| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Physiology (S.Z.H., A.D., M.T.) and the Cell and Molecular Biology Program (S.Z.H., K.S., M.A.), Michigan State University, East Lansing, Michigan 48824
Address all correspondence and requests for reprints to: Sandra Z. Haslam, Ph.D., Department of Physiology, 2201 Biomedical and Physical Sciences Building, Michigan State University, East Lansing, Michigan 48824. E-mail: shaslam{at}msu.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B ligand (RANKL), a progestin-induced paracrine factor, in mediating increased MEC proliferation. Quantitative RT-PCR showed that RANKL mRNA was induced by R5020 or HGF+R5020 and RANKL protein colocalized with PRs in LECs. The increased proliferation of MECs in response to HGF+R5020 could be blocked by neutralizing antibody to RANKL and reproduced by treatment with HGF plus exogenous RANKL in place of R5020. Neither R5020, nor exogenously administered RANKL increased proliferation of LECs. These results led us to conclude that RANKL, induced by progestin in PR-positive cells, is secreted and interacts with HGF to specifically increase proliferation of PR-negative MECs. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The rodent mammary gland provides a useful model to study LEC and MEC regulation and function. The focus of most studies has been on the regulation of LEC proliferation, morphogenesis, and differentiation. Fewer studies have focused on the regulation of MECs (1, 3). Most often the cell type-specific behavior and responses of LECs or MECs have been studied in vitro, separately (4, 5). This approach may not accurately reflect the coordinated responses of these two cell types in vivo. In vitro reconstitution experiments in which LECs have been cultured separately or recombined with MECs have demonstrated an important role of MECs in the establishment of LEC polarity (6, 7, 8). In the present report, we investigated the regulation and cell type-specific responses of LECs and MECs in vitro when both are present together in mammary organoids.
Hepatocyte growth factor (HGF) is a mesenchyme-derived growth factor that is synthesized in the stroma in vivo and stimulates the proliferation, motility, and morphogenesis of nearby epithelium (9, 10). HGF is important for normal mammary ductal development in vivo and has also been shown to be important for side branching leading to alveologenesis (11, 12). We have shown that HGF is produced by mammary stromal cells in vitro, in response to 17β-estradiol (E2) treatment (13) Thus, whereas E2 is not directly mitogenic in mammary cells in vitro, its mitogenic effects are likely mediated in part, indirectly through HGF. Whereas HGF is produced by stromal cells, it acts on mammary epithelial cells that express c-Met, the cognate receptor for HGF (11, 12, 14). Numerous in vitro studies have shown that HGF can induce proliferation and produce a tubulogenic response in a wide variety of epithelial cells and cell lines, including mammary epithelial cells, when the cells were cultured within a collagen gel matrix (11, 15, 16). We reported that organoids obtained from adult mice and containing both LECs and MECs proliferate and produce tubules in response to treatment with HGF. Neither treatment with the progestin, promogestone (R5020), nor R5020+E2 induce proliferation. When treated with R5020 alone, the organoids form cyst-like structures. Treatment with the combination of HGF+R5020 results in proliferation that was increased above treatment with HGF alone, and the organoids exhibited blunting of the tubulogenic response (17). However, in these studies it was not determined which cell types, LECs and/or MECs, proliferated in response to HGF or HGF+R5020. The purpose of the present study was to investigate the cell type-specific morphologic and proliferative responses of LECs and MECs to treatment with either HGF or R5020 by themselves or when combined (HGF+R5020).
In vivo in the adult virgin mammary gland, progesterone receptor (PR)-A is expressed in a subset of LECs but not in MECs; PRB expression is undetectable before pregnancy (14). PRA-containing cells rarely proliferate and the mitogenic effect of progesterone in the virgin gland is believed to be mediated by paracrine factors such as receptor activator of nuclear factor-B ligand (RANKL) (14, 18, 19). Thus, the role of RANKL in mediating increased proliferation in organoids treated with HGF+R5020 was also investigated. To identify the responses occurring in LECs vs. MECs, an immunocytochemical approach was used.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Preparation of primary mammary gland organoids
Mammary epithelial organoids were isolated using mechanical and enzymatic dissociation methods as previously described (17). Briefly, mouse mammary glands were excised, minced, and digested sequentially in collagenase (Worthington, Lakewood, NJ) and pronase (Calbiochem, San Diego, CA) solutions, followed by differential centrifugation and Percoll (Amersham Biosciences, Piscataway, NJ) gradient sedimentation to remove mammary stromal cells and obtain mammary epithelial organoids consisting of both LECs and MECs. At the time of plating, MECs represent 26 ± 4% of the total cells in organoids based on double-immunofluorescent labeling with anti-smooth muscle actin (SMA) and cytokeratin 18 (K18) antibodies to detect MECs and LECs, respectively. Cell viability was approximately 95% as determined by trypan blue exclusion. Total cell number was determined by counting nuclei in a hemocytometer after cell disruption with 2.1% citric acid containing 0.4% crystal violet (20).
