BMP-6 promotes E-cadherin expression through repressing deltaEF1 in breast cancer cells

Abstract

Background: Bone morphogenetic protein-6 (BMP-6) is critically involved in many developmental processes. Recent studies indicate that BMP-6 is closely related to tumor differentiation and metastasis.

Methods: Quantitative RT-PCR was used to determine the expression of BMP-6, E-cadherin, and deltaEF1 at the mRNA level in MCF-7 and MDA-MB-231 breast cancer cells, as well as in 16 breast cancer specimens. Immunoblot analysis was used to measure the expression of deltaEF1 at the protein level in deltaEF1-overexpressing and deltaEF1-interfered MDA-MB-231 cells. Luciferase assay was used to determine the rhBMP-6 or deltaEF1 driven transcriptional activity of the E-cadherin promoter in MDA-MB-231 cells. Quantitative CHIP assay was used to detect the direct association of deltaEF1 with the E-cadherin proximal promoter in MDA-MB-231 cells.

Results: MCF-7 breast cancer cells, an ER+ cell line that expressed high levels of BMP-6 and E-cadherin exhibited very low levels of deltaEF1 transcript. In contrast, MDA-MB-231 cells, an ER- cell line had significantly reduced BMP-6 and E-cadherin mRNA levels, suggesting an inverse correlation between BMP-6/E-cadherin and deltaEF1. To determine if the same relationship exists in human tumors, we examined tissue samples of breast cancer from human subjects. In 16 breast cancer specimens, the inverse correlation between BMP-6/E-cadherin and deltaEF1 was observed in both ER+ cases (4 of 8 cases) and ER- cases (7 of 8 cases). Further, we found that BMP-6 inhibited deltaEF1 transcription, resulting in an up-regulation of E-cadherin mRNA expression. This is consistent with our analysis of the E-cadherin promoter demonstrating that BMP-6 was a potent transcriptional activator. Interestingly, ectopic expression of deltaEF1 was able to block BMP-6-induced transactivation of E-cadherin, whereas RNA interference-mediated down-regulation of endogenous deltaEF1 in breast cancer cells abolished E-cadherin transactivation by BMP-6. In addition to down-regulating the expression of deltaEF1, BMP-6 also physically dislodged deltaEF1 from E-cadherin promoter to allow the activation of E-cadherin transcription.

Conclusion: We conclude that repression of deltaEF1 plays a key role in mediating BMP-6-induced transcriptional activation of E-cadherin in breast cancer cells. Consistent with the fact that higher level of deltaEF1 expression is associated with more invasive phenotype of breast cancer cells, our collective data suggests that deltaEF1 is likely the switch through which BMP-6 restores E-cadherin-mediated cell-to-cell adhesion and prevents breast cancer metastasis.

Figure 1

Expression of BMP-6, E-cadherin, and δEF1 in MCF-7 and MDA-MB-231 breast cancer cells is inversely correlated. (a) Transcript levels of BMP-6, E-cadherin, and δEF1 in MCF-7 and MDA-MB-231 breast cancer cells were detected by RT-PCR analysis. GAPDH was used as an internal control. (b) The agarose gel electrophoresis data was quantified by UV band scanning. Data represent three independent experiments.

Figure 2

Relative expression levels of BMP-6, E-cadherin, and δEF1 in 16 breast cancer specimens. Transcript levels of the three genes were determined in tumor samples from 16 breast cancer specimens by quantitative RT-PCR. GAPDH was used to normalize the individual expression levels.

Figure 3

BMP-6 down-regulates δEF1 and concurrently up-regulates E-cadherin transcription in MDA-MB-231 breast cancer cells. (a) BMP-6-induced (200 ng/ml) down-regulation of δEF1 mRNA in MDA-MB-231 cells was verified by quantitative RT-PCR. GAPDH was used to normalize the δEF1 level. Data represent three independent experiments. (b) BMP-6-induced (200 ng/ml) up-regulation of E-cadherin mRNA in MDA-MB-231 cells was verified by quantitative RT-PCR. GAPDH was used to normalize the E-cadherin level. Data represent three independent experiments. (c) Western blot with anti-ZEB antibody showing δEF1 expression in MDA-MB-231 cells after culture for 24 or 48 h with BMP-6 (200 ng/ml). Actin expression was used as an internal control. (d) Western blot with anti-E-cadherin antibody showing E-cadherin expression in MDA-MB-231 cells after culture for 24 or 48 h with BMP-6 (200 ng/ml). Actin expression was used as an internal control.

