miR-107 regulates tumor progression by targeting NF1 in gastric cancer

Our previous genome-wide miRNA microarray study revealed that miR-107 was upregulated in gastric cancer (GC). In this study we aimed to explore its biological role in the pathogenesis of GC. Integrating in silico prediction algorithms with western blotting assays revealed that miR-107 inhibition enhanced NF1 (neurofibromin 1) mRNA and protein levels, suggesting that NF1 is one of miR-107 targets in GC. Luciferase reporter assay revealed that miR-107 suppressed NF1 expression by binding to the first potential binding site within the 3′-UTR of NF1 mRNA. mRNA stable assay indicated this binding could result in NF1 mRNA instability, which might contribute to its abnormal protein expression. Functional analyses such as cell growth, transwell migration and invasion assays were used to investigate the role of interaction between miR-107 and its target on GC development and progression. Moreover, We investigated the association between the clinical phenotype and the status of miR-107 expression in 55 GC tissues, and found the high expression contributed to the tumor size and depth of invasion. The results exhibited that down regulation of miR-107 opposed cell growth, migration, and invasion, whereas NF1 repression promoted these phenotypes. Our findings provide a mechanism by which miR-107 regulates NF1 in GC, as well as highlight the importance of interaction between miR-107 and NF1 in GC development and progression.

Plasmids construction, transient transfection and luciferase assay. The NF1 Wild-type reporter plasmid was constructed by cloning a 490-bp DNA fragment of NF1 3′ -untranslated region (UTR), which is containing the two predicted miR-107 target sites, into the HpaI site of the pGL3-promoter vector (Promega, WI, USA; Fig. 1C, upper panel). Based on the Wild plasmid, the Mut-1, Mut-2 and Mut-both mutant constructs were made by removing the first, second and both predicted miR-107 target sites of the NF1 3′ -UTR fragment, respectively.
The plasmids or in combination with siRNAs were transiently transfected into cells using Lipofectamine 2000 (Invitrogen, CA, USA) following the manufacturer's protocol. As an internal standard, all plasmids were cotransfected with 10 ng pRL-SV40, which contained the Renilla luciferase gene. The pGL3-promoter vector without an insert was used as a negative control. After transfection at 48 h, luciferase activity in lysates was measured with a Dual-Luciferase Reporter Assay System (Promega, WI, USA). The reporter assay was performed with three biological replicates and three technical replicates.
Real-time PCR analysis. Total RNA was isolated from cell lysates according to the instructions provided by the manufacturer of TRIZOL (Invitrogen, CA, USA). Reverse transcription was performed using the TaqMan MicroRNA Reverse Transcription Kit (ABI, CA, USA). The expression level of miR-107 was assessed using the specific TaqMan MicroRNA Assay kit (ABI, CA, USA), and normalized to U6. To determine the expression levels of NF1 mRNA, the cDNA was amplified by real-time PCR with SYBR Green RT-PCR kit (Takara, Japan). The expression of GAPDH was used as an internal control. The following primers were used for amplification: 5′ -CGAATGGCACCGAGTCTT AC-3′ (F) and 5′ -GACCAGTTGGACGAGCCC -3′ (R) for NF1; 5′ -GCACCGTCA AGGCTGAGAAC -3′ (F) and 5′ -TGGTGAAGACGCCAGTGGA -3′ (R) for GAPDH. Real-time PCR was performed in triplicate on ABI 7900HT Real-Time PCR System (ABI, CA, USA). Relative expression was calculated using the comparative Ct method. The Real-time PCR assay was performed with three biological replicates and three technical replicates.
Cell growth assay. The transfected MGC803 cells were seeded in 96-well plates. Cell culture was continued for 24 h, 36 h and 48 h and subsequently incubated with MTT reagent (5 mg/ml) at 37 °C for 4 h. MTT assay was performed as described elsewhere 16 . Wound healing assay. The transfected MGC803 cells (5 × 10 5 ) were cultured in 6-well plates to monolayer.
The cells were then starved in serum-free medium for 12 h before a wound approximately 2 mm in width was made with a cell scraper. The wound was allowed to heal for 3 d in a fresh medium containing 1% fetal bovine serum. The wounded monolayer was photographed at the indicated day using a fluorescent microscopy (IX70, Olympus, Japan) with a 10 × objective. Wound closure was measured as a percentage of original wound width.
Transwell assay. Transwell assay was performed using 12-well Transwell chambers (Corning Costar, Cambridge, MA, USA) with a pore size of 8 μ m. For Transwell migration, Cells (1 × 10 5 ) were seeded in serum-free medium in the upper chamber and incubated at 37 °C for 8 h. Afterward, the cells remained in the upper chamber were carefully removed with a cotton swab, whereas the cells having traversed to reverse face of the membrane were fixed in methanol, stained with crystal violet (0.04% in water), and counted. Transwell invasion assay was done under the same conditions as the Transwell migration assays, but in Matrigel-coated transwells (BD Biosciences, MA, USA) and incubation for 24 h. Statistical analysis. Quantified data are presented as mean ± SEM. The difference between two independent means was assessed by t-test. All P-values are two-sided, and P < 0.05 was considered to statistically significant. Statistical analyses were carried out using SAS software (V.9.1.3; SAS Institute, Cary, NC, USA) and R software (V.2.15.0; The R Foundation for Statistical Computing).

