Glycogen synthase kinase 3 beta inhibits microRNA-183-96-182 cluster via the β-Catenin/TCF/LEF-1 pathway in gastric cancer cells

Glycogen synthase kinase 3 beta (GSK3β) is a critical protein kinase that phosphorylates numerous proteins in cells and thereby impacts multiple pathways including the β-Catenin/TCF/LEF-1 pathway. MicroRNAs (miRs) are a class of noncoding small RNAs of ∼22 nucleotides in length. Both GSK3β and miR play myriad roles in cell functions including stem cell development, apoptosis, embryogenesis and tumorigenesis. Here we show that GSK3β inhibits the expression of miR-96, miR-182 and miR-183 through the β-Catenin/TCF/LEF-1 pathway. Knockout of GSK3β in mouse embryonic fibroblast cells increases expression of miR-96, miR-182 and miR-183, coinciding with increases in the protein level and nuclear translocation of β-Catenin. In addition, overexpression of β-Catenin enhances the expression of miR-96, miR-182 and miR-183 in human gastric cancer AGS cells. GSK3β protein levels are decreased in human gastric cancer tissue compared with surrounding normal gastric tissue, coinciding with increases of β-Catenin protein, miR-96, miR-182, miR-183 and primary miR-183-96-182 cluster (pri-miR-183). Furthermore, suppression of miR-183-96-182 cluster with miRCURY LNA miR inhibitors decreases the proliferation and migration of AGS cells. Knockdown of GSK3β with siRNA increases the proliferation of AGS cells. Mechanistically, we show that β-Catenin/TCF/LEF-1 binds to the promoter of miR-183-96-182 cluster gene and thereby activates the transcription of the cluster. In summary, our findings identify a novel role for GSK3β in the regulation of miR-183-96-182 biogenesis through β-Catenin/TCF/LEF-1 pathway in gastric cancer cells.


INTRODUCTION
Glycogen synthase kinase 3 beta (GSK3b) is a serine/ threonine protein kinase whose function is required for the NF-kB-mediated anti-apoptotic response to tumor necrosis factor alpha (1). GSK3b also plays a critical role in numerous signaling pathways including Wnt/b-Catenin/ TCF/LEF-1 signaling pathway. GSK3b is constitutively active in cells and forms a complex with adenomatous polyposis coli (APC) and scaffold protein Axin in the absence of Wingless/Wnt signal. Phosphorylation of APC by GSK3b provides a docking site for b-Catenin binding. b-Catenin is a key component of both the cadherin cell adhesion system and the Wnt signaling pathway (2)(3)(4). GSK3b phosphorylates b-Catenin leading to its degradation by ubiquitin-proteasome pathway (5). Wnt signal inhibits GSK3b activity and increases free cytosolic b-Catenin level. b-Catenin translocates to the nucleus to act as a cofactor for the T cell factor (TCF) family of transcription factors, including TCF-1, TCF-3, TCF-4 and LEF-1 (leukemia enhancer factor 1). b-Catenin/TCF/ LEF-1 complex activates oncogenic target genes such as c-myc (6), c-jun (7) and cyclin D1 (8).
The miR-183-96-182 cluster is a critical sensory organspecific gene that locates to the short arm of chromosome 7 (7q32.2). The cluster is highly expressed in the retina and other sensory organs. Inactivation of the cluster results in early-onset and progressive synaptic defects of the photoreceptors, leading to abnormalities of scotopic and photopic electroretinograms (26). The products of miR-183-96-182 cluster gene, miR-183, miR-96 and miR-182, play important roles in a variety of cancers. For instance, miR-183 promotes cell growth and motility in prostate cancer cells by targeting Dkk-3 and SMAD4 (27). miR-96 promotes hepatocellular carcinoma (HCC) cell proliferation and colony formation by targeting FOXO1 and FOXO3a (28). miR-182 increases tumorigenicity and invasiveness in breast cancer by targeting the matrix metalloproteinase inhibitor RECK (29). The expression levels of the miR-183 family are upregulated in most cancer types (30). But the expression levels of miR-183 family in gastric cancer are controversial. Kong et al. (31) found that miR-182 was significantly downregulated in human gastric adenocarcinoma tissue samples. Li et al. (32) reported that miR-96, miR-182 and miR-183 were all upregulated in intestinal-type gastric cancers.
Previous reports have demonstrated the interaction between GSK3b and miRs in various human cancers. For instances, GSK3b increases miR-122 level through activating C/EBPa in HCC (33). Inhibition of GSK3b activates miR-181 expression through Wnt/beta-catenin signaling in HCC (34). MiR-26a promotes cholangiocarcinoma via reducing GSK3b expression, resulting in b-Catenin activation (35). The influence and mechanisms of GSK3b on miR biogenesis and function in gastric cancer remain unknown. Here we report that inhibition of GSK3b increases nuclear translocation of b-Catenin, which forms a complex with TCF/LEF-1 to enhance miR-183-96-182 cluster gene expression in gastric cancer cells. Our work identifies miR-183-96-182 cluster gene as a downstream target regulated by b-Catenin/TCF/LEF-1 pathway in gastric cancer cells.