To retain the three-dimensional (3-D) architecture of the organoids in vitro, they were cultured within a collagen I gel matrix and in serum-free media. Then 96-well culture dishes were coated with an underlay of 40 µl/well of neutralized rat tail collagen I (2 mg/ml; BD Biosciences, Bedford, MA) before plating organoids. Freshly isolated organoids were suspended in neutralized collagen I (2 mg/ml, 75 µl/well) at a density of 1 x 105 cells/well and allowed to set for 30 min at 37 C before the addition of media. All cultures were carried out in serum-free medium with or without growth factors or hormones [basal medium (BM): serum and phenol red free DMEM/F12, supplemented with 0.1 mM nonessential amino acids, 2 mM L-glutamine, 100 ng/ml human recombinant insulin, 1 mg/ml fatty acid-free BSA (fraction V), 100 µg/ml penicillin, and 50 µg/ml streptomycin]. Treatments with growth factors and hormones were added at the time of plating and included 50 ng/ml HGF (Calbiochem), 10 nM 17β-estradiol (Sigma, St. Louis, MO), 20 nM of the synthetic progestin, R5020 (PerkinElmer, Boston, MA) (13, 17), and 200 ng/ml RANKL (R&D systems; Minneapolis, MN) (21). Organoid cultures were maintained in 5% CO2 at 37 C for up to 3 d and culture media was replaced every 48 h. For RANKL inhibition experiments, rat antimouse RANKL-neutralizing monoclonal antibody (IgG2a clone IK22–5; BioLegend, San Diego, CA; 10 µg/ml final concentration) or control rat IgG2a (22) was added to the various treatment media at the time of plating and maintained in the media throughout the treatment period. For analysis of proliferation, 5-bromo-2'-deoxyuridine (BrdU; 10 mM; Sigma) was added for 18–20 h before termination of the cultures.
Antibody labeling of organoids in collagen gels
The method for antibody labeling of mammary organoids in collagen gels was adapted from O’Brien et al. (23). After 24, 48, or 72 h in culture, collagen gels containing mammary epithelial organoids were removed with forceps, placed in glass vials, and rinsed for 10 min in PBS (pH 7.2) containing 1 mM CaCl2 and 0.5 mM MgCl2 (PBS+). The gels were then fixed in 4% paraformaldehyde in PBS+ and permeabilized for 30 min in 0.025% saponin in PBS+ (P buffer). Gels were again rinsed in PBS+ for 10 min, followed by quenching for 10 min in PBS+ containing 75 mM NH4Cl and 20 mM glycine, and blocking for 10 min in a 0.025% saponin, 0.3% gelatin solution in PBS+ (B buffer). All rinsing and blocking steps were carried out at room temperature on a gently rocking platform. Gels were then incubated by gently rocking for 3 d at 4 C simultaneously with mouse monoclonal primary antibody against K18 (catalog no. ab668-100; Abcam, Cambridge, MA) diluted 1:200 and with rabbit polyclonal antibody against 
SMA (catalog no. ab5694; Abcam) diluted 1:400 in B buffer. Tissue sections from adult virgin mammary gland served as positive controls. For negative controls, samples were treated only with secondary antibody. Samples then were washed 4 x 15 min in P buffer and washed 1 x 15 min in B buffer. Samples were incubated overnight at 4 C with goat antimouse Alexa 488 and goat antirabbit Alexa 546 (Molecular Probes, Eugene, OR), each diluted 1:100 in B buffer. Samples then were washed 4 x 15 min in P buffer. Samples were postfixed in 4% paraformaldehyde in 0.1 M sodium cacodylate buffer (pH 7.2) and mounted on glass slides using fluorescence-mounting media. Images were captured using a Pascal laser scanning confocal microscope (Carl Zeiss Inc., Thornwood, NY). The images were captured in a 3-D Z plane series with a step size of 4 µm between planes. Individual planes were analyzed from the Z-stack.
