Ovarian hormones increase breast cancer risk by poorly understood mechanisms. We assess the role of progesterone on global stem cell function by serially transplanting mouse mammary epithelia. Progesterone receptor (PR) deletion severely reduces the regeneration capacity of the mammary epithelium. The PR target, receptor activator of Nf‐κB ligand (RANKL), is not required for this function, and the deletion of Wnt4 reduces the mammary regeneration capacity even more than PR ablation. A fluorescent reporter reveals so far undetected perinatal Wnt4 expression that is independent of hormone signaling. Pubertal and adult Wnt4 expression is specific to PR+ luminal cells and requires intact PR signaling. Conditional deletion of Wnt4 reveals that this early, previously unappreciated, Wnt4 expression is functionally important. We provide genetic evidence that canonical Wnt signaling in the myoepithelium required PR and Wnt4, whereas the canonical Wnt signaling activities observed in the embryonic mammary bud and in the stroma around terminal end buds are independent of Wnt4. Thus, progesterone and Wnt4 control stem cell function through a luminal–myoepithelial crosstalk with Wnt4 acting independent of PR perinatally.
This paper ascribes a new role for Wnt4 in pre‐pubertal mammary gland development while revealing luminal cells to respond to Wnt activation. During regeneration, Wnt4 interacts with progesterone receptor signaling, correcting previous notions on RANKL signaling in this context.
Wnt4 is an essential control factor for mammary epithelial stem cell function.
RANKL is not required for mammary gland regeneration potential.
Wnt4 activates canonical Wnt signaling in the basal/myoepithelial compartment.
Progesterone receptor signaling is required for mammary epithelial Wnt4 expression already during puberty.
The two key ovarian hormones, 17‐β‐estradiol (E2) and progesterone, regulate postnatal mammary gland development but also promote carcinogenesis in this organ. They act via nuclear receptors, the estrogen receptor α (ER) and the progesterone receptor (PR), respectively, which are expressed in between 30 and 50% of the mammary epithelial cells in the inner, luminal layer (Clarke et al, 1997). It was proposed that they impinge on ER‐/PR‐negative mammary stem cells by paracrine mechanisms (Tanos & Brisken, 2008). Experimental evidence for this model was provided with fluorescence‐activated cell sorting (FACS)‐based approaches (Asselin‐Labat et al, 2010; Joshi et al, 2010). The single‐cell‐based methods have been used to characterize mammary epithelial cell populations and to establish a cellular hierarchy within the mammary epithelium. Among dissociated CD24+ mouse mammary epithelial cells, the cell populations with high surface expression of integrin β1 (CD29hi) or integrin α6 (CD49fhi) were enriched for cells with the ability to establish new milk ducts in mammary fat pads surgically cleared of their endogenous epithelium and were hence considered bipotent mammary stem cells able to give rise to both luminal and basal/myoepithelial cell lineage (Shackleton et al, 2006; Stingl et al, 2006). These cells express basal/myoepithelial markers such as cytokeratin 5 and 14, smooth muscle actin, and laminin (Shackleton et al, 2006; Stingl et al, 2006) and their numbers increase during pregnancy and after stimulation with E2 and progesterone (Asselin‐Labat et al, 2010; Joshi et al, 2010) and decrease upon ovariectomy or anti‐estrogen treatment (Asselin‐Labat et al, 2010).
However, the physiological relevance of the dramatic expansion of these bipotent stem cells in response to hormones (Asselin‐Labat et al, 2010; Joshi et al, 2010) is questioned by lineage‐tracing experiments showing that postnatal mammary gland development is largely driven by luminal and basal/myoepithelial lineage‐restricted stem cells (Taddei et al, 2008; Zeng & Nusse, 2010; Van Keymeulen et al, 2011; Rios et al, 2014). The lineage‐restricted stem cells are not amenable to study by the single stem cell assays where the normal stem cell niche and its microenvironment are disrupted.
Recurrent peaks in serum progesterone levels are linked to menstrual cycles which are an important risk factor for breast carcinogenesis (Brisken, 2013). Moreover, most cell proliferation occurs in the luminal compartment and breast cancers arise from luminal and/or luminal progenitor cells (Molyneux et al, 2010). This begs the question how important progesterone is to mammary stem cell function in the intact tissue context and through which signaling pathways it impinges on different types of mammary stem cells, that is, bipotential, luminal restricted, and basal restricted.
