Connexin 43 Loss Triggers Cell Cycle Entry and Invasion in Non-Neoplastic Breast Epithelium: A Role for Noncanonical Wnt Signaling

Abstract

(1) Background: The expression of connexin 43 (Cx43) is disrupted in breast cancer, and re-expression of this protein in human breast cancer cell lines leads to decreased proliferation and invasiveness, suggesting a tumor suppressive role. This study aims to investigate the role of Cx43 in proliferation and invasion starting from non-neoplastic breast epithelium. (2) Methods: Nontumorigenic human mammary epithelial HMT-3522 S1 cells and Cx43 shRNA-transfected counterparts were cultured under 2-dimensional (2-D) and 3-D conditions. (3) Results: Silencing Cx43 induced mislocalization of β-catenin and Scrib from apicolateral membrane domains in glandular structures or acini formed in 3-D culture, suggesting the loss of apical polarity. Cell cycle entry and proliferation were enhanced, concomitantly with c-Myc and cyclin D1 upregulation, while no detectable activation of Wnt/β-catenin signaling was observed. Motility and invasion were also triggered and were associated with altered acinar morphology and activation of ERK1/2 and Rho GTPase signaling, which acts downstream of the noncanonical Wnt pathway. The invasion of Cx43-shRNA S1 cells was observed only under permissive stiffness of the extracellular matrix (ECM). (4) Conclusion: Our results suggest that Cx43 controls proliferation and invasion in the normal mammary epithelium in part by regulating noncanonical Wnt signaling.

Keywords: Wnt pathways; breast cancer; connexin 43; gap junctions; invasion; mammary epithelium; mammary gland; microenvironment; motility; proliferation.

Figure 1

Silencing Cx43 enhances the proliferation rate of S1 cells under 2-D and 3-D culture conditions. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D (A) or 3-D conditions (B). Proliferation rate was assessed by cell counting on days 6 and 10 in 2-D (A; upper panel) and by measurement of acinar diameter on days 4, 6, 9, and 11 in 3-D (B; right panel). An ocular micrometer calibrated against a stage micrometer was used, and acinar areas were then calculated and plotted as acinar size. Fifty acini were analyzed per group. The values depicted in histograms are the means (±S.D.) of cell counts (A; upper panel) or acinar sizes (B; right panel) from three independent experiments. Unpaired t-test; * p < 0.05, ** p < 0.01, *** p < 0.001. Representative images of cells on day 11 in 2-D (A; lower panel) and in 3-D (B; left panel) are shown. Nuclei were stained with Hoechst (blue; B; left lower panel).

Figure 2

Silencing Cx43 triggers cell cycle entry and upregulates the expression of cell cycle genes in S1 cells under 2-D and 3-D culture conditions. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D (A,C; left panel) or 3-D conditions (B,C; right panel). A and B. Cell cycle analysis was performed by flow cytometry on days 4, 6, 9, and 11 in 2-D (A) and on days 4 and 11 in 3-D (B). The values depicted in histograms are the means (±S.D.) of cell percentages in the different cell cycle phases from three independent experiments. Unpaired t-test; * p < 0.05, ** p < 0.01, *** p < 0.001. (C) Total proteins were extracted on days 4, 6, 9, and 11 in 2-D (left panel) and on day 11 in 3-D (right panel). Expression of c-Myc and cyclin D1 was assessed by Western blotting. Lamin B served as loading control. The values depicted in the histogram (right lower panel) are the means of fold change in c-Myc or cyclin D1 expression in 3-D normalized to that of Lamin B from three independent experiments. Fold change in normalized expression is set to 1 in S1 cells.

Figure 2

Silencing Cx43 triggers cell cycle entry and upregulates the expression of cell cycle genes in S1 cells under 2-D and 3-D culture conditions. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D (A,C; left panel) or 3-D conditions (B,C; right panel). A and B. Cell cycle analysis was performed by flow cytometry on days 4, 6, 9, and 11 in 2-D (A) and on days 4 and 11 in 3-D (B). The values depicted in histograms are the means (±S.D.) of cell percentages in the different cell cycle phases from three independent experiments. Unpaired t-test; * p < 0.05, ** p < 0.01, *** p < 0.001. (C) Total proteins were extracted on days 4, 6, 9, and 11 in 2-D (left panel) and on day 11 in 3-D (right panel). Expression of c-Myc and cyclin D1 was assessed by Western blotting. Lamin B served as loading control. The values depicted in the histogram (right lower panel) are the means of fold change in c-Myc or cyclin D1 expression in 3-D normalized to that of Lamin B from three independent experiments. Fold change in normalized expression is set to 1 in S1 cells.

Figure 3

Cx43 assembles GJ complexes in S1 cells under 2-D culture conditions. S1 cells were cultured under 2-D conditions. Total proteins were extracted on day 6. Association of Cx43 and β-catenin (upper panel) or ZO-2 (lower panel) was assessed by coimmunoprecipitation (IP) of Cx43 followed by Western blotting for detection of Cx43, β-catenin and ZO-2. The input served as a control. Three independent experiments were performed for Cx43-β-catenin association and two for Cx43-ZO-2 association.

