c-Ski activates cancer-associated fibroblasts to regulate breast cancer cell invasion

Abstract

Aberrant expression of c-Ski oncoprotein in some tumor cells has been shown to be associated with cancer development. However, the role of c-Ski in cancer-associated fibroblasts (CAFs) of tumor microenvironment has not been characterized. In the current study, we found that c-Ski is highly expressed in CAFs derived from breast carcinoma microenvironment and this CAF-associated c-Ski expression is associated with invasion and metastasis of human breast tumors. We showed that c-Ski overexpression in immortalized breast normal fibroblasts (NFs) induces conversion to breast CAFs by repressing p53 and thereby upregulating SDF-1 in NFs. SDF-1 treatment or p53 knockdown in NFs had similar effects on the activation of NFs as c-Ski overexpression. The c-Ski-activated CAFs show increased proliferation, migration, invasion and contraction compared with NFs. Furthermore, c-Ski-activated CAFs facilitated the migration and invasion of MDA-MB-231 breast cancer cells. Our data suggest that c-Ski is an important regulator in the activation of CAFs and may serve as a potential therapeutic target to block breast cancer progression.

Keywords: CAFs; P53; SDF-1; Tumor microenvironment; c-Ski.

Figure 1

Expression of c‐Ski in breast tumor fibroblasts. (A) Expression of c‐Ski in breast specimens was examined by IHC staining. Less or no staining of c‐Ski was found in normal breast. Moderate levels of c‐Ski were found in the fibroblasts of ducal carcinoma in situ (DCIS). The highest expression of c‐Ski was in the fibroblasts surrounding invasive and metastatic breast cancer (arrows). Scale bars, 200 μm. (B) The expression of c‐Ski in breast tumor cells and stroma were quantitatively displayed by pathology scoring. The data represent means ± s.d. from difference staining slides. (*P < 0.05, **P < 0.01, ANOVA followed by Student–Newman–Keuls test). (C) The data from Kaplan–Meier survival analysis shows the association of poor prognosis with c‐Ski expression.

Figure 2

Increase of c‐Ski promotes activation of breast NFs into CAFs. (A) Levels of c‐Ski and α‐SMA were determined by Western blot analysis. Higher levels of c‐Ski and α‐SMA were found in CAFs than in NFs. The data represent means ± s.d. from three different experiments. (*P < 0.01, Student t test). (B) c‐Ski expression in CAFs and NFs detected by immunofluorescence. Green: c‐Ski; Blue: DAPI. Scale bars, 200 μm. (C) Western blot analysis of c‐Ski, α‐SMA and FAP in parental NFs, parental CAFs, NFs with c‐Ski overexpression and corresponding control cells (Upper panel). β‐Actin is loading control. The relative fold change of c‐Ski, α‐SMA and FAP in cells is shown in the lower panel. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (D) Levels of FAP in fibroblasts were determined by immunohistochemistry staining. Scale bars, 200 μm.

Figure 3

c‐Ski overexpression enhances the proliferation, migration, invasion and contractile activities of fibroblasts. (A, B) Cellular proliferation assays were conducted by using either (A) MTT assay or (B) flow cytometric assay for NFs, NFs stably transfected with c‐Ski (NF/c‐Ski) or control vector (NF/Vec), and parental CAFs. The data represent means ± s.d. from three different experiments. (C, D) Cell invasion or migration were tested by Transwell Chamber (C) or wound healing assay (D) for parental NFs, parental CAFs, NF/c‐Ski and its control cells (NF/Vec). (**P < 0.01, Student t test.) (E) The contractile activity of each fibroblast was checked by contraction assays in collagen gel. Representative images of collagen gels are shown. (*P < 0.05, **P < 0.01, ANOVA followed by Student–Newman–Keuls test.)

Figure 4

c‐Ski activated NFs promote the migration and invasion of breast cancer MDA‐MB‐231 cells. (A) The migration of breast cancer MDA‐MB‐231 cells was tested by the wound healing assay. Cells were cultured in conditioned media (CM) derived from parental NFs, parental CAFs, c‐Ske‐activated NFs (NF/c‐Ski) and control cells (NF/Vec). The migration ability was determined as the percentage of non‐healed scratched area by using Image J software described in Material. The data represent means ± s.d. from three different experiments. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (B) The invasive potential was examined by the chambers coated with ECM Gel. MDA‐MB‐231 cells were seeded in the upper chambers of transwell, and CM derived from labeled fibroblasts was added into the lower chambers of transwell as chemoattractants. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test).