Mammary epithelium was shown to serially engraft cleared mammary fat pads for up to seven cycles (Daniel, 1973). This assay, which is based on the engraftment of an intact piece of mammary epithelium, preserves the epithelial architecture with its associated extracellular matrix, fibroblasts, and immune cells. It is currently the only way to assay comprehensively the mammary regeneration potential reflective of multiple types of stem and progenitor cells. We combine this assay with different genetic mutant strain to define the relative contributions of progesterone signaling and its downstream mediators RANKL and Wnt4 to the regenerative potential of the mammary epithelium.
PR signaling for mammary epithelial self‐renewal
To assess the role of PR signaling in mammary stem cell control, we serially engrafted pieces of intact mammary epithelium from the PR−/− and wild‐type (WT) littermates into contralateral mammary fat pads surgically cleared of their endogenous epithelium (Fig 1A). This approach rather than injection of limiting dilutions of dissociated cell populations was chosen so that the physiological interactions between the stem/progenitor cells and their microenvironment in the mammary epithelium niche would be preserved and the function of all types of stem cells, the bipotential, the luminal‐restricted, and the basal‐restricted stem cells, could be evaluated. To unequivocally distinguish the engrafted cells from the endogenous epithelium that may inadvertently be left after surgery, we used donors that ubiquitously express the enhanced green fluorescent protein (EGFP) (Okabe et al, 1997). To ensure that comparable amounts of mutant (MT) and WT donor tissue with comparable amounts of epithelium were engrafted, we dissected pinhead‐sized fragments from the inguinal glands near the lymph node on the side proximal to the teat. Eight to 12 weeks after grafting, recipients were sacrificed and the extent of outgrowth in the engrafted mammary glands was determined. Pieces of mammary tissue resulting from the contralateral PR−/− and WT grafts were dissected and retransplanted (Fig 1A). The WT epithelium completely reconstituted most of the fat pads over 4 serial transplant cycles, as expected, but the PR−/− epithelium ceased to reconstitute the mammary gland by the 3rd cycle (Fig 1B, C, and H).
Paracrine mediators of PR signaling in mammary epithelial self‐renewal
The TNF‐α family member, RANKL, was previously implicated in the paracrine control of mammary stem cells by hormones on the basis of use of dissociated individual stem cell assays (Asselin‐Labat et al, 2010). To determine the functional importance of RANKL in epithelial self‐renewal, we serially engrafted intact mammary epithelium derived from the RANKL−/− and the RANKL+/+ mice into cleared contralateral mammary fat pads. WT epithelium fully reconstituted fat pads in most hosts over 4 serial transplants and grew as expected (Fig 1D and E). Unexpectedly, RANKL−/− epithelia had the same regeneration capacity (Fig 1D and E). The only significant difference was that the MT grafts generated fewer side branches (Fig 1D and insets), consistent with the reported proproliferative activity of RANKL and its role in side branching (Beleut et al, 2010).
We tested the role of Wnt4 since evidence has accumulated that Wnt signaling is important for mammary stem cell function (Cai et al, 2014; Kessenbrock et al, 2013; Liu et al, 2004; van Amerongen et al, 2012; Wang et al, 2014; Zeng & Nusse, 2010) and Wnt4 is a key paracrine mediator of progesterone action (Brisken et al, 2000). As Wnt4−/− mice die at birth (Vainio et al, 1999), intact mammary epithelial buds from E12.5/E13.5 embryos were used for the initial engraftment. It has been reported that embryonic epithelia have more stem cells than postnatal epithelia (Spike et al, 2012). Notwithstanding, we noticed that epithelial tissue isolated from the Wnt4+/+ and the Wnt4+/− embryos reconstituted completely the mammary gland to the same extent as the WT epithelia from postnatal mammary glands through three transplantation cycles (Fig 1C, E, and G). However, the Wnt4−/− epithelium only established 50% of the fat pad in the first cycle and was reduced to 10% by the third cycle (Fig 1F and G). The much more significant impairment of reconstitution capacity of the Wnt4−/− versus the PR−/− grafts compared to their respective contralateral controls (Fig 1H) points to a key role for Wnt4 in the maintenance of the mammary stem cell function and indicates that PR is not exclusively controlling Wnt4 expression. RANKL does not appear to be essential in the control of the mammary stem cells under physiological conditions.
