Developmental Biology 391 (2014) 219–229

Contents lists available at ScienceDirect

Developmental Biology

journal homepage: www.elsevier.com/locate/developmentalbiology

Oestrogen -alpha regulates non-canonical Hedgehog-signalling in the mammary gland

Nadia Okolowsky a, Priscilla A. Furth b, Paul A. Hamel a,n a Department of Laboratory Medicine & Pathobiology, Faculty of Medicine, University of Toronto, Toronto, ON, Canada M5S 1A8 b Lombardi Comprehensive Cancer Center, Departments of Oncology and Medicine, Georgetown University, Washington, DC, USA article info abstract

Article history: Mesenchymal dysplasia (mes) mice harbour a truncation in the C-terminal region of the Hh-ligand Received 5 February 2014 receptor, Patched-1 (mPtch1). While the mes variant of mPtch1 binds to Hh-ligands with an affinity Received in revised form similar to that of wild type mPtch1 and appears to normally regulate canonical Hh-signalling via 22 March 2014 , the mes mutation causes, among other non-lethal defects, a block to mammary ductal Accepted 10 April 2014 elongation at puberty. We demonstrated previously Hh-signalling induces the activation of Erk1/2 and Available online 21 April 2014 c-src independently of its control of smo activity. Furthermore, mammary epithelial cell-directed Keywords: expression of an activated allele of c-src rescued the block to ductal elongation in mes mice, albeit with Hedgehog pathway delayed kinetics. Given that this rescue was accompanied by an induction in -alpha Patched1 (ERα) expression and that complex regulatory interactions between ERα and c-src are required for Estrogen receptor normal mammary gland development, it was hypothesized that expression of ERα would also overcome Mesenchymal dysplasia Mammary gland the block to mammary ductal elongation at puberty in the mes mouse. We demonstrate here that conditional expression of ERα in luminal mammary epithelial cells on the mes background facilitates ductal morphogenesis with kinetics similar to that of the MMTV-c-srcAct mice. We demonstrate further that Erk1/2 is activated in primary mammary epithelial cells by Shh-ligand and that this activation is blocked by the inhibitor of c-src, PP2, is partially blocked by the ERα inhibitor, ICI 182780 but is not blocked by the smo-inhibitor, SANT-1. These data reveal an apparent Hh-signalling cascade operating through c-src and ERα that is required for mammary gland morphogenesis at puberty. & 2014 Elsevier Inc. All rights reserved.

