Tumor-suppressive roles of ΔNp63β-miR-205 axis in epithelial-mesenchymal transition of oral squamous cell carcinoma via targeting ZEB1 and ZEB2.

We previously revealed that epithelial-to-mesenchymal transition (EMT) was mediated by ΔNp63β, a splicing variant of ΔNp63, in oral squamous cell carcinoma (OSCC). Recent studies have highlighted the involvement of microRNA (miRNA) in EMT of cancer cells, though the mechanism remains unclear. To identify miRNAs responsible for ΔNp63β-mediated EMT, miRNA microarray analyses were performed by ΔNp63β-overexpression in OSCC cells; SQUU-B, which lacks ΔNp63 expression and displays EMT phenotypes. miRNAs microarray analyses revealed miR-205 was the most up-regulated following ΔNp63β-overexpression. In OSCC cells, miR-205 expression was positively associated with ΔNp63 and negatively with zinc-finger E-box binding homeobox (ZEB) 1 and ZEB2, potential targets of miR-205. miR-205 overexpression by miR-205 mimic transfection into SQUU-B cells led to decreasing ZEB1, ZEB2, and mesenchymal markers, increasing epithelial markers, and reducing cell motilities, suggesting inhibition of EMT phenotype. Interestingly, the results opposite to this phenomenon were obtained by transfection of miR-205 inhibitor into OSCC cells, which express ΔNp63 and miR-205. Furthermore, target protector analyses revealed direct regulation by miR-205 of ZEB1 and ZEB2 expression. These results showed tumor-suppressive roles of ΔNp63β and miR-205 by inhibiting EMT thorough modulating ZEB1 and ZEB2 expression in OSCC.

We previously revealed that epithelial-to-mesenchymal transition (EMT) was mediated by ΔNp63β, a splicing variant of ΔNp63, in oral squamous cell carcinoma (OSCC). Recent studies have highlighted the involvement of microRNA (miRNA) in EMT of cancer cells, though the mechanism remains unclear. To identify miRNAs responsible for ΔNp63β-mediated EMT, miRNA microarray analyses were performed by ΔNp63β-overexpression in OSCC cells; SQUU-B, which lacks ΔNp63 expression and displays EMT phenotypes. miRNAs microarray analyses revealed miR-205 was the most up-regulated following ΔNp63β-overexpression. In OSCC cells, miR-205 expression was positively associated with ΔNp63 and negatively with zinc-finger E-box binding homeobox (ZEB) 1 and ZEB2, potential targets of miR-205. miR-205 overexpression by miR-205 mimic transfection into SQUU-B cells led to decreasing ZEB1, ZEB2, and mesenchymal markers, increasing epithelial markers, and reducing cell motilities, suggesting inhibition of EMT phenotype. Interestingly, the results opposite to this phenomenon were obtained by transfection of miR-205 inhibitor into OSCC cells, which express ΔNp63 and miR-205. Furthermore, target protector analyses revealed direct regulation by miR-205 of ZEB1 and ZEB2 expression. These results showed tumor-suppressive roles of ΔNp63β and miR-205 by inhibiting EMT thorough modulating ZEB1 and ZEB2 expression in OSCC.

