LncRNA RP11-79H23.3 Functions as a Competing Endogenous RNA to Regulate PTEN Expression through Sponging hsa-miR-107 in the Development of Bladder Cancer

Accumulating evidence indicates that the aberrant expression of long noncoding RNAs (lncRNAs) is involved in tumorigenesis and cancer development. However, the biological functions and underlying mechanisms of lncRNAs in bladder cancer (BC) remain largely unknown. Here, we analyzed the lncRNA and mRNA expression profiles in BC using a microarray assay. We found that lncRNA RP11-79H23.3 and phosphatase and tensin homolog (PTEN) were significantly downregulated in BC tissues and cells. Meanwhile, RP11-79H23.3 expression was negatively correlated with clinical stage in BC. Functionally, we found that overexpression of RP11-79H23.3 could suppress cell proliferation, migration, and cell cycle progression, rearrange the cytoskeleton, and induce apoptosis in vitro. Moreover, upregulation of RP11-79H23.3 inhibited the angiogenesis, tumorigenesis, and lung metastasis in vivo, whereas RP11-79H23.3 knockdown exerted a contrary role. Mechanistically, we identified that RP11-79H23.3 could directly bind to miR-107 and abolish the suppressive effect on target gene PTEN, which leads to inactivation of the PI3K/Akt signaling pathway. Taken together, we first demonstrated that RP11-79H23.3 might suppress the pathogenesis and development of BC by acting as a sponge for miR-107 to increase PTEN expression. Our research revealed that RP11-79H23.3 could be a potential target for diagnosis and therapy of BC.


Introduction
Bladder cancer (BC) is a common cancer worldwide, with an estimated 429,800 new cases diagnosed and 165,100 deaths occurring annually [1]. Bladder cancer patients can be treated by radiation, surgery, chemotherapy, and other methods of treatment to improve the survival rate of patients. However, BC has a poor prognosis and a high recurrence rate after surgery, with the 5-year overall survival rate being only 50-60% [2,3]. Therefore, discovering new molecular biomarkers and the molecular mechanism underlying the development and progression of BC are desperately needed.
Long noncoding RNAs (lncRNAs) are a class of RNA transcripts greater than 200 nt in length without the function of protein encoding [4]. Although only a small number of functional lncRNAs have been well characterized to date, they have been shown to regulate gene expression at various levels, including chromatin modification, transcription, and post-transcriptional

RP11-79H23.3 Is Downregulated in BC Tissues and Cells and Negatively Correlated with miR-107 Expression
To explore the biological functions of lncRNAs in BC, microarray analysis of lncRNA and mRNA expression profiles was performed as previously reported [15]. Compared with paracancerous normal tissues, we found that 6137 lncRNAs and 4416 mRNAs were differentially expressed, including 3217 upregulated and 2920 downregulated lncRNAs as well as 2472 upregulated and 3 of 20 1944 downregulated mRNAs, with a fold change ≥2.0, p < 0.05, and FDR (false discovery rate) <0.05 in four bladder cancer tissues ( Figure 1A). Among these, lnRNA RP11-79H23.3 was one of the most significantly downregulated lncRNAs and PTEN was one of the most markedly downregulated mRNAs. The qRT-PCR (Quantitative real time polymerase chain reaction) assays showed that RP11-79H23. 3 and PTEN expressions were significantly downregulated in BC tissues compared with adjacent normal tissues from 30 patients ( Figure 1B). Interestingly, the RP11-79H23.3 expression was negatively correlated with the tumor-node-metastasis (TNM) stage. Relationships between RP11-79H23.3 expression and clinical characteristics of the BC patients are shown in Table 1. Next, the expressions of RP11-79H23. 3 and PTEN were further determined in bladder cancer cell lines EJ, T24, and BIU87 and the normal bladder cell line SV-HUC-1 by qRT-PCR. The data also showed that the levels of RP11-79H23.3 were significantly downregulated in three kinds of BC cells. Moreover, PTEN expressions were remarkably downregulated in BC cells compared with normal bladder epithelial cells ( Figure 1C). Pearson correlation analysis revealed that the expression of RP11-79H23.3 was positively correlated with the level of PTEN in BC, r = −0.641 ( Figure 1D). The data suggest that the correlation between expression of RP11-79H23. 3  upregulated and 1944 downregulated mRNAs, with a fold change ≥ 2.0, p < 0.05, and FDR (false discovery rate) < 0.05 in four bladder cancer tissues ( Figure 1A). Among these, lnRNA RP11-79H23. 3 was one of the most significantly downregulated lncRNAs and PTEN was one of the most markedly downregulated mRNAs. The qRT-PCR (Quantitative real time polymerase chain reaction) assays showed that RP11-79H23. 3 and PTEN expressions were significantly downregulated in BC tissues compared with adjacent normal tissues from 30 patients ( Figure 1B). Interestingly, the RP11-79H23.3 expression was negatively correlated with the tumor-node-metastasis (TNM) stage. Relationships between RP11-79H23.3 expression and clinical characteristics of the BC patients are shown in Table 1. Next, the expressions of RP11-79H23. 3 and PTEN were further determined in bladder cancer cell lines EJ, T24, and BIU87 and the normal bladder cell line SV-HUC-1 by qRT-PCR.
The data also showed that the levels of RP11-79H23.3 were significantly downregulated in three kinds of BC cells. Moreover, PTEN expressions were remarkably downregulated in BC cells compared with normal bladder epithelial cells ( Figure 1C). Pearson correlation analysis revealed that the expression of RP11-79H23.3 was positively correlated with the level of PTEN in BC, r = −0.641 ( Figure 1D). The data suggest that the correlation between expression of RP11-79H23. 3 and PTEN might be involved in tumorigenesis and development of BC.  Data is shown as mean ± SD (standard deviation). * p < 0.05; ** p < 0.01; *** p < 0.001.

