The Long Non-Coding RNA SNHG16 Promotes NPC Cell Progression by Competitively Binding miR-23b-3p

Background: This study aims at verifying the effect of non-coding RNA SNHG16 on promotes NPC cell progression via binding miR-23b-3p. Methods: The expression of non-coding RNA SNHG16 was detected by qRT-PCR in cell lines including c666-1 and HONE-1. Si-MCM6 and si-SNHG16 are transfected to cells to verify their effects on cell proliferation and apoptosis. MTT is used to measure cell viability while ow cytometry assay and transwell assay were used for cell apoptosis, cell cycle and invasion respectively. The expression level of MCM6 was determined by western blot. Relationships between mRNA MCM6 and lncRNA SNHG16 were explored by qRT-PCR and nude mouse tumorigenicity assay. Results: The MCM6 was overexpressed in NPC tissues and lncRNA SNHG16 showed the same trend. Those two factors were correlated with high cancer stage. The expression of MCM6 was decreased after si-SNHG16 and dual luciferase reporter system demonstrated their combine with miR-23b-3p. Further we explored the down-regulation of lncRNA SNHG16 could inhibit NPC cell proliferation, colony formation and also accelerate cell apoptosis rate. And this result could be altered by adding miR-23b-3p inhibitor. Conclusion: The lncRNA SNHG16 is able to promote the NPC proliferation via binding miR-23b-3p, which has potential for future treatment.


Introduction
Nasopharyngeal carcinoma (NPC) is a highly chemo sensitive cancer, a malignancy derived from of the nasopharynx epithelial cells, is endemic in Southern China and Southeast Asia as well as association with Epstein-Barr virus (EBV). [1] Currently, concurrent radio chemotherapy is a mainstay of curative treatment for NPC. [2] Although excellent control has been achieved with the fast development of irradiation techniques and their combination with chemotherapy, local recurrence still mainly results in the failure of treatment, at a rate of approximately 8-10%. [3] The resistant to radiotherapy and chemotherapy remains the main cause of treatment failure of NPC. [4] Therefore, there is still an urgent need to develop effective therapeutic strategies to overcome chemoresistance and gure out its underlying molecular mechanism.
With the development of sequencing technologies, more than 90% of the human genome is actively transcribed, although only a minority of it is translated into proteins. Long non-coding RNA (lncRNA) is generally de ned as RNA transcripts contains over 200 nucleotides without protein coding potential. [5] Increasing evidence has indicated that lncRNAs are implicated in a variety of pathophysiological processes, such as gene expression, cell proliferation, apoptosis, and tumorigenesis. [6] Importantly, lncRNA was found to function as either oncogene or anti-oncogene to participate in the pathogenesis and development of many kinds of diseases including cancers. [7] It was recently identi ed that another lncRNA, Small Nucleolar RNA Host Gene 16 (SNHG16), might act as a potential cancer biomarker or functional cancer development modulator in various human cancers [8][9][10]. Yet, the function of SNHG16 in NPC still remains unclear and is necessary to explore.
MicroRNAs (miRNAs) belongs to small non-coding RNAs family, which can regulate the posttranscriptional level of gene expression through binding to target mRNAs, leading to target mRNA degradation or translation suppression. [11] As highly conserved, short non-coding RNAs contains [21][22][23][24][25] nucleotides, miRNAs negatively affect gene expression at the post-transcriptional level by binding to the 3′ untranslated region (3′-UTR) of target mRNA, leading to its degradation or translational repression. [12] Growing evidence implies that miRNAs have been related to the disease pathogenesis. Therefore, discerning the relationship between miRNAs and diseases has become an important goal in the biological context. [13] Furthermore, it has been well documented that lncRNAs function as molecular sponges for miRNA by negatively regulating its expression. [4] Nevertheless, the relationship of SNHG16 and miRNA on pathological process of NPC cells remains to be investigated.
Identifying the speci c lncRNAs and miRNAs related to NPC was the target of this study. To attain this object, we used microarray-based gene expression pro ling to characterize the differentially expressed mRNAs and lncRNAs in NPC cells compared with controls. We found that SNHG16 may mediate NPC progress through miR-23b-3p. Our present study provides a deeper understanding of the NPC pathogenesis and might indicate a new direction for the treatment of NPC.

