Human gastric cancer (GC) is currently the fourth most common cancer globally and remains the most prevalent cancer in Eastern Asia (Bray et al., 2013; Strong et al., 2015). Treatment for GC includes surgery, chemotherapy or a combination of therapies. Although new strategies have been proposed to improve patient survival rates, the median survival time for advanced GC patients remains approximately 12 months (Bang et al., 2010; Koizumi et al., 2008; 2014; Tanabe et al., 2010). Thus, new treatment approaches are needed (Shen et al., 2013). In the last decades, numerous researches focused on gene expression regulatory networks in GC with expectations of improving diagnosis, therapy and prevention.

MicroRNAs (miRNAs) are an abundant class of small non-coding RNA molecules that serve as gene regulators by base pairing and silencing mRNAs through either mRNA degradation or preventing mRNA translation. Different miRNAs can function as tumor suppressors or oncogenes in cancer cells, and the dysregulation of certain miRNAs may contribute to human cancer (Garofalo and Croce, 2011). An individual miRNA could potentially alter complex cellular processes such as cell growth, cell cycle, apoptosis and invasion. Aberrant levels of miRNAs, such as miR-16, miR-21, miR-20a, miR-372 and miR-181a, have been identified in GC and many other tumors (Cho et al., 2009; Lin et al., 2012; Shin et al., 2011; Wu et al., 2013). Recent studies have shown that miR-103a-3p is upregulated in GC, breast cancer and colo-rectal cancer and is thus considered an oncogene (Arabpour et al., 2016; Chen et al., 2012; Guo et al., 2015a; Hong et al., 2014; Kim et al., 2013; Nonaka et al., 2015; Rotkrua et al., 2013; Wang et al., 2012; Xia et al., 2016; Zheng et al., 2016). Based on these findings, miR-103a-3p may represent a potential diagnostic biomarker and therapeutic target in cancer. However, the regulatory mechanism remains unclear.

Activating transcription factor 7 (ATF7), a member of basic leucine zipper proteins family, is located on chromosome 12q13.13. ATF7 is phosphorylated by p38 at Thr-51 and Thr-53, leading to its transcriptional activation (Gozdecka and Breitwieser, 2012). ATF7 activities reportedly suppress tu-morigenesis in mouse lymphoma models (Walczynski et al., 2014). Immunohistochemistry results in 72 post-operation colorectal cancer tissue specimens have revealed a inverse correlation between ATF7 expression and pathological stage as well as a positive correlation between ATF7 expression and 5-year overall survival or 5-year progression-free survival (Guo et al., 2015b). However, the mechanisms of ATF7 on cell proliferation in GC remain to be elucidated.

In the present study, higher miR-103a-3p levels were frequently detected in GC tissues, consistent with The Cancer Genome Atlas (TCGA) stomach cancer cohort analysis. These results indicated that miR-103a-3p may act as a tumor promoter in cancer progression. To understand the underlying mechanism, additional experiments were performed in vitro. Overexpressing miR-103a-3p significantly improved the proliferation of gastric cancer cell lines MKN-45 and SGC-7901, while inhibiting miR-103a-3p had the opposite effect. Bioinformatics analysis and dual-luciferase reporter assays indicated that ATF7 is a novel target of miR-103a-3p. Overexpressing ATF7 suppressed MKN-45 and SGC-7901 cell growth and eliminated the effect of miR-103a-3p on GC cells, whereas ATF7 silencing via RNAi facilitated GC cell growth. Together, these results suggested that miR-103a-3p promotes cell proliferation in GC cell by targeting and suppressing ATF7. Thus, miR-103a-3p/ATF7 axis may be a promising molecular target for GC therapy.