Antibody labeling of organoid sections
To obtain mammary gland organoid sections for immunohistochemical analyses, organoid cultures were removed from 96-well plates and fixed for 1 h in 10% phosphate-buffered formalin [0.4% sodium phosphate monobasic and 0.65% sodium phosphate dibasic (anhydrous) in 10% formalin]. Fixed gels were then processed overnight through a series of dehydrating alcohols, followed by xylenes and a xylene/paraffin (1:1) mixture on a Tissuematon (Fisher Scientific, Vernon Hills, IL) before embedding in paraffin blocks for sectioning. Five-micrometer sections were mounted onto coverslips treated with 3-aminopropyl triethoxysilane. Whole mammary glands were fixed overnight in 10% phosphate-buffered formalin and processed in the same way to obtain sections for c-Met immunohistochemical analysis. Paraffin sections were first deparaffinized and rehydrated sequentially through xylenes and gradient alcohols (100–50% EtOH). Sections were then immersed in 10 mM sodium citrate solution (pH 6.0) and autoclaved at 15 lb/square in.2 for 20 min for antigen retrieval.
For double labeling of BrdU and SMA, sections were first incubated with goat antimouse IgG Fab fragments [Jackson Laboratories, West Grove, PA) (diluted 1:100 in PBS containing 1% BSA; Sigma; 1% PBSA) 60 min], rinsed with PBS, and then blocked with normal goat serum (Vector Laboratories, Burlingame, CA; diluted 1:1 in PBS, 30 min). BrdU immunoreactivity was detected using a mouse monoclonal anti-BrdU/nuclease solution (catalog no. RPN202; Amersham Biosciences, Buckinghamshire, UK; undiluted from kit, 1 h at room temperature), followed by incubation with goat antimouse Alexa 546 (Molecular Probes; diluted 1:400 in PBS, 30 min). Sections were then blocked again in a 2% PBSA solution and incubated with rabbit polyclonal antibody against SMA, (catalog no. ab5694; Abcam; diluted 1:400 in PBS, 60 min) followed by goat antirabbit Alexa 488 (Molecular Probes; diluted 1:200 in PBS, 30 min). Cell nuclei were counterstained using either 4',6'-diamino-2-phenylindole or ToPro 3-iodide (Molecular Probes), and coverslips were mounted using fluorescence-mounting media. Tissue sections from adult virgin mammary gland were used as a positive control. BrdU-positive LECs or MECs are expressed as a percentage of total LECs or MECs counted, respectively.
For cell type-specific detection of PR isoforms, the protocol was followed as above; however, either mouse monoclonal hPRa7 (PRA specific) or hPRa6 (PRB specific) (24) (Neomarkers, Fremont, CA; diluted 1:50 in PBS/0.5% Triton X-100, overnight at 4 C) was used in the place of the BrdU antibody. Tissue sections from adult virgin mice and midpregnant mice were used as positive controls for PRA- and PRB-specific staining, respectively (14).
Estrogen receptor (ER)- immunoreactivity was detected using mouse monoclonal anti-ER, NCL-L-ER-6F11 (Novocastra, Newcastle, UK; diluted 1:10 in PBS, overnight at 4 C) using identical blocking steps and secondary antibody conditions as previously described for BrdU, hPRa6, and hPRa7. Sections were then blocked with 2% PBSA and incubated with rabbit antihuman PR antibody (catalog no. A0098; Dako, Carpinteria, CA; diluted 1:200 in 2% PBSA, overnight at 4 C), followed by incubation with secondary antibody and nuclear counterstain, as previously described for rabbit polyclonal antibodies. Tissue sections from adult virgin mammary gland were used as a positive control.
For RANKL, receptor activator of nuclear factor-B (RANK), and PRA staining, nonspecific staining was blocked using normal rabbit serum (Vector Laboratories; diluted 1:1 in PBS, 30 min) and then incubated with goat antimouse RANKL antibody (R&D Systems) or rabbit antihuman RANK (Santa Cruz Biotechnology, Santa Cruz, CA; diluted 1:100 in PBS, overnight at 4 C), followed by incubation with rabbit antigoat Alexa 488 (Molecular Probes; diluted 1:100, 30 min) or secondary antibody and nuclear counterstain as previously described for rabbit polyclonal antibodies. Positive controls for RANKL and RANK were tissue sections from midpregnant mammary gland (25). Negative control for RANK was lack of stromal cell staining in tissue sections from virgin mammary gland. In the case of double staining with anti-RANKL and anti-PRA antibodies, sections were then blocked following the protocols above for mouse monoclonal antibodies and incubated with primary antibodies against PRA (hPRA7) and the appropriate secondary antibodies as described above.
c-Met and SMA double labeling of cultured organoids and mammary gland tissue sections was performed using the rabbit antimouse c-Met primary antibody SP260 (Santa Cruz Biotechnology) (diluted 1:50, 1 h at room temperature). Blocking steps, secondary antibodies, SMA staining, and nuclear counterstaining steps were identical with rabbit/mouse double labeling protocols as described above. For all antibody staining, negative controls received no primary antibody followed by the relevant fluorochrome-conjugated secondary antibody.