Control of Wnt4 expression
Previous studies suggested that Wnt4 is expressed only concomitant with high progesterone secretion in the adult and pregnant mammary gland (Gavin & McMahon, 1992; Weber‐Hall et al, 1994). The finding that deletion of Wnt4 affected mammary regeneration potential more severely than abrogation of PR signaling suggested that Wnt4 may have PR‐independent functions in stem cell stimulation. To detect Wnt4 expression at low levels, we crossed mice which express Cre from the Wnt4 locus (Wnt4::Cre) (Shan et al, 2009) to the mT/mG dual Cre reporter strain in which ubiquitous tomato expression is replaced by membrane EGFP upon Cre activation (Muzumdar et al, 2007). We failed to detect expression in embryos (E12.5 and E18.5) and were unable to detect EGFP in the mammary glands of newborn or 3‐day‐old double transgenic mice (Wnt4::GFP). EGFP expression was detected on postnatal day 5 by fluorescence stereomicroscopy (Fig 2A) prior to the onset of ovarian function. Immunofluorescence of histological sections revealed EGFP expression in scattered mammary epithelial cells on day 5 (Fig 2B) and day 10 (Fig 2C). PR expression was not detected at these stages (Fig 2B and C). Consistent with Wnt4 being a PR target, the double immunofluorescence for histological sections from postnatal double transgenic mice (Wnt4::EGFP) suggested that EGFP expression is restricted to PR+ luminal cells in the glands of pubertal mice (Fig 2D), adult (Fig 2E and inset), and pregnant females (Fig 2F and inset). To assess whether myoepithelial cells may express EGFP, we performed triple immunofluorescence for EGFP, PR, and the myoepithelial marker p63. In mammary epithelia from 5‐day‐old Wnt4::EGFP females (n = 7), rare double positive cells were detected, and in most sections, cells expressed either p63 or EGFP (Fig 2G). In 4‐ (Fig 2H) and 8‐week‐old (Fig 2I) females, p63 and EGFP staining labeled distinct cells. Thus, Wnt4 is expressed almost exclusively in luminal cells. The Wnt4 expressing cells appear to be terminally differentiated as no clonal clusters of EGFP+ cells are observed.
To assess whether trace amounts of estrogens and progesterone of maternal origin could account for this perinatal Wnt4 expression, we analyzed d15 mammary glands from the Wnt4::EGFP mice on WT, ERα−/−, and PR−/− genetic backgrounds by epifluorescence stereomicroscopy. At this stage, ductal outgrowth was comparable by red fluorescence (Fig 2J–L). Neither ERα nor PR deletion altered Wnt4::EGFP expression (Fig 2M–O) indicating that perinatal Wnt4 expression is largely independent of ERα and PR signaling.
To determine the respective roles of the two major ovarian hormones in control of Wnt4 expression, we pooled epithelial‐enriched organoids freshly isolated from mammary glands of pubertal and adult females (n = 3) and stimulated them for 6 h ex vivo (Fig 2P) (Ayyanan et al, 2011). Progesterone induced Wnt4 mRNA expression in pubertal and adult organoids to 8.7‐ and 4.5‐fold, respectively, whereas E2 elicited a 1.6‐fold induction of Wnt4 mRNA in the pubertal organoids only (Fig 2Q). To assess the physiological importance of PR signaling for pubertal Wnt4 expression, we grafted Wnt4::GFP epithelium derived from donors either PR−/− or PRWT to contralateral cleared fat pads of 3‐week‐old hosts. The engrafted glands were analyzed 3 weeks later when the recipients were pubertal. Epifluorescence stereomicroscopy for dTomato revealing unrecombined cells confirmed the presence of ductal outgrowth of PRWT and PR−/− grafts (Fig 2R and S). EGFP expression was readily detected in the PRWT graft (Fig 2T) but completely absent from some of the PR−/− grafts (Fig 2U). Double epifluorescence stereomicroscopy on a PRWT control graft reveals that EGFP is strongly enriched in the TEBs (Fig 2V and X), and in the contralateral PR−/− grafts, some EGFP expression is observed at the origin of the outgrowth (Fig 2W). These findings are consistent with perinatal ER‐/PR‐independent Wnt4 expression and indicate that pubertal Wnt4 induction is mediated by PR signaling.
