miR-203 Inhibits Frizzled-2 Expression via CD82/KAI1 Expression in Human Lung Carcinoma Cells

CD82/KAI1, a member of the tetraspanin superfamily, is a suppressor of metastasis and CD82 inhibits canonical Wnt signaling via downregulation of several Frizzled (FZD) isoforms, resulting in accumulation of β-catenin at the cell membrane. In this study, we investigated the mechanism through which CD82 inhibited FZD expression by examining the effects of microRNAs (miRNAs). The miRanda algorithm predicted 11 miRNAs from FZD sequences. Among these miRNAs, CD82 caused upregulation of miR-203 (by 2.095-fold) and downregulation of miR-338-3p (by 0.354-fold) as compared with control cells. Transfection with miR-203 and miR338-3p mimics or inhibitors revealed that miR-203 downregulated FZD2 mRNA (by 0.268-fold) and protein expression (by 0.701-fold). Moreover, transfection with the miR-203 mimic also inhibited cell migration. Therefore, these findings suggested that CD82 enhanced the expression of miR-203 and directly downregulate FZD2 expression, suppressing cancer metastasis by inhibition of the Wnt signaling pathway.


Introduction
Cancer metastasis is the most crucial event affecting patient prognosis. This multistep event involves the migration of cancer cells from the primary site, avoidance from host defense systems, and subsequent growth of cancer cells at secondary sites. These malignant characteristics of cancer cells are controlled by a complex network of cancer-specific signaling pathways. Thus, elucidation of the signaling mechanisms controlling cancer cell migration and malignant characteristics will provide important insights into the development of chemotherapeutic agents to improve prognoses.
Among the pathways involved in metastatic signaling, the Wnt signaling pathway is a primary target. The first step in this pathway is the binding of secreted Wnt ligands to their receptor, Frizzled (FZD), a seven-pass transmembrane-type receptor with 10 isoforms (FZD1-FZD10) in humans. Binding of Wnt ligands to the FZD-LRP5/6 complex leads the phosphorylation of LRP6 and recruits the Axin-related protein complex. This protein complex then releases β-catenin, which acts as an effector of the canonical Wnt signaling pathway. Released β-catenin in the cytosol accumulates and then transduces the Wnt signal to the nucleus thorough the TCF/LEF complex. In contrast, the noncanonical pathway, which is independent of β-catenin, proceeds through Wnt/planar cell polarity (PCP) and Wnt/protein kinase C (PKC). Both of these Wnt signal pathways are crucial for embryonic development and cancer progression [1]. For example, the expression of FZD1 and -2 is correlated with cancer malignancy and prognosis in breast cancer [2] and colon cancer [3]. However, despite the importance of the biological functions of these proteins, little is known about the gene regulatory mechanisms of Wnts and FZDs.
CD82/KAI1, a member of the tetraspanin superfamily, was originally recognized as a T-cell activation accessory molecule [4]. CD82 is also known to suppress metastasis during cancer progression [5]. Tetraspanins associate with cell surface receptors or proteins and modulate their functions. Moreover, our previous studies revealed a novel function for CD82 in E-cadherin-mediated cellular adhesion [6]. CD82 inhibitsβ-catenin tyrosine phosphorylation and increases the accumulation of E-cadherin/β-catenin complexes at the cell membrane by stabilization of the complex. This function strengthens hemophilic cancer cell adhesion in the primary cancer nest and inhibits invasion and metastasis. We have also revealed that CD82 attenuates Wnt signaling by downregulation of FZD2, -3, -5, -7, and -9 expression without regulating Wnt orβ-catenin expression, leading to inhibition ofβ-catenin nuclear translocation [7].
MicroRNAs (miRNAs) are a class of small noncoding RNAs (~22 nucleotides) that play important roles in the regulation of gene expression. miRNAs induce gene silencing by binding to target sites found within the 3 0 -untranslated region (UTR) of the target mRNA. This gene silencing leads to inhibition of protein production by suppressing protein synthesis and/or by mRNA degradation. Nearly 50% of human miRNAs are located in cancer-associated genomic regions or in fragile sites [8]. In many types of human cancer, miRNAs behave as oncogenes or anti-oncogenes, termed called onco-miRs and anti-onco-miRs, respectively [9][10][11]. However, whether miRNAs regulate the Wnt signaling pathway in lung cancer is not yet known.
In this study, we examined the roles of miRNAs in the regulation of FZD expression through CD82 in the Wnt signaling pathway.