RP11-79H23.3 Modulates BC (Bladder Cancer) Cell Proliferation, Migration, and Invasion
The expression of RP11-79H23.3 was examined in RP11-79H23.3 overexpression and RP11-79H23.3 knockdown BC cells by qRT-PCR. The result showed that the levels of RP11-79H23.3 were significantly upregulated in BC cells transfected with pIRES2-RP11-79H23.3. Conversely, the expressions of RP11-79H23.3 were remarkably decreased in BC cells transfected with si-RNA fragments (si-RP11-79H23.3I and si-RP11-79H23.3II) (Figure 2A,B). To investigate the functions of RP11-79H23.3, the effects of RP11-79H23.3 on cell proliferation, migration, and invasion were explored when RP11-79H23.3 was downregulated or upregulated. The CCK-8 results showed that cell viability with transfection of the pIRES2-RP11-79H23.3 was significantly decreased compared with empty vector group ( Figure 2C). EdU and colony formation assays further verified that upregulation of RP11-79H23.3 markedly inhibited the number of EdU-positive cells and colonies, while RP11-79H23.3 knockdown exhibited the opposite effects ( Figure 2D,E). Wound healing and transwell assays indicated that siRP11-79H23.3 could significantly accelerate the migration and invasion of EJ and T24 cells compared with vector control groups, whereas the number of migrating and invading cells in the pIRES2-RP11-79H23.3 groups were significantly decreased compared with vector control groups ( Figure 2F-I). It has been known that actin filaments are involved in adhesion and migration of tumor cells to provide support and motor activity. Cytoskeletal protein paxillin plays an important role in integrin signal transduction. Accordingly, F-actin and protein paxillin were detected with fluorescent phalloidin and immunofluorescence respectively. When RP11-79H23.3 was downregulated, more abundant actin filaments and a brighter fluorescent signal of paxillin were observed, whereas upregulation of RP11-79H23.3 significantly suppressed stress fiber formation and paxillin expression ( Figure 2J).