miR-107 is upregulated in GC tissues.
We detected the expression of miR-107 in 55 paired of cancer and normal tissues. The expression level of miR-107 in cancer tissues was significantly increased compared with normal tissues (P = 0.0003, Fig. 1A). The same result was found in TCGA (Fig. 1B). As shown in Table 1, the aberrant miR-107 expression in GC tissues was associated with tumor sizes and depth of invasion.
NF1 is a target of miR-107. miRNA carries out its biological function via regulating the expression of its target genes through base-pairing with endogenous mRNAs. As an initial step to identify putative miR-107 targets, four commonly used algorithms (i.e., TargetScan, PicTar, Microcosm Targets v5, and miRanda) were applied to predict miR-107 target genes; and finally, there were 22 genes predicted by all four algorithms (Table 2). Among these 22 predicted target genes, NF1 (Neurofibromin 1) stood out for the presence of two evolutionarily conserved binding sites, suggesting collaborative binding and biologically effective interaction ( Fig. 2A). NF1 is a tumor suppressor, and loss of NF1 expression has been linked to tumor development and progression 17,18 . To confirm that NF1 is a target of miR-107 in GC, the endogenous NF1 levels were measured in MGC803 cells at 48 h after miR-107 knockdown. The results showed that NF1 mRNA expression was significantly increased after miR-107 knockdown, compared with the negative control (Fig. 2B). Cichowski et al. 19 reported that NF1 degradation could be rapidly triggered in response to growth factors and re-elevated shortly after growth factor treatment. Replicate growth factor treatment of MGC803 cells with 10% goat serum revealed that NF1 was rapidly degraded within 5 min and quickly re-elevated in the NC processing group, whereas evident degradation of NF1 was seen in 10 min in the si-107 treated group, possibly attributing to a rise in the expression of NF1 induced by miR-107 knockdown (Fig. 2C). miR-107 downregulates NF1 expression by directly targeting its 3′-UTR. To establish a direct interaction between miR-107 and the 3′ -UTR of NF1, we cloned the NF1 3′ -UTR portion containing the two miR-107 target sites into a firefly luciferase reporter construct, designated as Wild-type Reporter (Fig. 3A), and used it for transient transfection into MGC803 cells. A significant reduction (87%) in the luciferase activity of the Wild-type Reporter was observed compared with the pGL3-promoter reporter (Fig. 3B). To evaluate whether the reduction of luciferase activity was associated with miR-107 targeting, the Wild-type Reporter was co-transfected with si-107 (miR-107 inhibitor) or NC (negative control) into the SGC7901 and MGC803 cells. The results demonstrated a significant increase in luciferase activity in both cell lines (1.48-fold for SGC7901 and 1.71-fold for MGC803) treated with si-107 compared with their NC-treated counterparts (Fig. 3C), indicating a direct interaction between miR-107 and the NF1 3′ -UTR. As mentioned above, there are two potential miR-107 target sites within the NF1 3′ -UTR. In order to identify the bona fide miR-107 target site, we mutated the first, second and both predicted target sites of the Wild-type Reporter, designated as Mut-1, Mut-2 and Mut-both reporter. Likewise, the three mutated reporters were co-transfected with si-107 or NC into MGC803 and SGC7901 cells, respectively, while the Wild-type Reporter was used as a positive control. The results revealed that in both cell lines, the luciferase activities were only increased in the cells co-transfected with the Wild and Mut-2 reporters after miR-107 knockdown compared with their NC-treated counterparts (Fig. 3D), indicating that the first predicted target site is the authentic miR-107 target site.
Regulation by miR-107 results in NF1 mRNA degradation. NF1 3′ -UTR sequence analysis revealed that there is an AU-rich element (ARE element, ATTTA) adjacent to the second predicted miR-107 target site (Fig. 4A). It was reported that miRNA could regulate the mRNA stability through binding to the ARE element in the 3′ -UTR of mRNA 20 . In order to determine whether the increased expression of NF1 mRNA was a consequence of the enhanced mRNA stability after miR-107 knockdown, we measured the mRNA stability of NF1. As  shown in Fig. 4B, the remaining NF1 mRNA expression was more in the si-107 treated group than in NC-treated group (1.08 ± 0.11 vs. 0.66 ± 0.09, P < 0.05) after ActD treatment for 2 h, implying that si-107 treatment significantly increased the stability of NF1 mRNA in contrast with the negative control (NC) and miR-107 may have an effect on NF1 mRNA decay and subsequent protein expression.