Introduction ERα regulates target through two indepen- dent activation functions (AF), AF-1 and AF-2 (Arao et al., 2011; Estrogens are a class of hormones that play crucial roles in Arnal et al., 2013; Lannigan, 2003; McGlynn et al., 2013). Tran- diverse aspects of biology, including sexual development scriptional activation by ERα can be promoted through functional and reproduction. Dysregulation of estrogen signalling has been cooperation between the two AFs or through either AF indepen- implicated in a variety of diseases, including breast and uterine dently (Arnal et al., 2013). AF-2 contains the ligand binding cancers, osteoporosis and cardiovascular disease. Estrogen receptor- domain (LBD), and its activity depends on E2 binding (Arao alpha (ERα) is one of the principal receptors responsible et al., 2011; McGlynn et al., 2013). Ligand-independent activation for mediating the actions of estrogen (Heldring et al., 2007). As a is facilitated by the AF-1 domain. This domain harbours several member of the large superfamily of nuclear hormone receptors, ERα sequences that facilitate phosphorylation by factors mediating a acts as a ligand-activated (Heldring et al., 2007; number of pathways. For example, MAPK and Mangelsdorf et al., 1995). Upon binding of its ligand, 17β-estradiol Cdk7 activates ERα by directly phosphorylating Ser118 within the (E2), ERα modulates transcription by either binding directly to AF-1 domain (Bunone et al., 1996; Chen et al., 2002). estrogen response elements (ERE) in the promoter of E2 regulated More recently, it was determined that E2 elicits changes in , or indirectly through protein–proteininteractionswithother cellular processes through ERα outside of the nucleus via activa- transcription factors (Heldring et al., 2007; Klinge, 2000). tion of protein kinases, phosphatases and secondary messengers (Haynes et al., 2003; Kousteni et al., 2001). One of these rapid “ ” n non-genomic signal transduction mechanisms includes the c-src Corresponding author at: Laboratory Medicine & Pathobiology, Faculty of Medicine, protein tyrosine kinase (Castoria et al., 2012; Li et al., 2007). ERα University of Toronto, 1 King's College Circle, Toronto, ON, Canada M5S 1A8. Fax: þ1 416 978 5959. transiently activates c-src in response to E2 binding by increasing 416 E-mail address: [email protected] (P.A. Hamel). phosphorylation of Tyr , resulting in downstream activation of http://dx.doi.org/10.1016/j.ydbio.2014.04.007 0012-1606/& 2014 Elsevier Inc. All rights reserved. 220 N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229 ras and MAPK (Migliaccio et al., 1998, 1996). The SH2-domain of fibroblasts (ATCC) were cultured in DMEM supplemented with c-src also interacts directly with ERα at Tyr537 (Arnold et al., 1995; 400 mg/ml G418 (Gibco) and 0.14 μg/ml Zeocin (Invitrogen). Migliaccio et al., 2000; Sun et al., 2012). Phosphorylation of Tyr537 Mice. Wild type C57Bl/6N mice (Charles River) or C57Bl/6N by c-src rapidly activates signal transduction pathways involving mice heterozygous for the mesenchymal dysplasia (mes) allele of Erk and phosphoinositide 3-kinase (PI3K) (Varricchio et al., 2007), Ptch1 (JacksonLabs) were crossed with CERM mice, which harbour as well as stimulating ERα target gene transcription, ERα ubiqui- a flag-tagged ERα transgene under the control of the tetO and the tylation and proteosomal degradation (Sun et al., 2012). “Tet-On” transgene under the control of the MMTV promoter Genetic deletion of ERα in mice established important roles for (Díaz-Cruz et al., 2011; Frech et al., 2008, 2005; Hruska et al., 2002; this receptor for normal adult mammary gland development. Miermont et al., 2010) (P. A. Furth, Georgetown University). Mammary glands of adult female ERα knock-out (αERKO) mice Compound mice were then backcrossed onto C57Bl/6 (CharlesRi- fail to respond to ovarian hormones resulting in a failure of ductal ver) for 44 generations. Transgene expression was induced by rudiments to elongate and fill the fat pad (Couse et al., 2000; constant administration (changed twice per week) of 2 mg/ml Lubahn et al., 1993; Pedram et al., 2009). Interestingly, animals doxycycline in sterile-filtered drinking water containing lacking detectable c-src also exhibit defects in mammary gland 5% sucrose. morphogenesis and primary mammary epithelial cells (MECs) For genotyping, tail DNA was extracted and 2 μlDNAwas derived from c-src-null mice fail to respond to exogenous estrogen amplified in a 25 μl polymerase chain reaction using Taq DNA stimulation (Kim et al., 2005). Thus, proper mammary gland Polymerase (Thermo Scientific). Forward (F) and reverse (R) primer morphogenesis not only requires E2-activated ERα, but also sequences were as follows: depends on ERα interaction with other signalling cascades. Mes:F50-TCCAAGTGTCGTCCGGTTTG-30 and R 50-GTGGCTTCCA- We and others demonstrated previously that Hedgehog (Hh)- CAATCACTTG-30 (Chang et al., 2012); FLAG-ERα:F50-CGAGCTCGG- signalling also plays a role in mammary gland development (Chang TACCCGGGTCG-30 and R 50-GAACACAGTGGGCTTGCTGTTG-30 et al., 2012; Moraes et al., 2009). Animals homozygous for the (Miermont et al., 2010); MMTV-rtTa: F 50-ATCCGCACCCTTGATGA mesenchymal dysplasia (mes)(Makino et al., 2001; Sweet et al., CTCCG-30 and R 50-GGCTATCAACCAACACACTGCCAC-30 (Miermont 1996) allele of the Hh-ligand receptor, Patched-1 (Ptch1), displayed et al., 2010). Reaction conditions for MMTV-rtTa and mes were 60 s a blocked mammary gland phenotype resembling that of the ERα each for denaturation, annealing and extension for 32 cycles, while knock-out mice. The lack of ductal outgrowth in mes mice was for FLAG-ERα they were 60 s denaturation, 90 s annealing, and associated with reduced expression of ERα and progesterone 120 s extension for 35 cycles. The annealing temperatures for receptor (PR) in epithelial cells (Chang et al., 2012; Moraes et al., mes, FLAG-ERα and MMTV-rtTa were 53 1C, 57 1C and 56 1C, 2009). The mes allele encodes a deletion in the second-last exon of respectively. mPtch1, resulting in a truncated protein that replaces the last 220 a. Whole mount analysis. Whole mount analysis of mammary a. with a random 68 a.a. polypeptide. We showed that this region of glands was performed as previously described (Chang et al., Ptch1 binds to factors containing SH3- and WW-domains, that the 2012; Rasmussen et al., 2000). Briefly, mammary glands were SH3-domain of c-src binds to the C-terminus of mPtch1 (Chang fixed in Carnoy's fixative (10% acetic acid, 30% chloroform, 60% et al., 2010) and that transiently expressed mPtch1 binds to endo- ethanol) overnight at 4 1C, and washed in 70%, 50% and 25% genous c-src in the absence of added Shh-ligand (Harvey et al., 2014). ethanol for 15 min each. The glands were rinsed in water for Using a genetic approach, we showed further that forced expression 5 min and then stained in Carmine alum overnight at room of an activated c-src (c-srcAct) transgene in luminal mammary temperature. The glands were washed in 70%, 95% and 100% epithelial cells rescued the blocked mammary morphogenesis arising ethanol for 15 min each, and then cleared in xylene for two in mes mice (Chang et al., 2012). This rescue was accompanied by a changes for 30 min each followed by mounting with Permount. strong increase in ERα expression. Immunofluorescence. Paraffin-embedded tissue sections were Interactions between the ERα activity and Hh-signalling has also deparaffinized in xylene and rehydrated gradually through 100%, been demonstrated. In ERα-positive gastric cancer cells, E2 induced 95%, 80%, 70% ethanol washed and water. Antigen retrieval was Shh expression and promoted cellular proliferation independent of performed by boiling sections for 15 min in 10 mM sodium citrate, smo-activity (Kameda et al., 2010). A similar result was reported in the pH 6.0, in a pressure cooker pre-heated in a microwave oven for ERα-positive breast cancer cell line, MCF-7. Here, E2 induced expres- 20 min before immersing slides. After boiling, sections were sion of Shh and Gli1. This activation is inhibited in cells treated with permeabilized with 0.2% TritonX-100 for 10 min. After washing, the anti-estrogen, ICI 182780 (Koga et al., 2008). Inhibiting smo- sections were incubated with primary antibodies for 1 h at room activity with the small molecule inhibitor, cyclopamine, significantly temperature in 3% BSA in PBS. Antibodies and dilutions are as suppressed proliferation of both ERα-positive and ERα-negative breast follows: 1:100 anti-ERα (Santa Cruz; sc-542), 1:100 anti-BrdU cancer cell lines, MCF-7 and MDA-MB-231, respectively (Che et al., (abcam: ab6326), 1:20 anti-Flag (abm;G191) and undiluted anti- 2013). Cyclopamine also significantly decreased ERα expression in Alx4 containing supernatant (produced in our lab). Sections were MCF-7 cells. Since c-src binds to mPtch1 (Chang et al., 2010; Harvey then washed and labelled with fluorescently-labelled secondary et al., 2014) and also activates ERα (Castoria et al., 2012; Sun et al., antibody added to samples for 1 h. Sections were imaged using a 2012), we sought to determine whether ERα-activity acts genetically Nikon Ellipse fluorescent microscope equipped with a QImaging downstream of Ptch1. We demonstrate here that conditional expres- Fast1394 digital camera and compiled using Qcapture Pro software sion of ERα overcomes the block to ductal elongation in mes mice with (QImaging). kinetics similar to those for the rescue of the mes phenotype by the Preparation of Shh-conditioned media. Shh-conditioned media MMTV-c-srcAct allele. Furthermore, we define a novel pathway was prepared as previously described (Capurro et al., 2012). Shh- stimulated by Shh that activates Erk1/2 and requires the activity of and pcDNA-conditioned media were prepared by transfecting 40% either ERα or c-src but not smo. confluent 10 cm plates of HEK293 cells in 10% FBS with 10 μgof pcDNA3.1-N-Shh (gift of J. Filmus, Sunnybrook Health Sciences Centre) or pcDNA3 using 2 mg/ml PEI at a 2:1 ratio. Cells were Materials and methods grown for 24 h before the media was switched to 5% FBS for 48 h. The media was collected, centrifuged at 2500 rpm for 5 min at 4 1C Cell culture. HEK 293 cells (a gift of S. Girardin) were cultured in and harvested. The supernatant was then sterile-filtered using a DMEM with 10% FBS and 1% penicillin-streptomycin. Shh Light II 0.22 μm syringe filter. Prior to use, conditioned media was diluted N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229 221 in serum free DMEM to a final serum concentration of 0.5%. bromophenol blue, 10% glycerol) was added to 50 mg of lysate and Activity of conditioned media was measured in a luciferase boiled for 5 min. Samples were resolved by 10% SDS-PAGE and assay in Shh Light II fibroblasts (Chang et al., 2010; Sasaki et al., blotted onto nitrocellulose membrane. Blots were probed with 1997). primary antibody overnight at 4 1C. Antibodies and dilutions used Primary cell culture. Primary mammary epithelial and mesench- are as follows: 1:1000 mouse α-p-ERK (Cell Signaling) and 1:1000 ymal cells were isolated as described previously (Chang et al., rabbit α-ERK (Cell Signaling). Blots to be re-probed were stripped 2012; Niranjan et al., 1995). Briefly, thoracic and inguinal mam- in stripping buffer (2% SDS, 62.5 mM Tris pH 6.8, 100 mM β- mary glands were dissected from euthanized mice and minced mercaptoethanol) for 30 min at 72 1C. with a razor blade into small fragments. The minced mammary Quantitative RT-PCR. For quantitative RT-PCR (qPCR), total RNA glands were incubated in a solution of 3 mg/ml collagenase A and was isolated from primary mouse mammary mesenchymal and trypsin (Sigma-Aldrich) in DMEM/F12 for 2 h at 37 1C. Mechanical epithelial cells using Trizol (Invitrogen). DNAse treatment (Fer- dissociation was then performed by slowly pipetting with a 5 ml mentas) was performed on 400 ng of RNA from each sample. pipette, adding FBS to a final concentration of 2%. Epithelial Reverse transcription was performed using SuperScript II Reverse (pelleted) and mesenchymal (supernatant) layers were separated Transcriptase (Invitrogen) and random hexamer primers (Fermen- by centrifuging at 200 rpm for 2 min. The mesenchymal layer was tas). Resulting cDNA was diluted 1:25 and qPCR was then per- collected and washed with DMEM/F12 and plated in DMEM/F12 formed using an iQ SYBR Green Supermix (Bio-Rad) in 10 μl containing 10% FBS supplemented with 10 μg/ml insulin, 10 μg/ml reactions. All reactions consisted of 40 cycles with 30 s denatura- transferrin and 20 ng/ml EGF. For organoid culture, the epithelial tion at 72 1C, 30 s annealing and 30 s extension at 60 1C. Primer pellet was resuspended in 3 mg/ml collagenase A in DMEM/F12 specificity was confirmed by a melt curve analysis. Results were and incubated for 30 min at 37 1C. The cells were pelleted by quantified using the ΔΔCt method with Arbp as a reference gene. centrifugation, washed and plated in primary culture media Primers adapted from previously published sequences (Webster mentioned above. The organoid containing media was removed et al., 1995; Zhang et al., 2009) were used as follows: and re-plated once the remaining mesenchymal cells attached to mPtch1 forward: 50-GGTGGTTCATCAAAGTGTCG-30; the plate (2 h). mPtch1 reverse: 50-GGCATAGGCAAGCATCAGTA-30; For Erk1/2 activation assays, primary mouse mammary epithe- mGli1 forward: 50-CCCATAGGGTCTCGGGGTCTCAAAC-30; lial cells and mesenchymal cells were serum starved in 0.2% FBS mGli1 reverse: 50-GGAGGACCTGCGGCTGACTGTGTAA-30; containing culturing media for 24 and 48 h, respectively. Cells mArbp forward: 50-GAAAATCTCCAGAGGCACCATTG-30; were then stimulated with pcDNA3-conditioned media, Shh- mArbp reverse: 50-TCCCACCTTGTCTCCAGTCTTTAT-30. conditioned media, or Shh-conditioned media with either 10 nM ICI 182 780, 20 nM SANT-1 (Toronto Research Chemicals), or 10 nM PP2 for 24 h. Cells were lysed in 1% NP40 lysis buffer and levels of Results phospho-Erk1/2 and total Erk1/2 determined by western blot. Western blotting. Cell lysates were prepared following addition Mice homozygous for the mes allele exhibit a block to ductal of 1% NP40 lysis buffer containing protease and phosphatase elongation in the mammary gland during puberty. This defect is inhibitors (0.57 mM PMSF, 10 μM leupeptin, 0.3 μM aprotinin, overcome, albeit with distinct kinetics relative to wild type 10 mM NaF, 1 mM sodium orthovanadate). For western blots, 4 animals, by constitutive expression of activated c-src under the SDS-loading buffer (50 mM Tris pH 6.8, 100 mM DTT, 2% SDS, 0.1% control of the MMTV-promoter. Given the increased expression of