Cells and tissues collection
Firstly, we collected 15 NPC tissues and 15 adjacent normal tissues from The A liated Hospital of Shaanxi University of Chinese Medicine. Then the samples were immediately encapsulated, put into liquid nitrogen and stored at -80 °C for subsequent analysis. Before the experiment, we received informed consents from each patient. The experimental protocol was approved by the Ethics Committee of West China Hospital, Sichuan University.

Bioinformatics analysis
The data of RNA expression microarray (GSE12452) from 41 samples including 31 NPC tissues and 10 non-NPC tissues was obtained from Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE12452). The differentially expressed mRNA and lncRNA were selected when |the logarithm of fold change value| > 1 and P < 0.05 based on R project (https://www.r-project.org/) analysis. Then GSEA v3.0 software (http://software.broadinstitute.org/gsea/index.jsp) was used for gene set enrichment analysis (GSEA). All results are shown by "joyplot" and "dotplot" packages through R project. Based on above research we chose cell cycle pathway to nd the deep relationship between differential mRNAs and lncRNAs.

Plasmid construction and cell transfection
On the GenePharma (Shanghai, China), the transfected materials MiR-23b-3p inhibitor, si-MCM, si-SNHG16, sh-MCM and sh-SNHG16 were purchased. Twenty-four hours before transfection, c666-1 and HONE-1 cells in the exponential phase were digested by pancreatin and made into cell suspension. Si-MCM6 and si-SNHG16 were used for cell experiment and the in vitro experiment. After trypsinization from asks, cells were cultured in six-pore plates, incubated at 37 °C with 5 % CO 2 for 18-24 h. Three hours before transfection, cells at about 80-90 % con uency were changed to the serum and antibiotic-free media. Then, cells were transfected using Lipofectamin 3000 reagent (Life Technologies, Gaithersburg, MD, USA) referring to manufacturer's instructions and incubated at the same conditions as above for 48 h. In addition to the in vivo experiment, based on the manual, concentrated lentiviral solutions of sh-SNHG16 and sh-MCM6 were mixed with two wells of nutrient solution contains c666-1 cells, respectively.
Finally, the cells were digested by pancreatin and injected into mice after incubation for 48 h.

RNA isolation and qRT-PCR
The RNA in the two cell lines was extracted with TRIzol reagent (Invitrogen) following the instruction of manufacturer

Invasion assays
Invasion assays were applied using cell culture inserts with 8 μm pore transparent polyethylene terephthalate lters (Becton Dickinson, Bedford, MA, USA) coated with Matrigel. A total of 3×10 4 cells disperse in 200 μl culture medium without serum were added to each insert, and 500 μl culture medium with 10 % FBS was added to the bottom chamber. Then the cells on the upper lter were removed at the end of incubation for 24 h at 37 °C, and those cells invaded the lower membrane were xed in methanol and dyed with crystal violet. Five optical elds for each lter with triplicate inserts were randomly selected to calculate the number of invaded cells.

Western blot
Western blotting was performed as described in previous study. The BCA Kit (Sigma-Aldrich) was used to quantify the protein concentration. Then samples were separated through electrophoresis on 10-12% sodium dodecyl sulfate polyacrylamide gels, and the separated proteins were moved to a polyvinylidene uoride membrane (Millipore, Billerica, MA, USA), which was blocked with 5 % bovine serum albumin in Tris-buffered saline (TBS) buffer for 30 min at room temperature later and incubated by rst antibody (MCM6 rabbit anti-human, 1:1000, Abcam, Cambridge, MA, USA; GAPDH rabbit anti-human, 1:2500, Abcam) at 4˚C overnight. After that, secondary antibody horseradish peroxidase-conjugated goat antirabbit IgG (1:5000, Abcam) were used, and immunoreactivity was visualized by the ECL western blotting detection system (GE Healthcare, Amersham, UK). Densitometric analysis of immunodetected bands was performed using Image Analysis software (Biorad).
Luciferase reporter assays 293T cells (l×10 5 ) were transferred to 24-well plates in each well and incubated before transfection. Then the cells were co-transfected using miR-23b-3p, the wild type or mutant SNHG16 and MCM6. After 48 h of transfection, the cells were washed by PBS before incubated with shake for 15 min at room temperature.
Then cell lysis buffer was collected at 4 °C and centrifuged for 2 min. The supernatant was collected and refrigerated at -80 °C refrigerator. Cell lysis buffer was added to 96-well plate with 10 μl per well. During the detection, 30 μl luciferase reagents II (Promega, Madison, WI, USA) were added. Then 30 μl stop buffer was added to terminate the activities. The luciferase activities were detected by the Dual-Glo luciferase reporter assay kit (Promega) for 48 h post transfection.