Primary antibodies and primers
GSK3b (3D10) mouse mAb, Lef-1 (C12A5) rabbit mAb, b-Catenin (6B3) rabbit mAb, CK1e polyclonal antibody, CK2a polyclonal antibody, FoxO1 rabbit mAb and b-Catenin (L87A12) mouse mAb were purchased from Cell Signaling Technology. GAPDH (0411) mouse monoclonal antibody, GAPDH (FL-335) rabbit polyclonal antibody, Lamin A/C (636) mouse mAb and b-actin (R-22) rabbit polyclonal antibody were purchased from Santa Cruz Biotechnology. All primers for mature miRNA detection were purchased from Applied Biosystems; all other primers were ordered from Integrated DNA Technologies. The sequences of the primers are listed in Supplementary  Table S1.

MiRNA array
Total RNA was extracted from WT and KO MEF cells using TRIZOL (Invitrogen). MiR expression profiling of both WT and KO cells (four replicates each) was performed using a commercial array (Dharmacon Inc of Thermo Scientific). Relative Intensity data for eight samples was subjected to statistical filtering, keeping miR probes with P 0.05 in at least three of the eight experiments. This resulted in 336 miR probes passing statistical filters. The remaining data were inter-array scaled and transformed to log2. The experiments were annotated with factor labels as indicated in Figure 1A. This annotated, filtered, scaled and log2 transformed data set was used for agglomerative hierarchical clustering using cosine correlation distance metric.

Cytoplasmic and nuclear fractionation
Cytoplasmic and nuclear fractionation was performed using EZ Nuclei Isolation Kit (Sigma) according to the manufacturer's instructions. Briefly, cells were harvested and washed once with cold phosphate buffered saline. Cells were then suspended in EZ Nuclei Isolation buffer and rotated at 4 C for 5 min. After centrifugation at 4 C for 5 min, supernatant was collected containing the cytoplasmic fraction. Cell lysis and centrifugation were repeated three times. The final pellets were collected as the nuclear fraction and lysed in Pierce IP lysis buffer.

Western blotting
Gastric cancer samples and the matched control gastric tissues were from Rhode Island Hospital Tissue Bank and their use was approved by Rhode Island Hospital institutional review board (IRB). MEF cell, AGS cell or gastric tissue lysates were prepared in Pierce IP lysis buffer, separated by 4-12% NuPAGE Õ Novex Õ 4-12% Bis-Tris gel electrophoresis and electroblotted to nitrocellulose membrane (Bio-Rad). Blotted membranes were probed with their respective primary antibodies, rotating at 4 C overnight. Membranes were washed three times in Tris-Buffered Saline with Tween 20 (TBST) buffer and probed with secondary antibody (Alexa Fluor 680 goat anti-rabbit IgG or IRDye800-conjugated Affinity Purified Anti-Mouse IgG, respectively) at room temperature for 1 h. Membranes were then washed three times in TBST buffer and direct infrared fluorescence detection was performed with a Licor Odyssey Õ Infrared Imaging System (36). The integrated intensities (counts-mm 2 ) of protein bands were quantified according to manufacturer's instructions. The relative protein level was normalized with the integrated intensity of respective GAPDH.

