SEARCH

SEARCH BY CITATION

Keywords:

  • Mammary gland;
  • Differentiation;
  • Stem cell;
  • Progenitor cell;
  • Hormones;
  • Receptor activator of NF-κB ligand

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

In mice, CD49fhi mammary stem cells (MaSCs) asymmetrically divide to generate CD49f+ committed progenitor cells that differentiate into CD49f phenotypes of the milk-secreting tissue at the onset of pregnancy. We show CD49f+ primary mammary epithelial cells (PMECs) isolated from lactating tissue uniquely respond to pregnancy-associated hormones (PAH) compared with CD49f+ cells from nonlactating tissue. Differentiation of CD49f+ PMEC in extracellular matrix produces CD49f luminal cells to form differentiated alveoli. The PAH prolactin and placental lactogen specifically stimulate division of CD49f luminal cells, while receptor activator of nuclear factor (NF)-κB ligand (RANKL) specifically stimulates division of basal CD49f+ cells. In nondifferentiating conditions, we observed a greater proportion of multipotent self-renewing cells, and RANKL treatment activated the RANK pathway in these cultures. Furthermore, we observed the deposition of calcium nodules in a proportion of these cells. These data imply that a MaSC unique to the lactating breast exists in humans, which generates progeny with discrete lineages and distinct response to PAH. STEM CELLS2012;30:1255–1264


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Expansion of the glandular epithelium during pregnancy is associated with upregulation of pregnancy-associated hormones (PAH) and activation of their signaling pathways in mammary epithelial cells [1–3]. During this time, CD49fhi mammary stem cells (MaSCs) undergo division producing CD49f+ and CD49f committed progenitor cells (CPCs) that lead to CD49f progeny that generates the lactating tissue phenotypes [4, 5].

Although the receptor activator of NF-κB (RANK) is best known for its role in osteoclast lineages in bone, it is also expressed in mammary epithelial cells as well as other epithelial lineages including prostate and lung [6]. Binding of RANK ligand (RANKL) to RANK activates the mitogen-activated extracellular-signal related kinase (MEK)/extracellular signal related kinase (ERK) pathway and promotes translocation of NF-κB to the nucleus where it regulates gene transcription [7, 8]. In the mammary gland, activation of RANK appears to be essential for alveolar differentiation during pregnancy, and it is proposed to do this by mediating ovarian hormone signaling resulting in cell division of the RANK expressing basal MaSC population [9–11]. Maintenance of tissue homeostasis during lactation requires paracrine crosstalk between the MaSC and CPCs to regulate tissue turnover [12]. In mice, it is believed that this is at least partly regulated by estrogenic signals originating in the estrogen receptor (ER) CPCs that feed back to promote proliferation in the ER+ MaSCs, potentially via RANKL/RANK activation [9]. This feedback mechanism is also demonstrated by experiments showing that RANKL is expressed in RANK CPCs that feed back to promote cell division in the RANK+ MaSCs [9, 13].

Proliferation of CPCs is shown to be driven by prolactin (Prl) through the Prl receptor (Prl-R) that acts via the signal transducer and activator of transcription (STAT)5 pathway to drive alveolargenesis during pregnancy and maintain tissue homeostasis during lactation [14, 15]. Placental lactogen (PL) also binds the Prl-R to promote proliferation of CPCs and is also required for alveolargenesis. Prl signaling in the mammary epithelium is suppressed by competitive inhibition of placental estrogen and progesterone that are upregulated during pregnancy [16]. Prl and PL activate STAT5, via Jak2 and ErbB4, and this pathway is known to be active in late pregnancy and is associated with the period of luminal cell differentiation [17]. Competitive suppression of Prl by estrogen and progesterone occurs during pregnancy and secretory differentiation, and it is alleviated at parturition when circulating placental hormone concentrations rapidly decrease [18]. At this stage, Prl signaling adopts an additional role where it is associated both with activation of milk synthesis in the luminal cells as well as promoting division of CPCs [3, 19].

The pathways that allow CPCs, but not MaSCs, to respond to pregnancy-induced proliferation and differentiation cues are not well known in humans. Likewise, the crosstalk that occurs between CPCs and their differentiated progenies to regulate tissue turnover and milk synthesis during lactation is poorly understood. While animal models, and in particular mice, have been useful in elucidating the endocrine regulation of mammary epithelial cells, there are significant differences that prevent direct extrapolation to humans. Even human explants in immune-compromised mice are unavoidably influenced by the innate mouse endocrine hormones that are different from those in humans [20–22].