Flow Cytometric Analysis of Apoptosis
A total of 2×10 5 cells of the transfection group and NC group were collected after post transfection, washed before resuspended with pre-cooled PBS following the instruction of Annexin V apoptosis detection kit (Life Technologies). Then the cells were incubated for 15 min before washed and resuspended in 500 μl of binding buffer. Ten microliter of Annexin-V-FITC and propidium iodide respectively were added and the mixture was analyzed using the FACS Calibur. The data was analyzed by FACS Diva software.
Flow cytometry analysis of cell cycle Cells in different groups were collected and washed in cols phosphate-buffered saline (PBS), xed in 70% ethanol, and stored at 4 °C. DNA was treated with RNase A solution (500 units/ml) at 37 °C for 15 min and stained by PI (50 μg/ml) in 1.12% sodium citrate at room temperature before analysis. Flow cytometry determination of DNA content was done by FACS Calibur. The fractions of the cells in different phases were analyzed using Multicycle (Phoenix Flow System, San Diego, CA) cell cycle analysis software.

Tumor xenograft in nude mice
All experimental procedures were acknowledged by the Animal Care and Use Committee of West China Hospital, Sichuan University, followed the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health. Twenty male BALB/C nude mice at the age of four weeks (n = 5 per group were purchased from Shanghai SLAC Laboratory Animal Co., Ltd, China) for xenograft tumor growth assay. Two hundred microliter of osteosarcoma cells (1x10 5 ) transfected with NC, sh-MCM6 and sh-SNHG16 were injected into the tibial medullary cavity of nude mice. Tumor size was measured and calculated regularly every week (7 days). After 35 days, the tumor tissues were excised and xed, followed by dehydration, para n-embedding, and cutting into sections.

Immunohistochemistry
Immunohistochemical staining was done on formalin-xed and para n embedded tissue sections from tumors tissue. The tissue was placed in the incubator at 65 °C for 2 hours before removing para n by xylene. After gradient elute the tissue by absolute ethyl alcohol, 95 % ethyl alcohol, 85 % ethyl alcohol, washed the tissue by redistilled water. In the third step, target retrieval was achieved with BOND Novocastra Epitope Retrieval Solution 1 (Leica Biosystems) at 100 °C for 20 min. Monoclonal Anti-ki67 antibody (2 µg/ml, Abcam) were used as primary antibody. After that, secondary antibody horseradish peroxidase-conjugated goat anti-rabbit IgG (1:5000, Abcam) were used. In this way, ki67 was stained in the tissue and the tissue was photographed by optical microscope.

Statistical analysis
Statistical analysis was nished using GraphPad Prism 6.0 software. The data was showed by mean value±standard deviation (SD) in each group. Student's t-test was applied to compare the differences between two groups, while the ANOVA was used to analyze differences among multi-samples. P < 0.05 was considered as statistically signi cance in the data.