To accommodate multiple protocols, in some experiments the fluorochromes Alexa 488 and Alexa 546 were switched; however, identical staining patterns with all antibodies were observed, regardless of which fluorochrome was used.
Image analysis and stain intensity quantification
Immunofluorescent images were captured using either a Pascal laser-scanning confocal microscope (Carl Zeiss) or a Nikon inverted epifluorescence microscope (Mager Scientific, Dexter, MI) with MetaMorph software (Molecular Devices Corp., Downington, PA). Confocal images were viewed and analyzed using the Zeiss LSM image browser program and National Institutes of Health Image J (Bethesda, MD). Epifluorescent images were viewed and analyzed using MetaMorph software. In both cases, to analyze fluorescence intensity, the average pixel intensity of all positively stained nuclei was determined. Images were thresholded to exclude background fluorescence and gated to include intensity measurements only from positively staining epithelial cells. In the case of cell type-specific c-Met intensity quantification, cells that were shown to be expressing SMA were analyzed separately from cells not expressing SMA.
RT-PCR analysis
Cultured whole mammary organoids, containing both MECs and LECs, were harvested at 24 h after treatment with BM, HGF, R5020, or HGF+R5020 as described above. Total RNA was extracted from organoids using TRIzol (Invitrogen, Carlsbad, CA) following the manufacturer’s suggested protocol. All extracted RNA samples were stored at –80 C until analyzed. RNA concentration was determined based on the ratio of UV absorbance read at 260 and 280 nm. cDNA was produced by reverse transcription with random hexamer primers using the Superscript III first-strand synthesis system for RT-PCR (Invitrogen) following the manufacturer’s instructions. Then 4 µl of cDNA from each treatment group were added to primers (Applied Biosystems, Foster City, CA) for either murine RANKL (assay no. Mm00441980_m1) or 18S rRNA (assay no. Hs99999901_s1). Real-time PCR for each sample was performed in triplicate using a Prism 7500 sequence detection system (Applied Biosystems). Cycling conditions were as follows: 10 min at 95 C for initial denaturation and enzyme inactivation and then 40 cycles of 15 sec at 95 C and 1 min at 60 C.
For quantification, a standard curve was first generated by making serial 10-fold dilutions of total cDNA. At each dilution, an amplification curve was generated for RANKL and 18S RNA to obtain the cycle threshold (CT). The slope of the 
CT,RANKL:18S vs. copy number line, where CT,RANKL:18S = CT,RANKL – CT,18S was less than 0.1, indicating that the amplification efficiency was approximately equal for RANKL and control 18S RNA amplification. Because amplification efficiency was the same for both RANKL and 18S RNA, we used the comparative CT method to calculate the fold change in RANKL expression. Briefly, the fold change in RANKL abundance in samples stimulated with 50 ng/ml HGF, 6.525 ng/ml R5020, or both 50 ng/ml HGF and 6.525 ng/ml R5020 over the abundance of RANKL expression in unstimulated samples (BM) was calculated by 2–CT,Stim:BM, where CT,Stim:BM = (CT,RANKL Stim – CT,18S Stim) – (CT,RANKL BM – CT,18S BM).
Statistical analysis
Quantitation of the percentage of labeled cells or immunofluorescence intensity was determined from a minimum of 1000 LECs or 500 MECs for each treatment from three to five separate culture experiments. In each experiment, each treatment was performed in triplicate. Results are expressed as mean ± SEM, and differences are considered significant at P < 0.05 by using Student’s t test or ANOVA where appropriate. Data were analyzed using SAS statistical analysis software (SAS Inc., Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A time course of labeling at 24, 48, and 72 h was carried out to examine morphological responses of LECs and MECs to the various treatments. In control BM-treated cultures, the organoids remained rounded in shape throughout the 72-h culture period (Fig. 1A
, BM). Centrally localized LECs (Fig. 1A, green) were surrounded basally by MECs (Fig. 1A, red). At 24 h, organoids treated with R5020 formed multicellular cyst-like structures comprised of an inner layer of LECs that were encircled by elongated MECs. The cysts contained a well-defined lumen (Fig. 1A, R5020). These cyst-like structures were maintained throughout the 72-h culture period.