Figure 4

Ectopic expression of δEF1 is sufficient to attenuate BMP-6-induced transactivation of the E-cadherin in MDA-MB-231 breast cancer cells. (a) Western blot with anti-myc antibody was performed to show δEF1-myc expression in MDA-MB-231 (Δ6B) and MDA-MB-231 (ΔδEF1) cells. Actin was used as a loading control. (b) MDA-MB-231 cells on a 6-well plate were co-transfected with a δEF1 expression plasmid (2 μg/well) and luciferase E-cadherin promoter constructs (2 μg/well) following treatment with 200 ng/ml rhBMP-6 after 24 h of transfection. The luciferase activity of the extracts was determined 24 h after BMP-6 induction using a Betascope analyzer. Luciferase values are normalized with Renilla activities. * indicates p < 0.05 in unpaired student t test when compared with vector alone. Data represent three independent experiments.

Figure 5

Scheme of the cloned promoter region of human E-cadherin. The putative δEF1-binding sites (E-box 1 and E-box 3) and mutations generated in the δEF1-binding sites are indicated.

Figure 6

Repression of endogenous δEF1 abolishes BMP-6-induced transactivation of the E-cadherin promoter in MDA-MB-231 breast cancer cells. (a) δEF1-specific siRNA plasmid (siRNA-δEF1) was introduced into MDA-MB-231 cells to generate δEF1-interfered stable transfectants. Control cells were treated with a scrambled siRNA. The efficiency of δEF1 protein knockdown was examined by western blot, using an anti-ZEB antibody. Actin was used as a loading control. (b) δEF1-interfered MDA-MB-231 cell were transiently transfected with luciferase E-caherin promoter constructs. After transfection for 24 h, cells were treated with 200 ng/ml rhBMP-6. The luciferase activity of the extracts was determined 24 h after BMP-6 treatment using a Betascope analyzer. Luciferase values are normalized with Renilla activities. * indicates p < 0.05 in unpaired student t test when compared with vector alone.

Figure 7

BMP-6-induced transactivation of E-cadherin is suppressed by δEF1 in a dose-dependent manner. Wild-type or mutated human E-cadherin promoter constructs were co-transfected with different amounts of δEF1 expression plasmid (1.5, 3.0, 6.0 μg/well) into MDA-MB-231 cells, followed by the treatment with 200 ng/ml rhBMP-6 after 24 h of transfection. The luciferase activity of the extracts was determined 24 h after BMP-6 treatment using a Betascope analyzer. Luciferase values are normalized with Renilla activities. Data represent three independent experiments. * indicates p < 0.05 in unpaired student t test when compared with vector alone. ** indicates p < 0.05 in one-way analysis of variance followed by Dunnett's test when compared with vector alone.

Figure 8

BMP-6 represses binding of endogenous δEF1 to the E-cadherin proximal promoter in MDA-MB-231 breast cancer cells. (a) Association of δEF1 with the E-cadherin promoter region was examined in vivo by CHIP analysis in MDA-MB-231 cells, using an unrelated anti-FLAG antibody or an anti-ZEB antibody directed against the N-terminal epitopes of δEF1. The amplified human E-cadherin promoter fragment is shown (-175/+21). (b) For quantitative CHIP assay, MDA-MB-231 cells were treated with or without 200 ng/ml rhBMP-6. Cell lysates were collected after 72 h. The IP was performed using anti-ZEB antibody (10 μg) with anti-FLAG antibody (10 μg) as a negative control. DNA fragments containing the E-cadherin promoter region (-175/+21) were amplified by quantitative PCR from anti-ZEB and anti-FLAG immunoprecipitated samples. Data represent three independent experiments. * indicates p < 0.05 in unpaired t-test when compared with un-treated group.