Epimorphin Is a Novel Regulator of the Progesterone Receptor Isoform-A

The mouse mammary gland is a reticulum of branching tubules and alveolar structures that undergoes most of its development postnatally. Stages of development involve branching morphogenesis that commences at puberty, functional differentiation in pregnancy and lactation, and involution of alveolar structures after weaning. These stages are reproduced during estrous cycles albeit to a less dramatic extent. Endocrine hormones such as 17β-estradiol, progesterone, prolactin, and glucocorticoids are essential for normal mammary gland development (reviewed in ref. 1). Genetic ablation of the receptors for 17β-estradiol and progesterone, Esr1 and Pgr, in mice, revealed that Esr1 is required for ductal elongation whereas Pgr is critical for alveologenesis to ensue (reviewed in ref. 2). Through Esr1 and Pgr activation, signals for mammary gland development are propagated by upregulation of transcription factors and release of local factors. Examples of local factors include EGF and keratinocyte growth factor for Esr1 (reviewed in ref. 3), and receptor activated NF-κB ligand (RANKL) for Pgr (4), and Esr1 is a primary transcriptional regulator of Pgr in the mammary gland (5). Here, we have made the surprising discovery that Pgr expression is also regulated transcriptionally by the local mammary morphogen, Epim.

Epim was first identified as a surface mesenchymal protein that was crucial for dermal and lung epithelial cell morphogenesis (6). Subsequently, Epim was rediscovered as a member of a family of integral membrane proteins renamed syntaxins (7) that were shown to have important intracellular functions in membrane fusion and exocytosis (reviewed in ref. 8). However, because of the distinct extracellular function of Epim, which was discovered before the syntaxin nomenclature and function, we continue to refer to the extracellular form as Epim and the intracellular function as Stx-2 (9, 10). Epim has been shown to regulate morphogenesis in a range of tissues including bile duct, intestinal crypts, hair follicle, skin, pancreas, gall bladder, testes, and lung, as well as in the mammary gland (reviewed in ref. 10). Epim is expressed in the mouse mammary gland by fibroblasts and myoepithelial cells in virgin mice (11), and additionally in luminal epithelial cells in pregnant and lactating mice (unpublished data). In three-dimensional (3D) collagen gel assays, we showed that Epim-stimulated mammary epithelial cell branching and lumen formation (11–13), and that these Epim functions were dependent upon binding to its cell surface receptor, integrin-αV (13). WAP-Epim transgenic mice (14) develop thickened mammary ducts, upregulate the transcription factor CCAAT enhancer binding protein β (Cebpb) in pregnancy, and develop hyperplasia and neoplasia at approximately 18-months after birth (15).

To determine why WAP-Epim mice are prone to mammary hyperplasia and neoplasia after 1 year of birth, we asked whether these mice exhibit histologic stages of cancer progression even earlier in life. We hypothesized that by analyzing young mice we could identify which molecular pathways become deregulated during this progression. We found that these mice develop also extensive lateral branches in nulliparous mice (week 8) and precocious alveolar structures in young pregnant mice (week 12) and increased mammary epithelial cell proliferation during alveologenesis. These phenotypes resembled those described for Pgr-a transgenic mice (16). In addition, WAP-Epim mice develop hyperplasia similar to Pgr-a transgenic mice. Because of the similarities between Epim and Pgr-a transgenic mice (15–18) and the fact that both Epim and Pgr upregulate Cebpb (14, 19, 20), we reasoned that there was a connection between the 2 molecules.

Interrogation of a downstream mechanism in both WAP-Epim transgenic mice and the ex vivo 3D assays showed that Epim indeed regulates Pgr-a, and that Epim controls Pgr-induced side branching, bud formation, and epithelial cell proliferation. We show also that Epim regulation of Pgr-a expression occurs through activation of integrin-αV. Furthermore, we show that Epim-induced proliferation and bud formation, which we showed previously to correlate with Mmp3 expression, occurs downstream of Pgr.

The generation of hemizygous WAP-Epim mice (abbreviated further as TG in figures), in which Epim is tagged with the mouse IL-2 signal peptide sequence and expressed under control of the whey acidic protein promoter, has been described previously (14). Transgene-negative littermates were used as controls. Animal use protocols were obtained and procedures were followed in strict accordance with guidelines established by the Lawrence Berkeley National Laboratory Animal Welfare and Research Committee (AWRC).