|
). At 24 h after HGF treatment LECs had formed extensions leading away from the organoid body (Fig. 1B, green arrow). Many MECs were rounded in shape and remained with the organoid body (Fig. 1B, red arrow). By 48 h the extensions of LECs had formed chains consisting of multiple cells arranged in a single layer of cells (Fig. 1B, chain). The chains also contained MECs. Some of the chains progressed to form cords, comprised of a bilayer of LECs, surrounded by elongated MECs (Fig. 1B, cord). LECs (Fig. 1B, green) formed the leading edge of extensions, chains, and cords, with MECs (Fig. 1B, red) located posterior and external to LECs. By 72 h, numerous fully formed tubules with lumens were present; elongated MECs were located exterior to the cuboidal LECs forming the tubules (Fig. 1B, tubule).
By 72 h treatment with HGF, very few MECs (Fig. 1B, red) were seen at the organoid body (Fig 1C, HGF). Organoids treated with HGF+R5020 produced the same general morphologic changes and exhibited the same cellular organization as organoids treated with HGF alone, with the exception that the tubules that formed were shorter, as previously reported (Fig. 1C, HGF+R5020) (17). In addition, there was a greater number of MECs (Fig. 1C, red) with a rounded morphology localized around the body of the organoid.
Analysis of proliferation in LECs and MECs
We had previously analyzed proliferation in cultured organoids by assaying 3H-thymidine incorporation into DNA (17). By that method we determined that BM-treated organoids exhibited only a low level of proliferation, and treatment with R5020 alone did not increase proliferation above that of BM. HGF induced significant proliferation and maximal proliferation was observed after treatment with HGF+R5020. Measurement of proliferation by that method did not permit analysis of cell-type-specific (LECs, MECs) proliferation. To address this question and determine the contribution of cell type-specific proliferation to the different morphological responses described above, organoids were treated with BrdU for 18 h before fixation at 72 h for all treatments. Paraffin sections of organoid cultures were dual labeled with antibody to SMA and BrdU, and BrdU-labeled cells were quantitated (Fig. 2). In BM-treated cultures, a low percentage of LECs (7.2 ± 0.5%) and MECs (2.4 ± 0.5%) were BrdU positive (BrdU+). Treatment with HGF alone produced significant 4-fold increases in BrdU+ LECs and BrdU+ MECs (P < 0.05). Treatment with R5020 did not cause a significant increase in proliferation of LECs or MECs above BM treatment. Treatment with HGF+R5020 did not increase the percentage of BrdU+ LECs, compared with treatment with HGF alone. However, treatment with HGF+R5020 produced a significant 3.5-fold increase in BrdU+ MECs over that observed with treatment with HGF alone (P < 0.05). These results demonstrate that HGF induces proliferation in both LECs and MECs. Furthermore, whereas R5020 alone did not cause significant proliferation of either LECs or MECs, when combined with HGF, proliferation was notably increased, specifically in MECs.
|
). Fluorescence intensity of c-Met staining was analyzed separately for LECs and MECs in all treatment groups at 24, 48, and 72 h of culture. c-Met expression was consistently higher in MECs than LECs. Neither length of time in culture nor the various treatments had a significant effect on the level c-Met expression in either LECs or MECs. Representative levels of c-Met expression as determined by fluorescence intensity are shown for the 72-h time point (Fig. 3B).
|
). In the adult mammary gland c-Met was expressed at a higher level in MECs than LECs. Fluorescence intensities of c-Met in LECs and MECs in the adult mammary gland were similar to that observed in vitro in cultured mammary organoids. Analysis of c-Met expression at different stages of mammary gland development revealed that the levels of c-Met expression in LECs were very similar across development and did not appear to be developmentally regulated. However, c-Met levels in MECs were significantly reduced in ducts during pregnancy and lactation (Fig. 3D).
