MicroRNA-182 targets cAMP-responsive element-binding protein 1 and suppresses cell growth in human gastric adenocarcinoma

Authors


H. Tang, Tianjin Life Science Research Center and Basic Medical School, Tianjin Medical University, Tianjin 300070, China
Fax: +86 22 23542503
Tel: +86 22 23542503
E-mail: htang2002@yahoo.com

Abstract

MicroRNAs (miRNAs) constitute a class of noncoding RNAs that post-transcriptionally regulate gene expression. Recent evidence indicates that many miRNAs function as oncogenes or tumor suppressors by negatively regulating their target genes. In our previous study, using miRNA microarray analysis, we found that miRNA-182 (miR-182) was significantly downregulated in human gastric adenocarcinoma tissue samples. Here, we confirmed the downregulation of miR-182 in a larger sample of gastric tissue samples. Overexpression of miR-182 suppressed the proliferation and colony formation of gastric cancer cells. An oncogene, encoding cAMP-responsive element binding protein 1 (CREB1), serves as a direct target gene of miR-182. A fluorescent reporter assay confirmed that miR-182 binds specifically to the predicted site of the CREB1 mRNA 3′-UTR. When miR-182 was overexpressed in gastric cancer cell lines, both the mRNA and protein levels of CREB1 were depressed. Furthermore, CREB1 was present at a high level in human gastric adenocarcinoma tissues, and this was inversely correlated with miR-182 expression. Ectopic expression of CREB1 overcame the suppressive phenotypes of gastric cancer cells caused by miR-182. These results indicate that miR-182 targets the CREB1 gene and suppresses gastric adenocarcinoma cell growth, suggesting that miR-182 shows tumor-suppressive activity in human gastric cancer.

Abbreviations
ASO

antisense oligonucleotide

CREB1

cAMP-responsive element binding protein 1

EGFP

enhanced green fluorescent protein

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

miR-182

microRNA-182

miRNA

microRNA

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide

Introduction

Gastric adenocarcinoma is a malignant tumor derived from gastric epithelium mucosa cells, which occupy 95% of the gastric malignant tumor. The interaction of multiple factors contributes to the development of gastric carcinogenesis, including regulatory changes in oncogenes and tumor suppressor genes [1,2]. The alterations of these genes are diverse. For example, the oncogenes k-ras and c-met are significantly amplified and overexpressed in gastric cancer. Abnormal expression of CD44v is closely related to the invasiveness and metastasis of gastric cancer. As tumor suppressor genes, p53, APC and DCC show absent expression or lower expression levels in gastric cancer. Recently, it has been shown that the regulation of oncogenes and tumor suppressor genes occurs not only at the transcriptional level, but also at the post-transcriptional level.

In the last decade, a class of novel noncoding RNAs called microRNAs (miRNAs) has been discovered in many species. miRNAs are 18–26 nucleotides in length, and post-transcriptionally regulate gene expression in multicellular organisms by affecting both the stability and translation of mRNAs [3]. miRNAs are transcribed by RNA polymerase II or III in the nucleus [4]. The primary transcripts (pri-miRNA) are capped, polyadenylated, and cleaved by the Drosha ribonuclease III enzyme to produce an ∼ 70-nucleotide stem–loop precursor miRNA [5–7], which is then transported to the cytoplasm by exportin 5 [8]. The RNase III enzyme Dicer processes precursor miRNAs into mature miRNAs (∼ 22 nucleotides) [9]. The mature miRNA is incorporated into an RNA-induced silencing complex, called miRNP [10,11], that recognizes target mRNAs through imperfect base pairing with the 3′-UTR of target gene mRNAs, and most commonly results in either repression of the translation of the target mRNA or its destabilization [11,12]. In addition, it is also reported that many miRNAs regulate the expression of target genes by destabilizing their target mRNAs via rapid deadenylation [13,14].