Consequences of Wnt4 ablation on cell proliferation
The observation that Wnt4 deletion impaired the regenerative capacity of the mammary epithelium more severely than PR deletion did, pointed to a role of Wnt4 before puberty. The fat pad grafting approach used to determine Wnt4 function (Brisken et al, 2000) assesses gene function from puberty onward because the donor epithelium is placed into a 3‐week‐old host. To determine the role of Wnt4 perinatally and at the onset of puberty, we conditionally deleted Wnt4 in the mammary epithelium by crossing mice with two conditional Wnt4 alleles (Wnt4fl/fl) (Shan et al, 2010) to mice that express Cre in the mammary epithelium under the control of the MMTV‐LTR (A‐strain) (Wagner et al, 2001). To identify cells in which Cre‐mediated recombination had occurred, the mice were crossed to the mT/mG dual Cre reporter strain (Muzumdar et al, 2007). Analysis of MMTV::Cre; mT/mG double transgenic females revealed widespread EGFP expression at postnatal day 10 both by stereo microscopy and (Fig 3A–E) immunofluorescence for EGFP (Fig 3F). In the Wnt4 depleted (MT) mammary glands, a 10% decrease in the number of branching points was observed compared to control littermates around day 10 (Fig 3G and H). In pubertal MT glands, the number of terminal end buds (TEBs) had decreased to 54% of the controls (Fig 3I and J). Similarly, the area of fat pad filled by ducts was 60% of that measured in littermates (Fig 3I and K). Cell proliferation, as assessed by BrdU incorporation, was reduced to 65% of that in the WT counterparts in TEBs of Wnt4 mutants (Fig 3M). The proliferative index of about 6% in the subtending ducts was not affected in the Wnt4‐deficient glands (Fig 3L). Thus, Wnt4 is required for perinatal and pubertal ductal expansion.
Activation of myoepithelial cells through canonical Wnt signaling
Wnt4 can activate its signaling, both canonical and non‐canonical Wnt signaling (Lyons et al, 2004; Heinonen et al, 2011). Canonical Wnt signaling activity can be assessed in vivo using the Axin2::LacZ reporter mouse strain (Leung et al, 2002; Lustig et al, 2002) and was reported in a subset of CD29fhi or CD49hi breast stem cells (Zeng & Nusse, 2010; van Amerongen et al, 2012). We detected reporter activity in the mammary buds of E12.5/13.5 embryos (Fig 4A) consistent with previous reports based on an artificial Wnt signaling reporter (Chu et al, 2004). Perinatal β‐galactosidase activity was confined to the nipple area (Fig 4B). Importantly, during puberty, LacZ expression was detected in the stroma surrounding the TEB necks (Fig 4C and D). In adulthood, β‐galactosidase activity was readily detected in the ducts where it peaked on day 8.5 of pregnancy (Fig 4E). At day 14.5 of pregnancy, β‐galactosidase activity was still detected in the ducts but not in the newly formed alveoli (Fig 4F). Histological sectioning suggested that X‐gal staining mapped to the myoepithelial layer (Fig 4G); immunostaining with the myoepithelial marker p63 (Fig 4H) confirmed that LacZ is exclusively expressed in myoepithelial cells. Thus, canonical Wnt signaling in the postnatal mammary epithelium is confined to myoepithelial cells.
The peak in myoepithelial β‐galactosidase activity during mid‐pregnancy suggested that serum progesterone levels and hence Wnt4 expression may control canonical Wnt signaling activation. Indeed, when mammary glands in individual mice were analyzed, during progesterone‐low estrous, they had lower β‐galactosidase activity (Fig 4I) and lower Wnt4 and Axin2 mRNA levels than the glands analyzed during progesterone‐high diestrous (Fig 4J); the fold differences varied between different animals (Fig 4I and J). To assess whether PR signaling induces canonical Wnt signaling, ovariectomized Axin2::LacZ females were treated with E2, to restore PR expression, and progesterone. Axin2 transcription reflected by β‐galactosidase activity was induced by this combination but not by solvent or E2 alone (Fig 4K). Thus, progesterone stimulation results in increased transcription of Axin2, in the context of ER‐dependent PR induction.