Cell lines and tissue specimens

Human GC cell lines, MKN-45 and SGC-7901, were obtained from the Key Laboratory of Environment and Genes Related to Diseases at Xi’an Jiaotong University (China). Cells were maintained in 1640 medium (GE, USA) supplemented with 10% FBS. A total of 33 paired gastric cancer and adjacent nontumor gastric tissues were collected from patients undergoing GC resection at the general surgery department of the First Affiliated Hospital, Xi’an Jiaotong University (China). The relevant characteristics of the studied subjects are shown in Table 1. No local or systemic treatment was administered prior to surgery. Both tumor and nontumor tissues were validated by pathologic examination. The study was approved by the Institute Research Ethics Committee at Cancer, Xi’an Jiaotong University, and all patients provided written informed consent. Tissue samples were immediately frozen in liquid nitrogen until subsequent RNA extraction.

RNA extraction and quantitative real-time PCR

Total RNA was extracted from prepared tissue or cell samples using TRIzol reagent (Invitrogen Life Technologies, USA), and cDNA was synthesized according to the manufacturer’s instructions (Takara, Japan). Quantitative real-time PCR (qRT-PCR) was performed using RealStar SYBR Green qPCR Power Mixture (GenStar, China) on an FTC-3000TM System (Funglyn Biotech Inc., Toronto, Canada) according to the manufacturer’s instructions. Relative gene expression (miR-103a-3p, U6, ATF7, β-actin, CDK2 and P27) was calculated using the 2^−ΔΔCt method (Livak and Schmittgen, 2001). The primers used are listed in Table 2.

Construction of expression plasmids

pEGFP-C1 was used to construct the ATF7 re-expression vector. ATF7 cDNA was chemically synthesized and cloned into pEGFP-C1 between the BglII and EcoRI sites. Vector was transfected using jetPRIME (Polyplus Co., France), and the medium was replaced 4 h after transfection.

miRNA and RNA interference

The miR-103a-3p mimics and miRNA negative control (double-stranded non-human miRNAs predicted to not target the human genome/transcriptome) were synthesized at GenePharma Biotech (China). Antisense miR-103a-3p oligonucleotides were synthesized as inhibitors at Sangon Biotech (Shanghai, China). The inhibitor control was also synthesized at Sangon Biotech (China). The siRNA duplexes targeting human ATF7 were synthesized and purified at GenePharma Biotech (China). siRNA duplexes predicted to not target the human genome/transcriptome were used as siRNA negative controls. RNA oligonucleotides were transfected using jetPRIME (Polyplus Co., France), and the medium was replaced 4 h after transfection. miRNA or siRNA was used at a final concentration of 30 nmol/L. All oligonucleotide sequences are listed in Table 2.

Cell viability assay

Cells were washed with warm 1640, 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)(Sigma, Germany) reagent was added, and the plates were incubated for 4 hours at 37°C. The supernatant was then discarded and replaced with dimethyl sulfoxide (DMSO). The absorbance of the converted dye was measured at 492 nm on a micro-plate reader (FLUOstar OPTIMA, BMG, Germany).

Colony formation assay

MKN-45/SGC-7901 cells were seeded in 6-well plates at 2000 cells per well, transfected 24 h later and incubated for ten days. Cells were washed with PBS and stained with 0.1% crystal violet for 30 min. Images were captured on a Bio-Rad Universal Hood Molecular Gel Imaging System (Bio-Rad, USA), and clones were counted using computer software (Bio-Rad Quantity One).

Cell-cycle analysis

For cell cycle analysis, the cells were detached via trypsinization, washed with PBS and fixed overnight in ice-cold ethanol at 4°C. Fixed cells were then washed with PBS and incubated in 0.5 ml of staining solution (0.1 mg/ml RNase A and 0.05 mg/ml propidium iodide) for 30 minutes at room temperature. Cell cycle distributions were detected by flow cytometry (FACSort; Becton Dickinson, USA) and analyzed using ModFit software.