Results
Difference analysis identi ed miR-23b-3p as a potential target for SNHG16 in NPC To investigate the differential expressed genes, bioinformatic methods were applied to analyze the data from GEO. As shown in the heat map, the results described part of lncRNAs and mRNAs that were upregulated and down-regulated in NPC tissues, compared with that in normal tissues (P < 0.05, Fig. 1A, 1B). SNHG16 and MCM6 were among the top 10 up-regulated lncRNAs and mRNAs respectively. Then KEGG analysis was used by GSEA to screen the relevant signaling pathways in NPC tissue. And then, we found that the cell cycle pathway was enriched in the 30 activated signaling pathway by analysis of R package joyplot and dotplot (Fig. 1C, 1D).
Gene ontology (GO) analysis was performed by GOplot through R project (Fig. 2). The results showed that cell cycle related cell functions were enriched. Biological process that focuses on cell cycle also gained a high mark of Z-score with a low P value, hence the cell cycle was high enrichment and reliable. Therefore, we chose cell cycle pathway for further investigation. To further detect the relationship between genes and cell cycle, a network between mRNA and lncRNA on cell cycle was performed by Cytoscape based on previous analysis (Fig. 3A). LncRNA SNHG16 and mRNA MCM6 were chosen to nd the potential links and from the Venn diagram, 2 common miRNAs were found including miR-23b-3p and miR-33b-3p (Fig. 3B). Since previous study has reported that miR-23b-3p is a tumor suppressor in hepatocellular carcinoma [14] and there is few studies focus on the effect of miR-23b-3p in NPC, miR-23b-3p was selected for following study. Overall,we made the hypothesis that SNHG16 help develop NPC cell progression by competitively binding miR-23b-3p.
LncRNA SNHG16 and mRNA MCM6 were overexpressed in NPC tissues and cell lines We next investigated how SNHG16 and MCM6 expressed in the NPC tissues and cell lines. MCM6 and lncRNA SNHG16 were up-regulated while miR-23b-3p was down-regulated in tumor tissues by qRT-PCR validation (P < 0.001, Fig. 4A, 4B, 4C). To set up appropriate cell model, a series of cell lines including human normal nasopharyngeal cell line (NP69) and carcinoma cell lines (HONE-1, c666-1, CNE, HNE) were used to detect the expression of MCM6, SNHG16 and miR-23b-3p by qRT-PCR. We found that the SNHG16 and MCM6 expressed remarkably higher in c666-1, HNE-1 and HONE-1 than that in normal cell lines (all P < 0.05), while the increase of SNHG16 and MCM6 level in CHE cell was not signi cant (P > 0.05). On the contrary, the expression of miR-23b-3p in the three cell lines above was notably declined (P < 0.05). But in the CHE cells, the miR-23b-3p expression showed slightly decrease (P > 0.05, Fig. 4E, 4F,  4G). Therefore, c666-1 and HONE-1 which had higher expression of SNHG16 and MCM6 than others were chosen for the lateral study.
Silence of SNHG16 inhibited the expression of MCM6 through up-regulation of miR-23b-3p To detect the function of SNHG16 and MCM6 in NPC cells, c666-1 and HONE-1 cells were transfected with si-SNHG16, si-MCM6 before detected by qRT-PCR, evaluating the transfection e ciency of si-SNHG16 and si-MCM6. Our results markedly showed that the expression level of SNHG16 and MCM6 decreased sharply in both cell lines transfected with si-SNHG16 and si-MCM6 (all P < 0.01), while the expression of control group was not affected obviously (P > 0.05, Fig. 5). And as shown in Fig. 6 the putative binding site of miR-23b-3p and SNHG16, miR-23b-3p and MCM6 were predicted by TargetScan and miRcode. Then the luciferase reporter assay indicated the interaction between miR-23b-3p and SNHG16, miR-23b-3p and MCM6 since miR-23b-3p could combine with both of the SNHG16 and MCM6 in wild type instead of in the mutant type (P < 0.01). Based on this conclusion, the expression level of MCM6, miR-23b-3p in c666-1 and HONE-1 cell lines were detected in groups of si-SNHG16, miR-23b-3p mimics and si-SNHG16 with miR-23b-3p mimics respectively. The miR-23b-3p was up-regulated notably when SNHG16 was silenced (P < 0.01). Rather, when the SNHG16 was silenced or the miR-23b-3p was increased, the MCM6 expression was down-regulated (P < 0.01), implying that miR-23b-3p could combine with MCM6, and SNHG16 could promote MCM6 expression level by combing with miR-23b-3p (all P < 0.05, Fig. 7A, 7B). The down-regulation of SNHG16 might promote miR-23b-3p to decline the expression of MCM6. To further understand the protein expression level of MCM6, western blot was performed and similar results were obtained compared with MCM6 mRNA expression (Fig. 7C). Overall, silence of SNHG16 expression limited the expression of MCM6 through up-regulation of miR-23b-3p.