Figure 4

Silencing Cx43 alters the localization of junctional and polarity proteins in S1 cells under 3-D culture conditions without affecting their expression levels. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D or 3-D conditions. (A) Total proteins were extracted on day 9 in 2-D (left panel) and on day 11 in 3-D (right panel). Expression of Scrib, ZO-2, β-catenin, and Cx43 was assessed by Western blotting. Lamin B served as loading control. Three independent experiments were performed. The values depicted in the histogram (right panel) are the means of fold change in Cx43 expression in 3-D normalized to that of Lamin B from three independent experiments. Fold change in normalized expression is set to 1 in S1 cells. (B) Localization of β-catenin (red) was assessed by immunofluorescence on day 9 in 2-D (left upper panel) and on day 11 in 3-D (left lower panel). Nuclei were counterstained with Hoechst (blue). The values depicted in the histogram are the means (±S.D.) of acini percentages with apicolateral β-catenin from three independent experiments. One hundred acini were analyzed per group. Unpaired t-test; ** p < 0.01. (C) Localization of Scrib (red) was assessed by immunofluorescence on day 12 in 2-D (left upper panel) and on day 11 in 3-D (left lower panel). Nuclei were counterstained with Hoechst (blue). The values depicted in the histogram are the means (±S.D.) of acini percentages with apicolateral Scrib from three independent experiments. One hundred acini were analyzed per group. Unpaired t-test; *** p < 0.001.

Figure 4

Silencing Cx43 alters the localization of junctional and polarity proteins in S1 cells under 3-D culture conditions without affecting their expression levels. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D or 3-D conditions. (A) Total proteins were extracted on day 9 in 2-D (left panel) and on day 11 in 3-D (right panel). Expression of Scrib, ZO-2, β-catenin, and Cx43 was assessed by Western blotting. Lamin B served as loading control. Three independent experiments were performed. The values depicted in the histogram (right panel) are the means of fold change in Cx43 expression in 3-D normalized to that of Lamin B from three independent experiments. Fold change in normalized expression is set to 1 in S1 cells. (B) Localization of β-catenin (red) was assessed by immunofluorescence on day 9 in 2-D (left upper panel) and on day 11 in 3-D (left lower panel). Nuclei were counterstained with Hoechst (blue). The values depicted in the histogram are the means (±S.D.) of acini percentages with apicolateral β-catenin from three independent experiments. One hundred acini were analyzed per group. Unpaired t-test; ** p < 0.01. (C) Localization of Scrib (red) was assessed by immunofluorescence on day 12 in 2-D (left upper panel) and on day 11 in 3-D (left lower panel). Nuclei were counterstained with Hoechst (blue). The values depicted in the histogram are the means (±S.D.) of acini percentages with apicolateral Scrib from three independent experiments. One hundred acini were analyzed per group. Unpaired t-test; *** p < 0.001.

Figure 5

Silencing Cx43 activates MAPK but not Wnt/β-catenin signaling in S1 cells under 3-D culture conditions. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D (A) or 3-D conditions (A–D). (A,B,D) Total proteins were extracted on days 4, 9, and 11 in 2-D (A; left panel) and on day 11 in 3-D (A; right panel, B,D). Expression of β-catenin, p-β-catenin, GSK-3α/β, p-GSK-3α/β, ERK1/2, and p-ERK1/2 was assessed by Western blotting. Lamin B and GAPDH served as loading controls. Three independent experiments were performed. The values depicted in the histogram (D; right panel) are the means of fold change in p-ERK1/2 levels in 3-D normalized to that of GAPDH from three independent experiments. Fold change in normalized levels is set to 1 in S1 cells. (C) Nuclear proteins were extracted on day 11 (S1 and Cx43 KO cells) or on day 7 in 3-D (T4-2 cells). β-catenin levels were assessed by Western blotting. Lamin B served as loading control. Purity of nuclear fractions was determined by analyzing cytosolic (GAPDH) and membrane markers (Tim23). Total proteins extracted on day 11 in 3-D (S1 cells) were co-analyzed as a control for the detection of GAPDH and Tim23. Three independent experiments were performed.

Figure 5

Silencing Cx43 activates MAPK but not Wnt/β-catenin signaling in S1 cells under 3-D culture conditions. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D (A) or 3-D conditions (A–D). (A,B,D) Total proteins were extracted on days 4, 9, and 11 in 2-D (A; left panel) and on day 11 in 3-D (A; right panel, B,D). Expression of β-catenin, p-β-catenin, GSK-3α/β, p-GSK-3α/β, ERK1/2, and p-ERK1/2 was assessed by Western blotting. Lamin B and GAPDH served as loading controls. Three independent experiments were performed. The values depicted in the histogram (D; right panel) are the means of fold change in p-ERK1/2 levels in 3-D normalized to that of GAPDH from three independent experiments. Fold change in normalized levels is set to 1 in S1 cells. (C) Nuclear proteins were extracted on day 11 (S1 and Cx43 KO cells) or on day 7 in 3-D (T4-2 cells). β-catenin levels were assessed by Western blotting. Lamin B served as loading control. Purity of nuclear fractions was determined by analyzing cytosolic (GAPDH) and membrane markers (Tim23). Total proteins extracted on day 11 in 3-D (S1 cells) were co-analyzed as a control for the detection of GAPDH and Tim23. Three independent experiments were performed.