Dual-luciferase reporter gene assay

The pmirGLO Dual-Luciferase Vector (Promega, USA) was used for the miRNA target detection assay. The putative miR-103a-3p-binding site in the ATF7 3′ UTR was used to construct wild-type or mutant reporter duplexes (named WT for wild type and MT for mutant type; the sequences are listed in Table 2). All duplexes were chemically synthesized and cloned into the pmirGLO vector between the SacI and XhoI sites according to the manufacturer’s instructions (Dual-Luciferase Reporter Assay System, Promega, USA). Luciferase reporter gene assays were performed using HEK293 cells seeded in 96-well plates at 90% confluence. For the ATF7 3′ UTR luciferase reporter assay, wild-type, mutant-type or empty pmirGLO vectors were cotransfected into HEK293 cells with 30 nmol/L miR-103a-3p using jetPRIME. Reporter gene assays were performed at 24 h post-transfection using the Dual-Luciferase Assay System (Promega, USA). The analysis was performed with three repeated experimental results.

Western blot analysis

For Western blotting, cells were lysed using RIPA buffer (Sigma-Aldrich, Germany). Lysates were then collected by centrifugation at 14, 000 × g for 20 min. Protein concentrations were determined using the BCA protein assay kit (Thermo Fisher, USA). Equal amounts of protein lysates were subjected to SDS-PAGE, transferred to methanol-activated PVDF membranes and blocked with 5% non-fat dry milk in Tris-buffered saline (pH 7.4) containing 0.1% Tween (TBST) for 2 h. A prestained ladder (Precision Plus Protein™ Dual Color Standards, Bio-Rad, USA) was used as a molecular size marker. The membranes were incubated overnight with primary antibodies (Abcam Biotechnology, UK) at 4°C. Anti-rabbit or anti-mouse (Santa Cruz Biotechnology, USA) secondary antibodies were then used for 1 h at room temperature. Protein expression was normalized to β-actin levels in each sample. All Western blotting analysis were repeated at least three times under similar conditions.

Bioinformatics analysis

The TCGA database with designed web tool (https://xena-browser.net/) was used to investigate the relative expression of miR-103a-3p and ATF7 and determine the correlation between their expression levels in stomach cancer. For miR-NA target prediction, two popular public databases, Targetscan (http://www.targetscan.org/vert_71/) and miRDB (http://mirdb.org/miRDB/), were used and crosschecked.

Statistical analysis

Data are expressed as the mean ± SEM from at least three separate experiments. The differences between groups were analyzed using Student’s t test, except where noted. All tests were two-sided. Differences were considered statistically significant at P < 0.05. All statistical analyses were performed using SPSS 13.0 software (SPSS Inc, USA). The linear correlation coefficient (Pearson’s r) was calculated to estimate the correlation between miR-103a-3p and ATF7 expression.

miR-103a-3p is frequently increased in both human gastric carcinoma tissue samples and in the TCGA database

To investigate the role of miR-103a-3p in GC, the expression of miR-103a-3p in 33 paired GC and adjacent nontumor stomach tissue samples was analyzed using qRT- PCR. Compared with their peritumor counterparts, a significant up-regulation of miR-103a-3p was observed in 26 of 33 gastric cancer samples (Fig. 1A). Additionally, TCGA stomach cancer cohort analysis showed that miR-103a-3p was significantly upregulated based on either paired or unpaired data (Figs. 1B and 2C), which matched the trend in the 33 paired gastric cancer tissues. In addition, we examined the correlation of miR-103a-3p levels with histological grade and tumor stage in the 33 paired GC tissue samples. miR-103a-3p expression was upregulated in 18 poorly differentiated tumor tissues (Table 3) or in 24 tumor stage III/IV samples (Table 4), indicating that miR-103a-3p may act as a tumor promoter in GC.

miR-103a-3p increases MKN-45/SGC-7901 cell growth, and miR-103a-3p overexpression promotes the S-G2/M transition in vitro

To explore the role of miR-103a-3p in GC, MKN-45/SGC-7901 cells were transfected with miR-103a-3p mimics or negative controls. qRT-PCR was performed to assess miR-103a-3p expression levels after transfection of miR-103a-3p mimics or the negative control. A greater than 70-fold increase of miR-103a-3p expression was observed in MKN-45/SGC-7901 cells transfected with 30 nmol/L miR-103a-3p relative to the cells transfected with 30 nmol/L miR-103a-3p negative control (Fig. 2A). MTT assays were used to investigate the role of miR-103a-3p in GC cell proliferation, showing that miR-103a-3p overexpression promoted the proliferation of MKN-45/SGC-7901 at 72 h after transfection (Fig. 2B).