Cell culture
The human cell line h1299 (a non-small cell lung carcinoma cell line) and its transfectant derivatives (h1299/zeo and h1299/CD82) were established in our laboratory by transfection of a control vector or CD82 cDNA and cell sorting-based clone selection techniques, as described previously [12]. h1299/zeo was a mock transfectants cell line exhibiting weak CD82 expression, while h1299/CD82 overexpressed CD82. The protein levels of CD82 in h1299/CD82 cells, as assessed by immunoblotting, were 20 times higher than that in wild-type and h1299/zeo cells, and its cell surface expression, as assessed by flow cytometry, was approximately 9-fold higher than that in wild-type and h1299/zeo cells. The cell lines used in this study were maintained in Dulbecco's modified Eagle's medium (DMEM; Sigma, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS; ICN Biomedicals, Aurora, OH, USA) and 2 mM L-glutamine at 37°C and in an atmosphere of 5% CO 2 .
Transfection with miRNA mimics or miRNA hairpin inhibitors h1299/zeo and h1299/CD82 cells were seeded at 2 x 10 5 cells per well in 6-well plates and transfected with miRIDIAN microRNA Mimics, miRIDIAN microRNA Hairpin Inhibitors, or controls (Thermo Fisher Scientific Inc., Waltham, MA, USA). All miRNA mimics, miRNA hairpin inhibitors, and controls were transiently transfected into cells using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions.

Prediction of possible miRNAs using a web-based application
We searched for miRNAs that potentially targeted FZDs and Wnts using a web-based prediction application, miRanda (http://www.microrna.org; released in August 2010) [13,14]. This program is a database search engine that is commonly used to search for target sites for miR-NAs. The algorithm detects partial complementary base sequences. We first searched for miR-NAs that would be expected to target FZD2, -3, -5, -7, and -9. From the possible target miRNAs, we excluded miRNAs that targeted other FZD isoforms and Wnts.

Real-time RT-PCR of mRNAs
Total RNA was extracted from h1299 cells using TRIzol reagent (Invitrogen) and used for firststrand cDNA synthesis. The mRNA levels were quantified in triplicate using a real-time PCR system with a LightCycler FastStart DNA Master SYBER Green 1 kit (Roche Diagnostics, Mannheim, Germany). Specific primers for FZDs were as follows: The PCR cycling conditions were 10 min at 95°C for 1 cycle, followed by 45 cycles at 95°C for 30 s, 60°C for 30 s, and 72°C for 60 s. Amplicons are confirmed that signals are unique by dissociation curve analyses. Expression levels were normalized to theβ-actin mRNA level of each sample, as obtained from parallel assays. The PCR cycling conditions were 15 min at 95°C for 1 cycle, followed by 45 cycles at 94°C for 15 s, 55°C for 30 s, and 70°C for 30 s. Amplicons are confirmed that signals are unique by dissociation curve analyses. RNU6B was used as miRNA endogenous control.

Target inhibition analysis of miRNA
We predicted the target site of miR-203 in the 3 0 UTR of FZD2 mRNA by miRanda software, and the specific complimentary sequence for the target site was synthesized using miScript Target Protector (Qiagen). An miR-203 mimic was transfected into h1299 cells with various concentrations of target protector according to the manufacturer's instructions. At 48 h after transfection, FZD2 mRNA levels were measured by real-time PCR.
Wound healing assay h1299/zeo and h1299/CD82 cells were seeded at 2 × 10 5 cells per well in 6-well plates and transfected with mimic miRNAs, hairpin inhibitor miRNAs, or controls. At 48 h after transfection, wounds were created by scratching with a 200-μL pipette tip. Cell monolayers were then washed with medium, and wounds were observed under a fluorescence microscope (BZ-8000; Keyence, Osaka, Japan). The wound regions were photographed again after 8, 16, or 24 h in order to measure the wound area. Wound area was standardized by using the following formula: wound area (% of control) = wound area after the indicated period × 100 / initial wound area. All experiments were carried out in triplicate and repeated three times.