| INTRODUCTION
Oral squamous cell carcinoma (OSCC) is one of the most common malignancies arising in oral cavity. Recent remarkable advancement of diagnostic modalities and reconstructive surgery led early detection of OSCC and radical operation for advanced cancer. However, the overall survival rate of OSCC patients has not risen satisfactory during last decade because of local aggressiveness and metastasis (Jerjes et al., 2010;Massano et al., 2006). Metastasis progresses through multi-step process which includes local cancer invasion, entry into the vasculature followed by the exit of cancer cells from the circulation and colonization at the distal sites (Spano, Heck, De Antonellis, Christofori, & Zollo, 2012;Woodhouse et al., 1997). Therefore, to clarify the molecular mechanism of cancer invasion as a first step of metastasis leads to improve the survival rate of OSCC patients by inhibiting metastasis.
Epithelial-to-mesenchymal transition (EMT) is a biological process characterized by biochemical and morphological changes that enable epithelial cell to acquire a mesenchymal cell phenotype during embryonic development and wound healing (Hay, 1995). Following EMT, migratory capacity is enhanced and production of extracellular matrix is greatly increased (Thiery, 2002). Recent reports have revealed that EMT is involved in cancer invasion and progression (Potenta, Zeisberg, & Kalluri, 2008). It has also been demonstrated that Snail, a zinc finger transcription factor involved in cancer cell EMT, directly suppresses E-cadherin expression (Batlle et al., 2000). Other zinc finger transcription factors, including zinc finger E-box binding homeobox (ZEB) 1, Slug and Twist have also been reported to repress E-cadherin expression and induce EMT (Adhikary et al., 2014;Sarrio et al., 2008;Yang et al., 2004).
Previously, we have also studied on the association of p63 gene, a homolog of the p53 tumor suppressor gene, with EMT in OSCC cells (Goto et al., 2014;Matsubara et al., 2011). p63 has two different promoter domains that generate two protein isoforms, TAp63 and ΔNp63 (Levine, Tomasini, McKeon, Mak, & Melino, 2011). In addition, each isoform yields three isotypes (α, β, and γ) generated by alternative splicing of the p63 COOH terminus (Yang et al., 2004). TAp63 transactivates p53 target genes that induce apoptosis by inhibiting cell proliferation in response to exposure to DNA-damaging agents (Crook, Nicholls, Brooks, O'Nions, & Allday, 2000). Conversely, ΔNp63 exerts dominant-negative activities against TAp63 and p53, and is thus considered an oncoprotein (Higashikawa et al., 2009). Our previous data demonstrated that down-regulation of ΔNp63 accompanied with EMT and that re-expression of ΔNp63β by stable cDNA transfection in OSCC cell lines reverted the EMT phenotype (Goto et al., 2014). However, the detail mechanisms of ΔNp63β-mediated EMT remain to be unclear. microRNAs (miRNAs) are small non-coding RNAs of 20-23 nucleotides in length that have a crucial role in post-transcriptional regulation of gene expression by binding to a target site in the 3ʹ-UTR of target mRNAs (Ambros, 2004). miRNAs are reported to regulate the expression of genes that mediate critical processes in tumorigenesis, such as differentiation, cell cycle regulation, apoptosis, and invasion (Esquela-Kerscher & Slack, 2006;Farazi, Hoell, Morozov, & Tuschl, 2013). Recent studies have shown that several miRNAs are also involved in EMT process of human cancer by targeting EMT-related transcription factors (Gregory, Bracken, Bert, & Goodall, 2008;Zaravinos, 2015). We thus sought to identify responsible miRNA associated with ΔNp63βmediated EMT using miRNA microarray analyses in this study. As samples for miRNA microarray, we used two clones established from an OSCC cell line SQUU-B, which lacks ΔNp63 expression and displays EMT phenotypes. One is SQUU-BO cell overexpressing ΔNp63β, and another is SQUU-BC cell transfected with empty vector (Goto et al., 2014). By comparing expression profiles of the two OSCC clones, we focused on miR-205, which is involved in epithelial differentiation (Ryan, Oliveira-Fernandes, & Lavker, 2006), because miR-205 expression was remarkably increased in SQUU-BO cells. Herein, we present some evidences that miR-205 is possibly involved in ΔNp63β-mediated EMT. whereas SQUU-A cells demonstrate expansive growth and have low metastatic ability (Morifuji, Taniguchi, Sakai, Nakabeppu, & Ohishi, 2000). SQUU-BO cells were generated by transfection with ΔNp63β expression vector, as previously described (Goto et al., 2014). All cell lines were maintained in a humidified atmosphere of 5% CO 2 at 37°C, and cultured in Dulbecco's modified Eagle's medium (DMEM)/F-12 (Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) and 100 units/ml penicillin/streptomycin.