Abnormal expression and the loss of the dynamic balance between oncogenes and tumor suppressor genes typically lead to tumor formation and the development of cancer [15]. As an important regulator of gene expression, miRNA is also involved in tumorigenesis. Considerable evidence has shown that miRNAs are important regulators in diverse biological processes of cancer, such as cell proliferation [16], apoptosis [17,18], angiogenesis [19,20], cell differentiation [21,22], adhesion, and metastasis [23]. Previous studies have identified cancer-specific miRNAs in many types of cancer, including breast cancer [24], lung cancer [25], hepatocellular carcinoma [26], gastric cancer [27], and many others.

In our previous study, we showed, by miRNA microarray analysis, that miRNA-182 (miR-182) was downregulated in gastric adenocarcinoma [28]. In this study, we confirm the downregulation of miR-182 in a larger sample of gastric adenocarcinoma tissue samples. miR-182 influenced the phenotypes of gastric adenocarcinoma cells in vitro. Subsequently, the cAMP-responsive element binding protein 1 gene (CREB1), an oncogene promoting tumor cell growth and proliferation [29–33], was determined to be a direct and functional target of miR-182. Our findings indicate that miR-182 functions as a tumor suppressor in gastric adenocarcinoma cells.

Results

miR-182 was downregulated in human gastric adenocarcinoma tissues

To confirm the downregulation of miR-182 in gastric adenocarcinoma, as suggested in our previous study, quantitative real-time PCR was used to detect the expression levels of miR-182 in 10 pairs of human gastric adenocarcinomas and the adjacent normal tissues. The results showed that miR-182 expression in gastric adenocarcinoma was significantly lower than that in the adjacent, normal tissues (Fig. 1A).

Figure 1.

 Deregulation of miR-182 in gastric adenocarcinoma tissues and suppression of miR-182 leads to gastric adenocarcinoma cell growth. (A) Differential expression of miR-182 in 10 pairs of human gastric adenocarcinoma tissues (Ca) and corresponding adjacent normal tissues (N) was detected with quantitative RT-PCR. U6 small nuclear RNA was regarded as an endogenous normalizer, and the relative miR-182 expression level is shown. (B) The relative level of miR-182 in MGC-803 cells transfected with pcDNA3/pri-miR-182 (pri-miR-182) or ASO-miR-182. (C) MGC-803, BGC-823 and SGC-7901 cells were transfected with pri-miR-182 or ASO-miR-182. Cell viability was determined at 48 h post-transfection with the MTT assay. The relative cell growth activity is shown. (D) The cell growth capacity in vitro was assessed with the colony formation assay. The colony formation rate is shown (*P < 0.05).

Overexpression of miR-182 suppresses proliferation of gastric adenocarcinoma cells

To determine the effects of miR-182 on gastric adenocarcinoma cellular proliferation, an miR-182 expression vector (pri-miR-182) and a commercially synthesized antisense inhibitor [antisense oligonucleotide (ASO)-miR-182] were used to alter the levels of miR-182 in gastric adenocarcinoma cells. The alteration of miR-182 was confirmed by quantitative RT-PCR (Fig. 1B). Next, we transfected the pri-miR-182 or ASO-miR-182 into the gastric cancer cell lines MGC-803, BGC-823, and SGC-7901. The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) assay showed that miR-182 could reduce the viability of MGC-803, BGC-823 and SGC-7901 cells by ∼ 20–40%. Conversely, ASO-miR-182 increased the viability of gastric cancer cells by ∼ 10–40% (Fig. 1C). The colony formation assay indicated that the colony formation rates of MGC-803, BGC-823 and SGC-7901 cells transfected with pri-miR-182 were decreased by 35–60% as compared with the control group, and ASO-miR-182 enhanced colony formation by ∼ 30% in these cells (Fig. 1D). These results demonstrate that miR-182 suppresses growth of gastric adenocarcinoma cells.