Nulliparous mice were analyzed at 1.5, 8, and 14 weeks after birth. For analysis of alveolar development, tissue was collected from pregnant wild-type and WAP-Epim animals at day 12 of pregnancy. To stage pregnancy, breeding mice were checked in the morning for vaginal plugs. If plugs were found, the female was separated from the male and this day was designated day 0 of pregnancy. For analysis of lactation, dams were allowed to nurse 6 pups to equalize suckling, tissue was collected on day 10 subsequent to parturition.

At the time of dissection, the stage of estrous was determined by vaginal lavage followed by cytological analysis. For each study, the thoracic and inguinal mammary glands were excised and frozen immediately on dry ice for RNA and protein isolation or they were formalin fixed for histologic analysis. One inguinal gland was fixed in Carnoy's solution overnight then stained with carmine alum to analyze ductal/alveolar morphology.

For genotyping, tail DNA was digested overnight in 50 μL proteinase K buffer, diluted 8×, and used as template for PCR reaction. The expression of the Epim transgene was confirmed by reverse transcriptase (RT)-PCR. For analysis of gene expression in mouse mammary glands, total RNA was extracted from frozen mammary glands using TRIzol (Invitrogen) or from mammary organoids using an RNeasy kit (Qiagen) then reverse transcribed using Superscript II First Strand Synthesis System (Invitrogen). qPCR was conducted using a LightCycler (Roche Diagnostics). Primers used in qPCR reactions are listed in Supplementary Table S1.

Histomorphometry to compare differences in epithelial density was conducted using a Mertz graticule on hematoxylin and eosin-stained 5 μm mammary gland paraffin sections generated by the UCSF Helen Diller Family Comprehensive Cancer Center Mouse Pathology Core. Five successive fields were examined for each mammary gland. The criteria included the presence or absence of epithelial structures or adipocytes. To quantify side branching, the 3 longest ducts were analyzed on each mammary gland whole mount beginning from the lymph node. The number of side branches was divided by the length to yield side branches/branch length.

From each animal, thoracic mammary glands were homogenized in 500 μL lysis buffer [10 mmol/L Tris (pH 7.6), 5 mmol/L EDTA, 50 mmol/L NaCl, 1% Triton-X] with 1× proteinase inhibitor cocktail I (CalBiochem, Merck KGaA) for immunoblotting. The homogenates were centrifuged at 12,000 × g for 20 minutes at 4°C, supernatant was isolated and stored at −70°C until needed. Protein was isolated from organoids as previously described (21). Protein concentration was determined using Bio-Rad DC protein assay reagents (Bio-Rad Laboratories Inc.).

For Western analysis, 10 μg protein lysate was added to loading buffer (250 mmol/L Tris-HCl pH 6.8, 10% SDS, 20% β-mercaptoethanol, 40% glycerol), boiled for 5 minutes, and electrophoresed on 12% polyacrylamide gels (Invitrogen). After transferring onto nitrocellulose membranes 0.45 μm (Bio-Rad Laboratories Inc.), total blotted protein was visualized by PonceauS staining. Blots were blocked overnight at 4°C using 5% nonfat dry milk in 1× Tris-buffered saline. Antibodies against β-casein, generated from hybridoma cells in the Bissell Lab used at a 1/10,000 dilution to test functional differentiation (22), the β-catenin mouse monoclonal antibody was from Sigma (C-7207), the Pgr antibody (Ab-8; Neomarkers), and lamin A/C (Lmna) and estrogen receptor antibodies were from Santa Cruz Biotechnology (H-110 and MC-20). Secondary horseradish peroxidase (HRP) conjugated antimouse and anti-rabbit antibodies were from Amersham Biosciences (GE Healthcare). Dilutions followed manufacturers' suggestion. Detection of HRP signal was achieved using Thermoscientific Supersignal West Dura with a FluorChem 8900 chemical imager (Alpha Innotech, Cell Biosciences).