Statistical analysis
Data are shown as means ± SDs. All data were analyzed via Student's t-test in SPSS 13.0 software. Differences with P values of less than 0.05 were considered significant.

Expression levels of the 11 selected miRNAs in h1299 cells
Next, we used real-time PCR to confirm the expression of the 11 selected miRNAs in h1299 cells. miR-338-3p was significantly downregulated in h1299/CD82 cells (0.354-fold), whereas miR-203 was significantly upregulated (2.095-fold), as compared with h1299/zeo cells. Additionally, knockdown of CD82 by shRNA in h1299/CD82 cells allowed recovery of miRNA levels to those in h1299/zeo cells. These results suggested that CD82 specifically regulated these miRNAs. No significant differences in miR-27a, miR-27b, miR-145, miR-185, miR-197, miR-221, miR-222, miR-376a, or miR-376b expression levels were observed between h1299/zeo and h1299/CD82 cells (Fig 1).  Effects of miR-338-3p and miR-203 on the expression levels of FZD mRNAs in h1299 cells To examine the functional effects of miR-338-3p and miR-203 on the regulation of FZDs, we transiently transfected h1299/zeo and h1299/CD82 cells with miR-338-3p or miR-203 mimics or inhibitors. Forty-eight hours after transfection, miRNA levels of miR-338-3p and miR-203 were significantly decreased in h1299/zeo cells transfected with the miR-338-3p inhibitor and h1299/CD82 cells transfected with the miR-203 inhibitor (Fig 2A and 2D). Transfection with miR-338-3p and miR-203 mimics significantly increased the miRNA levels in h1299/CD82 cells transfected with miR-338-3p mimic and in h1299/zeo cells transfected with the miR-203 mimic (Fig 2B and 2C). Next, we examined the expression levels of FZD mRNAs in these transfected h1299 cells. In cells transfected with miR-338-3p mimic (h1299/CD82/338-3p mimic) or hairpin inhibitor Total RNA was isolated from h1299 cells, and the 11 predicted miRNA levels were analyzed by real-time RT-PCR. RNU6B was used as an internal reference gene. Experiments were performed in triplicate, and relative miRNA levels (zeo = 1) were averaged. Asterisks indicate statistically significant differences (*p < 0.05, **p < 0.01) between the two values. Data are presented as the means ± SDs. (h1299/zeo/338-3p inhibitor), the mRNA levels of FZDs did not differ compared with control cells (Fig 3A and 3B). In contrast, h1299 cells transfected with miR-203 mimic showed significant downregulation of FZD2 mRNA levels compared to control cells (0.268-fold; Fig 3C). To confirm whether miR-203 targeted FZD2 directly, we perform target inhibition assays (Fig 4). The target protector designed for the miR-203 target site in FZD2 mRNA completely negated the inhibitory effects of the miR-203 mimic and CD82 on FZD2 mRNA expression.

Effects of miR-338-3p and miR-203 on the protein levels of FZDs in h1299 cells
Next, we examined the protein levels of FZDs by immunoblot analysis in h1299 cells transfected with miRNA mimics or inhibitors. h1299/zeo cells transfected with the miR-203 mimic showed significant downregulation of FZD2 protein levels compared to parental and control cells (0.701-fold; Fig 5A and 5B). In contrast, h1299/CD82 cells transfected with the miR-203 inhibitor showed a significant increase in FZD2 protein levels compared to parental and control cells (1.546-fold; Fig 5A and 5B).
In contrast, cells transfected with the miR-338-3p mimic or inhibitor showed no significant differences compared to parental and control cells (Fig 6). Effects of miR-338-3p and miR-203 on the migration ability of h1299 cells To elucidate the functional effects of FZD2 downregulation by miR-203, cell migration was quantitatively examined by wound healing assay. The miR-203 mimic inhibited cell migration in h1299/zeo cells (Fig 7A, 7B and 7G), while the miR-203 inhibitor induced cell migration in h1299/CD82 cells (Fig 7D, 7E and 7H). In contrast, miR-338-3p did not affect h1299 cell migration (Fig 7A, 7C, 7D and 7F). Fig 7I shows the inhibition of migration in h1299 cells.
Overexpression of CD82 reduced cell migration to approximately 60% that of h1299/zeo cells. The miR-203 mimic inhibited migration in h1299/zeo cells, and the miR-203 inhibitor induced migration in h1299/CD82 cells, reaching approximately 80% that of h1299/zeo cells.