Figure 6

Silencing Cx43 induces motility and invasion in S1 cells. (A) S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D conditions. Motility was assessed by time-lapse imaging on day 5. The total paths of representative S1 and Cx43 KO cells from time-lapse movies (upper panel) are shown (different colors represent different cells). Histograms show the quantification of cell motility. A total of 50 cells were analyzed per group. The values depicted are the means (±S.D.) of total path lengths (left lower panel) or migration speeds (right lower panel) from two independent experiments. Unpaired t-test; ** p < 0.01. (B) Invasion of S1 and Cx43 KO cells across diluted Matrigel (1:5) was assessed by transwell cell invasion assay (upper panel). The values depicted in the histogram are the means (±S.D.) of fold change in number of Matrigel-invading cells from three independent experiments. Unpaired t-test; ** p < 0.01. Representative images of cells cultured under 2-D conditions and stained for F-actin on day 5 (lower panel) are shown. Arrows indicate invadopodia-like actin-rich dots. (C) S1 and Cx43 KO cells were cultured under 3-D conditions atop of different Matrigel dilutions (undiluted, 1:5, 1:10 and 1:20 dilutions). Invasion was assessed by counting nonspheroid structures on day 11. One hundred structures were analyzed per group. The values depicted in the histogram (upper panel) are the means (±S.D.) of nonspheroid structures from three independent experiments. Unpaired t-test; *** p < 0.001. Representative images of cells (lower panel) are shown. Arrows indicate migrating Cx43 KO cells. Undil; Undiluted.

Figure 6

Silencing Cx43 induces motility and invasion in S1 cells. (A) S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D conditions. Motility was assessed by time-lapse imaging on day 5. The total paths of representative S1 and Cx43 KO cells from time-lapse movies (upper panel) are shown (different colors represent different cells). Histograms show the quantification of cell motility. A total of 50 cells were analyzed per group. The values depicted are the means (±S.D.) of total path lengths (left lower panel) or migration speeds (right lower panel) from two independent experiments. Unpaired t-test; ** p < 0.01. (B) Invasion of S1 and Cx43 KO cells across diluted Matrigel (1:5) was assessed by transwell cell invasion assay (upper panel). The values depicted in the histogram are the means (±S.D.) of fold change in number of Matrigel-invading cells from three independent experiments. Unpaired t-test; ** p < 0.01. Representative images of cells cultured under 2-D conditions and stained for F-actin on day 5 (lower panel) are shown. Arrows indicate invadopodia-like actin-rich dots. (C) S1 and Cx43 KO cells were cultured under 3-D conditions atop of different Matrigel dilutions (undiluted, 1:5, 1:10 and 1:20 dilutions). Invasion was assessed by counting nonspheroid structures on day 11. One hundred structures were analyzed per group. The values depicted in the histogram (upper panel) are the means (±S.D.) of nonspheroid structures from three independent experiments. Unpaired t-test; *** p < 0.001. Representative images of cells (lower panel) are shown. Arrows indicate migrating Cx43 KO cells. Undil; Undiluted.

Figure 7

S1 cells maintain their characteristic spheroid morphology, lumen-forming ability, and polarity under 3-D culture conditions of reduced ECM stiffness. S1 cells were cultured under 3-D conditions atop of different Matrigel dilutions (undiluted, 1:5, 1:10 and 1:20 dilutions). Lumen formation (A) and polarity (B), demonstrated by apicolateral Scrib localization (red; B; left panel), were assessed by immunofluorescence on day 11. (A) Fifty acini were analyzed per condition. The values depicted in the histogram (right panel) are the means (±S.D.) of acini percentages with lumen from three independent experiments. Unpaired t-test; * p < 0.05, *** p < 0.001. Representative images of acini (left panel) are shown. Nuclei were stained with Hoechst (blue; left lower panel). (B) Sixty-seven acini were analyzed per condition. The values depicted in the histogram (right panel) are the means (±S.D.) of acini percentages with apicolateral Scrib from two independent experiments. Unpaired t-test. Representative images of acini (left panel) are shown. Nuclei were counterstained with Hoechst (blue). Undil; Undiluted.

Figure 8

Silencing Cx43 upregulates the expression and activity of Rho GTPases in S1 cells under 2-D and 3-D culture conditions. S1 and Cx43-shRNA S1 cells (Cx43 KO) were cultured under 2-D (A), 3-D conditions of undiluted (B) or diluted Matrigel (1:5; C). Total proteins were extracted on day 9 in 2-D (A) and on day 11 in 3-D (B,C). Expression of RhoA, Cdc42, and Rac1 was assessed by Western blotting. Lamin B and GAPDH served as loading controls. Three independent experiments were performed. Activity of Rho GTPases was assessed by pulldown assay followed by Western blotting for detection of active Rho GTPases. Three independent experiments were performed.