Fig. 1. Forced expression of ERα rescues the mammary gland phenotype in mes mice. Mammary gland #4 was isolated from twenty week-old mice and whole mounts prepared. (A) and (A0) mes; (B) and (B0) mes/CERM. After twenty weeks, mammary ductal morphogenesis in mes mice is blocked. This phenotype is reversed in mes mice expressing ERα directed to mammary epithelium, albeit with delayed kinetics. Magnification of ducts reveal a non-proliferative rudimental ductal structure in mes mammary glands (A0). Terminal end buds are present and ductal structures have not reached the limits of the fat pads in mes/CERM mice (B0). 222 N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229

ERα in mes/MMTV-c-srcAct mice and the role of c-src in activating present. As we reported previously (Chang et al., 2012), develop- ERα, we hypothesized that forced expression of ERα would also ment of the mammary glands in mes littermates beyond the rescue the mes phenotype. CERM mice (Frech et al., 2005; Tilli primitive ductal structures established during embryogenesis et al., 2003), a compound transgenic mouse that harbours a tetO- was not observed, even at thirty weeks (Fig. 2E) nor were ducts ERαFlag allele under the control of MMTV-”Tet-On”, were bred to evident in mes/CERM animals not induced with doxycycline mes mice and backcrossed onto a C57Bl/6 background. ERα (Fig. 2F). expression in compound mes/CERM mice was then induced by To determine if the mammary ducts of mes/CERM mice long-term administration of the doxycycline beginning at 3 weeks exhibited normal epithelial organization, sections were probed of age. As Fig. 1A and A0 illustrate, at 20 weeks of age, mammary for ERα and the myoepithelial marker, p63 (Fig. 3). Nuclear ERα gland #4 in the adult mes females had not developed, remaining in expression was observed in mammary epithelial cells from ducts the state that developed during embryogenesis. However, as of adult virgin wild type, mes, wild type/CERM and mes/CERM Fig. 1B shows, ductal elongation was apparent in compound mice mice. Expression of ERα was restricted to the luminal epithelial homozygous for the mes allele and expressing MMTV-directed cells, evident by the lack of overlap between signals for ERα with ERα (mes/CERM). At this stage, terminal end buds were also p63. As we demonstrated previously (Chang et al., 2010; Moraes et apparent (Fig. 1B0) indicative of these ducts elongating into the al., 2009), reduced ERα expression was observed in mes mice fat pad similar to that observed for wild type mice at puberty. (Fig. 3E and H) compared to wild type littermates (Fig. 3A–D). Similar to the observation we reported previously for mes/ While approximately 40% of epithelial cells were positive for ERα MMTV-c-srcAct mice (Chang et al., 2012), the rescue by CERM mice staining in wild type animals, ERα staining was only observed in of the mes phenotype exhibited distinct kinetics of ductal elonga- 10–20% of epithelial cells in mes littermates. An approximate 2- tion relative to the development of wild type mammary glands fold increase in the percentage of cells demonsrating nuclear (Fig. 2). By ten weeks of age, only rudimentary ductal structures localized ERα expression was observed in mes mice expressing resembling those of their mes littermates were seen for MMTV-ERα, relative to mes alone. However, expression levels and doxycycline-fed mes/CERM mice (Fig. 2A). However, between 15 subcellular localization of ERα were indistinguishable between and 20 weeks, the mammary ducts in doxycycline-fed mes/CERM wild type (Fig. 3A–D), mes/CERM (Fig. 3M–P) and WT/CERM mice had penetrated the fat pad to various degrees (Fig. 2B). (Fig. 3I and L) littermates. Significantly, terminal end buds (TEBs) were present in these Immunofluorescence revealed that induction of the conditional animals (see Fig. 1B0). Complete morphogenesis was observed in ERα expression in both wild type and mes mice caused aberrant doxycyline-fed mes/CERM females between 24 to 30 weeks of age cell cycle progression in adult mammary glands (Fig. 4). Specifi- (Fig. 2C). For these mature animals, the ducts in these mammary cally, cells incorporating BrdU were essentially absent in the glands reached the limits of the fat pad and TEBs were no longer mammary glands of wild type adult animals (Fig. 4A and D).

Fig. 2. Delayed rescue of mammary morphogenesis in mes/CERM mice is complete by thirty weeks. Whole mounts of mammary gland #4 were prepared from glands isolated from age-matched littermates at different points during development. (A) Mammary glands from mes/CERM exhibit no ductal proliferation at ten weeks, while wild type littermates (D) are completely developed. (B) Terminal end buds with partial filling of the mammary gland fat pad by ductal elongation are found in mes/CERM females at twenty weeks of age. (C) Ducts have reached the limits of the fat pad by thirty weeks of age and terminal end buds are no longer present. (F) Mammary glands from uninduced adult mes/CERM mice fail to develop and resemble mammary glands from mes mice (E) that exhibit only a rudimentary ductal tree, even at thirty weeks. Rescue of the mes mammary defect by ERα in animals older than eighteen weeks is significant (p¼0.0278; Fisher's exact test). Ductal structures are magnified in corresponding inserts. Scale bar: 50 μm. N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229 223

Fig. 3. mes/CERM ducts demonstrate normal cellular organization and ERα expression. Immunofluorence staining for ERα and p63 in adult virgin wild type ((A)–(D)), mes/ mes ((E)–(H)), wild type/CERM ((I)–(L)) and mes/CERM ((M)–(P)) mice. No co-localization of ERα is observed with p63 ((D), (H), (L) and (P)). Reduced ERα expression is apparent in mes mice (F) compared to wild type (p¼0.02) (B). ERα expression is increased, however, in mes mice expressing MMTV-ERα (p¼0.021) (N), but expression levels do not differ between mes/CERM and wild type littermates. Scale bar: 25 mm. ERα-positive cell percentages were determined by cell counting; 10 fields per animal, n¼3 for all animal groups. Data was analyzed by one-way ANOVA followed by pairwise comparisons of means using Dunnett's multiple comparison's test.

