A reciprocal repression between ZEB1 and members of the miR-200 family promotes EMT and invasion in cancer cells

Abstract

The embryonic programme 'epithelial-mesenchymal transition' (EMT) is thought to promote malignant tumour progression. The transcriptional repressor zinc-finger E-box binding homeobox 1 (ZEB1) is a crucial inducer of EMT in various human tumours, and was recently shown to promote invasion and metastasis of tumour cells. Here, we report that ZEB1 directly suppresses transcription of microRNA-200 family members miR-141 and miR-200c, which strongly activate epithelial differentiation in pancreatic, colorectal and breast cancer cells. Notably, the EMT activators transforming growth factor beta2 and ZEB1 are the predominant targets downregulated by these microRNAs. These results indicate that ZEB1 triggers an microRNA-mediated feedforward loop that stabilizes EMT and promotes invasion of cancer cells. Alternatively, depending on the environmental trigger, this loop might switch and induce epithelial differentiation, and thus explain the strong intratumorous heterogeneity observed in many human cancers.

Figure 1

Knockdown of ZEB1 enhances expression of microRNAs. Growth patterns and staining for E-cadherin and β-catenin of characteristic short hairpin control (shCtl) and shZEB1 knockdown clones of breast (MDA-MB231) and colorectal (SW480 and HCT116) cancer cell lines used for the microRNA (miRNA) expression array screen. The mean relative ZEB1 expression levels of each two clones are indicated. The mean miRNA expression values (two hybridizations of each two independent sh control or shZEB1 clones per cell line) are shown. Knockdown of ZEB1 resulted in upregulation of the indicated miRNAs in all three cell lines. ZEB1, zinc-finger E-box binding homeobox 1.

Figure 2

ZEB1 directly suppresses transcription of miR-141 and miR-200c. (A) Schematic of the genomic organization of hsa-miR-141 and hsa-miR-200c on chromosome 12p13.31, and alignment of the highly conserved Z- and E-boxes. The putative ZEB1 binding sites, the cloned promoter and spacer sequence, as well as the region amplified for ChIP, are indicated. All numbers are relative to the start of the hsa-miR-200c stem–loop sequence. (B) Left panel: stable ZEB1 knockdown in the indicated undifferentiated colorectal (SW480), breast (MDA-MB231) and pancreatic (Panc-1) cell clones enhances the activity of the putative promoter, but does not or only weakly enhances the activity of the spacer sequence. The mean values of the corresponding short hairpin control (shCtl) clones were set to 1. Middle panel: overexpression of ZEB1 and to a lesser extent that of Snai1 in HCT116 cells repress the activity of the promoter but not the spacer sequence in a dose-dependant manner. The absolute transcriptional activity of the promoter was 55-fold higher than the activity of the spacer sequence. Right panel: mutations (mut) of the Z-boxes render the promoter less sensitive to suppression by ZEB1. (C) Electromobility shift assay using recombinant DNA-binding domain of ZEB1 (left panel) or nuclear extracts from SW480 colorectal cancer cells (right panel) and the indicated labelled probes. Mutation of the two conserved Z-boxes (Z1, Z2) strongly reduced the specific binding complex (asterisk). Three different antisera against ZEB1, but not control serum, supershifted (arrow) a specific complex (asterisk), which is missing in the mutated probes, indicating that it contains ZEB1. The known ZEB1-binding site of the interleukin-2 promoter (NRE) was used as a positive control. (D) ChIP with two different ZEB1 antisera shows in vivo binding of ZEB1 to the putative promoter in the indicated colorectal cancer cell lines. ChIP, chromatin immunoprecipitation; Gapdh, glyceraldehyde-3-phosphate dehydrogenase; WT, wild type; ZEB1, zinc-finger E-box binding homeobox 1.

Figure 3

miR-141 and miR-200c induce mesenchymal–epithelial transition. (A) Overexpression of both microRNAs (miRNAs) in undifferentiated cancer cell lines induces a mesenchymal–epithelial transition, indicated by upregulation of E-cadherin (green), partial membranous translocation of E-cadherin and restoration of basal–apical polarity (see z-mode). The effect was stronger for miR-200c and comparable with the effect of a ZEB1 knockdown. (B) Upper panel: miR-200c suppressed migration of MDA-MB231 cells. Lower panels: picture and quantification of a Matrigel invasion assay. Transfection with miR-200c strongly suppressed invasion of MDA-MB231 into Matrigel, comparable with short interfering (si) RNA-mediated knockdown of ZEB1. (C) Differentiated cancer cell lines transfected with miRNA inhibitors against miR-141 and miR-200c show increased cell scattering, upregulation of vimentin (HPAF2) or reduced expression of E-cadherin (DLD1). By contrast, the epithelial phenotype induced by stable knockdown of ZEB1 (SW480 shZEB1) cannot be reversed by miRNA inhibitors. sh, short hairpin; ZEB1, zinc-finger E-box binding homeobox 1.

Figure 4

miR-141 and miR-200c reduce expression of factors promoting epithelial–mesenchymal transition and invasion. (A) Predicted duplex formation of TGFβ2 3′UTR with the miR-141 stem–loop and of two ZEB1 3′UTR regions with the miR-200c stem–loop. The putative recognition sites are highly conserved among various species. (B) Real-time reverse transcription–PCR of various putative target factors after transfection of the indicated miRNAs or short interfering (si) RNA in undifferentiated cancer cells. Note strongest reduction of TGFβ2 expression after miR-141 transfection and that of ZEB1 after miR-200c transfection. (C) Confirmation of the results on protein levels by immunoblots. (D) Luciferase reporter assay showing reduced activity of 3′UTR-luciferase reporter constructs in the indicated SW480 short hairpin ZEB1 (shZEB1) clones with increased levels of miR-141 and miR-200c compared with shCtl clones. (E) Overexpression of miR-141 and miR-200c in SW480 cells leads to decreased activity of TGFβ2 3′UTR-luciferase and ZEB1 3′UTR-luciferase constructs, respectively. (F) Model of a feedforward loop linking ZEB1 and counteracting miRNAs, in particular miR-200c. As ZEB1 and both miRNAs repress each other, they are linked in a feedforward loop. Depending on the initial signal, this loop could stabilize either mesenchymal or epithelial differentiation. In tumours, environmental TGFβ could trigger upregulation of ZEB1, promoting a self-enhancing loop in tumour cells, finally resulting in EMT and invasion. Once the initial signal breaks down, the loop could induce a reversal of the epithelial phenotype. This might explain the strong phenotypic heterogeneity often seen within individual tumours and metastases. EMT, epithelial–mesenchymal transition; MET, mesenchymal–epithelial transition; miRNA, microRNA; TGF, transforming growth factor; UTR, untranslated region; ZEB1, zinc-finger E-box binding homeobox 1.