To investigate how human cells are regulated, we have been able to use a primary mammary epithelial cell (PMEC) culture derived from normal human breast cells obtained from milk [23, 24]. We isolated the basal stem cell marker CD49f expressing cells from this culture and assessed their proliferation and differentiation capacity in adherent and extracellular matrix (ECM) culture in the presence of PAH Prl, PL, and RANKL. The response of PMEC was compared to CD49f+ primary human mammary epithelial cells (HMEC) from nonlactating tissue, the nontumorigenic cell line MCF10A that is routinely used as the model of “normal” multipotent mammary cells [25–28], and MDA-231, a highly invasive tumor line.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Ethical Approval

All experimental work for the research reported here was approved by the University of Western Australia Human Research Ethics Committee and the King Edward Memorial Hospital Research Ethics Committee.

Cell Culture

The PMEC population was generated as previously described from human breastmilk cells [23, 24]. Briefly, cells isolated from human breastmilk were cultivated in HuMEC Ready Media (Life Technologies, NY) for selective growth of PMEC. Adherent cells were subcultured to obtain a pure population of PMEC that were stained for lineage markers CD49f, CD10, CD29, CD24, p63, CK5/6, EpCAM, Muc1, CK14, and CK18 to validate phenotypes. Primary HMECs were obtained from Life Technologies. Cell cultures were grown in HuMEC Ready Media (PMEC and MCF10A), Roswell Park Memorial Institute (RPMI) plus 10% fetal bovine serum (FBS) (Life Technologies; MDA-231), or non-hematopoietic (NH) media (Miltenyi Biotech, NSW, Australia) and maintained in a humidified incubator (37°C, 5% CO2). Three-dimensional alveolar cultures (which we also refer to as “ECM Cultures”) were generated by subcultivation of monolayer cells into Matrigel (Life Technologies, NY) by the “on-top” method [29]. Fresh media changes were carried out every 4 days with 10% Matrigel in HuMEC. For stimulation experiments, cells were cultivated in media containing 1 μg/ml purified Prl from human pituitary (Mybiosource, CA), 2.5 μg/ml purified PL from human placenta (Mybiosource), or 100 ng/ml recombinant glutathione s-transferase (GST)-RANKL [30].

Alveoli Extraction and Fixation

Cell structures were isolated from ECM culture by addition of 50 μg/ml dispase in phosphate buffered saline (PBS) in a volume equivalent to the culture media in the well. Cultures were incubated at 37°C for 30 minutes with gentle agitation every 5 minutes. When cell structures were clearly dissociated from ECM, the suspension was transferred to a 15 ml polypropylene centrifuge tube with 2× volumes of 10 mM EDTA in PBS. Suspensions were mixed by inversion for 1 minute then centrifuged at 300g for 10 minutes. The supernatant was removed and the cell pellet was resuspended in 1 ml of 10 mM EDTA, transferred to a Microfuge (Beckman Coulter, NSW, Australia) tube and centrifuged at 500g for 5 minutes to wash. Cells were fixed in 1 ml of 10% formaldehyde (Sigma, NSW, Australia) in PBS for 10 minutes at room temperature. Fixed cells were isolated by centrifugation at 500g for 5 minutes and washed once in PBS before antibody staining.

Immunofluorescent Antibody Staining

Fixed cells were stained in solution in 1.5 ml microcentrifuge tubes, and incubations were carried out on a rotating rocker at 4°C in the dark. Antibody dilutions and washes were carried out in staining solution consisting of 2% normal goat serum (NGS; Life Technologies) in 0.1% Triton X (Sigma) in PBS. Cells were initially blocked and permeabilized in 10% NGS in 0.1% Triton X in PBS for 30 minutes and then incubated in 500 μl of primary antibody diluted 1:250 overnight. Cells were washed twice for 20 minutes and then incubated in 500 μl of secondary antibody diluted 1:500 plus 2 mM Hoechst 33342 (Life Technologies) for 6 hours. Finally, cells were washed three times in 0.1% Triton X in PBS for 20 minutes and mounted in ProLong Gold Antifade Reagent (Life Technologies) on glass microscope slides and sealed with number 1.5 glass coverslips (ProSciTech, Qld, Australia). Monoclonal anti-human primary antibodies CD49f, CK18, CD10, CD29, CD24, p63, CK5/6, EpCAM, CK14, and Prl-R were purchased from AbCam, Muc1 was purchased from BD Bioscience (NSW, Australia), RANK was purchased from R&D Systems (MN), and pERK was purchased from Santa Cruz Biotechnology (CA). Fluorescent secondary antibodies anti-rat AF647, anti-mouse AF488, and anti-rabbit AF546 were purchased from Molecular Probes.