Multiple Wnts, some of which are secreted by mammary stromal cells, have been implicated in canonical Wnt signaling activation (Kessenbrock et al, 2013). To assess whether canonical Wnt signaling requires Wnt4 expression, we generated Axin2::LacZ transgenic mice in a Wnt4−/− background. In females engrafted with mammary buds from Axin2::LacZ+ female embryos, β‐galactosidase activity was readily detected in Wnt4+/+ grafts at pregnancy day 8.5 but abrogated in the contralateral Wnt4−/− epithelia (Fig 4L). Similarly, it was decreased in PR+/− and abrogated in PR−/− epithelia (Fig 4M), indicating that both PR and Wnt4 are required for canonical Wnt signaling activation in the myoepithelium.
Our finding that Wnt4 expression is detected only from postnatal day 5 onward suggested that other family members, possibly Wnt10b, might be responsible for canonical Wnt signaling activation in the embryonic mammary bud. In line with this scenario, β‐galactosidase activity was readily detected in the embryonic mammary buds of Wnt4−/−::axin2::LacZ female E14.5 embryos (Fig 5A). To test whether the stromal Axin2 expression around TEB necks depends on epithelial Wnt4 expression, we generated MMTV::Cre; axin2::LacZ females either Wnt4fl/fl or Wnt4fl/wt. Analysis of their mammary glands during puberty (6 weeks) showed that β‐galactosidase activity was comparable between the two genotypes (Fig 5B). Thus, canonical Wnt signaling activation in the mammary bud and stroma is Wnt4 independent, whereas specifically in the myoepithelium, canonical Wnt signaling activation requires epithelial Wnt4 expression.
Our data point to Wnt4 as a pivotal control factor of stem cell function for postnatal mammary gland development. We have uncovered a novel role for Wnt4 in perinatal development and puberty with progesterone as its major endocrine control factor. Progesterone, colloquially named ‘pregnancy hormone,’ appears as a primordial systemic factor in the postnatal mammary gland. It is the major proliferative stimulus to the adult mammary epithelium (Beleut et al, 2010) and controls the regenerative potential of the mammary gland by activating stem/progenitor cells throughout hormone‐dependent development. Surprisingly, while hormone stimulation experiments had shown that estrogens induce Wnt4 expression (Brisken et al, 2000; Cai et al, 2014), we find that genetic deletion of PR signaling completely abrogated Wnt4 expression during puberty. Yet, the two ovarian hormones remain intertwined in Wnt4 control with ERα signaling acting indirectly as an upstream regulator of PR expression (Haslam & Shyamala, 1979) (Fig 6A).
We had previously analyzed Wnt4 function in the mammary gland by grafting Wnt4−/− embryonic buds to cleared fat pads of pubertal mice and therefore failed to discern the prepubertal function of Wnt4 uncovered through the use of the conditional Wnt4 allele in the present work. The ability of Wnt4−/− epithelium to form alveoli, which was preserved in previous transplants (Brisken et al, 2000), was also observed upon serial transplantation when mice were impregnated in this study.
Our finding that RANKL is not important for stem cell potential is in apparent contradiction with previous results based on assays with dissociated cells (Asselin‐Labat et al, 2010). Compared to the widely used single‐cell‐based assays in which a defined number of cells are injected, grafting of intact epithelial fragments does not allow one to determine the number or fraction of cells endowed with regenerative potential. The approach does not disentangle the role of different types of stem and progenitor cells and their relative contributions to the outgrowth cannot be defined. Yet, the method is robust and multiple repeats of the experiment combined with several rounds of transplantation give a semiquantitative appreciation of an intact regeneration potential in a physiological tissue context that is missed by the use of dissociated cells. In fact, to our knowledge, this is currently the only stem cell assay in which at least part of the microenvironment remains intact that is so important to stem cell function. When single cells are injected into a cleared fat pad, they need to survive and to adhere to the stroma, an environment they are not exposed to physiologically. Both of these biological activities require integrin signaling and control each other through direct cadherin‐mediated cell–cell contacts. In the widely used FACS‐based stem cell assays, mammary stem cells are selected for high expression of integrin β1 or α6 (Shackleton et al, 2006; Stingl et al, 2006). It is conceivable that the selected cells are not intrinsically better ‘stem cells’ but that they have a strong advantage in adhering and surviving in the cleared fat pad, which is a conditio sine qua non for giving rise to de novo ducts. Alternatively, RANKL may be specifically required for the bipotential stem cells that are revealed under the challenging conditions of the single cell grafts. In the serial transplant approach, which is based on intact pieces of mammary tissue, the luminal‐ and the basal‐restricted stem cells likely account for most of the cellular proliferation during development and a role of RANKL for the bipotent progenitors may not be discerned.