Discussion
The mechanisms regulating FZD expression have not been elucidated. Recently, researchers have reported that FZD is regulated by miRNAs [16,17]. Therefore, in this study, we analyzed the effects of miRNAs on FZD expression. Using the miRanda algorithm [13,14], we predicted 11 candidate miRNAs that targeted specific FZDs. Of these, miR-203 and mir-338-3p were   Effects of miR-203 on Frizzled-2 Expression regulated by CD82 overexpression. After subsequent analyses, only miR-203 inhibited FZD2 expression at the mRNA and protein levels. As shown in Fig 8, FZD2 had one potential complimentary miR-203-binding site within its 3 0 -UTR, suggesting that FZD2 was targeted by miR-203. Additionally, target inhibition assays designed for this binding site revealed that miR-203 targeted FZD2 mRNA directly. Moreover, miR-203 also regulated cell migration, a critical component of metastatic progression, in cancer cells through CD82. Thus, our data provided important insights into the regulation of the Wnt signaling pathway by miRNAs in cancer cells.
CD82/KAI1 is recognized as an antimetastatic factor and has been reported to attenuate or modify cell membrane receptor signaling, including integrin, epidermal growth factor receptor (EGFR), and c-Met pathways [6,12,15]. We have previously reported that CD82 regulates canonical Wnt signaling by downregulating specific FZDs (FZD2, -3, -5, -7, and -9) [7]. These FZDs are mainly part of the noncanonical Wnt pathway (Wnt/PCP and Wnt/PKC), with the exception of FZD9. Therefore, CD82 inhibits both canonical Wnt signaling and noncanonical Wnt signaling by suppressing FZD receptor expression. Suppression of both Wnt signaling pathways is known to contribute to cancer progression via various mechanisms. Additionally, our analysis demonstrated that miRNAs were involved in the CD82-dependent regulation of FZDs, thereby providing insights into these complex regulatory mechanisms.
However, the mechanism through which CD82 downregulates miR-203 is still unclear. CD82 forms complexes with cell membrane receptors and modifies their signaling pathways. We previously reported that CD82 attenuates EGFR and c-Met signaling by binding to these receptors [6,12,15]. Recently, EGFR and c-Met receptor tyrosine kinase were shown to control miRNA expression [30]. In this previous report, overexpression of miR-203 in A549 cells was shown to reduce AKT phosphorylation and the expression of its substrate GSK3β. This finding was also consistent with our previous report showing that overexpression of CD82 downregulates GSK3β [7]. Furthermore, c-Met receptor inhibits the expression of miR-203, and forced expression of CD82 attenuates c-Met signaling in h1299 cells. These reports suggest that CD82 enhances the expression of miR-203 via attenuation of c-Met signaling. However, this indirect evidence needs to be confirmed in further studies.
Importantly, in this study, we confirmed the functional effects of miR-203 transfection in cancer cells. Transfection of h1299/zeo cells led to a 20% reduction in cell migration compared to parental h1299/zeo cells, and this level of reduction was half of that in h1299/CD82 cells (40%). These results suggested that FZD3, -5, and -7, which are also downregulated by CD82 expression, may contribute to CD82 function. These mechanisms should be investigated in further studies.
In conclusion, we found that CD82 enhanced the expression of miR-203 and directly downregulated FZD2 expression. Accumulating evidence has demonstrated the presence of abnormal miR-203 expression in cancer stem cell. Therefore, this novel mechanism for the regulation of FZD2 expression via miR-203 could be utilized for epigenetic therapy for cancer stem cells.