To further explore the tumor-contributor role of miR-103a-3p in GC cells, colony formation assays were used after a similar transient transfection. miR-103a-3p-transfected cells displayed a higher number and larger size of colonies compared with the negative control (Fig. 2C, quantification data for Fig. 2C is illustrated in Supplementary Fig. S3). Taken together, MTT and colony formation showed that miR-103a-3p promoted the proliferation of GC cells in vitro. To further investigate the mechanisms through which miR-103a-3p promotes GC cell proliferation, Cell cycle assays were performed to detect the percentage of cells in the G1, S, and G2 phases with flow cytometry. MKN-45/SGC-7901 cells were transfected with miR-103a-3p or the negative control. Results showed that miR-103a-3p overexpression induced the accumulation of the G2/M-population (Fig. 2D, FACS data for Fig. 2D is illustrated in Supplementary Fig. S1), suggesting that miR-103a-3p promotes the S-G2/M transition.

In addition, miR-103a-3p inhibitor was used to validate the proliferative role of miR-103a-3p by eliminating endogenous miR-103a-3p in human GC cells. Reducing the endogenous expression of miR-103a-3p in MKN-45/SGC-7901 cells using miR-103a-3p inhibitor resulted in decreased cell viability and colony formation potential (Figs. 2E–2G; Quantification data for Fig. 2G is illustrated in Supplementary Fig. S4). Furthermore, cell-cycle analysis showed that inhibitor-transfected MKN-45/SGC-7901 cells were preferentially blocked in S-phase (Fig. 2H, FACS data for Fig. 2H is illustrated in Supplementary Fig. S2). These results suggested that the baseline expression of miR-103a-3p is sufficient to initiate proliferation during S-G2/M in GC cells, thereby stimulating further malignancy progression.

miR-103a-3p promotes gastric cancer cell proliferation by directly targeting ATF7

To reveal the mechanisms by which miR-103a-3p promotes the S-G2/M transition, we searched for miR-103a-3p target genes. Among many putative targets, ATF7 is of particular interest because of its transcription factor role in cellular functions. As we all know, transcription factors play important roles in cell proliferation, division and death. Deregulation of transcription factors is a pervasive theme across many forms of human cancer. Targeting transcription factors can be an effective way in treating cancer. As shown in Fig. 3A, a region complementary to miR-103a-3p seed region was identified in the 3′ UTR of human ATF7. To validate whether ATF7 was a direct target of miR-103a-3p, a Dual-Luciferase Reporter System containing the wild-type or mutant 3′ UTR of ATF7 was used. Cotransfecting miR-103a-3p with wild-type ATF7 vector induced low luciferase activity, whereas cotransfection with the mutant ATF7 vector had no effect (Fig. 3B), suggesting that miR-103a-3p directly and specifically bound the predicted binding site in the 3′ UTR of ATF7. To identify whether miR-103a-3p expression affected endogenous ATF7 expression, miR-103a-3p, miR-103a-3p inhibitor or negative controls were transfected into MKN-45 cells. Decreased ATF7 protein levels were observed after transfection with miR-103a-3p, whereas cells transfected with miR-103a-3p inhibitor demonstrated an enhanced expression of ATF7 protein. No decrease was detected in ATF7 mRNA levels, indicating that miR103a-3p regulates ATF7 via translational silencing (Fig. 3C). To further investigate the relationship between ATF7 and miR-103a-3p, we also examined ATF7 expression levels in the TCGA stomach cancer cohort as well as the correlation between ATF7 and miR-103a-3p. We found that miR-103a-3p levels were inversely correlated with ATF7 expression (Fig. 3D). Taken together, the above data suggest that miR-103a-3p directly regulates ATF7 expression in GC cells.