miR-182 directly targets the CREB1 3′-UTR in gastric adenocarcinoma cells

To determine the target genes mediating the function of miR-182, bioinformatics methods were used to predict potential target genes. We found that the 3′-UTR of CREB1 mRNA contains miR-182-complementary binding sites (Fig. 2A). To confirm that CREB1 can be directly targeted by miR-182, we performed an enhanced green fluorescent protein (EGFP) reporter assay with EGFP reporter vectors that bear either the wild-type or the four-base mutated 3′-UTR fragment of CREB1 (Fig. 2A). When MGC-803 cells were cotransfected with pcDNA3/EGFP–CREB1 3′-UTR and pcDNA3/pri-miR-182, we found that miR-182 significantly repressed EGFP expression, whereas the intensity of EGFP fluorescence in cells transfected with ASO-miR-182 was significantly increased as compared with the control cells (Fig. 2B). In contrast, EGFP expression levels with the mutated CREB1 3′-UTR were not influenced by alteration of miR-182 (Fig. 2C). These data indicated that miR-182 could bind to the specific sites of the CREB1 mRNA 3′-UTR and negatively regulate the expression of the CREB1 gene.

Figure 2.

CREB1 is a direct target of miR-182. (A) The predicted binding sites of miR-182 on CREB1 mRNA are shown. The mutant UTR contains four bases in the complementary seed sequences, as indicated by the arrows. (B) MGC803 cells were transfected with the wild type of pcDNA3/EGFP–CREB1 3′-UTR reporter vector and either pri-miR-182 or ASO-miR-182. Pri-miR-182 suppressed the EGFP fluorescence intensity, and, ASO-miR-182 increased the EGFP fluorescence intensity. (C) MGC-803 cells were transfected with the mutated pcDNA3/EGFP–CREB1 3′-UTR (pcDNA3/EGFP–CREB1 3′-UTR-mut) reporter vector as well as either pri-miR-182 or ASO-miR-182. Mutation of the miR-182-binding site abolished the effect of miR-182 on the EGFP fluorescence intensity (*P < 0.05).

miR-182 plays a negative regulatory role for CREB1 expression

miRNAs regulate target genes at the post-transcriptional level by binding to 3′-UTRs of target genes to silence function [34]. To examine whether miR-182 depresses endogenous CREB1 expression, we transfected MGC803 cells with pri-miR-182 or ASO-miR-182, and examined the expression of CREB1 protein via western blot. The results showed that overexpression of miR-182 reduced the expression level of CREB1 by 30% (Fig. 3A). High expression levels of miR-182 in MGC-803 cells were capable of decreasing endogenous CREB1 mRNA as measured by real-time RT-PCR (Fig. 3B). Conversely, in MGC-803 cells transfected with ASO-miR-182, the level of endogenous CREB1 protein was significantly increased (Fig. 3A), and the CREB1 mRNA level was elevated (Fig. 3B). In addition, we found that the expression level of CREB1 mRNA in gastric adenocarcinoma tissues was significantly higher than that in matching adjacent, normal tissues in the 10 pairs of gastric tissues (Fig. 3C). All of these data indicate that miR-182 negatively regulates the expression of CREB1.

Figure 3.

 miR-182 negatively regulates CREB1. (A) MGC-803 cells were transfected with pri-miR-182 or ASO-miR-182, and the CREB1 protein expression level was evaluated by western blot. GAPDH protein was used as an endogenous normalizer, and the relative CREB1 protein quantity is shown. (B) MGC-803 cells were transfected with pri-miR-182 or ASO-miR-182, and expression of CREB1 mRNA was measured by quantitative RT-PCR. β-Actin mRNA was used as an endogenous normalizer, and the relative CREB1 mRNA expression level is shown. (C) The expression levels of CREB1 mRNA in the 10 pairs of gastric adenocarcinoma tissues (Ca) and the matched normal tissues (N) were assessed by quantitative RT-PCR (*P < 0.05).

Knockdown of CREB1 inhibits MGC-803 cell growth, and ectopic expression of CREB1 counteracts the cell growth suppression caused by miR-182 in vitro

Previous studies have shown that CREB1 plays an important role in promoting tumor cell proliferation and functions as an oncogene. Accordingly, to examine whether CREB1 affects MGC-803 cell growth, we constructed the pSilencer/siR-CREB1 (siR-CREB1) plasmid to knock down the expression of CREB1. Western blot analysis indicated that the CREB1 small interfering RNA could effectively suppress the expression of CREB1 protein in MGC-803 cells (Fig. 4A). The effects of CREB1 knockdown on cell viability and growth were evaluated in MGC-803, BGC-823 and SGC-7901 cells by the MTT and colony formation assays. At 48 h after transfection, CREB1 knockdown resulted in a reduction in cell viability as compared with the control group (Fig. 4B). Also, the colony formation rate of the cells transfected with siR-CREB1 vector was clearly lower than that of the control group (Fig. 4C).