Mammary sections (5 μm) were deparaffinized and rehydrated in graded alcohol to 1× PBS, then blocked for 10 minutes in 3% H₂O₂ in 70% ethanol, followed by antigen retrieval for 10 minutes in sodium citrate buffer (10 mmol/L sodium citrate, pH 6.0) using a microwave. Slides were incubated overnight in a humidified chamber with anti-Epim 1/50 dilution (MC-1 antibody, gift from Y. Hirai), anti-Cebpb at a 1/100 dilution (Santa Cruz Biotechnology), anti-cyclin D1(Ccnd1) at a 1/100 dilution (Thermoscientific), and anti-Pgr at a 1/50 dilution (DAKO) then washed in 1× PBS and incubated for 1-hour with biotinylated secondary antibodies. Avidin:biotinylated enzyme complex (ABC kit, Vector Laboratories) was used to amplify signal, followed by visualization using 3,3′-diaminobenzidine color substrate (Sigma).

To determine statistical significance, two-tailed paired T tests were used when treatments were to organoids or tissue from the same animal. For comparing differences in expression or morphology between various mice, two-tailed Mann–Whitney tests were used. A P-value of less than 0.05 was considered significant. All error bars represent SEM measurements.

We reported previously that WAP-Epim transgenic mice develop alveolar hyperplasia and adenocarcinoma with high incidence when they age to approximately 18 months; however, we did not know the molecular mechanism by which Epim overexpression could cause mammary tumors. To determine the mechanism(s) leading to malignancy, we investigated whether there were consequences to mammary gland development when Epim signaling becomes deregulated. To inhibit Epim, we used Pep7, a 10-amino acid peptide initially discovered to be the minimal portion of the H1 domain of Epim capable of inducing a telogen-to-anagen transition in mouse hair follicle morphogenesis (23), and that was found also to inhibit Eph4 mammary epithelial cell branching in collagen gels (24). We injected pep7, or a control scrambled peptide, into the contra-lateral mammary glands of wild-type mice and observed a significant decrease in side branching in pep7-injected mammary glands as compared with glands injected with control peptide (Fig. 1A and B). We did not, however, observe any difference in ductal length between pep7-treated mammary glands and contralateral scramble peptide-treated glands (not shown). Pgr is necessary also for side- branching but not ductal length in the mouse mammary gland (25). That we observed a decrease in side branching but not ductal length when using the Epim inhibitor suggested that Pgr might be involved. Thus, we proceeded to analyze young WAP-Epim transgenic mice to determine whether there are similar defects in ductal development as in Pgr transgenic mice.

Although the promoter of the WAP gene is activated most strongly from day 10 of pregnancy through lactation, with residual activity during involution (26), WAP is expressed also in nulliparous mice (27) and we observed upregulated Epim in the mammary epithelial cells of nulliparous WAP-Epim mice (Fig. 1C), associated with an increased ratio of epithelial cells (E) compared with adipocytes (A) as compared with the wild type (∼80% increased; Fig. 1D). To evaluate how overexpression of Epim-affected ductal outgrowth, mammary glands were isolated before, during, and after peak ductal outgrowth at 1.5, 8, and 14 weeks. At 1.5 weeks, WAP-Epim mammary glands were indistinguishable from wild type in terms of ductal length and the number of primary branches (n = 10, not shown). To control for hormonal differences between cycling 8- and 14-week-old wild-type and transgenic (TG) mice, they were compared only in the same estrous stage. By 8 weeks of age, mammary glands from WAP-Epim mice showed ∼80% more side branches than the wild type (Fig. 1E and F). In contrast, there were no statistical differences in ductal length (Fig. S1A). By 14 weeks, the wild-type mammary gland had caught up in the number of side branches compared with the WAP-Epim mouse (n = 11 wild type, n = 19 TG, not shown).

As a further control for the specificity of Epim action to induce side branching, we compared the effect of a recombinant soluble Epim with that of an engineered protein that is 82% homologous to Epim (rStx1a), but that does not activate Epim responses in culture (28). Using the same mouse, we injected rEpim peptide to one inguinal mammary gland and rStx1a peptide to the contralateral mammary gland. Whereas there were no significant differences in number of side branches or ductal length between the right and left inguinal mammary glands injected with PBS alone (not shown), the rEpim injected mammary gland showed a significant increase in side branches compared with the rStx1a treated contralateral side in the same mouse (Fig. 1G). Again, we did not observe a difference in ductal length as was the case also between WAP-Epim and wild-type mice, and the case between pep7 and mammary glands treated with scrambled peptides. Thus, attenuating Epim signaling early in ductal development hinders the elaboration of secondary branches, whereas overexpression of Epim stimulates precocious side branching without affecting ductal length. These phenotypes resemble both the defect in side branching in the Pgr knockout mice (25, 29) and the supernumerary side branching and hyperplasia in the Pgr-a transgenic mice (16).