Fluorescent Microscopy and Cell Analysis

Stained cells were visualized using a Nikon A1Si confocal microscope and data were collected with the Nikon NIS Elements software package. Raw images were exported to 8-bit TIFF files and analyzed using ImageJ software. Marker-positive cells were validated using light intensity cutoffs relative to positive and negative control cell samples. Automated positive-cell counts per optical confocal section were collected after adjusting the light intensity threshold for each wavelength empirically on a minimum of three positive and three negative control samples.

Cell Proliferation

CellScreen (Innovatis AG, Germany) real-time cell growth analysis was used to measure expansion of adherent cell populations by image capture followed by analysis of area occupied by cells. Control cell counts were carried out on representative fields of view to validate that the mean area of individual cells was consistent between measurements and that increase in area was directly proportional to cell number. CellTiter-Blue Cell Viability assay (Promega, NSW, Australia) and Cell Proliferation Reagent assay (Roche, Germany) were carried out according to the manufacturers' instructions. Briefly, CellTiter-Blue Reagent was added to live cultures at 0.2× media volume, incubated for 45 minutes, then fluorescence recorded at 560 nm(ex)/590 nm(em). WST-1 cell proliferation reagent was added to live cultures at 0.1× media volume, incubated for 1.5 hours, then absorbance recorded at 480 nm. All incubations were carried out at 37°C in 5% CO2. Fluorescence and absorbance values were measured on a Fluostar automated plate reader.

Von Kossa

Von Kossa staining for calcium deposition was carried out as previously described [31]. Cells were washed three times with PBS before fixation with 10% formalin for 1 hour at room temperature. Cells were washed five times with distilled water before adding 1 ml of 5% silver nitrate (Sigma) and exposing to UV light for 45 minutes. Cells were washed five times with distilled water. Sodium thiophosphate (5%) (Sigma) was added for 2 minutes followed by three washes in distilled water. Cells were counterstained with Nuclear fast red (Sigma) for 1 minute and then washed five times with distilled water. Calcium-phosphate deposits stain black, nuclei red, and other tissues pink. Images of the cells were pseudocolored digitally to generate a heat map of intensity of calcification staining.

Flow Cytometry

Changes in marker expression of early and late cultures were assessed by flow cytometry using a FACSCanto II flow cytometer (BD Biosciences, CA). The instrument is equipped with 405, 488, and 633 nm lasers and fluorescence was collected with 510/50BP (Live/Dead Fixable Aqua), 530/30BP fluorescein isothiocyanate (FITC), and 660/20BP allophycocyanin (APC). The antibody panel included directly conjugated antibodies, CD29 APC, CD49f FITC, CK18 APC (BD Biosciences), and Live/Dead Fixable Aqua (Invitrogen). Isotype controls were used to confirm the specificity of binding. Data were exported as flow cytometry standard (FCS) files and analyzed using FlowJo software (TreeStar, OR). Populations were defined with the following gating strategy; FSC-A versus SSC-A and FSC-A versus FSC-H were used to remove debris and cell doublets, and FSC versus Live/Dead Fixable Aqua to remove dead and autofluorescent cells. Bicolor dot plots (shown in Fig. 2) show expression characteristics of the antigens listed.

Statistical Analysis

Cell count data were expressed as the mean ± SD. Statistical comparisons were carried out by GraphPad Prism5 software (Graphpad Software, CA). The Student's t test was used for two-group comparisons. Two-way analysis of variance (ANOVA) using Bonferroni's model for multiple comparisons was used for comparison of multiple groups.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

CD49f and CK18 Are Expressed in Separate Populations and Coexpress Other Lineage Markers of Stem or Differentiated Phenotype, Respectively

CD49f and CK18 were clearly distinguishable as separate populations coexpressing mammary stem and differentiated markers, respectively, in all cell populations tested. To validate the phenotype of PMEC, coexpression of other lineage markers previously shown to be expressed in mammary epithelial stem and differentiated cells was measured and is summarized in 1 as a percentage (±SD) of CD49f+ or CK18+ cells. Coexpression counts are calculated from nine independent experiments. Each phenotype is described herein purely by expression of CD49f or CK18.