Figure 4.

 Ectopic expression of CREB1 rescues the suppression of cell growth caused by miR-182 in gastric adenocarcinoma cell lines. (A) Transfection of the pSilencer/siR-CREB1 reduced the CREB1 protein level in MGC-803 cells. GAPDH protein was used as an endogenous normalizer, and the relative CREB1 protein quantity is shown. (B) The MTT assay showed that transfection of the pSilencer/siR-CREB1 plasmid suppressed cell viability in MGC-803, BGC-823 and SGC-7901 cells. (C) Colony formation rates of MGC-803, BGC823 and SGC-7901 cells transfected with pSilencer/siR-CREB1 plasmid were lower than that of the control vector. (D) CREB1 ectopic expression was confirmed by western blot. MGC-803 cells were transfected with pcDNA3/CREB1 with or without pri-miR-182 vector. CREB1 protein levels were detected by western blot. (E) MGC-803 cells were transfected with pri-miR-182 with or without pcDNA3/CREB1, and cell viability was determined at 48 h post-transfection bwith the MTT assay. The relative cell growth activity is shown. (F) CREB1 ectopic expression reversed the cell growth suppression caused by miR-182 in the colony formation assay. MGC-803 cells were transfected with pri-miR-182 with or without pcDNA3/CREB1, and the colony formation rate is shown (*P < 0.05).

To further confirm that miR-182 suppresses the growth of gastric cancer cells through downregulation of CREB1, we constructed the expression vector pcDNA3/CREB1 without the 3′-UTR to avoid the miRNA interference. Western blot analysis was used to confirm CREB1 expression in transfected MGC-803 cells (Fig. 4D). Ectopic expression of CREB1 was able to rescue MGC-803 cells from growth inhibition caused by miR-182 in the MTT assay (Fig. 4E), as well as in the colony formation assay (Fig. 4F). These results may also explain why overexpression of miR-182 could inhibit the growth of gastric cancer cells. Altogether, CREB1, as an oncogene, is a direct target of miR-182 and a mediator of miR-182 function in gastric cancer cells.

Discussion

During the process of tumor formation, abnormal genetic expression and weakening of the dynamic balance between oncogenes and tumor suppressor genes usually lead to the development of cancer. As for many other genes, miRNAs are involved in the regulation of oncogenes and tumor suppressor genes [34]. Previous studies have shown that miR-182 is poorly expressed in human breast cancer stem cells, human and mouse normal mammary stem cells, and embryonic carcinoma [35]. It was also reported that miR-182 could deregulate RGS17 by targeting its 3′-UTR to ultimately suppress the occurrence of lung cancer [36]. Here, we focused on the regulation of miR-182 in gastric cancer cells and the mechanisms of regulation of its target genes. Initially, downregulation of miR-182 in human gastric adenocarcinoma tissue samples suggested that miR-182 might play an important role in the development of gastric adenocarcinoma as a tumor suppressor gene. Then, in vitro studies demonstrated that miR-182 had a negative effect on cellular proliferation in MGC-803, BGC-823 and SGC-7901 cells. Second, bioinformatics analyses predicted an miR-182-binding site on the CREB1 transcript. On the one hand, the ability of miR-182 to regulate CREB1 expression was probably direct, because the EGFP fluorescence intensity of EGFP–CREB1-3′-UTR was specifically responsive to miR-182 alteration. However, mutation of the miR-182-binding site abolished the effect of miR-182 on the regulation of EGFP fluorescence intensity. On the other hand, endogenous CREB1 expression, both mRNA and protein, was inversely correlated with miR-182 level. In addition, we also observed an inverse correlation between the expression of miR-182 and CREB1 in tumor tissues and corresponding normal gastric tissues. These results suggested direct and negative regulation of CREB1 by miR-182. We should also note that, in the in vitro studies, the effect of miR-182 blockade on the increase in CREB1 expression is smaller than in gastric cancer tissues. The main reason is that endogenous miR-182 in MGC-803 cells is present at a low level, owing to its downregulation in human gastric cancer cells, which leads to a moderate effect of its blockade.