| miRNA microarray analyses
To find miRNA associated with ΔNp63β-mediated EMT, total RNA including miRNA was extracted from SQUU-BO and SQUU-BC cells.
The 100 ng of total RNA from each sample was labeled using FlashTag™ Biotin HSR RNA Labeling Kit (Affymetrix, Santa Clara, CA), and hybridized to a Affymetrix GeneChip ® miRNA 4.0 Array according to the manufacturer's instructions. All hybridized microarray slides were scanned using an Affymetrix scanner. Relative hybridization intensities and background hybridization values were calculated using Affymetrix Expression Console™.

| Data analyses and filter criteria
We processed the raw CEL files for gene-level analysis with median polish summarization and quantile normalization by Affymetrix ® Transcriptome Analysis Console Software, and obtained normalized intensity values. To identify up or down-regulated genes, we calculated ratios (non-log scaled fold-change) from the normalized intensities of each gene for comparisons between SQUU-BC and SQUU-BO cells.
Then we established criteria for regulated genes with average signal levels of either more than 100: (up-regulated genes) ratio ≥2.0-fold, (down-regulated genes) ratio ≤0.5. In addition, prediction of target genes of miRNAs was performed using miRTarBase (http://mirtarbase.

| RNA extraction and complementary DNA (cDNA) synthesis
For mRNA analysis, total RNA was extracted from cultured cells using  Table 1. For miRNA detection, real-time PCR was performed with a miScript SYBR Green PCR Kit (QIAGEN). The primers for miRNAs were as follows: miR-205 (hsa-miR-205-5p, MS00003780, QIAGEN), and RNU6B (U6 small nuclear RNA 2, MS00033740, QIAGEN). For relative quantification, the 2 −ΔΔCt method, representing fold changes in the target genes normalized to a reference gene, was used. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and RNU6B were used as the internal control.

| Immunocytochemistry
Cultured cells were fixed in 75% methanol and then incubated with each primary antibody. The primary antibodies used were shown in Table 2.
The PVDF membrane was incubated with blocking buffer and then incubated separately with primary antibodies at 4°C overnight, followed by horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA) at room temperature for 1 hr. The primary antibodies used were shown in Table 2. The detection of specific proteins was carried out with enhanced chemiluminescence reagents (Chemi-Lumi One Super, Nacalai Tesque) and visualized using ImageQuant LAS 4000 (Fuji Film, Tokyo, Japan). β-actin was used as a positive control.

| miRNA mimic and inhibitor transfection
Cells were seeded at 2 × 10 5 cells per well in six-well plates and

| Target inhibition analyses of miRNA
We predicted the miR-205-binding sites in the 3ʹ-UTR of target gene by TargetScan (http://www.targetscan.org; August 2010 released) software (Agarwal, Bell, Nam, & Bartel, 2015;Lewis et al., 2005), and the specific complimentary sequence for the target site was synthesized using miScript Target Protector (QIAGEN). miR-205 mimic was co-transfected with the target or control protector in the SQUU-B cells according to the manufacturer's instructions. At 72 hr after transfection, cells were harvested for protein extraction.

| Wound healing assay
Cells were seeded in 24-well culture dishes and transfected. After 48 hr of the transfection, a wound was incised with a pipette tip in the central area of the confluent culture on the dishes. In order to inhibit cell proliferation, mitomycin C was added to cell cultures at 10 µg/ml for 2 hr after scratching. Detached cells were removed carefully with PBS and migration of cells into the wound areas was observed using a phase-contrast microscope (CKX41 NB-31PHP, Olympus, Tokyo, Japan). The area reduction rates were then calculated.

| Matrigel™ invasion assay
Cells were seeded at a density of 2 × 10 5 /well on 60-mm dishes, and

| Water-soluble tetrazolium (WST)-8 cell proliferation assay
Cell proliferation assays were performed using the Cell Count Reagent SF (Nacalai Tesque), according to the manufacturer's instructions. Cells were seeded at a density of 2 × 10 5 /well on 60-mm dishes, and then the cells were transfected with miR-205 mimic, miR-205 inhibitor, or control miRNA. After 48 hr, the transfected cells were replaced 2.0 × 10 3 cells/well into 96-well plates. After 12 hr of incubation, 10 µl of premixed reagent was added to each well. The plates were further incubated for 2 hr, and absorbance at 450 nm was then measured using a microplate reader (MULTISKAN FC, Thermo Fisher Scientific).