Table 1. Coexpression of phenotypic markers in PMEC
  1. * Values represent percentage mean ± SD of CD49f+ (top row) or CK18+ (bottom row) cells counted from 9 independent experiments.

inline image

Prl, PL, and RANKL Promote Proliferation of CD49f+ Cells from Lactating but Not Nonlactating Tissue in Monolayer Culture

Using real-time analysis of adherent culture growth, we show that in monolayer culture, PMEC number increased steadily before reaching a plateau at day 8 in the absence of PAH (Fig. 1A). In the presence of PAH, PMEC entered a second round of proliferation that continued to day 14. Entry into this second round of proliferation was at day 8 for PL and RANKL, and day 11 for Prl. HMEC, MCF10A cells, and MDA-231 did not respond to PAH, even at higher concentrations (not shown). We observed no synergistic or antagonistic effects when all or any pair of hormones were combined. Expansion of cell number was supported by quantitative cell proliferation assays WST-1 and CellTiter-blue at each day of growth for PMEC (Fig. 1B). We observed no major changes in the morphology or spreading pattern of the monolayer cells in treated cultures compared with controls (representative images treated with Prl for PMEC and HMEC, MCF10A, and MDA-231 controls in Fig. 1C).

thumbnail image

Figure 1. PMECs bypass cell-cycle arrest in response to pregnancy-associated hormone (PAH). (A): Percentage of growth area covered by PMEC, HMEC, MCF10A, and MDA-231 cells on days 1–14 of culture in the presence or absence of PAH. Graphs represent mean ± SD of 18 replicate cultures evenly distributed across three separate experiments. p = .001. (B): Cell proliferation analysis of PMEC by WST-1 (closed circles) on the left y-axis and CellTiter-Blue (open circles) on the right y-axis on days 1–12 of culture in the presence or absence of PAH. Graphs represent mean ± SD of 12 replicate cultures evenly distributed across three separate experiments. p = .001 (C): Bright-field images of PMEC, HMEC, MCF10A, and MDA-231 on days 1, 7, and 14 in monolayer culture plus representative images for PMEC in response to PAH (Prl). Scale = 20 μM. Abbreviations: HMEC, human mammary epithelial cells; PMECs, primary mammary epithelial cells; Prl, Prolactin; RANKL, receptor activator of NF-κB ligand.

Download figure to PowerPoint

PMEC Form a More Differentiated Alveolar Structure Compared with HMEC, MCF10A, and MDA-231 in ECM Culture

As we have previously reported [24], in ECM culture, single CD49f+ PMEC forms a differentiated alveolus-like structure with a basal layer of CD49f+CK18 cells representing 99% of the total cell population by day 4 of growth. The proportion of CD49f+CK18 cells regresses to 12.8% of the total population as a luminal layer of differentiating CD49fCK18+ cells form representing 87% of the total population by day 12 of growth (Fig. 2A). While we also observed small populations of CD49f+CK18+ and CD49fCK18 cells, there was no major change in their respective proportions between day 4 and day 12 of culture. HMEC and MCF10A cells also formed alveolus-like structures but failed to develop a CK18+ luminal layer and were on average smaller with less defined lumen, and MDA-231 cells invaded the ECM and formed disorganized stellate structures (Fig. 2B). This supports previous data by other researchers that show that MCF10A form spherical acinar structures with a basal (CD49f+) layer but not differentiated luminal cells and MDA-231 form stellate disorganized clumps in ECM culture [32, 33].

thumbnail image

Figure 2. Basal and luminal response to PAH. (A): Confocal cross-sections of primary mammary epithelial cell (PMEC) on day 4 and day 12 of growth with quantitative flow cytometry cell counts. (B): Confocal cross-sections of HMEC, MCF10A, and MDA-231 cells on day 12 of culture in extracellular matrix (ECM). (C): Single cell suspensions were prepared from dissociated alveolar cultures and subcultured into fresh ECM. Charts show the percentage of total cells capable of forming a new alveolus in the absence or presence of pregnancy-associated hormone (PAH). (D): Mean size of alveoli cultured in the absence or presence of PAH. (E): Confocal cross-sections of differentiated PMEC alveoli on day 12 of culture in the presence or absence of PAH with quantitative cell counts showing proportion of B (CD49f+) and L (CK18+) cells. (F): Bright-field images of morphology of alveolar structures after treatment with PAH. Note ductal-alveolar elongation in RANKL-treated cultures. Antibodies are CK18+ L (magenta) and CD49f+ B (green) cells with hoechst-labeled nuclei (blue). Scale = 100 μM and 20 μM (IF). Data are represented by the mean ± SD of eight replicate cultures evenly distributed across four separate experiments. p = .001. Abbreviations: Basal, B; HMEC, human mammary epithelial cells; L, luminal; PL, Placental Lactogen; Prl, Prolactin; RANKL, receptor activator of NF-κB ligand. * = 0.05; ** = 0.01; *** = .001.