Our finding that Wnt4 secreted by PR+ luminal cells activates canonical Wnt signaling exclusively in the neighboring basal/p63+ cells, points to a scenario, in which the myoepithelial/basal cells are a central component of the microenvironment or ‘niche’ that controls different types of stem cells (Fig 6B). The myoepithelial/basal cells are ultimately under control of progesterone and Wnt4; hence, stem cell activity controlled through the microenvironment/niche is linked to reproductive needs. Whether the rare bipotent mammary stem cells that are found in the basal layer (Wang et al, 2014) are directly and/or indirectly activated by Wnt4 is not addressed by our experiments but will be an exciting line of future work.
The finding that the myoepithelial/basal cells are the prime target of Wnt4 bears on a long‐standing conundrum. Wnt1 was long identified as an oncogene in the mouse mammary gland, and Wnt signaling is key for the development of the mammary gland. Yet, mutations in intracellular Wnt signaling components have not been found in breast carcinomas, which are of luminal origin. It is conceivable that Wnt signaling activation in the myoepithelial cells indirectly promotes tumorigenesis by inducing gene expression changes that result in the secretion of stimulatory signals and/or modulation of the extracellular matrix that result in the activation of luminal progenitor cells. Luminal progenitors with acquired oncogenic mutations could be expanded in response to Wnt4 stimulation of the myoepithelium (Fig 6C). In parallel, Wnt4, through its direct and/or indirect action on bipotent mammary stem cells, increases the number of stem cells. This in turn may result in an increased pool of luminal progenitors, which are more prone to oncogenic insults than more differentiated luminal cells.
These findings have clinical implications. The activation of the progesterone/Wnt4 pathway, which also operates in the human breast (Tanos et al, 2013; Pardo et al, 2014), may underlie the tumor promoting effects of recurrent menstrual cycles, oral contraception, and combined hormone replacement therapy with progestins. Selective progesterone receptor modulators and Wnt inhibitors, alone or together with RANKL inhibitors, may therefore be effective in breast cancer management, in particular, as preventive strategy in high‐risk premenopausal women.
Materials and Methods
Axin2::LacZ (Lustig et al, 2002), C57BL/6‐Tg(Act‐EGFP) (Okabe et al, 1997), ERα−/− (Dupont et al, 2000), MMTV::Cre (line A) (Wagner et al, 1997), mT/mG (Muzumdar et al, 2007), PR−/− (Lydon et al, 1995), RAG1−/− (Mombaerts et al, 1992), RANKL−/− (Wong et al, 1999), Wnt4−/− (Stark et al, 1994), Wnt4Cre (Shan et al, 2010), and Wnt4fl/fl mice (Shan et al, 2009) were kept on mixed genetic background 129SV/C57Bl6. All mice were maintained and handled according to the Swiss guidelines for animal safety. The ethic veterinary committee of canton of Vaud, Switzerland, approved all the animal experiments (Permit ID 1641.2 and 1641.3). To stage the estrus cycle, the vagina was flushed with 10 μl of PBS and vaginal secretions were collected, spread onto glass slides, and analyzed for different cell types (Caligioni, 2009). Mammary bud and mammary epithelial transplantations from 8‐week‐old donors were performed as described (Brisken et al, 2000). For serial transplantations, EGFP+ mammary duct outgrowth was visualized by stereo epi‐fluorescensce. Tissue fragments were prepared and retransplanted starting from at least three independent donors. Hormone treatments and BrdU injections were performed as described in Beleut et al (2010).
Mammary whole‐mounts and image analysis.
