Sema4C/PlexinB2 signaling controls breast cancer cell growth, hormonal dependence and tumorigenic potential

Abstract

Semaphorin 4C (Sema4C) expression in human breast cancers correlates with poor disease outcome. Surprisingly, upon knock-down of Sema4C or its receptor PlexinB2 in diverse mammary carcinoma cells (but not their normal counterparts), we observed dramatic growth inhibition associated with impairment of G2/M phase transition, cytokinesis defects and the onset of cell senescence. Mechanistically, we demonstrated a Sema4C/PlexinB2/LARG-dependent signaling cascade that is required to maintain critical RhoA-GTP levels in cancer cells. Interestingly, we also found that Sema4C upregulation in luminal-type breast cancer cells drives a dramatic phenotypic change, with disassembly of polarity complexes, mitotic spindle misorientation, cell-cell dissociation and increased migration and invasiveness. We found that this signaling cascade is dependent on the PlexinB2 effectors ErbB2 and RhoA-dependent kinases. Moreover, Sema4C-overexpressing luminal breast cancer cells upregulated the transcription factors Snail, Slug and SOX-2, and formed estrogen-independent metastatic tumors in mice. In sum, our data indicate that Sema4C/PlexinB2 signaling is essential for the growth of breast carcinoma cells, featuring a novel potential therapeutic target. In addition, elevated Sema4C expression enables indolent luminal-type tumors to become resistant to estrogen deprivation, invasive and metastatic in vivo, which could account for its association with a subset of human breast cancers with poor prognosis.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Figures

Fig. 1

Sema4C is widely expressed in human breast cancers and its elevated expression is associated with poor patient prognosis. a Kaplan–Meier analysis of overall patient survival (OS; follow-up at 12 years) of unselected breast cancer patients (n = 1402) correlation to Semaphorin 4C (Sema4C) high and low expression subgroups, based on online dataset analysis by http://kmplot.com; p = 0.015. b Western blot analysis of Sema4C and PlexinB2 expression in various breast cancer cell lines and immortalized human mammary epithelial cell line HMEC- hTERT; vinculin served as protein loading control. c Western blot analysis revealing full-length and cleaved form of Sema4C in total cell lysate (TCL) and conditioned medium (CM), respectively, of MCF-7 and MDA-MB-231 breast cancer cells

Fig. 2

Sema4C/PlexinB2 signaling blockade in breast cancer cells causes growth arrest. a–c Graphic representation of the growth curves of the indicated breast cancer cell lines (MCF-7, SKBR-3 and MDA-MB-231) upon knock-down of Sema4C or PlexinB2 (by targeted shRNA), or overexpression of a PlexinB2 dominant-negative construct (B2-DN). The cells were stained with crystal violet and absorbance was read at 595 nm. d Graphic representation of HMEC-hTERT cell growth curves upon silencing either Sema4C or PlexinB2 (by targeted shRNA). e, f MDA-MB-231 growth curves, upon silencing Sema4C (panel e) or PlexinB2 (panel f) in presence or absence of recombinant soluble Sema4C (1 μg/ml). Data are the mean ± SD from three separate experiments; **p < 0.01

Fig. 3

Sema4C/PlexinB2 signaling blockade in breast cancer cells causes cell senescence, cell cycle inhibition and cytokinesis defects. a Bright field images of HMEC-hTERT cells subjected to Sema4C or PlexinB2 silencing (or controls). b Bright field image of MDA-MB-231 cells either subjected to Sema4C or PlexinB2 silencing, or overexpressing PlexinB2-DN, or controls. Scale bar: 100 µm. c Quantification of area covered by engineered MDA-MB-231 described in panel b. d Fraction of SA-β-gal-positive senescent cells in the population of engineered MDA- MB-231 described in panel b. e Representative images of immunofluorescent staining of control (shCtrl) or PlexinB2-silenced (shPlxnB2) MCF-7 and MDA-MB-231 cells (as indicated), stained with Phalloidin to reveal F-actin (in red), and DAPI for nuclei (blue). f The percentage of multinucleated MCF-7 and MDA-MB-231 cells described in panel e was assessed by computer-assisted analysis of microscopic images. g Cytofluorimetric analysis of propidium iodide-stained MDA-MB-231 cells, either silenced for Sema4C or PlexinB2, showing relative accumulation in G2/M phase, compared with controls. h Quantitative RT-PCR analysis of mRNA levels of cell cycle genes CCNA2 and CDC2(CDK1) in MCF-7 and MDA-MB-231 cells stably silenced for Sema4C or PlexinB2 (fold change vs. control cells). i Immunofluorescence staining of p21 and p53 tumor-suppressor protein (green) revealing increased nuclear localization in MCF-7 cells depleted of PlexinB2, compared with controls; nuclei were stained with DAPI (blue). Data are the mean ± SD from three separate experiments; ***p < 0.001. Scale bar: 100 µm.

