ZEB1/miR-200c/AGR2: A New Regulatory Loop Modulating the Epithelial-Mesenchymal Transition in Lung Adenocarcinomas

Abstract

Epithelial-mesenchymal transition (EMT) is a process involved not only in morphogenesis and embryonic development, but also in cancer progression, whereby tumor cells obtain a more aggressive metastatic phenotype. Anterior gradient protein 2 (AGR2) maintains the epithelial phenotype and blocks the induction of EMT, thus playing an undeniable role in tumor progression. However, the mechanism through which AGR2 expression is regulated, not only during EMT, but also in the early stages of cancer development, remains to be elucidated. In the present study, we show an inverse correlation of AGR2 with ZEB1 (zinc finger enhancer binding protein, δEF1) that was verified by analysis of several independent clinical data sets of lung adenocarcinomas. We also identified the ZEB1 binding site within the AGR2 promoter region and confirmed AGR2 as a novel molecular target of ZEB1. The overexpression of ZEB1 decreased the promoter activity of the AGR2 gene, which resulted in reduced AGR2 protein level and the acquisition of a more invasive phenotype of these lung cancer cells. Conversely, silencing of ZEB1 led not only to increased levels of AGR2 protein, but also attenuated the invasiveness of tumor cells. The AGR2 knockout, vice versa, increased ZEB1 expression, indicating that the ZEB1/AGR2 regulatory axis may function in a double negative feedback loop. In conclusion, we revealed for the first time that ZEB1 regulates AGR2 at the transcriptional level, while AGR2 presence contributes to ZEB1 mRNA degradation. Thus, our data identify a new regulatory mechanism between AGR2 and ZEB1, two rivals in the EMT process, tightly associated with the development of metastasis.

Keywords: AGR2; EMT; ZEB1; cancer.

Figure 1

AGR2 negatively correlates with ZEB1 in human cancer cell lines and clinical samples. Determination of AGR2 mRNA expression with respect to (A) CDH1 and (B) ZEB1 mRNA levels extracted from CBioPortal database containing the Cancer Cell Line Encyclopedia. The value of Spearman’s and Pearson’s correlation coefficient was generated by the CBioPortal database using the default set-up. Scatter plots showed a positive correlation with epithelial CDH1 mRNA (Spearman r = 0.46; Pearson r = 0.47) and a negative correlation between AGR2 and ZEB1 mRNA levels (Spearman r = −0.43; Pearson r = 0.46) in cancer cell lines.

Figure 2

ZEB1 regulates AGR2 at mRNA and protein level. (A) The expression of ZEB1 was silenced with a specific siRNA (siZEB1) and the changes in AGR2 protein level were determined by Western blot analysis. β-actin served as a loading control. (B) Quantitative analysis of AGR2 mRNA was performed in A549 and H1299 cells transfected with specific siRNAs against ZEB1 or control siRNA (siCTR). 18S rRNA (ribosomal RNA) served as an endogenous control for data normalization. (C) Western blot analysis in A549 cells transfected with plasmid coding for ZEB1 or control plasmid (CTR). (D) A549 and H1299 cells were transfected with plasmid coding for ZEB1 to determine the effect of enhanced ZEB1 expression on AGR2 mRNA levels by Real-Time PCR (polymerase chain reaction). 18S rRNA served as an endogenous control for data normalization. Data plotted into the graphs are the mean ± standard deviation (SD) obtained from three independent experiments. * p ≤ 0.05, ** p ≤ 0.01, **** p ≤ 0.0001. Full Western Blots are shown on Figure S5.

Figure 3

ZEB1 directly binds to the promotor of the AGR2 gene to repress its expression under basal conditions. (A) Schematic representation of potential E-boxes located in the AGR2 promoter. Blue arrows indicate primer binding sites used for the chromatin immunoprecipitation (ChIP) assay. (B) ZEB1 associates with the AGR2 promoter in A549 cells as shown by ChIP analysis. Input DNA samples were used for normalization and rabbit ZEB1 antibody or control immunoglobulins (IgG) were used for ChIP experiments. * p ≤ 0.05. (C) AGR2 negative cells H1299 were transfected with specific siRNA against ZEB1 (siZEB1) or control siRNA (siCTR) for 24 hours and with plasmid coding for AGR2 promoter sequence from −1584 to +94 or with control pGL3-luc plasmid for additional 24 hours and the luciferase signal was determined by a spectrophotometer, whereas a Renilla luciferase empty plasmid was used for normalization of the transfection efficiency. Data plotted into the graphs are the mean ± standard deviation (SD) obtained from three independent experiments. ** p ≤ 0.01.

Figure 4

The knockout of AGR2 alters the mRNA expression of ZEB1. Comparison of ZEB1 mRNA (left) and ZEB1 protein levels (right) in (A) A549 scr and A549 KOAGR2 cells, (B) H1299 and H1299 AGR2 cells, (C) HEK-293 CTR and HEK-293 AGR2 cells, (D) A431 CTR and A431 AGR2 cells. The data were normalized using 18S rRNA for mRNA and β-actin for protein expression. The mRNA levels of ZEB1 were determined relative to its mRNA level in A549 scr cells. The results represent the mean ± SD of at least two independent experiments (A,B) and one experiment (C,D), respectively performed in technical triplicates; ** p ≤ 0.01, *** p ≤ 0.001. (E) Cells were transfected with Renilla luciferase reporter plasmid pLuc-CDS coding full-length ZEB1 to analyze the relative amount of ZEB1 in A549 scr and KOAGR2 cells either treated or untreated with TGF-β. (F) Immunochemical analysis of ZEB1 protein levels in cells either untreated or treated with TGF-β. (G) The subcellular fractionation and immunochemical analysis of ZEB1 protein in nuclear and cytoplasmic fraction. α-tubulin was used as a marker of cytoplasmic fraction and Lamin B1 was used as a marker for nuclear fraction. Full Western Blots of (A,B) are shown on Figure S6. Full Western Blots of (F,G) are shown on Figure S7.

Figure 5

AGR2 regulates the stability of the ZEB1 mRNA level. Analysis of the ZEB1 mRNA level in (A) A549 scr and A549 KOAGR2 cells and (B) H1299 and H1299 AGR2 cells exposed to actinomycin D as indicated. (C) RT-qPCR analysis of the miR-200c level in A549 scr and A549 KOAGR2 cells. RNU48 was used as housekeeping control for microRNA (miRNA) levels. ** p ≤ 0.01 (D) Protein-protein immunoprecipitation with a specific antibody recognizing AGR2, followed by immunochemical analysis of hnRNPU protein in A549 cells. A non-specific antibody was used as a negative control (NC) for the immunoprecipitation experiments. Full Western Blots are shown on Figure S8.

Figure 6

The alteration in AGR2 expression regulates an aggressive phenotype in vitro and in vivo. (A) Illustration of primary tumors developed in mouse xenografts. A549 scr (n = 5) and A549 KOAGR2 (n = 6) cells were injected into the flanks of mice with severe combined immunodeficiency (SCID). (B) Growth rate of tumors developed from A549 scr and A549 KOAGR2 cells. (C) The graph showing the frequency of metastasis development derived from Table S1. (D) An invasion assay was performed to analyze the number of cells that invaded through a semi permeable membrane. Invaded cells were stained and measured by spectrophotometer. The results of invasion experiments are an average of 3 technical replicates from 2 independent experiments, plotted as a mean ± SD where ** is ** p ≤ 0.01 and “ns” means statistically non-significant (p> 0.05).