Upregulation of RP11-79H23.3 Induces Apoptosis and Regulates Cell Cycle of BC Cells
To determine whether RP11-79H23.3 functions as a tumor suppressor, cellular apoptosis was evaluated by flow cytometry, Hoechst33342 staining, TUNEL (TdT-mediated dUTP Nick-End Labeling) assays, and Western blotting after transfection with pIRES2-RP11-79H23.3. Live cell images showed that upregulation of RP11-79H23.3 suppressed the malignant phenotype, including a smaller nucleus, fewer division phases, and slower growth compared with the control group, whereas siRP11-79H23.3 group cells indicated an enhanced malignant feature, including overlapping growth and more division phases ( Figure 3A). TUNEL assay revealed that upregulation of RP11-79H23.3 expression led to a significantly increased number of TUNEL-positive cells compared with control groups ( Figure 3B). Meanwhile, Hoechst33342 staining found that pIRES2-RP11-79H23.3 group cells displayed typical apoptotic morphology characteristics, such as chromatin condensation, nuclear shrinkage, apoptotic body, nuclear fragmentation, and brighter fluorescent. However, the apoptotic features did not appear in control cells ( Figure 3C). Furthermore, flow cytometry analysis with Annexin V/PI double staining showed that the percentages of apoptotic cells in pIRES2-RP11-79H23.3 cell groups were significantly higher than those of the control group cells, respectively ( Figure 3D

RP11-79H23.3 Suppresses Tumorigenesis, Metastasis, and Angiogenesis of BC Cells In Vivo
To further explore the biological function of RP11-79H23.3 in vivo, we established a human bladder carcinoma xenograft model in nude mice. 2.5 × 10 6 various kinds of EJ cells were subcutaneously inoculated into the back of nude mice. After 35 days, all the mice were sacrificed. We found that RP11-79H23.3 overexpression evidently inhibited tumorigenicity of BC cells, while the siRP11-79H23.3 cell group showed a significantly higher tumor growth rate and heavier tumor weight compared with control groups ( Figure 4A,B). Furthermore, upregulation of RP11-79H23.3 obviously suppressed tumor angiogenesis, whereas downregulation of RP11-79H23.3 led to more tumor microvessels ( Figure 4C,D). In addition, compared with control groups, RP11-79H23.3siRNA resulted in a significant enhancement of spontaneous pulmonary metastasis with apparent visible lung metastatic nodes, while upregulation of RP11-79H23.3 markedly attenuated metastasis with fewer invasive tumor cells ( Figure 4E,F). To further examine the effect of RP11-79H23.3 on angiogenesis, immunofluorescence assays for CD31 and S100A4 were implemented. CD31 is used primarily to demonstrate the presence of endothelial cells in histological tissue sections. This can help to evaluate the degree of tumor angiogenesis, which can imply a rapidly growing tumor. S100A4 is a member of the S100 calcium-binding protein family. It can support tumorigenesis by triggering angiogenesis. The results demonstrated that RP11-79H23.3 inhibited tumor angiogenesis, whereas si-RP11-79H23.3 led to a higher CD31 and S100A4 expression and more microvessels compared with the control group ( Figure 4G).   primarily to demonstrate the presence of endothelial cells in histological tissue sections. This can help to evaluate the degree of tumor angiogenesis, which can imply a rapidly growing tumor. S100A4 is a member of the S100 calcium-binding protein family. It can support tumorigenesis by triggering angiogenesis. The results demonstrated that RP11-79H23.3 inhibited tumor angiogenesis, whereas si-RP11-79H23.3 led to a higher CD31 and S100A4 expression and more microvessels compared with the control group ( Figure 4G).

RP11-79H23.3 Expression Correlates with PTEN and PI3K/AKT Signaling Pathway Molecules in BC
To understand the underlying relationship between RP11-79H23.3 and the PTEN/PI3K/AKT signaling pathway, we performed immunofluorescence and immunohistochemistry assays in vivo. First, representative HE-stained images are shown from 30 pairs of BC tissues and matched adjacent nontumor tissues in Figure 5A. Subsequently, we determined the expressions of PTEN and PI3K/AKT signaling pathway molecules in human BC tissues from 30 patients by immunofluorescence assay. The result showed that BC tissues displayed much weaker positive staining of PTEN level as well as stronger p-PI3K and p-Akt expressions compared with adjacent nontumor tissues ( Figure 5B,C). Next, immunofluorescence and immunohistochemistry assays revealed that siRP11-79H23.