Fig. 4

Sema4C/PlexinB2/LARG-dependent RhoA signaling is required to sustain breast cancer cell proliferation. a MDA-MB-231 cells either control or subjected to PlexinB2 silencing, or treated with ROCK kinase inhibitor H1152 (10 μM) or LARG inhibitor Y-16 (20 μM) for 5 days, were analyzed by immunofluorescence to reveal F-actin (with Phalloidin, in red) and the nuclei (with DAPI, in blue). Representative images are shown; scale bar: 100 µm. b Quantification of individual cell area of MDA-MB-231 cells treated as described in panel a, by computer-assisted analysis of microscopic images. c Long-term colony formation assay (10 days) of parental MDA-MB-231 cells treated with 10 μM H1152, or 20 μM Y-16, or with vehicle alone (DMSO); or the same cells subjected to stable silencing of Sema4C (shSema4C), or PlexinB2 (shPlxnB2), or shCtrl control cells. Fixed cells were eventually stained with crystal violet. Each condition was repeated in triplicate; representative images are shown. d Immunoblotting analysis of active RhoA-GTP pulled-down from lysates of MDA-MB-231 cells silenced for Sema4C or PlexinB2 or overexpressing PlexinB2-DN, or controls. Detection of total RhoA in input lysates served as a loading control. e Immunoblotting analysis of PlexinB2, CDC2 (CDK1) expression in MDA- MB-231 cells silenced for PlexinB2 (or controls), upon overexpression of constitutively active (CA) RhoA (GFP- Q63L). The expression of CA-RhoA was assessed with anti-GFP antibodies. f Graphic representation of MDA-MB-231 cell growth curves upon silencing PlexinB2, with and without overexpressing CA-RhoA construct (same cells described in panel e). Data are the mean ± SD from three separate experiments; **p < 0.01, ***p < 0.001

Fig. 5

Enhanced Sema4C signaling disrupts cell polarity and converts luminal breast cancer cells to acquire mesenchymal features. a Kaplan–Meier analysis of overall survival of patients (OS; follow-up at 12 years) bearing estrogen receptor positive (ER+) tumors (n = 548) correlation to Semaphorin 4C (Sema4C) high and low expression subgroups, based on online dataset analysis by http://kmplot.com; p = 6e-04. b Immunoblotting images revealing the expression of Sema4C-secr and Sema4C-full constructs in gene-transduced MCF-7 cells. c Graphic representation of MCF-7 cell growth curves, upon overexpression of Sema4C-secreted or full-length isoforms compared with Mock cells. The cells were stained with crystal violet and absorbance was read at 595 nm. d Microscopy images of MCF-7 described above stained with crystal violet. Mock-transduced cells display a cobblestone phenotype, whereas Sema4C-overexpressing cells show an elongated fibroblast-like morphology. Scale bars: 100 µm. e Representative immunofluorescence images showing the expression of E-cadherin or ZO-1 (in green) in Sema4C secr-overexpressing MCF-7 cells, compared with controls; nuclei were visualized with DAPI (in blue). Scale bars: 100 µm. f Immunoblotting analysis of E-cadherin and ZO-1 protein expression, in MCF-7 cells overexpressing Sema4C-secreted or full-length isoforms (same as in panel d). g Immunofluorescence analysis of F-actin (with Phalloidin staining) in Mock and Sema4C secr-overexpressing MCF-7 cells. Scale bars: 20 µm. On the right panel, the graph on the right shows the quantification of fluorescent signal from F-actin Phalloidin staining. h, i Graphic representation of the migratory (h) and invasive (i) capability of stably transfected MCF-7 cells expressing Sema4C-secr or full-length (or Mock), assessed using transwell chamber inserts. Data are the mean ± SD from three separate experiments; ***p < 0.001

Fig. 6

Enhanced Sema4C signaling disrupts cell polarity and reprograms luminal breast cancer cells in RhoA- and ErbB2-dependent manner. a Immunoblotting analysis revealing tyrosine phosphorylated ErbB2 in MCF-7 cells upon stimulation with recombinant Sema4C (300 ng/ml) or mock, for 20 min; total ErbB2 staining provided an internal reference. b MCF-7 cells were treated with recombinant Sema4C (300 ng/ml) or mock, for 20 min and protein lysates were immunopurified with anti-ErbB2 (upper panel) or anti-PAR-3 (lower panel) antibodies, followed by immunoblotting to reveal the interaction between PKC-Zeta and either of the two alternative partners. c GTP- bound active RhoA was pulled-down from lysates of serum-starved MCF-7 cells stimulated with recombinant Sema4C (300 ng/ml) for 20 min, and revealed by immunoblotting; detection of total RhoA served as an internal reference. d Graphic representation of transwell migration efficiency (fold changes, over a period of 16 h) of MCF-7 cell expressing Sema4C-secr (or mock), in the presence of the LARG inhibitor Y-16 (20 μM) or DMSO vehicle alone. e Serum-starved MCF-7 cells were incubated with or without the ErbB2 inhibitor lapatinib (10 μM) for 2 h prior to stimulation with 300 ng/ml Sema4C for 20 min. GTP-bound active RhoA was pulled-down from lysates and revealed by immunoblotting; detection of total RhoA provided an internal reference. f Serum-starved MCF-7 cells were incubated with or without the Rho-kinase Inhibitor Y-27632 (10 μM) for 30 min or with 10 μM Lapatinib for 2 h, or with DMSO, prior to stimulation with 300 ng/ml Sema4C for 20 min. Cell lysates were then immunopurified with anti-PAR-3 antibody (as in panel b) and the immunocomplexes were analyzed by immunoblotting with anti-aPKC and anti-PAR-3 antibodies. g Graphic representation of the fold change in mRNA levels of SNAIL, SLUG, ZEB1 and ZEB2 genes in MCF-7 cells stably expressing Sema4C-secr, compared with control cells, as determined by qRT-PCR analysis. h Graphic representation of the fold change in SNAIL and SLUG mRNA levels in Sema4C-secr-overexpressing (or control) cells treated with 10 μM Lapatinib or the ROCK kinase inhibitor H1152 (10 μM) for 24 h, as determined by qRT-PCR analysis. Data are the mean ± SD from three separate experiments; based on statistical analysis, ***p < 0.001