Numerous studies have demonstrated that the majority of patients with acute lymphoid and myeloid leukemia overexpress CREB in the bone marrow. CREB overexpression was associated with poor initial outcome of clinical disease in acute myelogenous leukemia patients. CREB is known to be involved in glucose homeostasis, growth factor-dependent cell survival, and memory [31], among other functions. The CREB1 gene is located on human chromosome 2q34, and encodes a transcription factor in the leucine zipper family of DNA-binding proteins [37,38]. This protein binds as a homodimer to the cAMP-responsive element, an octameric palindrome. When the level of intracellular cAMP or Ca2+ is elevated, the protein is phosphorylated by several protein kinases, and induces transcription of genes in response to hormonal stimulation of the cAMP pathway [37]. Alternative splicing of this gene results in two transcript variants encoding different isoforms. Phosphorylated CREB1 enhances the trans-activation potential and promotes aggregation of CREB1-binding protein and p300 [39,40]. Activated CREB1 can recognize the conserved cAMP-responsive element and regulate its downstream gene expression, which is associated with cell proliferation, differentiation, and survival signaling pathways [41]. Identified downstream genes include the oncogene bcl-2, the cell cycle-related genes cyclinA1, cyclinB1, and cyclinD2, signal transduction proteins, activated transcription factor 3, NF-κB, and other growth-related genes [42]. Therefore, CREB1 is regarded as an oncogene that promotes tumor cell growth and proliferation [43]. Our findings reveal the relationship between miRNAs and CREB1 in gastric adenocarcinoma cells. Knockdown of CREB1 suppresses the growth of MGC803 cells, which is consistent with the results of miR-182 overexpression. Overexpression of CREB1 rescues MGC803 cells from growth inhibition caused by miR-182. Thus, the low expression of miR-182 in gastric adenocarcinoma cells leads to an abnormally high expression level of the oncogene CREB1, resulting in cell proliferation.

In summary, we have demonstrated that miR-182 expression is downregulated in gastric adenocarcinoma tissues, and that miR-182 inhibits gastric adenocarcinoma cell viability and colony formation by targeting CREB1. These results suggest that miR-182 could have an important role in tumorigenesis by its regulation of CREB1. The elucidation of the mechanisms of miR-182 in gastric adenocarcinoma helps us to further understand the mechanism of gastric adenocarcinoma initiation and progression.

Experimental procedures

Tissue samples, cell lines, transfection, and RNA extraction

The human gastric tissue samples were obtained from the Tumor Bank Facility of Tianjin Medical University Cancer Institute and Hospital and National Foundation of Cancer Research, with the patients’ informed consent. This research was approved by the Ethics Committee of Tianjin Medical University. The 10 samples were from seven men and one women, ranging in age from 45 to 76 years. Postsurgery pathological reports diagnosed two well-differentiated adenocarcinomas, seven moderately differentiated adenocarcinomas, and one poorly differentiated adenocarcinoma. The histological grade of the samples was found to be as follows: one stage I case, three stage II cases, four stage III cases, and two stage IV cases.

MGC-803, BGC-823 and SGC-7901 cells were all maintained in RPMI-1640 (Gibco, Grand Island, NY, USA) with 10% heat-inactivated fetal bovine serum, 100 IU·mL−1 penicillin, and 0.1 mg·mL−1 streptomycin, in a humidified 5% atmosphere of CO2 at 37 °C. Transfection was performed with Lipofectamine 2000 Reagent (Invitrogen, Carlsbad, CA, USA), following the manufacturer’s protocol. Large and small RNAs were isolated from tissue samples with the mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA), according to the manufacturer’s instructions.