Administration of 17-β estradiol and progesterone to mice where both Pgr isoforms were ablated showed that Pgr is necessary for alveolar differentiation in the mouse mammary gland (25). Given that WAP-Epim mice also show increased side branching, ductal ectasia, and hyperplasia similar to the Pgr-a transgenic mice (16, 17), we hypothesized that the action of Pgr to induce alveologenesis during pregnancy could be regulated also by Epim. Whole mounts and histologic sections of wild-type and WAP-Epim mammary glands showed a dramatic increase in the number of alveolar structures at day 12 of pregnancy compared with wild type (Fig. 2A and B), a time point at which Epim was highly overexpressed in the WAP-Epim mammary gland (Fig. 2C). WAP-Epim mammary glands displayed larger, more abundant alveoli with larger lumena, increased lipid droplets (Fig. 2B, arrow) and elevated β-casein expression (not shown) compared with wild type. We reported previously that mammary glands from pregnant WAP-Epim mice developed enlarged mammary ducts (14), a phenotype observed also in the Pgr-a transgenic mice (16).

Esr1 is a well-known positive regulator of Pgr expression in the mammary gland (30). Therefore, we assessed whether Epim regulates Esr1 in the mouse mammary gland during midpregnancy (day 12 of gestation). We did not detect significant differences in Esr1 mRNA by qPCR (Fig. 2D). However, Pgr mRNA (Fig. 2E) and protein levels (Fig. 2F and G) were significantly elevated in the TG mice compared with wild type. Using a polyclonal antibody that detects both isoforms, we showed that Epim predominately regulates Pgr-a (Fig. 3F and G; band at 80 kDa for Pgr-a is compared with one at 110 kDa corresponding to Pgr-b), indicating that Epim regulates primarily Pgr-a. Therefore, the effects of Epim on ductal elongation and alveolar differentiation are consistent with the distinct effects of Esr1 and Pgr on these processes and the differential expression of Esr1 and Pgr-a in the WAP-Epim transgenic mouse mammary glands.

In nulliparous WAP-Epim mice, we observed also a sharp increase in Pgr immunoreactivity compared with wild type by immunohistochemistry (Fig. 3A and C). In addition, Pgr expression was less heterogeneous compared with wild type, often showing clustered expression along mammary ducts (Fig. 3A; WAP-Epim, arrow). We observed also upregulation of 2 Pgr-responsive genes (19): CCAAT enhancer binding protein b (Cebpb; not shown) and Cyclin D1 (Ccnd1; Fig. 3B and D). Ccnd1 has been shown to be involved in the first wave of proliferation induced by progesterone in Pgr positive luminal epithelial cells (4). To test whether Pgr expression is required for stimulation of side branching in the WAP-Epim mice, 6-week-old mice were injected subcutaneously with the Pgr antagonist RU486 or oil control every other day for 8 days. By whole mount analysis, the RU486-treated mice showed approximately 70% fewer side branches than controls (Fig. 3E). We conclude that the effects of Pgr effects on side branching in the mammary gland may be related to Epim regulation of Pgr-a.

Analysis of signaling cascades in vivo is complicated by the rapid development of the mammary gland, and the constantly changing composition of hormones and growth factors. To study these interactions in a more controlled microenvironment and to define further cellular mechanisms underlying Epim-driven precocious alveologenesis in the WAP-Epim mouse, we used an ex vivo assay established in our laboratory (21). We used fragments of mammary epithelium (organoids) generated from 8-week-old mice in 3D Lr-ECM. In this assay, addition of a growth factor, such as TGF-α, initiates alveolar bud-like morphogenesis. When the epithelial fragment is embedded into Lr-ECM, it forms a simple bilayered cyst with a hollow lumen. A growth factor such as TGF-α or fibroblast growth factor-2 stimulates epithelial cell proliferation and lumen filling, expansion of the multi-cell-layered cysts, and eventually budding and ductal extension (31). Cell proliferation was shown to be required early in the process of cyst formation, approximately a day after growth factor treatment (21).