However, following the 2 h pulse of BrdU, a fraction of cells in the MMTV promoter. As Fig. 5A and B illustrates, Alx4 expression every duct in doxycycline-fed WT/CERM (Fig. 3E and H) and mes/ was generally absent in the adult wild type stromal cells adjacent CERM (Fig. 3M–P) mice stained positively, indicating the presence the epithelial ducts. However, in wild type mice with induced ERα of cycling cells due to the induced expression of ERα. BrdU (Fig. 5E and F) or constitutive expression of c-srcAct (Fig. 5C and D) incorporation was not observed in WT/CERM mice not fed dox- in mammary epithelial cells, an induction of Alx4 expression in ycycline (Fig. 4I and L). Cells expressing detectable levels of the stromal fibroblasts in these adult animals was apparent. In con- flag-tagged ERα transgene were not associated with the cells that trast to CERM or MMTV-c-srcAct on the wild type background, had incorporated BrdU (Supplementary Fig. 1), although all BrdU- however, Alx4 expression was not induced by these two alleles on incorporating cells stained with an anti-ERα antibody (Fig. 4C and the mes background (Fig. 5I, J and K,L, respectively). Thus, the D), suggesting that transgene expression generated an overall luminal epithelial-restricted expression of either ERα or c-srcAct proliferative state in the mammary gland rather than driving appeared to generate a signal, emanating from mammary epithe- specific cells through the cell cycle. lial cells, that stimulated stromal fibroblasts to express Alx4. The To further investigate the overall proliferative state of the induction of Alx4 expression in these fibroblasts was dependent, mammary gland, changes in the expression of the stromally- however, on the presence of wild type mPtch1. restricted, homeodomain protein, Alx4 (Hudson et al., 1998; Qu E2 rapidly activates the MAPK signalling pathway in the human et al., 1998, 1997a, 1997b), were assessed. Alx4 is induced in breast cancer cell lines, MCF-7 and T47D (Castoria et al., 1999; stromal fibroblasts during puberty and pregnancy but is not Improta-Brears et al., 1999; Migliaccio et al., 1996; Song et al., 2002), expressed in the resting adult mammary gland (Hudson et al., in a c-src-dependent manner (Di Domenico et al., 1996; Song et al., 1998; Joshi et al., 2006). Furthermore, administration of E2 to pre- 2002). Shh-ligand also stimulates the activation of both Erk1/2 pubescent female mice induces Alx4 expression in these fibro- (Chang et al., 2010) and c-src (Harvey et al., 2014) in a human blasts while the absence of Alx4 in stromal fibroblasts in Strong's mammary epithelial cell line, MCF10a, that lacks detectable expres- luxoid mice causes defects in mammary ductal morphogenesis sion of smo (Chang et al., 2012, 2010; Mukherjee et al., 2006; Zhang during puberty. Thus, we examined the expression of Alx4 et al., 2009). As Fig. 6A illustrates, phosphorylation of Erk1/2 was expression in stromal cells in wild type and mes adults that induced in wild type, serum-starved primary mammary epithelial conditionally expressed ERα (CERM mice) and compared these cells within 15 min of stimulation by N-Shh. Stimulation occurred in to mammary glands expressing activated c-src under the control of the presence of the smo-antagonist, SANT-1, which we showed 224 N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229

Fig. 4. Forced ERα expression induces cell cycle progression in adult mammary glands. Twenty four week-old wild type ((A)–(D)), induced wild type/CERM ((E)–(H)), uninduced wild type/CERM ((I)–(L)) and mes/CERM ((M)–(P)) littermates were injected with BrdU 2 h prior to mammary gland excision and incorporation was determined through immunofluorescence. BrdU incorporation was absent in the mammary glands of 24 week-old adult wild type mice (B). Cells positive for BrdU staining are observed in ducts from both induced wild type/CERM (F) and mes/CERM mice (N), but not in wild type/CERM that were not fed doxycycline (J). BrdU positive cells co-localize with cells expressing ERα ((H) and (P)). Scale bar: 25 mm.

Fig. 5. Conditional expression of ERα induces stromal expression of Alx4 in adult female mice. Mammary glands of wild type ((A)–(B)), wild type/c-srcAct ((C)–(D)), wild type/ CERM ((E)–(F)), mes ((G)–(H)), mes/c-srcAct ((I)–(J)), mes/CERM (induced) ((K)–(L)) virgin 24 week-old adult female mice were stained for Alx4. Staining is absent in wild type mammary ducts. Expression of constitutive c-srcAct or conditional ERα in ductal epithelial cells induces stromal expression of Alx4. Scale bar: 25 mm. N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229 225

5 + Shh Media 4

Control 15min 30min 3

phospho-Erk 2

1

total-Erk 0

15min 30min Control

15min 5

4

Control pcDNA3 Shh ICI + Shh 3

phospho-Erk 2

Fold DifferenceFold 1 Difference Fold total-Erk 0

Shh Control pcDNA3 ICI + Shh 15min

* 5 T-1 + Shh pcDNA3 Shh SAN PP2 + Shh 4 Control phospho-ERK 3

2 Fold Difference Fold total-ERK 1

0 Control pcDNA3 Shh Sant-1 PP2

Fig. 6. Shh activates Erk1/2 in primary mammary epithelial cells is an ERα- and c-Src- dependent manner. Primary mammary epithelial cells isolated from wild type mice at three months of age were serum starved for 24 h and then treated with Shh-conditioned media, pcDNA3-control media, or Shh-conditioned media with ICI 182 780, SANT-1 or PP2 for 1 h. Blots were probed for activated (phospho)-Erk1/2 and then re-probed for total Erk1/2. (A) Activation of Erk1/2 is observed 15 min following stimulation by . (C) Addition of ICI 182 780 1 h prior to Shh stimulation partially inhibits Erk1/2 activation. (E) Cells were treated with 20 nM SANT-1 or5μM PP2 for 1 h before Shh stimulation. PP2, but not SANT-1 treatment inhibits Shh-dependent Erk1/2 activation. Quantification of Erk1/2 activation are shown ((B), (D) and (F)). Data were analyzed by two-way ANOVA followed by pairwise comparison of means using Dunnett's multiple comparisons test. Data are displayed as mean7SEM, n¼3 for all groups, except PP2þShh, n¼2. *po0.05.

previously inhibited canonical Hh-signalling in primary mammary not been reported. Thus, primary epithelial and mesenchymal cells mesenchymal cells (Harvey et al., 2014). Activation of Erk1/2 from wild type mammary glands were stimulated with N-Shh occurring through a smo-independent mechanism (Fig. 6E) is also peptide and changes in expression of both Hh-pathway targets, consistent with our previous observations using the smo-deficient Gli1 and Ptch1 were determined. As Fig. 7A reveals, stimulation of MCF10a breast cancer line (Chang et al., 2010)aswellasinsmo- primary mammary mesenchymal cells resulted in a greater than deficient MEFs (unpublished observation). Thus, the dependence of 25-fold increase in Gli1 expression. Likewise, Fig. 7B illustrates the c-src activity on N-Shh stimulation of Erk1/2 was determined in 4-fold increase in Ptch1 expression in these same cells. We primary mammary epithelial cells treated with the c-src inhibitor demonstrated previously that Gli1 and Ptch1 induction in primary PP2 (Hanke et al., 1996). As Fig. 6E shows further, a complete block mouse mesenchymal cells through N-Shh stimulation was depen- of Erk1/2 activation occurs when c-src activity is inhibited. dent on the activity of smo, since induction is abolished in cells The role of ERα for N-Shh activation of Erk1/2 was then treated with N-Shh in the presence of SANT-1 or by using heat- determined by stimulating cells in the presence of the anti- killed N-Shh peptide (Harvey et al., 2014). However, despite the estrogen, ICI 182780 (Wakeling et al., 1991). As illustrated in apparent response of these mesenchymal cells to N-Shh ligand, a Fig. 6C, ICI 182780 partially inhibited Erk1/2 activation by N-Shh very limited or no induction of Gli1 and Ptch1, respectively, were ligand. Specifically, treatment of primary epithelial cells with N- observed for primary epithelial cells. Specifically, a small but Shh-conditioned media caused a 4-fold increase in Erk1/2 phos- significant 3-fold increase in Gli1 levels was observed after 24 h phorylation. This level of activation was reduced by 50% in the while no induction of Ptch1 was observed. Thus, mammary presence of ICI 182780. Thus, N-Shh activated Erk1/2 in the epithelial cells are relatively refractory to stimulation through presence of the smo-inhibitor, SANT-1, but required the activities the canonical Hh-pathway but are able to respond through non- ERα and c-src. canonical pathway, determined by the activation of Erk1/2 by While N-Shh stimulates non-canonical Hh-signalling cascades N-Shh. in both primary mammary epithelial and mesenchymal cells, smo- Taken together, using both a genetic approach and experiments dependent canonical Hh-signalling in primary epithelial cells has using primary mammary cells in culture, these data suggest that a 226 N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229