Quantitative RT-PCR

To detect the relative levels of the CREB1 transcripts, a real-time RT-PCR assay was performed. Briefly, a cDNA library was generated through reverse transcription with Moloney murine leukemia virus reverse transcriptase (Promega, Madison, WI, USA), with 5 μg of the large RNA. The cDNA was used for the amplification of the CREB1 gene, and the β-actin gene was used as an endogenous control for the PCR reaction. The PCR primers were as follows: CREB1-3′-UTR-S, 5′-CGCGGATCCAAATGGACTGGCTTGG-3′; CREB1-3′-UTR-A, 5′-CGGAATTCTGCCCTATGGAAGAGCTG-3′; β-actin-S, 5′-CGTGACATTAAGGAGAAGCTG-3′; and β-actin-A, 5′-CTAGAAGCATTTGCGGTGGAC-3′. PCR cycles were as follows: 94 °C for 4 min, followed by 40 cycles of 94 °C for 30 s, 58 °C for 30 s, and 72 °C for 30 s.

To detect the expression level of miR-182, a real-time RT-PCR assay was performed. Briefly, 2 μg of small RNA was reverse transcribed to cDNA with Moloney murine leukemia virus reverse transcriptase (Promega), with the following primers: miR-182-RT, 5′-GTCGTATCCAGTGCAGGGTCCGAGGTGCACTGGATACGACAGTGTGA-3′; and U6-RT, 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACAAAATATGGAAC-3′. The following PCR primers were used in the following PCR reactions: miR-182-Fwd, 5′-TGCGGTTTGGCAATGGTAGAAC-3′; U6-Fwd, 5′-TGCGGGTGCTCGCTTCGGCAGC-3′; and Reverse (universal), 5′-CCAGTGCAGGGTCCGAGGT-3′. PCR cycles were as follows: 94 °C for 4 min, followed by 40 cycles of 94 °C for 30 s, 56 °C for 30 s, and 72 °C for 30 s.

An SYBR Premix Ex Taq Kit (TaKaRa, Otsu, Shiga, Japan) was used, following the manufacturer’s instructions, and real-time PCR was performed and analyzed with the iQ5 Real-Time PCR Detection system (Bio-Rad, Hercules, CA, USA). All primers were purchased from AuGCT.

Plasmid construction

To construct the pcDNA3/pri-miR-182 (pri-miR-182)-expressing vector, we first amplified a 256-bp DNA fragment carrying pri-miR-182 from genomic DNA with the following PCR primers: pri-miR-182-S, 5′-CGGAATTCGGAAGGACCTTGTCGCAGTTGC-3′; and pri-miR-182-A, 5′-GACACTCGAGCCAGTTCCCTCACTCCTCGA-3′. The amplified fragment was cloned into pcDNA3.1(+) at the BamHI and EcoRI sites. We also commercially synthesized the 2′-O-methyl-modified ASOs of miR-182 (ASO-miR-182, 5′-AGTGTGAGTTCTACCATTGCCAAA-3′) as the inhibitor of miR-182.

The EGFP reporter vector pcDNA3/EGFP was constructed as follows. The 3′-UTR of CREB1 mRNA containing the miR-182-binding site was amplified by PCR with the following primers: CREB1-3′-UTR-S, 5′-CGCGGATCCAAATGGACTGGCTTGG-3′; and CREB1-3′-UTR-A, 5′-CGGAATTCTGCCCTATGGAAGAGCTG-3′. The amplified fragments were cloned into pcDNA3/EGFP at EcoRI and XhoI sites downstream of the EGFP-coding region. Also, four nucleotides within the miR-182 seed sequence binding site of the CREB1 3′-UTR were mutated by use of the PCR side-directed mutagenesis assay. The two additional primers used in the mutation assay were as follows: CREB1-3′-UTR-MS, 5′-GAAGTGTTGAATCCGATATTGACATGTTG-3′; and CREB1-3′-UTR-MA, 5′-CAACATGTCAATATCGGATTCAACACTTC-3′. The fragment of CREB1 3′-UTR mutant was also cloned into pcDNA3/EGFP at the same sites.