We confirmed that the Epim transgene was expressed in organoids embedded in Lr-ECM (Supplementary Fig. S2A). At the time of organoid isolation and embedding into Lr-ECM, there was a nonsignificant size difference between wild-type and WAP-Epim organoids (Supplementary Fig. S2E). However, 1 day after embedding into Lr-ECM, WAP-Epim organoids or wild-type organoids treated with rEpim formed larger cysts than wild-type or BSA controls, respectively (Supplementary Fig. S2D and S2F), indicating that Epim-TG mammary organoids grew at a faster rate than wild type. This observation was made before the organoids were treated with TGF-α, the growth factor that is added to initiate branching. Labeling organoids with 5-ethynyl-2′-deoxyuridine (EdU) to detect cells in S-phase showed that WAP-Epim organoids, on average, contained 2.8-fold greater number of EdU positive epithelial cells compared with wild type in Lr-ECM on the first day after embedding into Lr-ECM (Fig. 4A and B).

Pgr regulates alveologenesis in the mouse mammary gland (25) and we had found that Pgr activity was elevated in Epim-induced side branching (Fig. 2E and G). As Pgr also stimulates MEC proliferation (4, 32), we asked whether Epim stimulates MEC proliferation through Pgr. We observed an increase in total Pgr expression in the WAP-Epim organoids in Lr-ECM compared with wild type (Supplementary Fig. S2B) as observed also in vivo. We also observed a trend for upregulation of Pgr in wild-type organoids treated with rEpim (Supplementary Fig. S2C and S2G), suggesting that this regulation occurs without the requirement of the WAP promoter. To test whether Epim controls MEC proliferation through Pgr, we treated wild-type and WAP-Epim mammary glands in Lr-ECM with RU486 and found that this treatment abrogated the increased proliferation in WAP-Epim organoids (Fig. 4A and B). In addition, we found that increased bud formation in WAP-Epim organoids was blocked also by RU486 (Fig. 4C and D), implicating an Epim-Pgr signaling axis in both proliferation and bud formation.

Epim was shown previously to induce MEC branching in a 3D collagen gel culture assay through upregulation of matrix metalloproteinases (MMP; ref. 12), and to increase hepatocellular carcinoma invasion and metastasis through activation of MMP9 (33). Here, we found also that WAP-Epim mice upregulate MMP3 and MMP9 (Fig. 5A and B). We hypothesized that MMP activity is required for increased proliferation and bud formation in mammary glands of WAP-Epim mice. To test this, we used the broad spectrum chemical inhibitor of MMP activity, GM6001, and found this indeed to be true (Fig. 5C–E). We also tested whether Pgr activity regulates MMPs in this context by treating wild-type and WAP-Epim organoids with RU486 and analyzed MMP expression by qPCR. RU486 treatment downregulated MMP3 (Fig. 5F) but not MMP9 (not shown) in both wild-type and WAP-Epim organoids. MMP2 has been shown previously to signal downstream of Pgr in Pgr-a transgenic mice (17). Our results show that Epim-stimulated MMP3 expression is also downstream of Pgr in WAP-Epim mice.

We asked whether integrin-αV, known to bind to Epim to elicit downstream responses in 3D-collagen gels (13), was required to stimulate precocious bud formation in WAP-Epim mice. Treatment of organoids in the ex vivo 3D Lr-ECM assay with integrin function blocking antibodies against integrin-αV or IgG control did not show any morphological differences when induced with TGFα (Fig. 6A and C). In contrast, WAP-Epim mammary organoids treated with integrin-αV inhibitory antibodies showed a 50% decrease in the number of organoids with more than 3 alveolar buds compared with the anti-IgG-treated control antibody (Fig. 6A and C). Chemical inhibitors of the mitogen-activated protein kinase and the phosphoinositide-3-kinase pathways, downstream of integrin-αV, were even more potent inhibitors of alveolar development and proliferation in both wild-type and WAP-Epim transgenic mice (Fig. 6B and D). Finally, we showed that Epim controls Pgr expression through activation of integrin-αV after by downregulation of Pgr expression following treatment with the inhibitor of integrin-αV (Fig. 6E).