MiR-107 Directly Targets RP11-79H23.3/PTEN in BC
To clarify the molecular mechanism underlying RP11-79H23.3, firstly, bioinformatics analysis was executed by miRcode and TargetScan. The data showed that both RP11-79H23.3 and PTEN contain conserved target sites of miR-107 ( Figure 6A). Next, dual-luciferase reporter assays were performed to detect the binding between RP11-79H23.3 and miR-107. RP11-79H23.3/PTEN wild and mutant dual-luciferase plasmids were constructed. The 3 UTR of RP11-79H23.3 and PTEN were subcloned to the pmirGLO dual-luciferase reporter vectors ( Figure 6B,C). The data showed that transfection of miR-107 mimics could obviously reduce the activity of a luciferase reporter carrying the 3 -UTR of RP11-79H23.3 of WT but not mutant of 3 -UTR, while transfection of the miR-107 inhibitor (inh-107) could remarkably increase the luciferase activity of WT reporters but not the mutant one compared with controls ( Figure 6D,E). Moreover, the luciferase reporter activity was significantly decreased in EJ cells co-transfected with miR-107 mimic and pmirGLO-PTEN-WT vector but not the mutant one compared with other controls ( Figure 6F). In addition, we investigated the subcellular localization of RP11-79H23.3 and miR-107 in BC cells by FISH. The results indicated that RP11-79H23.3 (red) and miR-107 (green) mainly distributed in cytoplasm, the orange region showed that RP11-79H23.3 and miR-107 are colocalized in the EJ and T24 cells ( Figure 6G). Subsequently, to further confirm the endogenous binding between RP11-79H23.3 and miR-107 at the cellular level, we performed RNA immunoprecipitation (RIP) to pull down miRNAs connected with RP11-79H23.3 and confirmed this with qRT-PCR. The data demonstrated that the RIP of RP11-79H23.3 was remarkably enriched for miR-107 in EJ cells compared with MS2 the empty vector and RP11-79H23.3 mutations with miR-107 targeting site vector groups ( Figure 6H,I). The miRNAs negatively regulate gene expression in an AGO2-dependent manner. Therefore, we further executed anti-AGO2 RIP in EJ cells transfected with miR-107. Endogenous RP11-79H23.3 pull-down by AGO2 was specifically enriched in the overexpressing miR-107 cells ( Figure 6J). Furthermore, we co-transfected the luciferase reporter of pmirGLO-PTEN 3 UTR-wt with pIRES2-RP11-79H23.3, and the data revealed that the luciferase activity was significantly enhanced compared with the control group of co-transfection with WT luciferase reporter and vector, suggesting that upregulation of RP11-79H23.3 increased luciferase activity of pmirGLO-PTEN 3 UTR-wt reporter by competitively binding endogenous miR-107. However, this effect could be obviously abrogated by overexpression of miR-107 ( Figure 6K). More importantly, ectopic expression of miR-107 significantly counteracted the proliferation and invasion-suppressing effects of RP11-79H23.3 overexpression in EJ cells, whereas miR-107 inhibitor attenuated the promoting effects mediated by RP11-79H23.3 knockdown in T24 cells by EdU and transwell assays ( Figure 6L-O). Collectively, these data demonstrated that miR-107 could directly target RP11-79H23.3 and bind to PTEN in BC cells.

