Significance of the NOR1-FOXA1/HDAC2-Slug regulatory network in epithelial-mesenchymal transition of tumor cells

Abstract

The epithelial-mesenchymal transition (EMT) process is believed to play a crucial role in nasopharyngeal carcinoma (NPC) progression, a squamous cell carcinoma of the head and neck with the tendency to metastasize early. At present, much attention has been given to the inducer of EMT involved in NPC progression, while antagonists have been less intensively characterized. In this study, unbiased analysis of EMT-associated gene expression patterns was performed using data mining of global gene expression profiles derived from NPC samples, leading to the successful identification of NOR1, FOXA1, and Slug, all of which showed aberrant expression during NPC progression. The effect of tumor suppressor NOR1 on Slug-induced NPC cells during the EMT process was investigated by use of ectopic expression and RNA interference methods. The molecular mechanisms underlying the tumor-suppressing effect of NOR1 on Slug-induced EMT were thought to be dependent on the cooperation of NOR1 with the FOXA1-HDAC2 complex. We also showed that FOXA1 and HDAC2 bind the slug promoter and directly repress its transcription. Our data revealed a previously unrecognized role of the NOR1-FOXA1/HDAC2-Slug network in the regulation of the EMT process and aggressiveness of NPC.

Keywords: FOXA1; NOR1; epithelial-mesenchymal transition; nasopharyngeal carcinoma.

Figure 1. Unbiased analysis of EMT-associated gene mRNA levels by data mining of the NPC GEO dataset

(A) Box plot showing the mRNA levels of EMT-associated molecules in NPC tissues. These data were collected from the global gene expression profile data set GSE12452, which contains 31 NPC and 10 normal nasopharyngeal tissue samples examined with a Human Genome U133 Plus 2.0 Array (HG-U133 Plus 2) from Affymetrix. (B) Correlations between differentially expressed EMT-associated genes in NPC. The levels of Slug mRNA were shown to be inversely correlated with keratin 4 and NOR1 mRNA levels. NOR1 mRNA levels were shown to be positively correlated with keratin 4 and FOXA1.

Figure 2. Lost expression of NOR1 and FOXA1 protein and increased immunoreactivity of slug protein in NPC cells

Intensive nuclear immunohistochemical staining for Slug protein was found in NPC samples, while only weak staining for Slug protein was found in NPE or Ad-NPE cells. In contrast, intensive cytoplasmic immunohistochemical staining for NOR1 and intensive nuclear immunohistochemical staining for FOXA1 were found in NPE or Ad-NPE cells, but expression of NOR1 and FOXA1 proteins was lost in NPC samples.

Figure 3. NOR1 suppresses the slug-induced EMT process and invasiveness of NPC cells

(A) qRT-PCR analysis of the EMT transcription factor Slug and the epithelial marker keratin in NPC 5–8F, and HNE1 cells expressing control vectors pIRES or pIRES/NOR1. (B) Tumor cell migration assay. Ectopic expression of NOR1 suppresses migration of NPC 5-8F and HNE1 cells, but migration was rescued by concomitant transfection with a Slug expression plasmid. The graph represents summarized data of five independent experiments. Expression of exogenous Slug protein in transfected cells was determined by western blotting. (C) Tumor cell invasion assay. The same cells described in (B) were subjected to a tumor cell invasion assay. The graph provides summarized data of five independent experiments. **P < 0.01, ***P < 0.001 relative to the control cell data.

Figure 4. Stable silencing of endogenous NOR1 induces EMT-like properties, migration, and invasion of HeLa cells via upregulation of slug

(A) qRT-PCR analysis of the EMT transcription factor Slug and the epithelial markers keratin in HeLa cells with or without NOR1 expression. (B) Tumor cell migration assay. Stable silencing of NOR1 promotes migration of HeLa cells, but migration was alleviated by concomitant inhibition of Slug. (C) Tumor cell invasion assay. The same cells described in B were subjected to a tumor cell invasion assay. The graph provides summarized data of five independent experiments. (D) Scrambled shRNA or NOR1 shRNA along with Slug siRNA-transfected HeLa cells were subjected to western blotting. **P < 0.01, ***P < 0.001 compared to the control cells.