Further evidence of Hh-pathways operating independently of

smo was revealed in the mesenchymal dysplasia (mes) mouse

40 * Mesenchymal cells (Sweet et al., 1996). The 32-bp deletion in this animal in the 35 Epithelial cells sequence encoding the mPtch1 C-terminus causes a truncation in 30 this region of mPtch1 and the addition of a small amount of an 25 unrelated, random sequence (Makino et al., 2001). A number of

Expression 20 non-lethal defects arise in this mouse, including a block to

15 mammary gland morphogenesis during puberty. As we have Gli1 shown, this defect is not due to the lack of ovarian hormones 10 * but rather to the apparent unresponsiveness in cells of the 5 mammary gland to these factors (Chang et al., 2012; Moraes

Relative 0 et al., 2009). Despite these defects, the mes variant of the mPtch1

Shh Shh protein binds Shh-ligand with an affinity similar to the wild type Control Control mPtch1 (Nieuwenhuis et al., 2007) and appears to facilitate Heat Killed Sant-1 + Shh signalling through the canonical Hh-signalling pathway similar to cells with the full length protein (Harvey et al., 2014). Thus, in

4 Mesenchymal cells the context of the canonical Hh-pathway, the basis for blocked Epithelial cells morphogenesis in the mes mammary gland is not apparent. 3 We hypothesized and showed previously, however, that the defect in the mes mammary gland could be rescued, albeit with