Silence Of Snhg16 Suppressed Npc Cell Progression
To understand the functions of SNHG16-miR-23b-3p-MCM6 axis in NPC, MTT assay and transwell assay were performed to detect cell proliferation and the ability of invasion separately. The results demonstrated that SNHG16 or MCM6 knockdown respectively signi cantly inhibited the cell viability (P < 0.001, Fig. 8A, 8B) and invasion ability (P < 0.001, Fig. 8C, 8D) in both cell lines when compared with control group. But the situation could be rescued when miR-23b-3p inhibitor was added even though SNHG16 and MCM6 were silenced before. The cell propagation ability was enhanced in si-SNHG16 + miR-23b-3p inhibitor group and si-MCM6 + miR-23b-3p inhibitor group compared with si-SNHG16 group and si-MCM6 group (all P < 0.05, Fig. 8).
Furthermore, the ow cytometry assay on cell apoptosis analysis in HONE-1 and c666-1 cells showed the apoptosis in uenced by the same operations as above. When SNHG16 or MCM6 was silenced, apoptosis could be activated most signi cantly in comparison to the control (all, P < 0.01, Fig. 9A, 9B). When the miR-23b-3p inhibitor was added, apoptosis showed a diversity with the silent groups. Following this, to determine whether the lncRNA SNHG16 and MCM6 attributed to cell cycle arrest, we performed cell cycle analysis by ow cytometry. As shown in Fig. 10A and 10B, silence of SNHG16 and MCM6 resulted in G0/G1 phase accumulation in HONE-1 cells (P < 0.05), whereas adding the miR-23b-3p inhibitor led to lower accumulation level of G0/G1 phase (P < 0.05). Similar results were obtained in c666-1 cells (Fig. 10C, 10D). SNHG16 or MCM6 silence induced a signi cant accumulation of cells in the G0/G1 phase with concomitant decrease of cells at the G2/M and S phases. All these results showed that the decrease of SNHG16 could inhibit NPC cell progression through cell cycle pathway.

Silenced Snhg16 Or Mcm6 Inhibited Tumor Growth
Tumor cell transplantation experiment was performed to determine the potential of SNHG16 on NPC in alleviating the symptoms of the disease. C666-1 cells after transfection by lentivirus were injected to nude mice and the tumor was excised after 35 days (Fig. 11A). Tumor volume was signi cantly smaller in the sh-SNHG16 and sh-MCM6 groups compared to the control (P < 0.001, Fig. 11B). The tumor weight followed the same pattern that was lower in the sh-SNHG16 and sh-MCM6 groups than that in others groups (all P < 0.01, Fig. 11C). Moreover, we detected the content of SNHG16, MCM6 and miR-23b-3p in tumors. Just as we predicated, the expression of MCM6 was well suppressed by silence of SNHG16 (P < 0.01, Fig. 11E), while the SNHG16 expression was not affected by MCM6 level (P > 0.05, Fig. 11D). The miR-23b-3p showed a higher expression level than the control (P < 0.05, Fig. 11F). To make the result of in vivo experiment more intuitively, we made para n section to detect the tumor cell activity by ki67 score through IHC. Compare to the control, IHC revealed that the ki67 expression was decreased in sh-SNHG16 and sh-MCM6 groups (P < 0.05, Fig. 11G, 11H). Animal experiment further con rmed our hypothesis that rising SNHG16 could promote the NPC cell progression by competitively binding miR-23b-3p.