In this article, we have defined a role for Epim in postnatal formation of the mammary gland. We found that overexpression of exogenous Epim or introduction of a recombinant Epim peptide lead to precocious side branching and bud formation in nulliparous mice. We found also that a peptide inhibitor of Epim decreased side branching in wild-type mice. Because WAP-Epim mice eventually develop mammary carcinoma, these results suggest that Epim-induced precocious mammary gland development may be a factor predisposing these mice to mammary cancer later in life.

We believe this is the first demonstration that Epim is a regulator of Pgr-a expression. In the mouse mammary gland, regulation is through integrin-αV. The finding could have broad clinical implications for how Pgr is regulated in the mammary gland during normal development as well as in cancer progression by molecules in the cellular microenvironment. Increased Pgr-a expression has been shown previously to be involved in mammary hyperplasia (16, 17) and mammary tumors in BRCA1/P53 null transgenic mice (34). Our results add to this literature by showing that the high propensity of mammary cancer we have reported in this WAP-Epim mouse model (15) is essentially because of increased Pgr.

Pgr is required for side branching and alveologenesis in the mouse mammary gland (25, 29). We observed a striking increase in Pgr mRNA and protein expression in WAP-Epim mammary glands compared with wild type as well as upregulation of Pgr target genes such as Ccnd1, Cebpb, and keratin 14 (not shown; ref. 19). Treatment of WAP-Epim mice with the Pgr antagonist, RU486, or addition of RU486 to the culture medium of WAP-Epim organoids reduced side branching and alveolar bud formation significantly, showing that Pgr is downstream and required for the observed Epim-induced effects.

The WAP-Epim mice share some similarities to the WAP-Mmp3 mice, including precocious alveolar development (35) and spontaneous development of mammary cancer (15, 36, 37). The mammary glands of Mmp3 knockout mice have decreased lateral branching without any effect on ductal length (38), similar to the phenotype we observed using the Epim inhibitor in the mouse mammary gland. Treatment with rEpim was shown to upregulate Mmp3 in mammary organoids in 3D collagen gels (12). We found that RU486 blocked Mmp3 expression in both wild-type and WAP-Epim mice showing that Pgr regulates Mmp3 in the mouse mammary gland. These results implicate Mmp3 as an important effector downstream of Pgr-induced processes during mammary gland development, and potentially also during mammary cancer progression.

In Brca1/P53 null mice, Pgr was shown to be upregulated, because of decreased turnover of Pgr (34). In this mouse model, upregulation of Pgr caused mammary tumor formation, and inhibition of Pgr using RU486 decreased tumor formation (34). Pgr is a well-established transcriptional target of Esr1, we did not detect differences in Esr1 mRNA or protein expression between the mammary glands of wild-type and WAP-Epim pregnant mice. Therefore, Epim should be included as a novel regulator of Pgr-a mRNA and protein. In addition, our study shows that Epim-induced increase in Pgr-a expression also leads to tumor formation.

It is important to note that we observed Epim induction of Pgr without treating mice with either estrogen or progesterone or adding the hormones to the organoid culture medium. There is prior evidence for unliganded Pgr activity in development and cancer processes. One mechanism for Pgr unliganded activity is through activation of MAP kinase and/or PI3kinase pathways (reviewed in refs. 39, 40). Because Epim upregulates both of these pathways (13), it is likely that this could explain the mode by which Epim activates Pgr-a.

We reported previously a profound morphological difference between the mammary glands of the homozygous WAP-Epim mice and those of the wild type. In addition, we reported that the former have difficulty nursing their pups (14). However, we did not observe any morphological differences between wild-type and WAP-Epim mammary glands of the hemizygous mice during lactation by whole mount and histologic analysis (day 10). This included the number and size of alveolar structures (Supplementary Fig. S1C). In addition, there were no measurable differences in β-casein expression at lactation (not shown), and WAP-Epim mice were able to support full litters (10–12 pups, on average). The absence of an obvious morphological or functional difference between wild-type and hemizygous transgenic mice at lactation is likely due to less transgene expression compared with the homozygous WAP-Epim mice. Also, serum progesterone level is known to drop precipitously at the onset of lactation and remain low throughout lactation (reviewed in ref. 41). This may diminish any effects of Epim-induced Pgr-a effects in WAP-Epim hemizygous mice at this developmental stage (14). These results further emphasize the importance of the Epim-Pgr signaling axis.

