EpCAM overexpression prolongs proliferative capacity of primary human breast epithelial cells and supports hyperplastic growth

Abstract

Introduction: The Epithelial Cell Adhesion Molecule (EpCAM) has been shown to be strongly expressed in human breast cancer and cancer stem cells and its overexpression has been supposed to support tumor progression and metastasis. However, effects of EpCAM overexpression on normal breast epithelial cells have never been studied before. Therefore, we analyzed effects of transient adenoviral overexpression of EpCAM on proliferation, migration and differentiation of primary human mammary epithelial cells (HMECs).

Methods: HMECs were transfected by an adenoviral system for transient overexpression of EpCAM. Thereafter, changes in cell proliferation and migration were studied using a real time measurement system. Target gene expression was evaluated by transcriptome analysis in proliferating and polarized HMEC cultures. A Chicken Chorioallantoic Membrane (CAM) xenograft model was used to study effects on in vivo growth of HMECs.

Results: EpCAM overexpression in HMECs did not significantly alter gene expression profile of proliferating or growth arrested cells. Proliferating HMECs displayed predominantly glycosylated EpCAM isoforms and were inhibited in cell proliferation and migration by upregulation of p27(KIP1) and p53. HMECs with overexpression of EpCAM showed a down regulation of E-cadherin. Moreover, cells were more resistant to TGF-β1 induced growth arrest and maintained longer capacities to proliferate in vitro. EpCAM overexpressing HMECs xenografts in chicken embryos showed hyperplastic growth, lack of lumen formation and increased infiltrates of the chicken leukocytes.

Conclusions: EpCAM revealed oncogenic features in normal human breast cells by inducing resistance to TGF-β1-mediated growth arrest and supporting a cell phenotype with longer proliferative capacities in vitro. EpCAM overexpression resulted in hyperplastic growth in vivo. Thus, we suggest that EpCAM acts as a prosurvival factor counteracting terminal differentiation processes in normal mammary glands.

Figure 1

Characterization of primary human mammary epithelial cells (HMECs). Paraffin sections (A) or HMECs cultivated on collagen-coated cover slides (B) were stained with antibodies (markers) specific for myoepithelial (ASMA), basal/stem cells (p63), basal cells (CK5/14), and luminal cells (CK18). Immunohistochemical stainings revealed that in vivo EpCAM was expressed in basal and luminal cells of the breast, but not in myoepithelial cells. Immunofluorescence analysis revealed that HMECs stain for basal marker CK5/14, but were consistent negative for EpCAM, p63 and ASMA (n = 3). Magnification 200 ×; bars indicate 50 μm.

Figure 2

Adenoviral overexpression of EpCAM in HMECs inhibits cell proliferation and migration. HMECs (n = 3) were adenovirally transfected to overexpress EpCAM and GFP or GFP alone. A multiplicity of infection of 100 viruses/cell was used for all experiments. Overexpression of EpCAM was confirmed 48 h after transfection by Immunofluorescence (A). EpCAM was expressed on cell surface and cytoplasm (Phycoerythrin, red signal). Nuclei were counterstained with DAPI (blue signal, magnification 400×, bars indicate 25 μm). EpCAM overexpression was analyzed by real time PCR using GAPDH as housekeeping gene for normalization and GFP transfected cells as controls (B). As expected, overexpression resulted in a more than hundred-fold induction of EpCAM gene expression even 5 days after transfection. Protein expression was confirmed by Western Blot analysis (C). In comparison to control cells EpCAM was overexpressed as glycosylated isoform in proliferating cells and primarily as not glycosylated isoform in growth arrested HMECs. EpCAM glycosylation has been analyzed by enzymatic deglycosylation experiments with PNGaseF and subsequent Western Blot analysis. In all samples we observed a reduction of the 40-42 kDa glycosylated isoforms to the 35 kDa not glycosylated EpCAM isoform (D). Cell proliferation was analyzed in real time by the use the xCelligence system. EpCAM overexpression significantly inhibited cell proliferation (E). Western Blot analysis of cell cycle inhibition 48 h after EpCAM overexpression; p53 and p27^KIP1 proteins were upregulated in EpCAM overexpressing cells. (F). Cell migration was monitored by xCelligence CIM plate system after adenoviral transfection of EpCAM or GFP (G). EpCAM overexpression also inhibited cell migration; stars indicate p values < 0.05.

Figure 3

Analysis of EpCAM target genes in HMECs. HMECs were cultivated as mitotic subconfluent cultures (A) or for 10 days on Matrigel-coated transwells to induce polarization of epithelial monolayers (B). Then, cells were adenovirally transfected (MOI = 100) to overexpress GFP alone or EpCAM/GFP. Polarized HMECs cells did not express immunoreactive EpCAM protein, but the gap junction protein ZO-1, E-cadherin and β-catenin as determined by Immunofluorescence analysis (C, bars indicate 25 μm). Polarized cells and mitotic cultures (log-phase) of three donors (HMEC 1–3) were adenovirally transfected and EpCAM gene expression quantified 24 h later by real time PCR using GAPDH as internal house-keeping gene (D). MA-plot of genes regulated by EpCAM as determined by Affymetrix Gene analysis in mitotic standard cultures (E) or confluent polarized cultures of HMECs (F). Besides EpCAM no additional genes were significantly regulated in all three donors analyzed 20 h after transfection. Mean M-values indicate differential gene expression between EpCAM over-expressing and GFP transfected cells (log2 scale), Mean A-values indicate average expression of a gene in all microarrays. Stars indicate p values <0.05.

Figure 4

EpCAM inhibits TGF-β1 induced terminal growth arrest and differentiation in HMECs. Adenovirally transfected HMECs were stimulated with 1 ng/mL TGF-β1 to undergo terminal growth arrest and differentiation in vitro. (A) In comparison to control cells TGF-β1 treated GFP expressing cells got growth arrested, flat and enlarged. Populations of EpCAM transfected cells were protected from TGF-β1 and acquired a small cell body (white arrows). (B) In comparison to proliferating control cells TGF-β1 treated cells stained positive for senescence-associated beta galactosidase (SA-β-Gal, blue color), a marker for terminally arrested cells. EpCAM transfected cells were predominantly negative and acquired a more spindle shaped morphology (black arrows). (C) Long term cultures of transfected HMECs in the presence of TGF-β1. EpCAM transfected cells showed a higher proliferative capacity within the observed time window of 6 days than GFP controls. Bars indicate 25 μm. (D) Cell numbers were analyzed 1, 3 and 6 days after TGF-β stimulation by counting in a Buerker-Tuerk chamber. EpCAM transfected cells displayed significantly higher proliferative activities, i.e. higher cell counts after 6 days of growth. (E) Western Blot analysis of differentiation markers for epithelial mesenchymal transition (vimentin, E-cadherin). (F) EpCAM transfected HMECs show a downregulation of E-cadherin (E-cad) but no significant upregulation of vimentin (Vim) protein. Stars indicate p values <0.05.

Figure 5

EpCAM overexpression in vivo leads to hyperplastic cell growth without ductal lumen formation. HMECs were transplanted together with matrigel into the chorioallantoic membrane (CAM) of chicken embryos and cell growth and morphology analyzed after 6 days. (A) Chicken embryos with HMEC xenografts. White arrows indicate transplants of HMECs in matrigel plugs (B) Fluorescence stereo microscope picture of a HMEC graft 6 days after in vivo growth. The green clusters (adenoviral GFP) indicate human cell-cell aggregates that growth inside the CAM (magnification 20×) (C) Immunohistochemical analysis of cross-sections of HMEC grafts in the CAM. Sections were stained with an antibody specific for human cadherin, thus detecting only human cells. In contrast to GFP transfected controls, EpCAM overexpressing grafts show bigger glandular structures that lack formation of lumen. (D) Immunohistochemical analysis of p63^high progenitor cells. Noteworthy, HMECs at the glandular base express the progenitor marker p63. EpCAM overexpressing clusters are surrounded by significant bigger clusters of chicken leukocytes (black arrows). Bars indicate 50 μm, asterisks indicate lumen.

Figure 6

EpCAM overexpression leads to upregulation of c-myc and increased cell proliferation in immortalized MCF10A human breast epithelial cells. (A) Flow cytometric analysis of EpCAM expression in adenovirally transfected MCF10A cells. In comparison to GFP transfected controls, only EpCAM transfected cells show the immunoreactive protein on the cell membrane. (B) Overexpression of EpCAM results in a significant increase in cell proliferation under serum-reduced conditions. (C) Real time PCR analysis of TP53, p27kip1 and c-myc gene expression in EpCAM transfected cells. Overexpression of EpCAM upregulated c-myc gene expression. (D) Western Blot analysis of EpCAM overexpression and upregulation of c-myc protein levels. Tubulin alpha served as internal loading control. (E) Densitometric analysis of c-myc to tubulin protein ratio. MCF10A cell lines were generated by a lentiviral system to have a stable expression of a non-silencing control (ns/crtl) or an EpCAM specific (E#2) shRNA. MCF10A ^ns/crtl and MCF10A ^E#2 cells were adenovirally transfected to overexpress GFP or EpCAM/GFP. (F) In comparison to MCF10A ^ns/crtl cells MCF10A ^E#2 cells were significantly downregulating EpCAM transcript levels 24 and 48 h after adenoviral transfection. (G) Real time cell proliferation of MCF10A ^E#2 cells was significantly lower than those of MCF10A ^ns/crtl after adenoviral EpCAM overexpression. Stars indicate p values <0.05.