RP11-79H23.3 Regulates the Expressions of PTEN and PI3K/AKT Signaling Pathway Molecules In Vitro
To further elucidate the underlying mechanism of RP11-79H23.3 in BC, we determined whether RP11-79H23.3 could regulate the target gene PTEN and the PI3K/AKT signaling pathway. The results of an immunofluorescence assay for PTEN indicated that upregulation of RP11-79H23.3 markedly enhanced the expression of PTEN and the increased impact could be attenuated by ectopically expressing miR-107, whereas downregulation of RP11-79H23.3 observably decreased the PTEN level and the reduced role could be reversed by miR-107 inhibitor ( Figure 7A,B). Furthermore, we found that ectopic expression of miR-107 significantly inhibited the level of PTEN, whereas knockdown of miR-107 obviously increased expression of PTEN by qRT-PCR and Western blot assays in BC cells ( Figure 7C,D). Additionally, our data showed that the expression of PTEN was significantly upregulated in BC cells transfected with the pIRES2-RP11-79H23.3, whereas siRP11-79H23.3 could remarkably reduce the expression of PTEN with qRT-PCR compared with control groups ( Figure 7E). The co-transfection of RP11-79H23.3 and miR-107 could diminish the expression of PTEN compared with RP11-79H23.3 alone in BC cells, while co-transfection of siRP11-79H23.3 and miR-107 inhibitor could abolish the miR-107 inhibitor-induced upregulation of PTEN in BC cells with qRT-PCR ( Figure 7F). Subsequently, the Western blot assay showed that upregulation of RP11-79H23.3 significantly decreased the levels of p-PI3K, p-AKT, and p-GSK3β in BC cells as well as increased the expression of PTEN, whereas si-RP11-79H23.3 led to the activation of the PI3K/AKT pathway. Consistently, the co-transfection of miR-107 and RP11-79H23.3 attenuated the expression of PTEN as well as enhanced the phosphorylation level of PI3K, AKT, and GSK3β compared with RP11-79H23.3 alone in EJ cells, while si-RP11-79H23.3 abolished the effects caused by inhibition of miR-107 in T24 cells ( Figure 7G,H). Collectively, these data demonstrated that RP11-79H23.3 might function as a ceRNA to regulate the PTEN/PI3K/AKT pathway in the pathogenesis and development of BC (Figure 8).

Discussion
It has been documented that lncRNAs play important roles in cancer development and progression, but most of them have not yet been studied in functional and mechanistic detail. In the present study, we identified a lncRNA RP11-79H23.3 which was downregulated in BC based on microarray gene-expression profile analysis. RT-qPCR analysis showed that RP11-79H23.3 levels were significantly downregulated in BC tissues and cells. We first demonstrate that RP11-79H23.3 could function as a ceRNA to regulate PTEN expression by sponging miR-107. Our data suggest that RP11-79H23.3 might play a role as a tumor suppressor in the tumorigenesis and progression of BC.
The tumor suppressor PTEN dephosphorylates the D3 position of phosphatidylinositol-3,4,5 triphosphate (PIP3) to negatively control PI3K activity and thus inhibits a panel of cellular responses mediated by the PI3K/Akt pathway including cell growth, mobility, and invasion [20]. Tsuruta et al. found that PTEN expression was significantly reduced in bladder cancer patients, and this decrease in PTEN correlated with disease stage and grade. Thus, PTEN deficiency may contribute to initiation and progression of bladder cancer [21]. Recently, some miRNAs have been demonstrated to promote tumorigenesis and metastasis by downregulating PTEN expression. MiR-130b could target PTEN to mediate drug resistance and proliferation of breast cancer cells via the PI3K/Akt signaling pathway [22]. MiR-106b and miR-93 were reported to regulate cell progression by suppression of PTEN via PI3K/Akt pathway in breast cancer [23]. MiR-21 promoted proliferation and invasion of triple-negative breast cancer cells through targeting PTEN [24]. However, whether miR-107 can target PTEN in BC has not yet been reported. On the basis of bioinformatics prediction and luciferase reporter assay, we further showed miR-107 mimics significantly reduced the PTEN level, whereas miR-107 inhibitor observably enhanced PTEN expression by qRT-PCR and Western blot assays. We provided the evidence that PTEN was a direct target of miR-107 in BC.
In recent years, a ceRNA hypothesis has been proposed that RNA transcripts of coding RNAs and noncoding RNAs can communicate with each other to regulate gene expression by competing for binding to shared miRNAs [11]. It has been reported that the abnormally expressed lncRNAs could act as ceRNAs for miRNAs to regulate the expression of miRNA target genes and disrupt the equilibrium of ceRNAs and miRNAs in tumor development. Long noncoding RNA MEG3 could function as a competing endogenous RNA for miR-181a to regulate carcinogenesis and progression of gastric cancer [25]. LncRNA HOTAIR (HOX transcript antisense RNA) regulated HIF-1α (hypoxia inducible factor 1 subunit alpha)/AXL (AXL receptor tyrosine kinase) signaling and promoted tumorigenesis through acting as a ceRNA of miR-217 in renal cell carcinoma [26]. Long