| Statistical analyses
All statistical analyses were performed with JMP software version 11 (SAS Institute, Tokyo, Japan). The Mann-Whitney U-test was also used to compare relative mRNA and miRNA expression levels by real-time PCR methods. It was also used to assess the significance of the differences between each group in the wound healing assay, cell proliferation assay, and invasion assay. A p-value of less than 0.05 was considered statistically significant.

| Identification of the miRNAs involved in ΔNp63β-mediated EMT in OSCC cells
To identify the responsible miRNAs involved in ΔNp63β-mediated EMT, miRNA microarray analyses were performed by using SQUU-BC and SQUU-BO cells. miRNA microarray analyses revealed that 16 miRNAs were significantly up-regulated and 12 miRNAs were down-regulated in SQUU-BO cells compared with SQUU-BC cells among 6599 miRNAs (see Table 3). The heat map and scatterplot showed differentially expressed miRNAs between two clones (Figure 1). In the results of miRNA microarray analyses, miR-205 was remarkably overexpressed in SQUU-BO cells compared with SQUU-BC cells. We thus focused on miR-205 for the potential target involved in ΔNp63β-mediated EMT.   (Figures 3a and 3b). Conversely,   (Figure 2e).

| Effects of miR-205 overexpression on EMT phenotype of OSCC cells
To determine the functional roles of miR-205 in ΔNp63β-mediated EMT, miR-205 mimic was transfected into SQUU-B cells for and ZEB2 were significantly decreased at both the gene and protein levels, while those of E-cadherin were increased (Figures 4a and 4b).
Although up-regulation of cytokeratin (CK) 19 and down-regulation of N-cadherin ( Figure 4a) were also found, expression of vimentin and fibronectin remained unaffected.
In the wound healing assay and Matrigel™ invasion assay, SQUU-B cells transfected with miR-205 mimic showed significantly lower migration and invasion abilities compared with the control cells (Mann-Whiteney U-test, p < 0.05) (Figures 4c and 4d). However, in the WST-8 assay, no significant difference in the cell proliferation was observed between the cells with overexpression of miR-205 and without (Figure 4e).

| Effects of miR-205 knockdown on EMT phenotype of OSCC cells
We

| Interfering of miR-205-binding sites in ZEB1 or ZEB2 mRNA
The results in this study suggested that miR-205 regulates ZEB1 and ZEB2 expression in OSCC cells. However, it remains unclear whether miR-205 directly regulates these expressions. Therefore, we predicted the binding sites of miR-205 in the 3ʹUTR of ZEB1 and RNAs as protectors, and performed target inhibition analyses ( Figure 6a). By the co-transfection of miR-205 mimic and ZEB1 or ZEB2 target protector into SQUU-B cells, the expression level of ZEB1 or ZEB2 proteins was recovered, though ZEB1 and ZEB2 expression was decreased when miR-205 mimic and control protector were co-transfected. (Figures 6b and 6c).