Immunohistochemistry
Using the same gastric cancer samples and their matched controls, immunohistochemistry (IHC) was performed on paraffin-embedded tissues sectioned at four microns. Sections were deparaffinized and rehydrated through graded alcohols to distilled water and treated with heat and pressure induced antigen retrieval in 10 mM citrate buffer (pH 6.0) for 10 min and blocked using Peroxidase Block (Dako, Carpinteria, CA) for 5 min. Sections were incubated with GSK3b Rabbit mAb (Cell Signaling) or b-Catenin Rabbit mAb (Cell Signaling) overnight at 4 C. The Envision+Dual Link System Peroxidase (Dako) was used as the secondary antibody followed by Liquid DAB+Substrate ChromogenSystem (Dako). Counterstaining was performed with hematoxylin. The slides were dehydrated and cleared through xylene then coverslipped.

Real-time reverse transcriptase-polymerase chain reaction
Total RNA was extracted by TRIZOL (Invitrogen) and 1 mg of total RNA was used for cDNA synthesis using MMLV reverse transcriptase (New England Biolabs) as described in the manufacturer's manual. TaqMan realtime reverse transcriptase-polymerase chain reaction (RT-PCR) miRNA detection kits (Applied Biosystems) that include RT primers and TaqMan probes were used to quantify the levels of mature miRNAs, and 18 S RNA was used for normalization. All PCR reactions were run in triplicate.

Luciferase assay
A DNA fragment of 2340 base pairs from the upstream region of the miR-183-96-182 cluster containing the putative TCF/LEF-1 binding elements (TBEs) was amplified from the genomic DNA of AGS cells and subcloned into the pSwitchlight_Prom Promoter Reporter Vector (SwitchGear Genomics) between SacI and HindIII sites (sense primer: ACCTGAGCTCTCTC GACTTTC; antisense primer: AGTTAAGCTTCCTGC GCCGG). The newly cloned construct was named pmiR-96 cluster promoter. AGS cells were transfected with pmiR-96 cluster promoter plus indicated constructs or the empty reporter. A b-Gal plasmid was cotransfected with the reporter constructs, respectively, to control for transfection efficiency. Twenty-four hours after transfection, the cells were harvested for luciferase assay. Renilla luciferase activities were quantified using LightSwitch Luciferase Assay Reagent LS010 (SwitchGear Genomics), and Renilla luciferase activity was normalized to b-Gal activity. For each experiment, a control using an empty vector (EV) was used and corrected luciferase values were averaged, arbitrarily set to a value of '1' and served as a reference for comparison of fold differences in experimental values.

Chromatin immunoprecipitation assay
Chromatin immunoprecipitation (ChIP) assays were performed using a SimpleChIP Õ Enzymatic Chromatin IP Kit (Magnetic Beads) from Cell Signal Technology following the manufacturer's protocol. Briefly, AGS or Hela cells were fixed with 1% formaldehyde for 10 min to cross-link proteins to DNA. Nuclei were prepared and treated with Micrococcal Nuclease for 20 min at 37 C to digest the chromatin into 150-900 bp DNA/protein fragments. b-Catenin rabbit mAb and ChIP Grade Protein G Magnetic Beads were used to immunoprecipitate b-Catenin/TCF/LEF-1 bound DNA fragments. Normal Rabbit IgG was used as a negative control. After chromatin was eluted from the beads, the cross-links were reversed by adding NaCl and Proteinase K and incubating for 2 h at 65 C. DNA was purified with spin column and used for standard PCR and quantitative real-time PCR. We used Native Pfu DNA Polymerase (Stratagene) for standard PCR and RT 2 Real-Time TM SYBR Green PCR Master Mix (Thermo Fisher/Fermentas) for quantitative real-time PCR according to the manufacturer's instructions.

Cell Proliferation and migration assays
To suppress the miR-183-96-182 cluster, AGS cells were transfected with miRCURY LNA TM Inhibitors (Exiqon). Their respective sequences are: miRCURY LNA TM miRNA Inhibitor Negative Control A: GTGTAACACG TCTATACGCCCA; miRCURY LNA TM miR-183 inhibitor: AGTGAATTCTACCAGTGCCAT; miRCURY LNA TM miR-96 inhibitor: GCAAAAATGTGCTAGTG CCAA; miRCURY LNA TM miR-182 inhibitor: TGTGA GTTCTACCATTGCCAA. To knock down GSK3b, AGS cells were transfected with GSK3B Pre-design Chimera RNAi or negative control Naito 1 Pre-design Chimera RNAi (Abnova). Forty-eight hours after transfection, the cells were trypsinized and cultured for another 24 h in either 96-well flat-bottom plate for cell proliferation assay, in Boyden Chamber 12-well Cell Culture Insert (BD Falcon TM ) for migration assay, or in 12-well plate for western blot. A cell proliferation assay was performed with a colorimetric WST-1 assay kit (Roche Applied Science) according to the manufacturer's instructions. In the Boyden Chamber migration assay, cells migrated from the upper chamber (5% FBS) to the lower one (10% FBS) were collected and counted. We set the control as '1' arbitrarily to quantify the proliferation or migration of the cells.