miR-107 regulates cell proliferation, migration and invasion by targeting NF1. It has reported
that NF1 acts as a tumor suppressor to arrest cell growth, migration, and invasion 18,21 . We next examined the role of the interaction between miR-107 and NF1 on cell growth, migration, and invasion of GC cells. Apart from si-107 and its negative control, the MGC803 cells were further treated with si-NF1 (Fig. 5A, left panel) to evaluate whether the biological effect of miR-107 is through directly targeting NF1. The cell growth assay indicated that the cells treated with si-107 illustrated a moderate reduction in cell growth (25%) compared with its negative control (Fig. 5A, right panel). Furthermore, the cells with miR-107 knockdown displayed a significant reduction in cell migration using wound healing (46%, Fig. 5B) and Transwell migration (44%; Fig. 5C, panel a) assays. A strong decrease in invasive ability of the cells transfected with si-107 was also observed by Transwell invasion assay (70%; Fig. 5C, panel b). It was noteworthy that si-NF1 could reverse inhibition of cell proliferation (Fig. 4A, right panel), migration and invasion (Fig. 5, panel a and b, respectively) by miR-107 knockdown, indicating that miR-107 exerted its function by directly targeting NF1.

Discussion
Recent study indicates that the role of miR-107 in tumor development and progression is contradictory and differs in a context-dependent manner. miR-107 could function as tumor suppressor gene by inducing cell cycle arrest in lung cancer and glioma 11,12 , whereas serve as oncogene for promoting tumor invasion and metastasis in breast and gastric cancer 13,14 .
In GC, Li et al. 14 found that miR-107 was upregulated in GC and promoted tumor invasion and metastasis by negatively regulating DICER1. Song et al. 22 showed that miR-107 was capable of advancing proliferation of GC cells by targeting CDK8. However, Feng et al. 23 reported an opposite role of miR-107 that it was down-regulated in GC and acted as a tumor suppressor to oppose proliferation and invasion of GC cells. In our present study, functional analyses revealed that miR-107 stimulated cell growth, migration, and invasion, in agreement with Li et al. 14 and Song et al. 22 , indicating that miR-107 could function as onco-miRNA in GC through multiple targets, including Dicer, CDK8, and NF1. The biological processes and molecular mechanisms underlying miR-107 still remain unclear in GC, and further functional studies are warranted to address these unsolved issues.   NF1 is a GTPase which converts active Ras-GTP to its inactive form, thereby negatively regulating Ras signaling 24 . Loss of NF1 expression by mutation 25 , copy number alteration 26 , or miRNA regulation 18 can result in constitutive activation of Ras, which can mediate signal transduction via multiple pathways, including Ras/Raf/MEK/ERK pathway, leading to various cancer phenotypes like decreased apoptosis and increased proliferation and migration 27 . It is biologically plausible that the observed miR-107 phenotype in GC may be attributable to ERK activation induced by decreased NF1 expression. Recently, Lenarduzzi et al. 18 reported that miR-193b enhanced tumor progression via down regulation of NF1, which in turn leading to activation of ERK, resulting in proliferation, migration, invasion, and tumor formation. miRNAs exert their function by repressing translation and/or triggering degradation of mRNA targets 28 . It is reported that miRNA is involved in the ARE-mediated mRNA instability 20 . By bioinformatics analysis, we identified an ARE element adjacent to the second predicted miR-107 target site in NF1 3′ -UTR. Our results showed that NF1 mRNA stability was enhanced after miR-107 knockdown. We proposed that miR-107 was involved in the ARE-mediated mRNA instability. Further studies on the role of direct interaction between miR-107 and ARE element in the regulation of NF1 mRNA stability are warranted. There is a predicted HuR binding site adjacent to the second predicted miR-107 target site in NF1 3′ -UTR (Fig. 3A). Haeussler et al. 29 found that mRNA binding protein HuR could interact with ARE element in the 3′ -UTR of NF1, thereby negatively regulating the expression of mRNA on the posttranscriptional level. Therefore, the role of HuR in the regulation of NF1 mRNA stability by miR-107 also merits further investigation.
Some limitations in this study should be addressed. First, the function of miR-107 was evaluated through miR-107 knockdown in a loss-of-function model. Gain-of-function studies via overexpression of miR-107 in GC cell lines are needed to verify our findings. Second, future in vivo studies are needed to validate the role of miR-107 in tumor development and progression. It is reported by Li et al. that silencing the expression of miR-107 could suppress the migration and invasion of GC cell in nude mice 30 . We propose that the miR-107 may play an important role in vivo in GC carcinogenesis through multiple targets including NF1.
In summary, our data demonstrated that miR-107 targeted NF1, and suppression of miR-107 enhanced proliferation, migration and invasion of GC, whereas repression of NF1 promoted these phenotypes. We propose that miR-107 may serve as a useful therapeutic strategy for advanced GC. Means ± SEM from three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001.