Figure 5

Silence of c‐Ski abrogates CAFs activation and function. (A) Levels of c‐Ski, α‐SMA, FAP in CAFs stably transfected with pLVX‐shRNA1‐Ski (CAF‐sh/c‐Ski) or control vector (CAF‐sh/Vec) were determined by Western blot analysis (Left panel). β‐Actin was used as loading control. The relative fold changes of proteins levels were quantified by density value relative to the loading control (Right panel). The data represent means ± s.d. from three different experiments. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (B) The invasion of breast cancer MDA‐MB‐231 cells was analyzed by transwell assays. CM collected from parental CAFs, parental NFs, CAFs‐sh/c‐Ski and its control cells (CAFs‐sh/Vec) were separately added into lower chambers of transwell as chemoattractants. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test.)

Figure 6

c‐Ski activates stromal fibroblasts through upregulating SDF‐1 and repressing p53. (A) Levels of 9 cytokines in parental NFs, parental CAFs, NFs stably transfected with c‐Ski (NF/c‐Ski) or control vector (NF/Vec) were analyzed using qRT‐PCR. The data represent means ± s.d. from three experiments. (*P < 0.01, Student t test). (B) SDF‐1 expression was determined by qRT‐PCR in CAFs, CAFs treated with Chalcone 4, CAFs with silenced c‐Ski (CAF‐sh/c‐SKi) and control cells (CAF‐sh/Vec). The relative fold change was compared with CAFs. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (C) Western blot analysis was used to test the levels of α‐SMA, FAP in cells treated with or without Chalcone 4 (5 μg/mL) for 48 h β‐Actin was used as a loading control. The relative fold change was compared with CAFs. (*P < 0.01, Student t test). (D) Levels of P53 and PTEN in cells were checked by qRT‐PCR. The relative fold change of P53 and PTEN in each cell was compared with NFs. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (E) qRT‐PCR was used to analyze the P53 and SDF‐1 expression in parental CAFs, CAFs transfected with pCMV‐HA‐p53 (CAF/p53) and control vector cells (CAF/Vec). The relative fold change of P53 and SDF‐1 in each cell was compared with parental CAFs. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (F) Western blot analysis was used to determine the expression of p53, c‐Ski, α‐SMA and FAP in parental CAFs, CAFs transfected with pCMV‐HA‐p53 (CAF/p53) and control vector cells (CAF/Vec). β‐Actin was used as an internal control.

Figure 7

c‐Ski governs invasion of MDA‐MB‐231 cells through SDF‐1 and p53. (A) Transwell assay was employed to test the invasion of MDA‐MB‐231 cells in conditional medium (CM) from each fibroblast. NFs: CM from parental NFs; NFs + SDF‐1: CM from NFs added with SDF‐1; NFs/c‐Ski: CM from NFs transfected with c‐Ski; NFs/c‐Ski + Anti‐SDF‐1: CM from NFs/c‐Ski treated with anti‐SDF1 neutralizing antibody; CAF: CM from parental CAFs; CAFs + Anti‐SDF‐1: CM from CAFs treated with anti‐SDF1 neutralizing antibody; CAF‐sh/c‐Ski: CM from CAFs with silenced c‐Ski by c‐Ski shRNA. CAF‐sh/c‐Ski + SDF‐1: CM from CAF‐sh/c‐Ski treated with human SDF‐1. (B) A quantitative graph of invasive cells measured in (A). Each bar represents the mean ± S.D. of three different experiments. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (C) Western blot analysis was used to test p53 level in cells. β‐Actin was used as an internal control. (D) The effect of fibroblast‐associated p53 on cancer cell invasion was tested. CM was added into the transwell lower chamber. The labeled cells were listed as: NF or CAF means parental cells; NF/p53 siRNA: NF with silenced p53 by siRNA; CAF/p53: CAF transfected with p53; NF/c‐Ski: NF transfected with c‐Ski; NF/c‐Ski/p53: NF with overexpressing c‐Ski and p53. The quantitative graph showed the invasive MDA‐MB‐231 cells. (*P < 0.01, ANOVA followed by Student–Newman–Keuls test). (E) A model depicting the role of c‐Ski in activation of NFs into CAFs and promotion of breast cancer invasion via p53 and SDF‐1.