Discussion
It has been documented that lncRNAs play important roles in cancer development and progression, but most of them have not yet been studied in functional and mechanistic detail. In the present study, we identified a lncRNA RP11-79H23.3 which was downregulated in BC based on microarray gene-expression profile analysis. RT-qPCR analysis showed that RP11-79H23.3 levels were significantly downregulated in BC tissues and cells. We first demonstrate that RP11-79H23.3 could function as a ceRNA to regulate PTEN expression by sponging miR-107. Our data suggest that RP11-79H23.3 might play a role as a tumor suppressor in the tumorigenesis and progression of BC.
The tumor suppressor PTEN dephosphorylates the D3 position of phosphatidylinositol-3,4,5 triphosphate (PIP3) to negatively control PI3K activity and thus inhibits a panel of cellular responses mediated by the PI3K/Akt pathway including cell growth, mobility, and invasion [20]. Tsuruta et al. found that PTEN expression was significantly reduced in bladder cancer patients, and this decrease in PTEN correlated with disease stage and grade. Thus, PTEN deficiency may contribute to initiation and progression of bladder cancer [21]. Recently, some miRNAs have been demonstrated to promote tumorigenesis and metastasis by downregulating PTEN expression. MiR-130b could target PTEN to mediate drug resistance and proliferation of breast cancer cells via the PI3K/Akt signaling pathway [22]. MiR-106b and miR-93 were reported to regulate cell progression by suppression of PTEN via PI3K/Akt pathway in breast cancer [23]. MiR-21 promoted proliferation and invasion of triple-negative breast cancer cells through targeting PTEN [24]. However, whether miR-107 can target PTEN in BC has not yet been reported. On the basis of bioinformatics prediction and luciferase reporter assay, we further showed miR-107 mimics significantly reduced the PTEN level, whereas miR-107 inhibitor observably enhanced PTEN expression by qRT-PCR and Western blot assays. We provided the evidence that PTEN was a direct target of miR-107 in BC.
In recent years, a ceRNA hypothesis has been proposed that RNA transcripts of coding RNAs and noncoding RNAs can communicate with each other to regulate gene expression by competing for binding to shared miRNAs [11]. It has been reported that the abnormally expressed lncRNAs could act as ceRNAs for miRNAs to regulate the expression of miRNA target genes and disrupt the equilibrium of ceRNAs and miRNAs in tumor development. Long noncoding RNA MEG3 could function as a competing endogenous RNA for miR-181a to regulate carcinogenesis and progression of gastric cancer [25]. LncRNA HOTAIR (HOX transcript antisense RNA) regulated HIF-1α (hypoxia inducible factor 1 subunit alpha)/AXL (AXL receptor tyrosine kinase) signaling and promoted tumorigenesis through acting as a ceRNA of miR-217 in renal cell carcinoma [26]. Long noncoding RNA HNF1A-AS1 (HNF1A antisense RNA1) acted as a ceRNA for miR-30b-5p to promote proliferation and suppress apoptosis of bladder cancer cells through upregulating Bcl-2 [27]. Here, we report a lncRNA, lnc-RP11-79H23.3, that might function as a tumor suppressor in human BC. The data showed that the expression of RP11-79H23.3 was positively correlated with the level of PTEN. The luciferase reporter assays found that RP11-79H23.3 and PTEN could compete for binding to miR-107. In addition, RIP assay verified that RP11-79H23.3 could directly bind to miR-107 in an AGO2-dependent manner. More importantly, RP11-79H23.3 might regulate PTEN expression by antagonizing miR-107. Together, these data confirm that RP11-79H23.3 might function as a competing endogenous RNA for miR-107 in the development of BC.
In conclusion, we found that RP11-79H23.3 was downregulated in bladder cancer and could liberate miR-107 via its function as a ceRNA to suppress PTEN expression and activate the PI3K/Akt signaling pathway, which consequently contributes to the pathogenesis and progression of BC. Taken together, our findings suggest that the RP11-79H23.3 could play a potential tumor suppressor role in the progression and development of BC. The RP11-79H23.3/miR-107/PTEN axis might serve as a novel clinical marker and therapeutic target for BC.