The pcDNA3/CREB1 expression vector contains the full-length ORF of CREB1 without 3′-UTR to avoid any miRNA-induced suppression. The PCR product was amplified with the primers 5′-CGGAATTCATGACCATGGAATCTGGAGC-3′; and 5′-ATCCCTCTAGAATCTGATTTGTGGCAGTAAAGG-3′, and was cloned into pcDNA3.1(+) at the EcoRI and XbaI restriction sites.

Cell proliferation assay

MGC-803, BGC-823 and SGC-7901 cells were seeded in a 96-well plate at 6000, 7000, and 8000 cells per well, respectively, 1 day prior to transfection. The cells were transfected with pri-miR-182 or control vector at 0.15 μg per well. The MTT assay was used to determine cell viability 48 h after transfection. The absorbance at 570 nm was measured with a μQuant Universal Microplate Spectrophotometer (BioTek, Winooski, VT, USA).

Colony formation assay

After transfection, MGC-803, BGC-823 and SGC-7901 cells were counted and seeded in 12-well plates (in triplicate) at 50, 60 and 75 cells per well, respectively. Culture medium was replaced every 3 days. Colonies were counted only if they contained more than 50 cells, and the number of colonies was counted at the sixth day after seeding. The rate of colony formation was calculated with the following equation: colony formation rate (%) = (number of colonies/number of seeded cells) × 100%.

miRNA target prediction

The target genes of miR-182 were predicted by the following two computer-aided algorithms: pictar (http://pictar.mdc-berlin.de/cgi-bin/new_PicTar_vertebrate.cgi) and targetscanhuman Release 5.2 (http://www.targetscan.org).

Fluorescent reporter assay

Cells were transfected with the reporter vector pcDNA3/EGFP–CREB1 3′-UTR or pcDNA3/EGFP–CREB1 3′-UTR-mut along with pri-miR-182 or the control vector pcDNA3 in 48-well plates. The red fluorescence protein expression vector pDsRed2-N1 (Clontech, Mountain View, CA, USA) was spiked into the samples, and served as the normalization. The intensities of EGFP and red fluorescence protein fluorescence were detected with an F-4500 Fluorescence Spectrophotometer (Hitachi, Tokyo, Japan).

Western blot

Cultured cells were lysed with RIPA lysis buffer (0.1% SDS, 1% Triton X-100, 1 mm MgCl2, 10 mm Tris/HCl, pH 7.4) at 4 °C for 30 min. Lysates were collected and cleared by centrifugation at 14 000 g for 10 min, and the protein concentration was determined. Total cell lysates (50 μg) were fractionated by SDS/PAGE. Proteins were electroblotted onto nitrocellulose membranes. Nonspecific binding sites of membranes were saturated with 5% skimmed milk in NaCl/Tris-T solution (100 mm Tris/HCl, pH 7.5, 150 mm NaCl, 0.1% Tween-20). The membranes were then incubated with antibodies against CREB1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) at room temperature for another 2 h. After being washed with NaCl/Tris-T four times, the membranes were incubated with secondary antibody [hosrseradish peroxidase-conjugated goat anti-(rabbit Ig)] (Sigma-Aldrich, St Louis, MO, USA) in 5% skimmed milk in NaCl/Tris-T solution for 1 h at room temperature. Reactions were developed by use of enhanced chemiluminescence (Perkin-Elmer Life Sciences, Boston, MA, USA).

Statistical analysis

Data are expressed as means ± standard deviation, and P < 0.05 is considered to be statistically significant by Student’s t-test.

Acknowledgements

We would like to thank the Tumor Bank Facility of Tianjin Medical University Cancer Institute and Hospital and National Foundation of Cancer Research for providing human gastric tissue samples. We also thank the College of Public Health, Tianjin Medical University, for technical assistance with fluorescence detection. This work was supported by the National Natural Science Foundation of China (No. 30873017, No. 91029714, and No. 31071191) and the Natural Science Foundation of Tianjin (No. 08JCZDJC23300 and No. 09JCZDJC17500).

Ancillary