PR isoform expression in mammary organoids
PR isoforms, PRA and PRB, are temporally and spatially separated at different stages of mouse mammary gland development (24). PRA is the major isoform expressed in the adult, virgin mammary gland in vivo; PRB is undetectable before pregnancy. PRA and PRB expression is restricted to LECs (24). To investigate the role of progestin and PR in the morphological and proliferative responses observed, we examined PR isoform expression in cultured mammary organoids. Immunofluorescent analyses of PRA and PRB expression were carried out using antibodies that detect only PRA or only PRB (Fig. 4). Organoids derived from adult, virgin mammary gland expressed only PRA, and only LECs were PRA positive (PRA+) (Fig. 4A). In vivo, ER colocalizes with PRA in LECs in intact mammary gland (18). ER was also expressed in the organoids and was highly colocalized with PRA (Fig. 4B). Treatment of organoids with E2 did not increase the number of PRA+ cells or intensity of PRA staining (data not shown). No PRB positive cells were detected (data not shown). Time-course analysis revealed that there were no significant differences in the percentage of PRA+ cells under all treatment conditions and that the percentage of PRA+ cells was similar to the percentage present in organoids at the time of plating (time 0 = 30 ± 3%) (Fig. 4C). We also analyzed the effect of the various treatments on PRA expression levels (Fig. 4D). Analysis of immunofluorescence intensity of anti-PR antibody staining revealed that treatment with R5020 or HGF+R5020 significantly reduced overall PRA expression.
|
). RANKL mRNA was increased 1000-fold in R5020- and HGF+R5020-treated organoids, compared with treatment with BM or HGF alone. Protein expression was analyzed by immunofluorescent staining with anti-RANKL antibody (Fig. 5B). Intense RANKL staining was seen in LECs in organoids treated with R5020 or HGF+R5020 but not in organoids treated with BM or HGF alone. Additionally, RANKL was expressed specifically in PRA+ cells. We also examined the expression of RANK, the receptor for RANKL (Fig. 5C). Similar expression levels of RANK were detected in all treatment groups.
|
). Treatment with RANKL+HGF produced the same degree of MEC proliferation observed in organoids treated with HGF+R5020. MEC proliferation was significantly reduced by 70% (P < 0.05) in organoids treated with HGF+R5020+anti-RANKL-neutralizing antibody. RANKL alone or anti-RANKL antibody did not significantly affect proliferation of either LECs or MECs under any other treatment. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mechanism of tubule formation by HGF
We found that under all treatment conditions, LECs were centrally located and MECs were located external to LECs. This in vitro organization of the two cell types is similar to that observed in the intact mammary gland in vivo. Time-course analysis of HGF-induced tubule formation revealed that the initial event was the formation of LEC cellular extensions followed by the formation of chains of LECs, which then developed into cords composed of two layers of LECs. The final stage was the formation of a lumen within the cord of LECs. MECs were observed to migrate behind the leading edge of LEC during all stages of tubule formation.
The observation that LECs produced a morphologic response leading to tubule formation is in contrast to a previously published report of the behavior of primary cultures of purified primary human mammary LECs (4). In that study only purified MECs, and not purified LECs, exhibited a morphologic/motogenic response to HGF. However, in other studies of tubulogenesis using mouse mammary epithelial lines devoid of MECs, such as NMuMg and EpH4, tubulogenic responses to HGF have been reported (15, 29). The reasons for the differences reported for human and mouse LEC response to HGF could be due to differences between the responses of primary cells and cell lines or due to species differences in cell-type-specific responses in the human vs. the mouse.
The process of tubule formation observed in HGF+R5020-treated organoids followed the same basic process observed in HGF-treated organoids; however, the tubules were shorter in length and frequently larger in diameter as previously reported (17), suggesting that combined treatment with HGF+R5020 had an inhibitory effect on tubule elongation. The mechanism(s) mediating the R5020-induced inhibition of tubule elongation is currently under investigation. Additionally, there was a higher concentration of MECs around the organoid body, and the MECs were rounded rather than elongated in appearance. The increased number of MECs under this treatment condition is likely due to the increased proliferation of MECs.
Studies of in vitro tubule formation have identified five general mechanisms to date: 1) wrapping (neural tube), 2) cavitation (salivary gland), 3) cell hollowing, 4) budding (ureteric bud) and 5) cord hollowing (30). The sequence of events observed herein during HGF-induced tubule formation in mammary organoids most closely resembles the process described by cord hollowing. Cord hollowing is seen when a linear array of cells, a chain of cells, polarizes to delineate an apical membrane and a central lumen. Although tubule formation in mouse mammary organoids composed of LECs and MECs in 3-D collagen gels have previously been investigated (31), the behavior of MECs during tubule formation has not been previously described. To the best of our knowledge, this is the first report of an analysis of LECs and MECs proliferative and morphological responses in cultured mouse mammary organoids that contain both LEC and MEC.