Fig. 7

Sema4C-overexpressing cells attain altered mitotic spindle orientation, detachment from the substrate and growth in suspension. a Bright field images revealing the morphology of MCF-7 mock or Sema4C-secr-overexpressing cells, in lower and higher confluency conditions. Scale bars: 100 µm. b Representative immunofluorescence analysis of Sema4C-secr-overexpressing (or Mock) MCF-7 cells, stained for alpha-tubulin (in red), gamma-tubulin (in green) and DAPI (blue). The rightmost panel shows corresponding X–Z axis stack projections generated from laser-scanning confocal images taken at 0.5-μm steps. Confocal micrographs display parallel orientation of the mitotic spindle (relative to the substrate) in Mock-transduced MCF-7 cells and a misoriented spindle in Sema4C-secr-overexpressing cell, with axis perpendicular rather than parallel to the substrate. Scale bars: 10 µm. c Graphical representation of distribution of mitotic spindle angles of MCF-7 cells (see Materials and methods for details), upon expression of Sema4C-secr or Sema4C full-length, or mock-transduced (n > 50 cells in three independent experiments). d Relative frequency and average size of mammospheres formed by MCF-7 cells upon Sema4C-secr or Sema4C full-length expression, or controls. Below the graph shown are representative bright field images of the different conditions; scale bars: 20 µm. e qRT-PCR analysis of mRNA expression of stemness markers SOX-2 and ALDH3A1, in Sema4C-secr and Sema4C full-length expressing MCF-7 cells, compared with Mock. f qRT-PCR analysis of mRNA expression of SOX-2, ALDH3A1 in MCF-7 cells upon silencing Sema4C, PlexinB2 or overexpressing PlexinB2-DN, compared with controls. g Relative frequency of mammospheres formed by MCF-7 cell silencing Sema4C, PlexinB2 or overexpressing PlexinB2-DN, compared with controls. Data are the mean ± SD from three separate experiments; ***p < 0.001

Fig. 8

Sema4C overexpression in luminal-type breast cancers cells leads to tamoxifen resistance, estrogen-independent growth and metastatic tumor formation in vivo. a Graphic representation of the growth curves of MCF-7 cells overexpressing Sema4C-secr (or Mock) in estrogens-depleted medium (phenol red-free DMEM with 5% charcoal-stripped FBS). The cells were stained with crystal violet and absorbance was read at 595 nm. b Graphical representation of relative fold change in the cell viability of Mock and Sema4C-secr-overexpressing cell treated with the indicated increasing concentrations of 4-OH tamoxifen for 5 days (or with vehicle alone). c qRT-PCR analysis of mRNA expression of estrogen receptor genes (ER-alpha and ER-beta) in MCF-7 cells overexpressing Sema4C-secr, compared with Mock cells. d Immunoblotting analysis of ER-alpha protein expression in Mock and Sema4C- secr-overexpressing cells; beta-actin staining provided a protein loading control. e Immunocytochemistry analysis of ER-alpha expression and nuclear localization in Mock and Sema4C-secr-overexpressing MCF-7 cells. Scale bars: 100 µm. f qRT-PCR analysis of mRNA expression of progesterone receptor (PGR) in MCF-7 cells overexpressing Sema4C-secr, compared with Mock. g Z-score correlation plot between Sema4C and progesterone receptor (PGR) expression, based on the analysis of TCGA breast cancer dataset (Breast Invasive Carcinoma, TCGA, Cell 2015-cBioPortal; n = 421 cases); p < 0.0001. h MCF-7 cells overexpressing Sema4C-secr (or mock-transduced) were orthotopically injected into the mammary fat pad of female nude mice (n = 5 per condition), without any estrogen supplementation, and tumor burden was assessed at the indicated time points. i Representative immunohistochemical images showing Sema4C, ER-alpha and PGR expression in Sema4C-secr-overexpressing (and Mock) tumors described in panel h. Scale bars: 100 µm. j Quantification of spontaneous lung macro-metastasis detected in tumor-bearing mice described in panel h. Data are the mean ± SD from three separate experiments; **p < 0.01, ***p < 0.001