Expression delayed kinetics, by expressing an activated version of c-src 2 (Chang et al., 2012). Among changes in the mammary glands of Gli1 the mes/MMTV-c-srcAct mice was a significant increase in the 1 levels of expression of ERα relative to mes mice alone and its accumulation in the cytoplasm of mammary epithelial cells. These Relative 0 data suggested that Hh-signalling in the mammary gland may require, at least in part, signalling through pathways that con- Shh Shh Control Control trolled the expression and activities of ERα and that forced Heat Killed α Sant-1 + Shh expression of ER in mes mice might also rescue the mes mammary phenotype. As we demonstrated here, conditional Fig. 7. Limited canonical Hh-signalling in primary mammary epithelial cells. expression of an ERα transgene under the control of the MMTV Primary mammary epithelial and mesenchymal cells isolated from three month- old wild type mice were serum starved for 24 h and 48 h, respectively. Cells were promoter also overcame the block to mammary morphogenesis Act then stimulated with heat-killed N-Shh peptide, active N-Shh peptide, or N-Shh with kinetics similar to that of the MMTV-c-src allele. Thus, peptide with 20 nM SANT-1 for 24 h. (A) qPCR reveals that Gli1 is induced 25-fold these two distinct mouse models have revealed a novel Hh- fi in mesenchymal by N-Shh. Only a small but signi cant 3-fold increase in Gli1 was signalling pathway, involving c-src and ERα, that plays an impor- seen in epithelial cells from the same animals, however. (B) N-Shh peptide induced Ptch1 expression 3.5-fold in mesenchymal cells. No induction of mPtch1 was tant role in postnatal mammary gland morphogenesis. observed for epithelial cells. Data were annalyzed by one-way ANOVA followed by We note that the phenotype of the mes mice we reported in pairwise comparisons of means by Dunnett's multiple comparisons test. Data are this and our previous paper (Chang et al., 2012) differs from a displayed as mean7SE, *po0.05, n¼3. previous report (Moraes et al., 2009). In the latter paper, the blocked development of the mammary gland at puberty was not as robust as we have observed. The reason for the differences between these phenotypes is not apparent, particularly given that novel Hh-signalling pathway involving the activities of ERα and both labs backcrossed their animals onto C57Bl/6 backgrounds, c-src is required for regulating mammary gland development. albeit, using C57Bl/6 mice from colonies that first diverged in 1951 (Charles River (C57Bl/6N) versus Jackson Labs (C57Bl/6J)). We have characterized almost 40 mes mice on the C57Bl/6N background Discussion and have found only 3 with partial ductal development of mammary gland #4. Furthermore, for both the mes/MMTV-c-srcAct The Hedgehog-signalling pathway regulates morphogenesis in and mes/CERM animals, ductal growth was not apparent until after a large number of tissues. Detailed genetic and molecular studies about 15 weeks, again supporting the stronger penetrance of mes have shown further that control of these processes by the Hedge- mammary phenotype in our colony. Similar to Moraes et al., hog ligands is typically mediated by control of the activities of the however, we observed that mixing FVB onto the mes/C57Bl/6N Patched-1 receptor that, in turn, indirectly controls the activity of background had no effect on the mes phenotype (unpublished smo. Interestingly, the primary sequence of mPtch1 predicts that, observation). while potentially habouring a sterol-sensing domain (Carstea Our data also showed that stimulation by N-Shh of primary et al., 1997; Loftus et al., 1997), signalling cascades involving the mammary epithelial cells activated Erk1/2 in both an ERα and c-src- direct association of SH3 or WW-domains with Ptch1 may also be dependent manner. Inhibition of activation of either factor, but not involved in Hh-signalling. Our previous studies showed that a inhibition of smo, reduced or blocked the ability of N-Shh to number of factors harbouring these domains can associate with stimulate Erk1/2. Thus, the non-canonical Hh-signalling pathway the cytoplasmic C-terminal region of mPtch1 (Chang et al., 2010). we identified previously appears to involve changes in the activities These include the SH3-domains of signalling factors such as c-src, of ERα and c-src in mammary epithelial cells. The precise mechan- Grb2 and PIK3R2 (p85α) as well as factors harbouring WW- ism remains undefined. It is clear, however, that the activities of domains such as the E3 ubiquitin ligases, WWP2 and Smurf2. both c-src and ERα are required for the N-Shh-dependent activation Supporting this was the recent observation that Drosophila Smurf of Erk1/2. In the presence of inhibitors that blocked the activities of (dSmurf) complexes Drosophila Patched (dPtch) (Huang et al., ERα or c-src, respectively, N-Shh-dependent Erk1/2 activation was 2013). attenuated or blocked completely while treatment with the small N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229 227 molecule inhibitor of smo, SANT-1, had no effect. Thus, these data (for reviews, see Che et al., 2013; Cui et al., 2010; Hui et al., 2013; support the existence of a pathway stimulated by the Hh-ligands in O’Toole et al., 2009). However, these studies have typically not mammary epithelial cells that operates through both c-src and ERα described the activation of the canonical Hh-pathway upon resulting in the activation of Erk1/2. stimulation by Hh-ligand. Indeed, evidence suggests that the “Hh In both MMTV-c-srcAct and in CERM mice on the mes back- repression state”, that is the state in which Gli3 acts as a ground, a significant increase in the levels of ERα relative to mes (Gli3rep), is required for normal mammary gland development mice alone are observed in the mammary epithelial cells. The (Hatsell and Cowin, 2006; Hui et al., 2013). It might be expected, altered expression of ERα suggested that the level of its expression therefore, that changes to the expression of downstream factors required for ductal morphogenesis at puberty in response to 17β- that overcome the of the Gli3rep might lead to altered estradiol requires the normal activity of the mPtch1 C-terminus. cell characteristics and might contribute to cell transformation as This model does not preclude the possibility that the canonical occurs, for example, in the MMTV-Gli1 mice (Fiaschi et al., 2009). Hh-pathway, signalling through smo, is also required. This possi- However, our data suggest further that, while the mammary bility has not been strictly tested in the mouse mammary gland (i.e. epithelial cells appear refractory to Hh-signalling through the conditional or tissue-specific knock out of smo in mammary canonical pathway, non-canonical signalling cascades can be epithelial and/or mesenchymal cells). However, mice harbouring invoked. conditional c-srcAct or ERα alleles are sufficient to overcome the To conclude, combined with our data characterizing the effect block to the pubescent mammary morphogenesis and is associated of the MMTV-c-srcAct allele on development in the mes mammary with increased levels of ERα relative to mes mice alone. gland, our data suggest that a novel Hh-signalling pathway Our data also revealed the mPtch1-dependent activation of the involving the activities of ERα and c-src are required for mammary mammary stromal cells in vivo. Specifically, Alx4-expression was gland morphogenesis at puberty. These data suggest further that, probed in mammary glands of wild type mice expressing mam- for mammary epithelial cells, non-canonical pathways rather mary epithelial-restricted ERα or c-srcAct. In adult mice taken at 24 than the canonical Hh-pathway operating through smo are weeks, the stromally-restricted paired-like homeodomain tran- required for branching morphogenesis at this stage of mammary scription factor, Alx4, was induced in mammary fibroblasts gland development. whereas the wild type adults do not express Alx4 in these cells. We conclude that ERα or c-src expression in the luminal cells induced the expression of factors that stimulated the surrounding Acknowledgements stromal cells. Interestingly, despite the rescue by ERα and c-src in the mes mice, induction of Alx4 expression was not observed in This project was funded by a grant to PAH by the Canadian mes stromal cells. Thus, wild type mPtch1 appears to be required Institutes of Health Research (MOP-97929). for stromal induction of Alx4 by adjacent mammary epithelial cells. The basis for the difference between mes and wild type mice has not been defined. We propose the possibility that ERα and c- Appendix A. Supporting information src induce the expression of a Hh-ligand in mammary epithelial cells. This ligand then stimulates adjacent mesenchymal cells to Supplementary data associated with this article can be found in express Alx4 but is dependent on the activities of wild type the online version at http://dx.doi.org/10.1016/j.ydbio.2014.04.007. mPtch1in these cells. This model can be tested in reciprocal transplantation experiments between mes and wild type mam- mary glands. Regardless, these data offer further evidence for the References role of components of the Hh-pathway in mediating signalling interactions between stromal and epithelial compartments in the developing mammary gland. Arao, Y., Hamilton, K.J., Ray, M.K., Scott, G., Mishina, Y., Korach, K.S., 2011. Estrogen receptor α AF-2 mutation results in antagonist reversal and reveals tissue While the focus of this study was on Hh-signalling that selective function of estrogen receptor modulators. Proc. Nat. Acad. Sci. U.S.A. operates, at least in part, through novel non-canonical signalling 108, 14986–14991. pathways, we also tested the response of primary mammary Arnal, J.-F., Fontaine, C., Abot, A., Valera, M.-C., Laurell, H., Gourdy, P., Lenfant, F., 2013. Lessons from the dissection of the activation functions (AF-1 and AF-2) of epithelial cells in culture to signal through the smo-dependent the in vivo. Steroids 78, 576–582. canonical pathway. Under the conditions of primary cell culture, Arnold, S.F., Obourn, J.D., Jaffe, H., Notides, A.C., 1995. Phosphorylation of the we observed that primary mammary epithelial did not induce human estrogen receptor on tyrosine 537 in vivo and by src family tyrosine kinases in vitro. Mol. Endocrinol. 9, 24–33. expression of the Hh-target gene, mPtch1, and only weakly Bunone, G., Briand, P.A., Miksicek, R.J., Picard, D., 1996. Activation of the unliganded induced mGli1. In contrast, primary mesenchymal cells from the estrogen receptor by EGF involves the MAP kinase pathway and direct same animals responded robustly to the same N-Shh ligand, as phosphorylation. EMBO J. 15, 2174–2183. Capurro, M.I., Shi, W., Filmus, J., 2012. LRP1 mediates the Shh-induced endocytosis evident by mPtch1 and mGli1 expression levels being strongly of the GPC3-Shh complex. J. Cell Sci. induced. The lack of signalling in primary epithelial cells through Carstea, E.D., Morris, J.A., Coleman, K.G., Loftus, S.K., Zhang, D., Cummings, C., Gu, J., the canonical pathway is consistent with the lack of response we Rosenfeld, M.A., Pavan, W.J., Krizman, D.B., Nagle, J., Polymeropoulos, M.H., Sturley, S.L., Ioannou, Y.A., Higgins, M.E., Comly, M., Cooney, A., Brown, A., observed for human cell lines derived from breast cancers or Kaneski, C.R., Blanchette-Mackie, E.J., Dwyer, N.K., Neufeld, E.B., Chang, T.Y., untransformed cells, including MCF7, MDA-MB-231 and MCF10a Liscum, L., Strauss 3rd, J.F., Ohno, K., Zeigler, M., Carmi, R., Sokol, J., Markie, D., (Chang et al., 2010 and unpublished observations). For some of O’Neill, R.R., Van Diggelen, O.P., Elleder, M., Patterson, M.C., Brady, R.O., Vanier, these latter lines, specifically MCF10a and MDA-MB-231, the M.T., Pentchev, P.G., Tagle, D.A., 1997. Niemann-Pick C1 disease gene: homology to mediators of cholesterol homeostasis. Science 277, 228–231. expression of smo message appears to be absent, at least as Castoria, G., Barone, M.V., Di Domenico, M., Bilancio, A., Ametrano, D., Migliaccio, A., determined by sensitive rtPCR studies of smo expression in these Auricchio, F., 1999. Non-transcriptional action of oestradiol and progestin – cells (Chang et al., 2010; Mukherjee et al., 2006; Zhang et al., triggers DNA synthesis. EMBO J. 18, 2500 2510. Castoria, G., Giovannelli, P., Lombardi, M., De Rosa, C., Giraldi, T., De Falco, A., 2009). The failure to stimulate the canonical pathway by N-Shh is Barone, M.V., Abbondanza, C., Migliaccio, A., Auricchio, F., 2012. Tyrosine distinct from the myriad of studies describing the consequences of phosphorylation of estradiol receptor by Src regulates its hormone- altered expression of the downstream Hh-pathway components dependent nuclear export and cell cycle progression in breast cancer cells. Oncogene 31, 4868–4877. on cell growth, mobility, transformation or metastasis in mam- Chang, H., Balenci, L., Okolowsky, N., Muller, W.J., Hamel, P.A., 2012. Mammary mary epithelial cell lines or, more often, breast cancer cell lines epithelial-restricted expression of activated c-src rescues the block to 228 N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229