Statistical analysis
Quantitative data were analyzed by unpaired Student's t-test. The miR array data were analyzed by textbook analysis of variance (ANOVA), with FDR multiple test correction, across the 'Group' factor (KO versus WT). The raw ANOVA results are reported in the form of agglomerative hierarchical clustering graphic.

KO of GSK3b changes miR expression differentially
The raw ANOVA miR array results are reported in the form of agglomerative hierarchical clustering graphic ( Figure 1A). Of the 336 measured miRs, 55% (185 of 336) were upregulated and 45% (78 of 336) downregulated ( Figure 1B). The top 20 differentially expressed miRs by fold change are listed in the Table 1, where the direction of change is relative to factor level WT. These hits have been highlighted on the scatter plot with all 336 miR data points ( Figure 1C).

KO of GSK3b increases protein level and nuclear translocation of b-Catenin
GSK3b phosphorylates b-Catenin that is primed by other kinases such as casein kinases 1 and 2, a necessary prerequisite to its entry into the ubiquitin-proteasome pathway for degradation (5). We first quantified protein levels of b-Catenin, GSK3b, CK1e and CK2a in WT and GSK3b KO MEF cells. As expected, GSK3b KO increased b-Catenin expression level by 2-fold but had no effects on CK1 and CK2 expression (Figure 2A). To determine if b-Catenin protein translocation into the nucleus was increased in GSK3b KO MEF cells, we fractionated the cytoplasmic and nuclear parts of MEF cells and found, as expected, that the nuclear b-Catenin protein levels were also increased by 2-fold in GSK3b KO MEF cells ( Figure 2B). Our previous studies have shown that phosphorylation of Drosha by GSK3b facilitates its nuclear localization (9,10). Unexpectedly, GSK3b KO also increased some miR expression. Of the miRs that were increased the most by GSK3b KO, miR-96, miR-182 and miR-183 are all from the same miR gene cluster. The miR array data revealed that they were increased 6-, 5-or 3-fold, respectively (Table 1 and Figure 2C), suggesting that GSK3b may suppress the generation of miR-96, miR-182 and miR-183. To further verify this, we ectopically expressed a GSK3b construct in human gastric epithelial AGS cells. Compared with EV, overexpression of GSK3b inhibited the expression   Figure 3A, the overall GSK3b protein level in gastric cancer samples was 50% of that in the matched normal samples (n = 8, P < 0.05). b-Catenin levels were increased 2-fold in gastric cancer samples compared with matched normal gastric tissue samples ( Figure 3B). We further confirmed the changes of the expression levels of GSK3b and b-Catenin by IHC (Figure 4). The levels of miR-96, miR-182 and miR-183 in gastric cancer were increased by 2-fold ( Figure 3C). Surprisingly, the primary miR-183-96-182 cluster (pri-miR-183) levels were higher in gastric cancer tissues than that in the matched normal tissues, indicating that GSK3b regulates the production of miR-96, miR-182 and miR-183 through b-Catenin at the transcription level.
b-Catenin/TCF/LEF-1 binds to and activates the promoter of miR-183-96-182 cluster gene The gene encoding miR-96, miR-182 and miR-183 locates to human chromosome 7q32.2. In silico screening identified seven potential TBEs in the promoter region of miR-96-182-183 cluster gene ( Figure 5A). To determine if these TBEs are bona fide binding sites for b-Catenin/ TCF/LEF-1 complex, we performed ChIP experiments using a SimpleChIP Õ Enzymatic Chromatin IP Kit and a rabbit mAb against b-Catenin. We confirmed that all the TBEs upstream of the putative core promoter were bona fide binding sites for b-Catenin/TCF/LEF-1 complex in AGS cells ( Figure 5B). In HeLa cells, we also confirmed another TBE downstream of the core promoter ( Figure 5B). To determine if the binding of b-Catenin/TCF/LEF-1 complex to TBEs is functional, we generated a renilla luciferase construct by subcloning the upstream TBEs containing DNA fragment into a luciferase vector. Cotransfection of a construct encoding b-Catenin together with the luciferase vector in AGS cells increased the renilla luciferase activity by 3-fold (compared with EV, P < 0.05), while cotransfection of a construct encoding GSK3b had the opposite effect (2-fold reduction; compared with EV, P < 0.05) ( Figure 5C). To further confirm the effect of b-Catenin and GSK3b on promoter function, we knocked down b-Catenin or GSK3b with respective specific siRNA molecules. Knockdown of b-Catenin decreased renilla luciferase activity by 2-fold (compared with control siRNA, P < 0.05), while knockdown of GSK3b increased renilla luciferase activity by 2.5-fold (compared with control siRNA, P < 0.05) ( Figure 5D).  Figure 7A). Suppression of miR-183-96-182 cluster decreased proliferation and migration of AGS cells ( Figure 7B and C). To investigate the effects of GSK3b knockdown on gastric cancer phenotype, we transfected control siRNA or GSK3b-specific siRNA into AGS cells. Compared with control siRNA, GSK3b siRNA specifically downregulated GSK3b protein ( Figure 7D). Knockdown of GSK3b increased AGS cell proliferation ( Figure 7E), but had no significant effect on AGS cell migration ( Figure 7F).    Figure 4C was cotransfected with b-Catenin siRNA or GSK3b siRNA, respectively, into AGS cells (compared with control siRNA, *P < 0.05 by Student's t-test). All experiments were repeated three times with similar results.