Mammary gland whole‐mounting and (5‐bromo‐4‐chloro‐3‐indolyl‐beta‐D‐galacto‐pyranoside) X‐gal staining were performed as described (Brisken et al, 1998). Images were acquired either using a LEICA MZ FLIII stereomicroscope with PixeLINK (PL‐A662) camera or LEICA M205FA/MZ16F fluorescent stereomicroscope with Leica DFC 340FX or Leica DC300F camera. Area of the fat pad filled by the mammary ducts was quantified by drawing a contour around the mammary ductal tips using Axiovision Rel 4.7 software. TEBs and branching points were quantified on images of whole‐mounted mammary glands and on fluorescence stereo micrographs.
Mammary glands were fixed in 4% paraformaldyde for 2 h at room temperature and embedded in paraffin. Five‐micrometer sections were used for nuclear red and for immunohistochemical or immunofluorescence staining. Anti‐p63 Molecular Probes 4A4 (1:100), anti‐BrdU Oxford biotechnology, OBT0030, 1:300), anti‐PR Thermo Fisher Scientific Pierce‐MA1‐411 (1:500), anti‐GFP Molecular Probes A11122 (1:800). Images were acquired on LEICA DM 2000 microscope with a PixeLINK (PL‐A662) camera or on Zeiss Axioplan 2‐imaging fluorescence microscope with Axiocam MRm camera.
RNA extraction and semiquantitative RT–PCR
Mammary glands were homogenized in TRIzol® (Invitrogen). Total RNA was isolated from fragments using RNeasy (Qiagen). cDNA was synthesized using random p(dN)6 primers (Roche Diagnostics) and MMLV reverse transcriptase (Invitrogen). Semiquantitative real‐time PCR analysis in triplicates was performed with SYBR Green PCR Core Reagents system (Qiagen)/PerfeCTa® SYBR® Green SuperMix for iQ™ (Quanta) on Realplex2 (Eppendorf) or 7900HT Fast Real‐Time PCR System (Applied Biosystems) qRT–PCR detection systems. All reactions were performed in triplicate. The following primers sequences were used: Wnt4, AGG AGT GCC AAT ACC AGT TCC, TGT GAG AAG GCT ACG CCA TA; Axin‐2, GGC AGT GAT GGA GGA AAA TG, TGG GTG AGA GTT TGC ACT TG; CK18 (Schroeder & Lee, 1998); 18S rRNA, GCA ATT ATT CCC CAT GAA CG, GGC CTC ACT AAA CCA TCC AA.
Two‐tailed, paired Student's t‐test was used to calculate statistical significance; data are shown as means ± SD. Statistical analyses were carried out using Microsoft Excel. For serial transplantation, the extent of fat pad filling in percentage at each generation was compared by Wilcoxon signed rank test between contralateral outgrowths. The statistical software R was used for analysis. The statistical test used and P‐values are indicated in each figure legend. P < 0.05 was considered to indicate statistical significance. *P < 0.05, **P < 0.01, ***P < 0.001.
RRD and CB designed experiments. MC, AA, and OYO performed experiments. RRD and DB performed experiments and analyzed data. JR and CB analyzed data. JS and SV provided mouse models. SV and CB contributed in writing the manuscript.
Conflict of interest
The authors declare that they have no conflict of interest.
The authors would like to thank Drs. T. Tanos, A. Wilson, R. Jeitziner, and J. Dessimoz for advice and V. Simanis for critical reading of the manuscript; Drs. Y. Choi, J. Lydon, and T. Okabe for providing RANKL+/−, PR+/−, and C57BL/6‐Tg(Act‐EGFP) strains. This work was supported by funds from the Swiss National Center of Competence in Research (NCCR) in Molecular Oncology, SNF3100A0‐112090, Oncosuisse, to C.B., a Marie Heim‐Vögtlin fellowship from Swiss National Science Foundation (MHV ‐ PMPDP3_129024) to R.D.R., and grants from the Academy of Finland (206038, 121647), Centre of Excellence Grant 2012–2017 of the Academy of Finland and FiDiPro (SQ) (251314), the Sigrid Jusélius Foundation, Novonordisk Fonden, and the European Union (LSHG‐CT‐2004‐005085; HEALTH‐F5‐2008‐223007 STAR‐T REK, EURenOmics) to S.V.
FundingSwiss National Center of Competence in Research (NCCR) in Molecular Oncology SNF3100A0‐112090
- © 2015 The Authors