| DISCUSSION
In our previous study, we showed that down-regulated vimentin and downregulated E-cadherin expression was found in the oral cancer cells at the invasive front. Interestingly, the vimentin positive rate or the presence of decreased intensity of ΔNp63 was positively associated with the frequencies of metastases and poor prognosis in the OSCC patients (Goto et al., 2014 (Zaravinos, 2015). In this study, we thus focused on miR-205, remarkably increased in OSCC cells overexpressed ΔNp63β by miRNA microarray analyses.
miR-205 is transcribed from the MIR205HG gene that is located in the second intron of LOC642587 locus in chromosome 1 (Lim, Glasner, Yekta, Burge, & Bartel, 2003). It has already reported that miR-205 promotes human epidermal keratinocytes and corneal epithelial keratinocytes via the lipid phosphatase SHIP2 (Yu et al., 2010). On the other hands, growing evidences show that miR-205 plays a key role in a variety of tumors and functions as an oncogene or a tumor suppressor gene determined by the cancer context or its target genes (Qin et al., 2013). In gynecological malignancies, miR-205 was remarkably up-regulated in endometrial carcinoma tissues, and enhance tumor proliferation and invasion (Jin & Liang, 2015). Li et al. (2015) showed Previously, several studies demonstrated miR-205 expression and the significance as a biomarker in the head and neck squamous cell carcinoma (HNSCC) (Tran et al., 2007;Kimura et al., 2010). miR-205 was firstly identified as an up-regulated miRNA in HNSCC by microarray analysis (Tran et al., 2007). After that, Kimura et al. (2010) also demonstrated that miR-205 could be a specific biomarker for HNSCC. Although there were a few studies on function of miR-205 in HNSCC, most of these studies reported that miR-205 suppresses oncogenic activities such as antiapoptosis and proliferation (Kim et al., 2013(Kim et al., , 2014. Interestingly, Childs et al. (2009) showed that low-level expression of miR-205 significantly associated with loco-regional recurrence independent of disease severity.
They also described that combined low expression levels of miR-205 and let-7d, which target the Ras oncogene, are significantly correlated with poor prognosis in HNSCC. These data indicate close association of lowlevel miR-205 expression with cancer progression. In this study, miR-205 knockdown led to elevated cell motility in OSCC cells, suggesting that our result may be supportive for Childs's data.
We further showed that enhancing cell motility by miR-205 knockdown was associated with ΔNp63β-mediated EMT. Recent study demonstrated that loss of ΔNp63 and miR-205 enhanced cell migration and metastasis via targeting expression of ZEB1 in prostate cancer (Tucci et al., 2012). Furthermore, ΔNp63α-mediated expression of miR-205 contributed to the regulation EMT in bladder cancer cells, and that miR-205 prevented EMT through suppression of ZEB1 and ZEB2 (Tran et al., 2013). Matsushima et al. (2011) also reported that miR-205 modulates migration and invasion by regulating ZEB2 expression in esophageal SCC. Using miRNA target prediction algorithms, ErB3, E2F5, ZEB1, ZEB2, and protein kinase Cε have been identified as putative miR-205 targets (Gandellini et al., 2009). In fact, knockdown of miR-205 or ΔNp63 led to up-regulation of ZEB1 and ZEB2 in OSCC cells, and the expression level of ZEB1 or ZEB2 was recovered by the co-transfection of miR-205 mimic and ZEB1 or ZEB2 target protector into SQUU-B cells in this study. Previous studies employing a reporter assay confirmed miR-205 binding to the each 3ʹ-UTR of ZEB1 or ZEB2 (Matsushima et al., 2011;Niu, Shen, Zhang, Zhao, & Lu, 2015). Together, these data demonstrated that miR-205 directly inhibits ZEB1 or ZEB2 in OSCC cells.
As described above, it was revealed that ΔNp63α mediated miR-205 expression in bladder cancer (Tran et al., 2013). In this study, however, the miR-205 expression is regulated by not ΔNp63α but ΔNp63β. Previously, different effects of the three isoforms of ΔNp63 were showed by gene expression profiling of HNSCC cells overexpressing each ΔNp63 isoform (Boldrup, Coates, Gu, & Nylander, 2009). Thereby, ΔNp63β was most efficient activator of gene expression than ΔNp63α and ΔNp63γ despite low expression levels in HNSCC, whereas ΔNp63γ was most effective at repressing gene expression. These results thus indicated that miR-205 expression was induced by both ΔNp63α and ΔNp63β.
In conclusion, we clarify that ΔNp63β regulates miR-205 and that these effects contribute to EMT suppression through inhibiting ZEB1 and ZEB2 expression. To the best of our knowledge, this is the first report that elucidated the association of miR-205 with ΔNp63β-mediated EMT. Elucidating the functions of miR-205 in ΔNp63β-mediated EMT will facilitate the identification of valuable HASHIGUCHI ET AL..