Proliferation of LECs and MECs
Treatment with HGF or HGF+R5020 produced the same amount of proliferation of LECs and was due to the action of HGF because treatment with R5020 alone did not induce proliferation in LECs. In contrast, treatment with HGF vs. HGF+R5020 significantly increased MEC proliferation. These results reveal a novel role for progestin in promoting MEC proliferation. Additionally, MECs had a more rounded morphology and were concentrated around the cell body in organoids treated with HGF+R5020 indicating that R5020 also had an affect on MEC morphology in addition to increasing MEC proliferation.
It was not surprising that HGF stimulated proliferation in both LECs and MECs because both cell types express c-Met, the receptor for HGF in intact mammary gland and in our cultured mammary organoids. The increased proliferation observed with R5020+HGF was not due to an increase in c-Met levels in either LECs or MECs because c-Met levels were not changed by any of the treatment conditions and, more specifically, were not increased by the addition of R5020.
We also observed that in the intact mammary gland, c-Met expression in LECs was relatively constant from puberty through pregnancy and lactation, suggesting that c-Met expression is not developmentally regulated in LECs. In contrast, we observed that during pregnancy and lactation c-Met expression in ductal MECs was significantly decreased, compared with alveolar MECs. We speculate that this may be a mechanism to promote HGF-induced proliferation of alveolar MECs concomitant with the specific expansion of alveoli number and size during pregnancy and lactation, respectively.
PR and MEC proliferation
To determine the role of PR expression in MEC proliferation, we analyzed PR isoform expression. We found that PRA was the only isoform detected and localization of PRA was confined to LECs under all treatment conditions. ER was also detected in cultured mammary organoids and colocalized with PRA. Whereas the percentage of PRA-positive cells was not increased or decreased by any treatment condition, treatment with R5020 or HGF+R5020 significantly reduced the overall intensity of PRA immunofluorescent antibody staining, indicative of a decrease in PRA level. The percentage of PRA-positive LECs, the level of PRA expression, colocalization of PRA with ER, lack or PRB expression, and progestin-induced decrease in PRA levels in organoids were similar to that reported for the adult intact mammary gland from which the organoids were derived (18). Taken together, these results indicate that our mammary organoid culture model provides a relevant in vitro system for the study of progesterone action in normal mammary cells.
Interestingly, treatment of organoids with E2 did not increase the number of PRA+ cells or intensity of PRA staining. This was not surprising because we previously reported that E2 does not increase PR levels in primary monolayer culture of mouse mammary epithelial cells. E2 up-regulation of PR requires coculture with mammary fibroblasts (32). Because our organoid cultures were devoid of mammary stromal cells, the lack of E2 regulation of PR herein agrees with our previous findings of a requirement of mammary stroma for E2-induced PR up-regulation. The intact mammary gland is known to contain ER+ stromal cells, and it is plausible that estrogen up-regulation of PR in vivo occurs through an indirect mechanism involving estrogen action in ER+ stromal cells (33). In this regard, preliminary studies of mouse PR promoter regulation also suggest that E2 has no significant positive effect on promoter activation (Haslam, S. Z., unpublished observations).
RANKL and MEC proliferation
Previous observations of the lack of colocalization of PR with proliferation markers in intact rodent and human mammary glands has formed the basis for the concept that progestins can act on PR+ cells to produce a paracrine factor(s) that acts on PR– cells to promote their proliferation (34, 35). Because MECs and a significant percentage of LECs do not contain PR, we considered the possibility that R5020 induced a factor in PR+ LECs that acted as a paracrine factor to promote proliferation in PR-negative (PR–) MECs. Embedded in this hypothesis is the concept that this paracrine factor was not mitogenic by itself in either PR– MECs or PR– LECs because treatment with R5020 alone did not increase proliferation of LECs or MECs.
Based on previous reports, we considered RANKL a possible candidate for the putative progestin-induced paracrine factor (19). Both RANKL mRNA and protein were highly up-regulated in R5020- and HGF+R5020-treated organoids. Because RANK expression was similar in all treatment groups, it was most likely that increased RANKL expression was responsible for increased MEC proliferation in response to HGF+R5020 treatment. The addition of exogenous RANKL+HGF stimulated MEC proliferation to the same degree observed after treatment with R5020+HGF, and cotreatment with HGF+R5020 plus a neutralizing antibody to RANKL produced a 70% inhibition of MEC proliferation. Thus, we concluded that progestin-induced RANKL synergizes with HGF to increase MEC proliferation and that activation of both RANKL/RANK and HGF/c-Met signaling pathways are required for the synergistic promotion of MEC proliferation.