Download figure to PowerPoint

In PMEC Cultures, RANKL Promotes Proliferation of Basal CD49f+ Cells and Inhibits Differentiation, While Prl and PL Promote Proliferation of Luminal Cells and Differentiation

When PMEC alveoli were dissociated and subcultured, the proportion of cells capable of producing serial alveolar cultures corresponded to around 45%–50% of the population and until passages 10-12 when self-renewal capacity began to decrease (Fig. 2C). Addition of either Prl or PL (which both act via the Prl-R) reduced average size of differentiated alveoli (Fig. 2D) but increased the proliferation rate of cells accompanied by earlier expansion of cell number, earlier formation of alveoli, and enrichment of the CK18+ phenotype (Fig. 2E). While treatment with RANKL did not affect mean differentiated alveolar size, we observed some changes in alveolar morphology including branched and ductal-like structures, an enrichment of CD49f+ cells, and a marked reduction in CK18+ cells (Fig. 2E). This was associated with an increase in the proportion of cells that were capable of forming a new alveolus in serial subcultures of dissociated alveoli (Fig. 2C). The relative size and morphology of alveolar structures were distinctly visible by bright-field microscopy (Fig. 2F).

Prl-R and RANK Are Expressed in Histologically Separate Subsets of Proliferating PMEC

We have recently shown that specific subsets of luminal cells in PMEC generated alveoli treated with Prl (but not PL) express hallmark milk proteins α-lactalbumin, β-casein, and human milk fat globule protein [24]. To investigate the dual role of Prl to also promote proliferation in the luminal compartment, we stained for the Prl-R in differentiating alveoli. Prl-R+ cells appeared on day 8 and the receptor was expressed in clustered subpopulations migrating away from the basement membrane to form the luminal layer. After an initial increase in the proportion of Prl-R+ cells from day 8 to 10, the population decreased to a static level that was maintained in the differentiated cultures (Fig. 3A). RANK expression was visible at day 4 of culture in the CD49f+ population and was maintained through to day 8 at the same level, and it appeared spread across the cell. By day 12, the total proportion of cells expressing RANK decreased and expression appeared restricted to the basolateral membrane, where it was coexpressed with CD49f in approximately 4% of the total cell population and this was maintained through to day 24 (Fig. 3B).

thumbnail image

Figure 3. Prl-R and RANK expression in differentiating primary mammary epithelial cell (PMEC) alveoli. (A): Confocal cross-sections of PMEC alveoli showing basal CD49f+ cells (green) and Prl-R+ cells (magenta) on days 4, 10, and 12 of culture with corresponding cell counts showing proportion of total cells that are Prl-R+. (B): Confocal cross-sections of PMEC alveoli showing basal CD49f+ cells (green) and RANK+ cells (magenta) on days 4, 10, and 12 of culture with corresponding cell counts showing proportion of total cells that are RANK+. Scale = 20 μM. Data represent the mean ± SD of 240 alveoli evenly distributed across four separate experiments. p = .001. Abbreviations: Prl-R, Prolactin receptor; RANK, receptor activator of NF-κB.

