARγ transgene is under the control of the MMTV long terminal repeat, which is regulated by the GR (32), the expression of Pax8PPARγ mRNA was assessed (Fig. 2C). There was no significant change in Pax8PPARγ mRNA levels by mifepristone treatment. Another notable change specifically associated with the antitumor response of WT mice to 2.5 mg mifepristone was the 11-fold increase in Dkk2 mRNA (Fig. 2A) and protein (Supplementary Fig. S1E), suggesting the possible involvement of Wnt pathway inhibition in the antitumor response to mifepristone.

The aim of the present study was to assess the chemopreventive efficacy of mifepristone in a progestin-dependent mammary carcinogenesis model and to correlate therapeutic response with the gene expression profile of tumors induced in WT and Pax8 mice. Pax8 mice were previously found to express a stem/progenitor cell phenotype in the mammary gland and to form ER⁺ adenocarcinomas exhibiting activation of Ras/extracellular signal-regulated kinase (ERK) and ER signaling following progestin/DMBA-induced carcinogenesis (26). Interestingly, several genes previously identified with an ER⁺ cluster in human breast cancer (33) were preferentially expressed in Pax8 mice, namely, Fgfr2, Gata3, Slc39a8, and Stc2 (Supplementary Tables S3 and S4). Pax8 mice also expressed Mmp11 identified in tumors with myoepithelial characteristics (34). Apart from these associations, it is difficult to draw further inferences about Pax8 mice with respect to a specific subtype of breast cancer. Nevertheless, Pax8 mice were more resistant to the lower dose of mifepristone than WT mice, whereas the higher dose led to marked suppression of tumorigenesis in both groups of animals. It was previously reported that resistance to mifepristone in a progestin-dependent mouse mammary tumor correlated with high activation of ERK (35), a finding consistent with resistance in Pax8 mice to the lower dose of mifepristone. Mifepristone is known to block activation of insulin-like growth factor (IGF)-1 signaling in MCF-7 breast cancer cells (36), and inhibition of this and other growth factor–dependent pathways may have also contributed to its chemopreventive effect. In this context, expression of the Wnt pathway inhibitor, Dkk2 (37), was markedly elevated in WT tumors sensitive to the lower dose of mifepristone (Fig. 2A and Supplementary Fig. S1E and Table S2). This finding is consistent with the induction of Wnt4 expression in the mammary gland by progesterone and its role in ductal branching (38). In the endometrium, the progesterone-dependent luteal phase is associated with increased Wnt and reduced Dkk2 gene expression (39, 40) and is in agreement with the reduction of Dkk2 in mammary tumors from mifepristone-sensitive WT mice.

One of the most notable features of mifepristone sensitivity in WT mice was the marked reduction in expression of genes associated with invasion and inflammation, including Klk6, Klk7, Mmp13, Saa1, and Saa2 (Table 1 and Fig. 2B). Kallikreins are secreted proteases involved in extracellular matrix degradation and metastasis and represent negative prognosticators for several malignancies, including breast cancer (41, 42). Not only is Klk6 a PR target gene (43) but it and other kallikrein members are involved in the regulation and activation of matrix metalloproteinase (MMP; ref. 41), including Mmp12 and Mmp13. The latter MMPs are associated with metastatic breast cancer (44) and Mmp13 has been linked to several inflammatory conditions and malignancies (45). In association with the reduction in invasive gene expression by mifepristone in WT tumors was the marked decrease in Saa1 and Saa2 expression. The serum amyloid A family are acute-phase proteins that induce several MMPs (46) and are prognostic for reduced survival of patients with breast cancer (47). Thus, a previously unrecognized feature of mifepristone appears to be its ability to suppress the expression of several inflammatory and metastatic genes, and it will be of interest to assess whether it has antimetastatic activity in an appropriate animal model or patient setting.

Mifepristone may also act through non–PR-dependent pathways, such as those related to GR and ERK signaling. In the context of our progestin-dependent mammary carcinogenesis model (27, 48), the involvement of GR and ERK is unlikely. This conclusion is based, in part, on the lack of an effect by mifepristone on Pax8PPARγ mRNA expression (Fig. 2C) and the regulation of the MMTV long terminal repeat by the GR (32, 49). In addition, mifepristone has been reported to inhibit ERK activation (50), which is associated with mammary tumors derived from MMTV-Pax8PPARγ mice (26), which were resistant to the lower dose of mifepristone (Fig. 1). Overall, these results suggest that anti-GR activity is not a significant factor in the antitumor effects of low-dose mifepristone treatment

To obviate the potential side effects of PR antagonists used for contraception, selective PR modulators (SPRM) were developed with little or no antagonism of other receptors, such as GR by mifepristone (51). However, in contrast to the higher potency of SPRMs in uterine tissue, little change in potency or efficacy was observed for PR-selective agents, such as ORG31710, in comparison with mifepristone in DMBA-induced breast tumors in rats (18, 52). This suggests that the antitumor effects of PR antagonists may be dissociated from its antiprogestational effects (22, 53) and is consistent with the lesser sensitivity of Pax8 mice with greater levels of PR (Fig. 2C, Supplementary Fig. S1D), in comparison with WT mice.

Previous clinical trials of mifepristone in patients with relapsed postmenopausal breast cancer with PR⁺ metastatic disease who had received prior adjuvant hormone/chemotherapy produced only a partial response in approximately 10% of patients (54). The use of mifepristone as a second-line intervention in patients previously treated with tamoxifen (55) or presenting with bone metastases (56) produced a similar outcome. Nevertheless, mifepristone exhibited synergistic inhibitory activity with tamoxifen and an aromatase inhibitor on mammary tumorigenesis (18), and similar effects were noted on breast cancer cells in culture (57, 58). The possible use of PR antagonists is further suggested by the efficacy of mifepristone in an experimental model of Brca1-mediated tumorigenesis, where increased survival was attained (21). Thus, SPRMs have the potential to reduce the emergence of resistance to antihormone monotherapy and treat cancers highly resistant to most therapeutic modalities.

No potential conflicts of interest were disclosed. This investigation was conducted using the Animal Research, Genomics and Epigenomics, and Microscopy and Imaging Shared Resources of the Lombardi Comprehensive Cancer Center. The content does not necessarily represent the official views of the National Cancer Institute or the NIH.

Conception and design: H. Yuan, L. Kopelovich, R.I. Glazer.

Development of methodology: H. Yuan, L. Kopelovich, R.I. Glazer.

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Yuan, G. Upadhyay, R.I. Glazer.

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Yuan, G. Upadhyay, J. Lu, L. Kopelovich, R.I. Glazer.

Writing, review, and/or revision of the manuscript: H. Yuan, L. Kopelovich, R.I. Glazer.

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H. Yuan, R.I. Glazer.

Study supervision: H. Yuan, L. Kopelovich, R.I. Glazer.

The study was supported by NIH grant 1NO1 CN43302-WA19, and award P30CA051008 from the National Cancer Institute, NIH, to the Lombardi Comprehensive Cancer Center.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.