Binding of RANKL to its receptor RANK initiates a signaling cascade that activates nuclear factor-B, resulting in p65 phosphorylation and nuclear translocation (21). However, in vivo administration of exogenous RANKL to virgin mice or RANKL treatment of primary epithelial cell cultures derived from virgin mammary gland fails to cause P65 nuclear translocation, indicating a lack of epithelial cell response to RANKL (21). The same study showed that RANK, the receptor for RANKL, is expressed at low levels in basally located cells in ducts of the virgin gland. In that study it was not determined whether the cells expressing RANK were LECs or MECs. RANK and RANKL expression is significantly increased at d 14–16 of pregnancy and is restricted to lobuloalveolar structures and coincides with the timing of alveolar proliferation (21). The emerging picture is that RANK and RANKL expression is temporally and spatially regulated, being most highly expressed during pregnancy to promote alveolar proliferation and survival. In agreement with this interpretation is impaired alveologenesis in RANKL gene-deleted mice (25).
Although RANKL was highly up-regulated in LECs in R5020-treated organoids, R5020 did not cause proliferation of LECs. RANKL was also up-regulated in LECs in HGF+ R5020-treated organoids; however, proliferation of LECs was not increased above that of treatment with HGF alone. RANK, the receptor for RANKL was expressed in LECs under all treatment conditions and was not an apparent limiting factor for RANKL-induced LEC proliferation. There are several possibilities to explain the lack of an observed LEC proliferative response to RANKL in our studies. First, our organoids were derived from adult virgin mammary gland, and the alveologenic, proliferative response to RANKL occurs only in more differentiated LECs induced by pregnancy. In this regard, we previously reported that the B isoform of PR is undetectable in the adult virgin mouse mammary gland and is highly expressed only during pregnancy (14). It is possible that RANKL stimulates proliferation only in PRB-containing LECs. Another possibility is that an interaction between the RANKL signaling pathway and another, as-yet-undefined, ligand and its signaling pathway are required for a proliferative response in LECs, much the same way that RANKL and HGF coordinated signaling are required to promote MEC proliferation. In this regard, HGF is a stroma-derived growth factor that acts in conjunction with R5020 to promote MEC proliferation. It is possible that another stroma-derived factor is required to act in conjunction with R5020 to promote LEC proliferation. Alternatively, it is possible that RANKL is a paracrine factor that acts specifically on MECs to promote their proliferation. Currently the detailed mechanism(s) of RANKL action in the murine mammary gland have not been defined. Our results demonstrate a novel role for RANKL acting in conjunction with HGF to promote MEC proliferation.
In summary, we have defined the morphological and proliferative responses of LECs and MECs during the processes of tubule or cyst formation in response to HGF and/or R5020-treatments, respectively, in 3-D, primary organoid culture system. The tubulogenesis response is initiated and carried out in a stepwise progression by LECs. MECs do not appear to have an active role in this process. HGF by itself mainly causes proliferation of LECs and to a lesser extent of MECs. However, HGF+R5020 produce a synergistic increase in proliferation of MECs. This response appears to be mediated by RANKL, a progestin-induced paracrine factor produced in LECs that interacts with HGF to specifically increase proliferation of MECs.
In addition to their roles in normal mammary gland function MECs also appear to play an important role in the development and progression of mammary cancers. In this regard, it has been proposed the MEC may have tumor suppressor properties (3, 36, 37, 38). Thus, understanding how coordinated regulation of LECs and MECs occurs during normal mammary gland development and function may provide information relevant to understanding the basis of the aberrant organization observed in mammary cancers.
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online January 24, 2008
Abbreviations: BrdU, 5-Bromo-2'-deoxyuridine; CT, cycle threshold; 3-D, three-dimensional; E2, 17β-estradiol; ER, estrogen receptor; HGF, hepatocyte growth factor; K18, cytokeratin 18; LEC, luminal epithelial cell; MEC, myoepithelial cell; PBSA, PBS containing BSA; PR, progesterone receptor; R5020, progestin; RANK, receptor activator of nuclear factor-B; RANKL, RANK ligand; SMA, smooth muscle actin.
Received October 12, 2007.
Accepted for publication January 11, 2008.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
provides an essential link between RANK signaling and cyclin D1 expression during mammary gland development. Cell 107:763–775[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  M. D. Aupperlee, A. A. Drolet, S. Durairaj, W. Wang, R. C. Schwartz, and S. Z. Haslam Strain-Specific Differences in the Mechanisms of Progesterone Regulation of Murine Mammary Gland Development Endocrinology, March 1, 2009; 150(3): 1485 - 1494. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