Download figure to PowerPoint

PMEC Distinctly Shows a Change in Differentiation Pattern and Response to RANKL in the Absence of Mammary-Specific Culture Media

To investigate the multilineage potential of PMEC, we grew monolayer cultures in nonlineage-specific (NH) media with and without RANKL. Compared with normal mammary (HuMEC) media (Fig. 1C), PMEC, but not HMEC, grown in NH media formed small, isolated colonies of transparent, spread cells (Fig. 4A). Compared with HuMEC cultures (Fig. 1A), PMEC grown in NH media was also less proliferative and failed to re-enter a second distinct proliferative phase in the presence of RANKL (Fig. 4B). HMEC also showed a slight reduction in proliferation compared with HuMEC cultures (Fig. 1A) but still showed no difference in growth in response to RANKL (Fig. 4B). We observed no changes in growth or proliferation in MCF10A or MDA-231 cells between NH and HuMEC media (data not shown for NH media, Fig. 1 for HuMEC data). Notably, PMEC from NH cultures without RANKL that were replated into ECM with HuMEC media showed a higher proportion of alveolus-forming cells, and this was abrogated by addition of RANKL to the NH culture (Fig. 4C). HMEC showed a distinct loss in alveolus-forming cells in consecutive passages, and this was not affected by addition of RANKL (Fig. 4C). This was mirrored in the presence of HuMEC media (not shown). In addition, NH cultures treated with RANKL developed clusters of cells by day 8 that stained strongly with silver nitrate, an indicator of calcium deposition (Fig. 4D). To support our assumption that RANKL treatment was indeed activating the RANK pathway, we used sensitive cell-based ELISA to quantify the relative expression of phosphorylated ERK (a marker for RANK activation [13]). We found a significant correlation between proliferation in RANKL-treated cultures and ERK activity (Fig. 4E).

thumbnail image

Figure 4. PMEC can differentiate outside MEC lineage. (A): Bright-field images of PMEC in monolayer culture in HuMEC (mammary) or NH (stem cell) media in the presence or absence of RANKL on days 1 and 14 of culture. Scale = 50 μM. (B): Percentage of growth area covered by PMEC and HMEC in NH media on days 1–14 of culture in the presence or absence of RANKL. Data represent the mean ± SD of nine replicate cultures evenly distributed across three separate experiments. (C): Monolayer cells isolated from sequential passages of PMEC and HMEC grown in NH media in the presence or absence of RANKL were subcultured into extracellular matrix. Charts show the percentage of total cells capable of forming a new alveolus. Data are represented by the mean ± SD of eight replicate cultures evenly distributed across four separate experiments. p = .001. (D): Bright-field image (left) and corresponding heat map of Von Kossa calcium stain (right) of RANKL-treated PMEC in monolayer culture with NH media on day 14. (E): Relative quantitation of phosphorylated ERK. Abbreviations: HMEC, human mammary epithelial cells; PMEC, primary mammary epithelial cell; RANKL, receptor activator of NF-κB ligand.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

We have previously demonstrated that PMECs isolated from lactating human mammary epithelium possess self-renewal capacity and are multipotent as indicated by the expression of lineage-specific markers [23, 24]. Based on this, we hypothesized that the PMEC would contain distinct populations capable of functioning in a manner similar to the murine CPC phenotype that arises by differentiation of the more primitive CD49f+ MaSC in pregnancy. In this study, we have specifically examined the response of PMEC to PAH, and in particular RANKL, in comparison to nonlactating tissue-derived HMEC and the cell lines MCF10A and MDA-231. These data have revealed that the PMECs, unlike cells from nonlactating tissue, are able to generate a range of cellular phenotypes that can be driven to proliferation or differentiation into cells capable of producing milk products.

Treatment with Prl or PL (that both bind Prl-R) promoted proliferation of only the CD49f luminal population, indicating that these hormones target these more differentiated cell types. This aligns well with our observation that the transition from the CD49f+ to CD49f phenotype coincides with expression of Prl-R and apparent orientation of these cells into a luminal position. It has been shown previously that CD10 is required in MaSCs to prevent proteolytic cleavage of CD29 and differentiation [34–36]. However, we show that CD49f+ cells are enriched for the expression of CD10 but not CD29. Whether or not this supports our evidence that the PMEC phenotype we describe is unique to lactating tissue is a topic for future investigation.

In contrast to PAH treatment, RANKL stimulated the cultures to maintain a greater proportion of CD49f+ cells, and these were demonstrated to have a higher self-renewal capacity in subculture experiments. This suggests that RANKL specifically targets the less differentiated phenotype and is supported by the demonstrated expression of RANK in the CD49f+ population early in the proliferative phase of alveolar growth compared with later cultures where alveoli reach growth stasis [10]. We thereby propose that the CD49f+ cells we observed correspond more closely to murine MaSC and that differentiation of these cells results in a phenotype that correspond more closely to murine CPC.