Discussion
In this study, lncRNA SNHG16 expression level is increased in NPC cells and its up-regulation can promote NPC cell proliferation and invasion. Additionally, our study veri ed that SNHG16 can posttranscriptionally affect the expression of MCM6 and further promote the progression of NPC by binding with miR-23b-3p. lncRNAs play vital roles in cancer development. To date, only a small number of lncRNAs have been characterized at functional level. [15] Previous studies have proved that lncRNA differentially expressed patterns can be represented by the microarray probers and lnc-BCL2L11-3 can be a distinguishing lncRNA identi ed in the recurrent NPC. [16] The lncRNA NEAT1 regulates epithelial to mesenchymal transition (EMT) and radio resistance in NPC through miR-204/ZEB1 has been reported. [17] Song et al. nd that SNHG16 which is overexpressed in osteoarthritis chondrocytes is known to inhibit cell survival ability and is aberrantly expressed in several cancers and stimulates proliferation by regulating cyclin-dependent kinase 6. [18] Zhao et al. nd that increase of SNHG16 in neonatal hypoxic brain injury identi es a novel mechanism of lncRNA-mediated neural injury. [19] Furthermore, the different expression of SNHG16 may suggest the differential role of SNHG16 in carcinogenesis of different tissues. [20] For example, SNHG16 is overexpressed in esophageal cancer cells. Nevertheless, few studies focus on SNHG16 in NPC. We identify that SNHG16 may act as an oncogene in NPC rst based on the article research.
MiRNAs play key roles in many complex diseases, including various cancers. For instance, miR-139-5p induced timeless down-modulation and inhibited cell proliferation and apoptosis. [21] Nowadays mounting evidence has proved that miR-23b-3p not only functions as stated above, but also works as a biomarker for diagnosis for various cancers. For example, its expression is notably decreased in hepatocellular carcinoma tissues compared with their paired adjacent tissues. [22] Besides, overexpression of miR-23b-3p reverses cancer cell resistance to plenty of chemotherapy drugs in vitro, and sensitizes tumors to chemotherapy in vivo. [23] Francesco et al. also nd a signi cant downregulation of miR-23b-3p in Duchenne muscular dystrophy patients. [24] All previous studies prove that miR-23b-3p is important in cancer procession and consistent with our results that miR-23b-3p may suppress the NPC development.
A novel computational approach suitable to explore the potential role of lncRNAs as miRNA sponges for preserving homeostasis and preventing disease has already been proposed. [25] In previous study, Paci et al. have identi ed a sponge interaction network between long non-coding RNAs and messenger RNAs in human breast cancer. [26] In present research, we predicted the possible miRNAs for SNHG16 using bioinformatics analysis and found miR-23b-3p sharing the complementary binding sites to mediate the expression of MCM6. Among others proliferation markers, the mini-chromosome maintenance (MCM) proteins as the key proteins in the initiation of DNA synthesis and DNA replication, are considered to be linked with histological grades in various neoplastic development. [27] A recent report indicates that MCM is associated with histological grade and survival ability of cells in endometrioid endometrial adenocarcinoma, which may be a new marker in early diagnosis. [28] In our study, we verify that SNHG16 accelerates NPC progression by acting as a sponge of miR-23b-3p and activating MCM6 indirectly, which may provide a new insight for NPC prediction or prognosis.
However, several limitations exist in our study. For instance, despite the relative large sample size, the total number of samples in this study is still small. This study is based on only two cell lines and more experiments should be applied. In addition, due to the di culty to investigate the downstream metabolic network of MCM6, more researches should be executed in future.
In summary, this study provides a comprehensive report of the relationship between SNHG16, miR-23b-3p and MCM6 axis. Highly expressed SNHG16 promotes NPC cell growth by competitively binding with miR-23b-3p. Further studies regarding the mechanism of SNHG16 in NPC may provide clues as to whether they can serve as potential therapeutic targets. Taken together, our study has demonstrated the importance of SNHG16 expression in NPC, which may be a novel predictive indicator for NPC patients. Informed Consent All samples were collected with the informed consent of the patients and the study was approved by West China Hospital, Sichuan University.
Con ict of Interest: The authors con rm that there are no con icts of interest.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.
Authors' contributions: Shixi Liu: critical revision of the manuscript; Jiaojiao Peng: substantial contribution to the conception and design of the work, manuscript drafting; Feng Liu and Hong Zheng: acquisition, analysis, and interpretation of the data; Qi Wu: revising the manuscript critically, nal approval of the version to be published. All authors have read and approved the nal article. Primer