Knocking down ATF7 mimics the effects of miR-103a-3p overexpression in gastric cancer cells

Two siRNAs were designed to silence ATF7 and were used to identify whether ATF7 is involved in the tumor-promoting effects of miR-103a-3p. ATF7 was specifically knocked down by at least 90% at both the mRNA and protein level using siRNA in MKN-45/SGC-7901 cells (Figs. 4A and 4F). The results of MTT and colony formation assays revealed that in MKN-45 cells, ATF7 silencing stimulated cell proliferation, and cell-cycle analysis results showed that siATF7 increased the accumulation of cells at the G2/M phase (Figs. 4B–4D, quantification data for Fig. 4C is illustrated in Supplementary Fig. S5), consistent with the effects of miR-103a-3p overexpression. Similar proliferation effects after ATF7 silencing were also observed in SGC-7901 cells, which is consistent with miR-103a-3p overexpression as well (Figs. 4G–4I, quantification data for Fig. 4H is illustrated in Supplementary Fig. S5). In addition, Western blotting was used to analyze S-G2/M-related regulators after miR-103a-3p overexpression or ATF7 silencing in MKN-45 and SGC-7901 cells. As shown in Figs. 4E and 4J, both miR-103a-3p overexpression and ATF7 silencing promoted the expression of CDK2, an essential regulator of the S-G2/M transition. Furthermore, miR-103a-3p overexpression and ATF7 knockdown obviously decreased p27. Moreover, reducing the expression of miR-103a-3p using miR-103a-3p inhibitor showed adverse effects at the protein level. These results demonstrated that miR-103a-3p and its target gene ATF7 share similar cellular and molecular effects in both MKN-45 and SGC-7901 cells.

ATF7 overexpression eliminated the effects of miR-103a-3p

To further understand the role of miR-103a-3p in GC as a tumor promoter via ATF7, we constructed an ATF7 overexpression vector. MKN-45/SGC-7901 cells were transfected with the ATF7 overexpression vector, and an empty vector was used as a control. We observed a significant increase in both ATF7 protein and mRNA levels after transfection with the ATF7 overexpression vector (Fig. 5A). Upon cotransfection with miR-103a-3p and the ATF7 vector, ATF7 overexpression counterbalanced the tumor-stimulating effect of miR-103a-3p in GC cell proliferation (Fig. 5B). Cell-cycle assays were used to investigate the effect of ATF7 overexpression on cell-cycle progression in MKN-45/SGC-7901 cells. ATF7 overexpression blocked the S/G2 transition in MKN-45/SGC-7901 cells. Moreover, compared with the cotransfection of miR-103a-3p and ATF7-ctrl, cotransfection of miR-103a-3p and ATF7 also blocked the S/G2 transition in GC cells (Fig. 5C). Furthermore, the expression of cell-cycle regulators was investigated using Western blotting after cotransfection. Compared with miR-ctrl and ATF7-ctrl cotransfection group, miR-ctrl and ATF7 cotransfection resulted in the downregulation of CDK2 in GC cells. Notably, compared with miR-103a-3p and ATF7-ctrl cotransfection, the expression of regulators was disrupted after cotransfection with miR-103a-3p and ATF7 (Fig. 5D).

Taken together, these results further suggest that miR-103a-3p has an oncogenic role in GC by directly targeting ATF7.

Over the past decades, studies have shown that the dysregulation of miRNAs can control cell proliferation in GC development and progression. miR-103a-3p, previously referred to as miR-103 or miR-103a, shows an increased expression in multiple cancers. In liver cancer, miR-103 potentially functions as an oncogene in hepatocellular carcinoma by inhibiting AKAP12 expression (Xia et al., 2016). In addition, miR-103 promotes cell proliferation and migration by targeting DICER and PTEN in colorectal cancer cells (Geng et al., 2014). These data suggest that miR-103a-3p is an oncogene. In GC, although there is debate over the expression level and function of miR-103a-3p in human GC (Liang et al., 2015), most research revealed that miR-103a-3p is upregulated in human GC studies (Shrestha et al., 2014). Additionally, mouse model study indicated that miR-103 was significantly upregulated in serum in both early and advanced-stage and considered promising biomarkers for early detection of gastric cancer (Rotkrua et al., 2013). Moreover, another study also demonstrated that high expression of miR-103 in advanced GC tissues may be a high risk factor associated with tumor penetration and poor long-term survival (Kim et al., 2013). Until now, few reports have investigated the function and underlying mechanisms of miR-103a-3p in GC.