We further propose that these populations are differentially responsive to Prl/PL and RANKL. This is supported by our observation that there is a delay in Prl compared with RANKL stimulation of the second proliferative phase of adherent cultures. In these cultures, RANKL treatment increases proliferation of CD49f+ cells prior to the emergence of differentiated phenotypes, which respond to Prl. Although we have not provided direct experimental evidence here, this observation is consistent with probable asymmetric division of the MaSC to give rise to CPC. Since it is clear that these molecules activate different cellular pathways, this effect may in part be mediated by activation of gene transcription that promotes either self-renewal or differentiation. Indeed, Prl and PL are upregulated during the later stages of pregnancy when luminal differentiation is taking place in the alveolar lobules, supporting the hypothesis that these molecules specifically target and drive differentiation.

We have also previously shown that Prl treatment promotes expression of hallmark milk proteins in almost all luminal cells of the PMEC but not in any of the other mammary epithelial cell cultures [24]. However, expression of Prl-R was seen only in a small subset of luminal cells, and this implies that activation of milk protein synthesis may operate either via a paracrine signaling mechanism or that Prl binds an alternative receptor in these cells to activate milk synthesis. This would in part explain the dual role of Prl in both promoting proliferation and activating milk synthesis during the process of differentiation and into lactation. A better understanding of the interaction between Prl and PL during the phase of Prl-R-mediated proliferation may help to explain this and studying the Prl-R negative but Prl-responsive cells present in PMEC will enable this.

It has been proposed by Asselin-Labat et al. that alveolar development during pregnancy occurs by proliferation and differentiation of a CPC generated from the multipotent MaSC [9, 37, 38]. We propose that committed progenitors that failed to lose expression of RANK as they differentiate and acquire receptors for proproliferative signals (Prl/PL) may be similar to the cancer stem cell proposed in previous research [9, 39, 40] (Fig. 5). While it is shown that MaSC and CPC subtypes appear to underlie distinct subtypes of breast cancer [10, 41], it is clear that the mammary microenvironment plays an important role in dictating stem cell fate and we hesitate to assume which subtype of cancer the CPC we identify here may give rise to. Nonetheless, the evidence that the RANKL/RANK pathway is involved in tumor cell invasion and migration indicates that aberrant expression of RANK in any CPC may favor tumor cell metastasis to bone [7].

thumbnail image

Figure 5. Mammary differentiation (top) and cancer (bottom). Failure of CD49f+ cells to downregulate RANK as they acquire the Prl-R after asymmetric division generates a population of committed progenitors receptive to both self-renewal and proliferation signals. Invasion of this population into the basement membrane may restimulate expression of basal ligands such as CD49f and result in a heterogeneous tumor of more and less differentiated cells. Abbreviations: PL, placental lactogen; Prl-R, prolactin receptor; RANKL, receptor activator of NF-κB ligand.

Download figure to PowerPoint

The specific role of RANKL in this process is supported by our observation that in the absence of mammary differentiating media, PMEC is less proliferative and maintains a multipotent phenotype even after more than 2 weeks of culture. Addition of RANKL to these cultures produced calcium-rich colonies of cells that resemble calcified nodes in the breast. RANKL-induced calcification has also been demonstrated in other human tissue hierarchies, where calcification occurs in the less differentiated endothelial compartment of the tissue. In vascular tissue, RANKL activation of the ERK pathway is required for insulin-induced osteoblastic differentiation and subsequent calcification, and in aortic tissue, RANKL-induced calcification is inhibited by estrogen [42–44]. This supports our observation in the PMEC and implies RANKL may be involved in breast epithelial calcifications.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

We show that multipotent cells isolated from the lactating human mammary gland are differentially regulated compared with the nonlactating gland. These cells are likely to correspond to the committed progenitor population that generates differentiated alveoli. Further investigation of these cell types and their function may provide additional information about human breast cell hierarchies. Since it has been proposed that breast cancer may originate from breast stem cells, these cells may also be useful in investigating to what extent RANKL is involved in the development of such cancers.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Infrastructure for the laboratory of Professor Peter Hartmann that hosted part of this research is financially supported by Medela AG, Switzerland. The company has had no input on the study design, collection or interpretation of data, in writing the report, or the decision to publish the manuscript. The authors acknowledge the facilities, scientific and technical assistance of the Australian Microscopy, and microanalysis research facility at the Centre for Microscopy, Characterization and Analysis, The University of Western Australia, a facility funded by the University, State and Commonwealth Government.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES