TGF-β1-mediated EMT in NMuMG/E9 cells. (A) NMuMG/E9 cells were stained for E-cadherin (a,b), ZO-1 (c,d), paxillin (e,f) and F-actin (phalloidin; g,h) in the absence of TGF-β1 (a,c,e,g) or in the presence of TGF-β1 (5 ng ml^–1 for 1 day; b,d,f,h). In the presence of TGF-β1, the signals for E-cadherin and ZO-1 became weak and punctate, and focal adhesions (f) and stress fibers (h) were induced. Photographs were taken using a 63× oil objective; scale bar, 10 μm. (B) NMuMG/E9 cells were treated with TGF-β1 and cultured without passage for the indicated times. TNE extracts were immunoblotted for N-cadherin, E-cadherin, β-catenin and GAPDH at the indicated times from t=0 hour (h) to t=9 days (d). Immunoblots were quantified and normalized to GAPDH as shown on the graph. ZO-1 (an epithelium marker) and fibronectin (a mesenchyme marker) were examined in the same time course.

E-Cadherin continued to function after morphological EMT was induced. (A) Permeabilized (0.2% Triton X-100, Tx100; a) and unpermeabilized (b) NMuMG/E9 cells that had been treated with 5 ng ml^–1 TGF-β1 for 1 day were stained with antibodies against the extracellular portion of E-cadherin (ECCD2). E-Cadherin was localized to the cell surface (b). Photographs were taken using a 40× dry objective; scale bar, 10 μm. (B) Extracts of untreated NMuMG/E9 cells (lanes 3, 4) or NMuMG/E9 cells treated with 5 ng ml^–1 TGF-β1 for 1 day (lanes 5, 6) were immunoprecipitated (IP) with anti-E-cadherin (EC; lanes 4, 6) or control antibodies (Ctr; lanes 3, 5) and immunoblotted for E-cadherin, β-catenin and α-catenin. Comparable amounts of β-catenin and α-catenin were coimmunoprecipitated with E-cadherin in treated (lane 6) and untreated (lane 4) NMuMG/E9 cells, indicating that E-cadherin junctional complexes were functional. Lanes 1, 2 are immunoblots of treated and untreated cells to show N-cadherin and fibronectin were induced in this experiment, as expected, and to show the input levels of E-cadherin, β-catenin and α-catenin in the immunoprecipitations.

Snail and SIP1 were induced during TGF-β1-mediated EMT in NMuMG/E9 cells. (A) RT-PCR using total RNA extracted from untreated (lane 1) and TGF-β1-treated (5 ng ml^–1 for 2 days) NMuMG/E9 cells (lane 2) was done for snail, SIP1, slug, E12/E47 and β-actin (as a control). The identity of the slower migrating band in the E12/E47 lane is not known. (B) RT-PCR using total RNA extracted from NMuMG/E9 cells over the same time course as Fig. 2B. E-Cadherin, snail, SIP1 and GAPDH were analysed. (C) The snail and SIP1 mRNA levels were analysed by quantitative real-time RT-PCR using the same total RNA as in B.

Cadherin switching was reversible in TGF-β1-treated NMuMG/E9 cells. (A) Untreated NMuMG/E9 cells (a,d,g), NMuMG/E9 cells treated with 1 ng ml^–1 TGF-β1 and 20 ng ml^–1 EGF for 30 days (b,e,h), and NMuMG/E9 cells treated with TGF-β1 and EGF for 23 days and subsequently cultured upon removal of both reagents for an additional 7 days (c,f,i) were stained for E-cadherin (a-c), ZO-1 (d-f) or F-actin (phalloidin) (g-i). Photographs were taken using a 40× dry objective; scale bar, 10 μm. (B) TNE extracts from untreated NMuMG/E9 cells (lane 1), TGF-β1- and EGF-treated NMuMG/E9 cells for 60 days (lane 2) or reversibly treated NMuMG/E9 cells (lane 3) were examined for N-cadherin, E-cadherin, fibronectin and GAPDH by immunoblot. (C) RT-PCR was performed using total RNA extracted from untreated NMuMG/E9 cells (lane 1), TGF-β1- and EGF-treated NMuMG/E9 cells (for 1 day, lane 2, or for 30 days, lane 3) or reversibly treated NMuMG/E9 cells (lane 4) to examine the levels of mRNAs encoding E-cadherin, N-cadherin, snail, SIP1 and GAPDH.