DISCUSSION
The Wnt signaling plays a pivotal role in tumorigenesis in various cancers including gastric cancer (37,38). Given that the CK1 and CK2 protein kinase families play important roles in Wnt signaling pathway (39,40), we wondered whether KO GSK3b deregulated the expression of these kinases. We found, however, that knocking out GSK3b did not change the expression of CK1 and CK2, ruling out deregulated activity of these kinases in GSK3b KO cells. As a key component of this pathway, GSK3b has emerged as a potential therapeutic target for cancer treatment (41). Because GSK3b is a multifunctional protein kinase, inhibition of GSK3b may have serious side effects. To reduce these side effects, miR-183-96-182 cluster could serve as a potential downstream target of the Wnt signaling pathway for treatment of gastric cancer and deserves further exploration.
b-Catenin/TCF/LEF-1 complex binds to a region near the core promoter of the miR-183-96-182 cluster gene. Various other transcription factors bind to this region as well, indicating that the cluster gene is potentially regulated by many other transcription factors in addition to TCF and LEF-1 (Supplementary Figure S1).
We measured pri-miR-183 and mature miR-96, miR-182, miR-183 expression levels in gastric cancer and matched normal gastric tissue by qRT-PCR. Our results showed that both the primary and mature miR-96, miR-182, miR-183 expression levels were significantly upregulated in gastric cancer tissues compared with the adjacent normal control gastric tissues. By means of western blotting and IHC techniques, we found that GSK3b protein expression decreased and b-Catenin protein level increased significantly in gastric cancer. We hypothesized that GSK3b regulates miR-183-96-182 cluster through b-Catenin/TCF/LEF-1 pathway in gastric cancer cells. Using miR array, ChIP assay, luciferase assay, qRT-PCR, we confirmed our hypothesis and identified miR-183-96-182 cluster as a novel target of the b-Catenin/TCF/LEF-1 pathway in gastric cancer cells.
Gastric cancer, the fourth most common cancer and the second leading cause of cancer-related deaths in the world, is one of the major threats to human health. According to the World Health Organization, gastric cancer annually claims 800 000 lives worldwide, metastatic disease being uniformly fatal (42). In this study, we found that miR-183-96-182 cluster inhibitors decrease the proliferation and migration of gastric cancer AGS cells and provide a functional link between GSK3b, the miRNA-183-96-182 cluster and the b-Catenin/TCF/LEF-1 pathway in gastric cancer.