Figure 5. NOR1 expression induces chromatin remodeling of slug promoter regions in cancer cells

ChIP-qPCR assay: tumor cells with or without NOR1 expression were subjected to ChIP with antibodies specific to acetyl-H3K9 (H3K9Ace), trimethyl-H3K9 (H3K9Me3), or control IgG, followed by amplification with three PCR primers sets(HS-P1, HS-P2 and HS-P3) specific for the slug promoter regions. Data show the representative results from three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 compared to the control cells.

Figure 6. NOR1 inhibition of slug expression and malignant behaviors of cancer cells are dependent on FOXA1 and HDAC2

(A) qRT-PCR assay; ectopic expression of NOR1 or silencing of endogenous NOR1 leads to upregulation or downregulation of FOXA1 and HDAC2 mRNAs in cancer cells, respectively. (B) Western blot assay; ectopic expression of NOR1 or silencing endogenous NOR1 leads to up-regulation or down-regulation of FOXA1, HDAC2 proteins in cancer cells respectively. (C) Western blot assay; silencing either FOXA1 or HDAC2 using siRNAs results in recovery of Slug protein in NOR1-expressing NPC cells. Conversely, ectopic expression of FOXA1 protein leads to decreased Slug protein levels in NOR1-silenced HeLa cells. (D) Tumor cell migration assay; silencing either FOXA1 or HDAC2 using siRNAs rescues the migration of NOR1-expressing NPC 5–8F and HNE1 cells, while ectopic expression of FOXA1 suppresses the migration of NOR1-deficient knockdown HeLa cells. (E) Tumor cell invasion assay. The same cells described in (D) were subjected to a tumor cell invasion assay. The graph provides summarized data of five independent experiments. *P < 0.05, ***P < 0.001 compared to the control cells.

Figure 7. Either FOXA1 or HDAC2 negatively regulates slug transcription

(A) A schematic illustration of slug gene structure and the potential binding sites for different primer sets. Seven primer sets named as F1S-P1~ F1S-P7 were used for validation of FOXA1 binding sites on Slug promoter regions and downstream 3-UTR region. Three primer sets named as HS-P1~ HS-P3 were used for amplification slug gene DNA fragments associated with acetylated-H3K9. (B) ChIP was performed by using anti-FOXA1 antibody, and qPCR following ChIP was performed to identify the binding sites in the Slug promoter and downstream 3-UTR region in 5–8F cells by using seven primer sets(F1S-P1~ F1S-P7). Normal rabbit IgG was used as a negative control. The quantification of the band intensity of PCR products is shown on the right side. (C) Activation of slug promoter activity by co-transfection with FOXA1-targeted siRNAs in 5–8F cells. FOXA1 protein levels were measured by western blotting. Relative luciferase activities of the slug promoter are shown. Values are the mean of three independent assays, and the error bar is the standard deviation of the mean. (D) ChIP was performed using anti-acetylated-H3K9 antibody, and qPCR following ChIP was performed to compare the changes of slug promoter-associated H3K9-acetylation levels after silencing HDAC2 in 5-8F cells. Increased slug promoter-associated H3K9-acetylation levels was observed by silencing of HDAC2 in 5–8F cells. HDAC2 protein levels were measured by western blotting. (E) Activation of slug promoter activity by co-transfection with a HDAC2-targeted siRNAs in 5–8F cells. Relative luciferase activities of the slug promoter are shown. Values are the mean of three independent assays, and the error bar is the standard deviation of the mean. **P < 0.01, ***P < 0.001 compared to the control cells.

Figure 8. Schematic illustration of the NOR1-FOXA1/HDAC2-SLUG regulatory network in the EMT process of NPC

In normal nasopharyngeal epitheliums, high levels of NOR1 expression lead to optimal expression of FOXA1; FOXA1 then translocates to the nucleus, recruits HDAC2, and binds to the slug promoter region, resulting in de-acetylation and secondary tri-methylation of Slug-associated H3K9. The transcription of Slug is repressed and Slug protein levels are decreased, so the cells are maintained as epithelial cells. During NPC development, NOR1 is silenced by DNA hypermethylation or deletions, which in turn results in a decrease in the expression of FOXA1 and HDAC2, thus removing the inhibitory effect on the transcription of slug. The transcription of slug is activated and Slug protein production is upregulated, subsequently leading to the Slug-induced EMT process. HMTs, histone methyl transferases. JMJD, Jumonji domain–containing protein histone demethylases. HATs, histone acetyltransferases.