Manipulation of N-cadherin expression in NMuMG/E9 cells did not interfere with TGF-β1-induced morphological change. (A) N-Cadherin-knockdown NMuMG/E9 cells (NMuMG/E9ΔNcad cells) showed similar morphology to parental NMuMG/E9 cells in phase-contrast microscopy (a,b). Photographs were taken of living cells using a 10× phase objective; scale bar, 50 μm. Parental NMuMG/E9 cells and NMuMG/E9ΔNcad cells were stained with N-cadherin mAb in the absence of TGF-β1 (c,d). Cell-cell border staining of N-cadherin was significantly reduced. Photographs were taken using a 40× dry objective; scale bar, 10 μm. (B) TNE extracts of NMuMG/E9 cells and NMuMG/E9ΔNcad cells in the absence and presence of TGF-β1 (5 ng ml^–1 for 1 day) were immunoblotted for N-cadherin, E-cadherin, GAPDH and fibronectin. N-Cadherin and E-cadherin levels were quantified and normalized to GAPDH as shown on the bar graph. (C) NMuMG/E9ΔNcad cells in the absence (a,c) or presence of TGF-β1 (5 ng ml^–1 for 1 day; b,d) were stained for paxillin (a,b) and F-actin (c,d). Photographs were taken using a 40× dry objective; scale bar, 10 μm. (D) Transwell motility assays were done using control knockdown NMuMG/E9 cells (ctr KD) and NMuMG/E9ΔNcad cells in the absence of TGF-β1 showed that NMuMG/E9ΔNcad cells were significantly less motile than control cells (P<0.05). (E) NMuMG/E9 cells expressing 2×-Myc-tagged human N-cadherin (NMuMG/E9 Ncad cells) showed similar morphology to parental NMuMG/E9 cells in phase-contrast microscopy (a,b). Photographs were taken using a 40× dry objective; scale bar, 10 μm. (F) TNE extracts of parental NMuMG/E9 cells and NMuMG/E9 Ncad cells were immunoblotted for N-cadherin (top) and E-cadherin (bottom).

E-Cadherin and N-cadherin expression in MCF10A cells was dependent on cell confluence. (A) Phase-contrast micrographs of MCF10A cells in sparse culture (a) or confluent culture (b). Photographs were taken of living cells using a 10× objective; scale bar, 50 μm. (B) MCF10A cells in sparse culture (a,c) or confluent culture (b,d) were stained for E-cadherin (a,b) or N-cadherin (c,d). Photographs were taken using a 40× dry objective; scale bar, 10 μm. (C) TNE extracts of MCF10A cells in increasing confluence were immunoblotted for N-cadherin (top), E-cadherin (middle) and β-catenin (bottom).

TGF-β1-mediated EMT in MCF10A cells. (A) MCF10A cells in the absence (a) or presence (b) of TGF-β1 [2 ng ml^–1 for 1 day (d)] were stained for F-actin. Photographs were taken using a 63× oil objective; scale bar, 10 μm. (B) Phase-contrast micrographs of TGF-β1-treated MCF10A cells (t=0-3 days). Photographs were taken of living cells using a 10× objective; scale bar, 50 μm. (C) TNE extracts of TGF-β1-treated parental MCF10A cells (t=0-3 days; lanes 1-4) control knockdown cells (ctr KD cells) (t=0-3 days; lanes 5-8), N-cadherin knockdown cells (ΔNcad cells) (t=0-3 days; lanes 9-12) and untreated N-cadherin overexpressing cells (Ncad cells) (lane 13) were immunoblotted for N-cadherin (regular and longer exposure), E-cadherin, GAPDH and fibronectin. The slower migrating bands in the N-cadherin immunoblot are pro-region-containing precursor forms of N-cadherin. (D) Quantification of N- and E-cadherin levels using normalization with GAPDH. (E) MCF10AΔNcad in the absence (a,c) or presence (b,d) of TGF-β1 (2 ng ml^–1 for 2 days) were co-stained for paxillin (a,b) and F-actin (c,d). Photographs were taken using a 63× oil objective; scale bar, 10 μm.

Knockdown of N-cadherin in MCF10A cells decreased cell motility. (A) Wound-healing assays. Phase-contrast micrographs of living cultures of parental MCF10A cells (a-f), MCF10A/ctr KD cells (g-l) and MCF10AΔNcad cells (m-r) in a wound-healing assays performed in the absence (a-c,g-i,m-o) or presence (d-f,j-l,p-r) of TGF-β1 (2 ng ml^–1). TGF-β1 treatment began 2 days before initiation of the assays. Photographs were taken just after the incision (0 hour) (a,d,g,j,m,p) and at 5 hours (b,e,h,k,n,q) and 13 hours (c,f,i,l,o,r) after incision using a 10× objective; scale bar, 50 μm. (B) Motility was quantified by measuring the decrease in the denuded area at 5 hours and 13 hours, and presented as the average decrease in the number of pixels with standard deviation in three independent experiments. TGF-β1 treatment increased motility in all cell lines tested. In the presence of TGF-β1, MCF10AΔNcad cells were significantly slower than parental or control knockdown cells at 5 hours (h) and 13 hours. (C) Transwell motility assays of parental MCF10A cells, MCF10A/ctr KD cells and MCF10AΔNcad cells in the absence (–) or presence (+) of TGF-β1 (2 ng ml^–1). TGF-β1 treatment began 2 days before initiation of the assays. Data are presented as the average number of migrating cells in nine random fields of view in three independent experiments. In the presence of TGF-β1, MCF10AΔNcad cells were significantly less motile than parental or control knockdown cells.