mammary gland morphogenesis due to the deletion of the C-terminus of Kousteni, S., Bellido, T., Plotkin, L.I., O’Brien, C.A., Bodenner, D.L., Han, L., Han, K., Patched-1. Dev. Biol. 370, 187–197. DiGregorio, G.B., Katzenellenbogen, J.A., Katzenellenbogen, B.S., Roberson, P.K., Chang, H., Li, Q., Moraes, R.C., Lewis, M.T., Hamel, P.A., 2010. Activation of Erk by Weinstein, R.S., Jilka, R.L., Manolagas, S.C., 2001. Nongenotropic, sex- sonic Hedgehog independent of canonical Hedgehog signalling. Int. J. Biochem. nonspecific signaling through the estrogen or androgen receptors: dissociation Cell Biol. 42, 1462–1471. from transcriptional activity. Cell 104, 719–730. Che, J., Zhang, F.-Z., Zhao, C.-Q., Hu, X.-D., Fan, S.-J., 2013. Cyclopamine is a novel Lannigan, D.A., 2003. Estrogen receptor phosphorylation. Steroids 68, 1–9. Hedgehog signaling inhibitor with significant anti-proliferative, anti-invasive Li, L., Hisamoto, K., Kim, K.H., Haynes, M.P., Bauer, P.M., Sanjay, A., Collinge, M., and anti-estrogenic potency in human breast cancer cells. Oncol. Lett. 5, Baron, R., Sessa, W.C., Bender, J.R., 2007. Variant estrogen receptor-c-Src 1417–1421. molecular interdependence and c-Src structural requirements for endothelial Chen, D., Washbrook, E., Sarwar, N., Bates, G.J., Pace, P.E., Thirunuvakkarasu, V., NO synthase activation. Proc. Nat. Acad. Sci. U.S.A. 104, 16468–16473. Taylor, J., Epstein, R.J., Fuller-Pace, F.V., Egly, J.-M., Coombes, R.C., Ali, S., 2002. Loftus, S.K., Morris, J.A., Carstea, E.D., Gu, J.Z., Cummings, C., Brown, A., Ellison, J., Phosphorylation of human estrogen receptor alpha at serine 118 by two distinct Ohno, K., Rosenfeld, M.A., Tagle, D.A., Pentchev, P.G., Pavan, W.J., 1997. Murine signal transduction pathways revealed by phosphorylation-specific antisera. model of Niemann-Pick C disease: mutation in a cholesterol homeostasis gene. Oncogene 21, 4921–4931. Science 277, 232–235. Couse, J.F., Curtis Hewitt, S., Korach, K.S., 2000. Receptor null mice reveal contrast- Lubahn, D.B., Moyer, J.S., Golding, T.S., Couse, J.F., Korach, K.S., Smithies, O., 1993. ing roles for estrogen receptor alpha and beta in reproductive tissues. J. Steroid Alteration of reproductive function but not prenatal sexual development after Biochem. Mol. Biol. 74, 287–296. insertional disruption of the mouse estrogen receptor gene. Proc. Nat. Acad. Sci. Cui, W., Wang, L.-H., Wen, Y.-Y., Song, M., Li, B.-L., Chen, X.-L., Xu, M., An, S.-X., Zhao, U. S. A. 90, 11162–11166. J., Lu, Y.-Y., Mi, X.-Y., Wang, E.-H., 2010. Expression and regulation mechanisms Makino, S., Masuya, H., Ishijima, J., Yada, Y., Shiroishi, T., 2001. A spontaneous of Sonic Hedgehog in breast cancer. Cancer Sci. 101, 927–933. mouse mutation, mesenchymal dysplasia (mes), is caused by a deletion of the Di Domenico, M., Castoria, G., Bilancio, A., Migliaccio, A., Auricchio, F., 1996. most C-terminal cytoplasmic domain of patched (ptc). Dev. Biol. 239, 95–106. Estradiol activation of human colon carcinoma-derived Caco-2 cell growth. Mangelsdorf, D.J., Thummel, C., Beato, M., Herrlich, P., Schütz, G., Umesono, K., Cancer Res. 56, 4516–4521. Blumberg, B., Kastner, P., Mark, M., Chambon, P., Evans, R.M., 1995. The nuclear Díaz-Cruz, E.S., Sugimoto, Y., Gallicano, G.I., Brueggemeier, R.W., Furth, P.A., 2011. receptor superfamily: the second decade. Cell 83, 835–839. Comparison of increased aromatase versus ERα in the generation of mammary McGlynn, L.M., Tovey, S., Bartlett, J.M.S., Doughty, J., Cooke, T.G., Edwards, J., 2013. hyperplasia and cancer. Cancer Res. 71, 5477–5487. Interactions between MAP kinase and oestrogen receptor in human breast Fiaschi, M., Rozell, B., Bergstrom, A., Toftgard, R., 2009. Development of mammary cancer. Eur. J. Cancer 49, 1176–1186. tumors by conditional expression of GLI1. Cancer Res. 69, 4810–4817. Miermont, A.M., Parrish, A.R., Furth, P.A., 2010. Role of ERalpha in the differential Frech, M.S., Halama, E.D., Tilli, M.T., Singh, B., Gunther, E.J., Chodosh, L.A., Flaws, J.A., response of Stat5a loss in susceptibility to mammary preneoplasia and DMBA- Furth, P.A., 2005. Deregulated estrogen receptor α expression in mammary induced carcinogenesis. Carcinogenesis 31, 1124–1131. epithelial cells of transgenic mice results in the development of ductal Migliaccio, A., Castoria, G., Di Domenico, M., De Falco, A., Bilancio, A., Lombardi, M., carcinoma in situ. Cancer Res. 65, 681–685. Barone, M.V., Ametrano, D., Zannini, M.S., Abbondanza, C., Auricchio, F., 2000. Frech, M.S., Torre, K.M., Robinson, G.W., Furth, P.A., 2008. Loss of cyclin D1 in Steroid-induced -oestradiol receptor beta-Src complex trig- concert with deregulated estrogen receptor alpha expression induces DNA gers prostate cancer cell proliferation. EMBO J. 19, 5406–5417. damage response activation and interrupts mammary gland morphogenesis. Migliaccio, A., Di Domenico, M., Castoria, G., De Falco, A., Bontempo, P., Nola, E., Oncogene 27, 3186–3193. Auricchio, F., 1996. Tyrosine kinase/p21ras/MAP-kinase pathway activation by Hanke, J.H., Gardner, J.P., Dow, R.L., Changelian, P.S., Brissette, W.H., Weringer, E.J., estradiol-receptor complex in MCF-7 cells. EMBO J. 15, 1292–1300. Pollok, B.A., Connelly, P.A., 1996. Discovery of a novel, potent, and Src family- Migliaccio, A., Piccolo, D., Castoria, G., Di Domenico, M., Bilancio, A., Lombardi, M., selective tyrosine kinase inhibitor study of Lck- and FynT- dependent T cell Gong, W., Beato, M., Auricchio, F., 1998. Activation of the Src/p21ras/Erk activation. J. Biol. Chem. 271, 695–701. pathway by via cross-talk with estrogen receptor. EMBO Harvey, M.C., Fleet, A., Okolowsky, N., Hamel, P.A., 2014. Distinct effects of the J. 17, 2008–2018. mesenchymal dysplasia variant of murine patched-1 on canonical and non- Moraes, R.C., Chang, H., Harrington, N., Landua, J.D., Prigge, J.T., Lane, T.F., Wain- canonical Hedgehog-signalling pathways. J. Biol. Chem. wright, B.J., Hamel, P.A., Lewis, M.T., 2009. Ptch1 is required locally for Hatsell, S.J., Cowin, P., 2006. Gli3-mediated repression of Hedgehog targets is mammary gland morphogenesis and systemically for ductal elongation. Devel- required for normal mammary development. Development 133, 3661–3670. opment 136, 1423–1432. Haynes, M.P., Li, L., Sinha, D., Russell, K.S., Hisamoto, K., Baron, R., Collinge, M., Sessa, Mukherjee, S., Frolova, N., Sadlonova, A., Novak, Z., Steg, A., Page, G.P., Welch, D.R., W.C., Bender, J.R., 2003. Src kinase mediates phosphatidylinositol 3-kinase/Akt- Lobo-Ruppert, S.M., Ruppert, J.M., Johnson, M.R., Frost, A.R., 2006. Hedgehog dependent rapid endothelial nitric-oxide synthase activation by estrogen. signaling and response to cyclopamine differ in epithelial and stromal cells in J. Biol. Chem. 278, 2118–2123. benign breast and breast cancer. Cancer Biol. Ther. 5, 674–683. Heldring, N., Pike, A., Andersson, S., Matthews, J., Cheng, G., Hartman, J., Tujague, Nieuwenhuis, E., Barnfield, P.C., Makino, S., Hui, C.