The present research confirmed that miR-103a-3p was up-regulated in 33 paired gastric cancer tissues, consistent with our findings in the TCGA stomach cancer cohort, suggesting that miR-103a-3p may be a tumor promoter in GC. To further explore the effect of miR-103a-3p in vitro, gain- and loss-of-function studies were performed. Overexpression of miR-103a-3p in GC cells promotes cell proliferation, clonogenicity, and enhances the S-to-G2/M transition, while inhibition of miR-103a-3p suppressed GC cell proliferation, clonogenicity, and blocked the S-G2/M transition. This result verified the results previously reported in literatures, provided a more comprehensive understanding of the oncogenic role of miR-103a-3p during gastric cancer development.

To investigate the underlying mechanisms of miR-103a-3p in GC, bioinformatics and dual-luciferase reporter assays were used to identify the potential target through which miR-103a-3p regulates cell proliferation. An inverse correlation was observed between miR-103a-3p and ATF7 expression in the TCGA stomach cancer cohort, and dual-luciferase reporter assays confirmed that miR-103a-3p suppressed ATF7 expression by binding directly to its 3′ UTR.

Studies on ATF7 are relatively scarce. ATF7 is primarily mentioned alongside ATF2, which is structurally very similar to ATF7, especially within the bZIP (basic region/leucine zipper) DNA-binding and dimerization domains (Vinson et al., 2002). ATF7 can form homodimers as well as heterodimers with c-jun or c-fos through its C-terminal leucine zipper region. Depending on the cellular context, the composition of the dimeric complexes determines the regulation of growth, survival or apoptosis (Vlahopoulos et al., 2008). In mice, B cell-specific deletion of ATF2 and ATF7 significantly accelerates the onset of Em-Myc-induced lymphoma. In addition, loss of ATF2/7 desensitizes Em-Myc lymphoma cells to spontaneous and stress-induced apoptosis (Walczynski et al., 2014). In colorectal cancer, ATF7 was negatively correlated with pathological stage and positively correlated with 5-year overall survival or 5-year progression-free survival (Guo et al., 2015b), suggesting that ATF7 may play a potential tumor suppressor role in colorectal cancer. To reveal the functions of ATF7 in GC, we overexpressed ATF7 and observed that ATF7-treated MKN-45/SGC-7901 cells showed a significant accumulation of ATF7 protein, consequently resulting in restricted cell proliferation, colony formation and the inhibition of the S-G2/M transition in both MKN-45 and SGC-7901 cells. Furthermore, ATF7 overexpression eliminated the effect of miR-103a-3p on GC cells, and ATF7 silencing improved the tumorigenic properties of GC cells, similar to that of miR-103a-3p overexpression. These results suggested that ATF7 may also act as a tumor suppressor in GC. Further gain- and loss-of-function experiments revealed that miR-103a-3p expression was positively correlated with the cell cycle-related protein CDK2 and that ATF7 expression was negatively correlated with CDK2 in GC cells. Moreover, P27 protein expression showed a positive trend compared with ATF7. These findings suggest that miR-103a-3p may contribute to cell-cycle progression by targeting ATF7 and regulating ATF7 related signaling pathways in GC, presenting the first evidence that increased ATF7 can indirectly downregulate CDK2 expression and upregulate P27 expression in vitro.

In summary, for the first time, we investigated the roles of miR-103a-3p and its target gene, ATF7, in controlling the cell cycle and their potential implication in pathological processes in GC cell lines. Moreover, we provided novel evidence that miR-103a-3p, as a tumor promoter, can enhance the growth of GC cells, suppress ATF7 expression and regulate related cell-cycle signaling pathways. These findings highlight the function of miR-103a-3p in GC cells, offering new insights into the regulatory network of the cell cycle, which may be a potential therapeutic strategy for the treatment of GC in the future.