In normal adult human breast cells, only 7% to 10% of luminal epithelial cells express PGR and ESR1. In contrast, 70% of breast tumors express PGR and ESR1 (39). Cells that express PGR in both normal human breast and mouse mammary gland are often growth arrested (34). However, during pregnancy the proportion of cells that express PGR are increased, often coexpress CCND1, and acquire the ability to proliferate (reviewed in ref. 39); whereas human breast cancers that lack ESR, PGR, and HER-2 are well known to have a poor prognosis (reviewed in ref. 42), breast cancers that lack ESR but express PGR have been reported to be more aggressive and have a poorer prognosis than ESR+/PGR+, ESR+/PGR−, and ESR−/PGR− breast cancers (43, 44). Our finding that Epim induces Pgr-a in the absence of Esr1 upregulation highlights a potential role for Epim in more aggressive tumors. These and other data in the literature highlight the potential danger of deregulation of Pgr and progesterone signaling in breast cancer and underscore the possibility that the culprit in many of the breast cancers may be progesterone rather than estrogen (45). It is therefore surprising that in clinical settings more attention is not paid to antiprogesterone receptor therapy because in animal models the data for tumor inhibition is indeed quite impressive (34).

In summary, we have shown that Epim controls mammary branching morphogenesis in vivo through transcriptional regulation of Pgr-a. Transgenic expression of Epim caused increased expression of Pgr-a, and inhibition of Epim in vivo blocked Pgr-a expression. Pgr-a activity was essential for mediating the effects of Epim on ductal side branching, epithelial cell proliferation, and bud formation. Pgr-a expression was shown to be downstream of the association of Epim with its extracellular receptor, integrin-αV, and upstream of MMP-3, a key mediator of Epim-induced branching morphogenesis. Our results also extend literature-linking Pgr-a expression and mammary hyperplasia in transgenic mouse models, by showing that the high propensity for developing mammary cancer we have reported in the WAP-Epim mouse model is essentially because of increased Pgr-a. Finally, our finding that Epim induces increased Pgr-a expression in the absence of Esr1 upregulation highlights a potential role for Epim in more aggressive, Esr1-negative breast cancers.

No potential conflicts of interest were disclosed.

Conception and design: J.L. Bascom, D.C. Radisky, M.J. Bissell

Development of methodology: J.L. Bascom, J.E. Fata, A. Lo, M.J. Bissell

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.L. Bascom, E. Koh, A. Lo, N. Roosta, M.J. Bissell

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.L. Bascom, D.C. Radisky, E. Koh, M.J. Bissell

Writing, review, and/or revision of the manuscript: J.L. Bascom, D.C. Radisky, A. Lo, H. Mori, M.J. Bissell

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.L. Bascom, E. Koh, N. Roosta, Y. Hirai

Study supervision: J.L. Bascom, H. Mori, Y. Hirai, M.J. Bissell

The authors thank Susan B. Komen for the Cure for postdoctoral fellowship awards to J.L. Bascom (KG080306) and to H. Mori (02-1591).

D.C. Radisky is supported by the NCI (CA122086), the Susan B. Komen foundation (KG110542), and the Mayo Clinic Breast Cancer SPORE (CA116201). Y. Hirai was supported by grant-in aid for Scientific Research (JSPS: KAKENHI 24590365). The work from M.J. Bissell's laboratory is supported by grants from the U.S. Department of Energy, Office of Biological and Environmental Research and Low Dose Scientific Focus Area (contract no. DE-AC02-05CH1123); by National Cancer Institute (awards R37CA064786, R01CA140663, U54CA112970, U01CA143233, and U54CA143836—Bay Area Physical Sciences-Oncology Center, University of California, Berkeley, CA); by U.S. Department of Defense (W81XWH0810736); and in part by a grant from The Breast Cancer Research Foundation.

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.