-C., 2007. Epidermal hyperplasia M., Strom,̈ A., Treuter, E., Warner, M., Gustafsson, J.-A., 2007. Estrogen and expansion of the interfollicular stem cell compartment in mutant mice receptors: how do they signal and what are their targets. Physiol. Rev. 87, with a C-terminal truncation of Patched1. Dev. Biol. 308, 547–560. 905–931. Niranjan, B., Buluwela, L., Yant, J., Perusinghe, N., Atherton, A., Phippard, D., Dale, T., Hruska, K.S., Tilli, M.T., Ren, S., Cotarla, I., Kwong, T., Li, M., Fondell, J.D., Hewitt, J.A., Gusterson, B., Kamalati, T., 1995. HGF/SF: a potent cytokine for mammary Koos, R.D., Furth, P.A., Flaws, J.A., 2002. Conditional over-expression of estrogen growth, morphogenesis and development. Development 121, 2897–2908. receptor alpha in a transgenic mouse model. Transgenic Res. 11, 361–372. O’Toole, S.A., Swarbrick, A., Sutherland, R.L., 2009. The Hedgehog signalling path- Huang, S., Zhang, Z., Zhang, C., Lv, X., Zheng, X., Chen, Z., Sun, L., Wang, H., Zhu, Y., way as a therapeutic target in early breast cancer development. Expert Opin. Zhang, J., Yang, S., Lu, Y., Sun, Q., Tao, Y., Liu, F., Zhao, Y., Chen, D., 2013. Ther. Targets 13, 1095–1103. Activation of Smurf E3 ligase promoted by smoothened regulates Hedgehog Pedram, A., Razandi, M., Kim, J.K., O’Mahony, F., Lee, E.Y., Luderer, U., Levin, E.R., signaling through targeting patched turnover. PLoS Biol. 11, e1001721. 2009. Developmental phenotype of a membrane only estrogen receptor alpha Hudson, R., Taniguchi-Sidle, A., Boras, K., Wiggan, O., Hamel, P.A., 1998. Alx-4, a (MOER) mouse. J. Biol. Chem. 284, 3488–3495. transcriptional activator whose expression is restricted to sites of epithelial- Qu, S., Li, L., Wisdom, R., 1997a. Alx-4: cDNA cloning and characterization of a novel mesenchymal interactions. Dev. Dyn. 213, 159–169. paired-type homeodomain protein. Gene 203, 217–223. Hui, M., Cazet, A., Nair, R., Watkins, D.N., O’Toole, S.A., Swarbrick, A., 2013. The Qu, S., Niswender, K.D., Ji, Q., Van der Meer, R., Keeney, D., Magnuson, M.A., Hedgehog signalling pathway in breast development, carcinogenesis and Wisdom, R., 1997b. and ectopic ZPA formation in Alx-4 mutant cancer therapy. Breast Cancer Res. 15, 203. mice. Development 124, 3999–4008. Improta-Brears, T., Whorton, A.R., Codazzi, F., York, J.D., Meyer, T., McDonnell, D.P., Qu, S., Tucker, S.C., Ehrlich, J.S., Levorse, J.M., Flaherty, L.A., Wisdom, R., Vogt, T.F., 1999. Estrogen-induced activation of mitogen-activated protein kinase 1998. Mutations in mouse Aristaless-like4 cause Strong's luxoid polydactyly. requires mobilization of intracellular calcium. Proc. Nat. Acad. Sci. U.S.A. 96, Development 125, 2711–2721. 4686–4691. Rasmussen, S.B., Young, L.J.T., Smith, G.H., 2000. Preparing Mammary Gland Whole Joshi, P.A., Chang, H., Hamel, P.A., 2006. Loss of Alx4, a stromally-restricted Mounts from Mice. Plenum Publishers, New York. homeodomain protein, impairs mammary epithelial morphogenesis. Dev. Biol. Sasaki, H., Hui, C., Nakafuku, M., Kondoh, H., 1997. A binding site for Gli is 297, 284–294. essential for HNF-3beta floor plate enhancer activity in transgenics and can Kameda, C., Nakamura, M., Tanaka, H., Yamasaki, A., Kubo, M., Tanaka, M., Onishi, H., respond to Shh in vitro. Development 124, 1313–1322. Katano, M., 2010. Oestrogen receptor-alpha contributes to the regulation of the Song, R.X.-D., McPherson, R.A., Adam, L., Bao, Y., Shupnik, M., Kumar, R., Santen, R.J., Hedgehog signalling pathway in ERalpha-positive gastric cancer. Br. J. Cancer 2002. Linkage of rapid estrogen action to MAPK activation by ERalpha-Shc 102, 738–747. association and Shc pathway activation. Mol. Endocrinol. 16, 116–127. Kim, H., Laing, M., Muller, W., 2005. c-Src-null mice exhibit defects in normal Sun, J., Zhou, W., Kaliappan, K., Nawaz, Z., Slingerland, J.M., 2012. ERα phosphor- mammary gland development and ERalpha signaling. Oncogene 24, ylation at Y537 by Src triggers E6-AP-ERα binding, ERα ubiquitylation, 5629–5636. promoter occupancy, and target . Mol. Endocrinol. 26, Klinge, C.M., 2000. Estrogen receptor interaction with co-activators and co- 1567–1577. . Steroids 65, 227–251. Sweet, H.O., Bronson, R.T., Donahue, L.R., Davisson, M.T., 1996. Mesenchymal Koga, K., Nakamura, M., Nakashima, H., Akiyoshi, T., Kubo, M., Sato, N., Kuroki, S., dysplasia: a recessive mutation on 13 of the mouse. J. Hered. Nomura, M., Tanaka, M., Katano, M., 2008. Novel link between estrogen 87, 87–95. receptor alpha and Hedgehog pathway in breast cancer. Anticancer Res. 28, Tilli, M.T., Frech, M.S., Steed, M.E., Hruska, K.S., Johnson, M.D., Flaws, J.A., Furth, P.A., 731–740. 2003. Introduction of estrogen receptor-alpha into the tTA/TAg conditional N. Okolowsky et al. / Developmental Biology 391 (2014) 219–229 229

mouse model precipitates the development of estrogen-responsive mammary Webster, M.A., Cardiff, R.D., Muller, W.J., 1995. Induction of mammary epithelial adenocarcinoma. Am. J. Pathol. 163, 1713–1719. hyperplasias and mammary tumors in transgenic mice expressing a murine Varricchio, L., Migliaccio, A., Castoria, G., Yamaguchi, H., Falco, A. de, Domenico, M. mammary tumor virus/activated c-src fusion gene. Proc. Nat. Acad. Sci. U.S.A. D., Giovannelli, P., Farrar, W., Appella, E., Auricchio, F., 2007. Inhibition of 92, 7849–7853. estradiol receptor/Src association and cell growth by an estradiol eeceptor α Zhang, X., Harrington, N., Moraes, R.C., Wu, M.-F., Hilsenbeck, S.G., Lewis, M.T., – tyrosine-phosphorylated peptide. Mol. Cancer Res. 5, 1213 1221. 2009. Cyclopamine inhibition of human breast cancer cell growth independent fi Wakeling, A.E., Dukes, M., Bowler, J., 1991. A potent speci c pure antiestrogen with of Smoothened (Smo). Breast Cancer Res. Treat. 115, 505–521. clinical potential. Cancer Res. 51, 3867–3873.