Patient and Tissue Samples
Bladder cancer tissues and pair-matched adjacent tissues used in this paper were obtained from the First Affiliated Hospital of Chongqing Medical University and the Affiliated Hospital of Southwest Medical University during 2010-2015 in accordance with the Helsinki Declaration. None of the patients had undergone radiation treatment or chemotherapy before surgery. Informed consent was obtained from these patients. Tumors were classified according to the tumor-node-metastasis (TNM, 2010) system of classification. The clinicopathological characteristics of the BC patients are summarized in Table 1. This study was approved by the Ethics Committee of Chongqing Medical University.

Cell Lines, Plasmid Construction, and Transfection
Human normal bladder epithelial cell line (SV-HUC-1) and bladder cancer cell lines (EJ, T24, BIU87) were obtained from the American Type Culture Collection (Manassas, VA, USA). Cells were maintained in a humidified incubator at 37 • C in an atmosphere of 5% CO 2 . EJ, T24, BIU87 and SV-HUC-1 cells were maintained in RPMI-1640 and DMEM/F12K medium, respectively.

LncRNA and mRNA Microarray
Briefly, as previously reported (Huang et al., 2016 [15]), four pairs of human BC and the adjacent normal tissues were applied for microarray assay to determine differentially expressed lncRNAs and mRNAs by the Arraystar Human lncRNA Array v2.0 (KangChen Bio-tech, Shanghai, China). The transcript of the significant differential expression was reserved by fold change ≥2.0, p < 0.05, and FDR < 0.05. Hierarchical clustering was implemented to produce an outline of expression patterns according to the expression value of all the transcripts with or without significant differences. Microarray data were available through Gene Expression Omnibus (GEO) with accession code GSE89006.

qRT-PCR Assay
Total RNA was extracted from the frozen tissues and cultured cells with TRIzol (TAKARA, Dalian, China). The quantity and quality of RNA were detected by Nano Drop 2000 spectrophotometer (Nano Drop Thermo, Wilmington, DE, USA), and the integrity of RNA was evaluated by agarose gel electrophoresis. cDNA was generated with a reverse-transcription kit (Takara, Dalian, China) according to manufacturer's protocol. The qRT-PCR analysis was implemented by the ABI 7900HT Sequence Detection Machine (Bio-Rad, Hercules, CA, USA) using SYBR Green chemistry. U6 or GAPDH was used as an endogenous control. The specific primers sequences were as follows: RP11-79H23.3, 5 -TGGCCTCAGTTAGGACTGCT-3 and 5 -CTGCTTCCGCTCTCTTTCTC-3 ; PTEN, 5 -GCTATGGGATTTCCTGCAGAA-3 and 5 -GGCGGTGTCATAATGTCTTTCA-3 ; GAPDH, 5 -GAAGGTGAAGGTCGGAGTC-3 and 5 -GAAGATGGTGATGGGATTTC-3 . PCR was executed with an initial denaturation step at 95 • C for 5 min, followed by amplification with 40 cycles at 95 • C for 10 s and 60 • C for 35 s, the melt curve step at 60 • C to 95 • C, and the increment at 0.5 • C for 5 s. The genes were amplified in separate wells in triplicate. The relative expressions levels were calculated with the (2 −∆∆Ct ) method.

Cell Proliferation, Viability, and Colony Assays
The cell proliferation ability was determined by CCK8, EdU and colony formation assays. For CCK-8 assay, Cell Counting Kit-8 was bought from DingGuo (Beijing, China), and cells were seeded into 96-well plates (5000 cells/well) with complete growth medium. After incubation for 24, 48, 72, and 96 h, respectively, 10 µL CCK-8 was added into each well, and then the cells were cultured for an additional 2 h. The absorbance value was tested by a plate reader at 450 nm (Bio-Rad, Hercules, CA, USA). 1 × 10 5 cells were put into 24-well plates for EdU incorporation with an EdU detection kit (Ribobio, Guangzhou, China). The percentage of EdU-positive cells was quantified from four random fields per well. For colony formation assays, 2.5 × 10 2 cells, were plated onto six-well plates and incubated for 14 days, fixed with 4% paraform, then stained with 0.1% crystal violet, and the number of colonies in four random fields were counted under an inverted microscope. The experiments were performed in triplicate.

Wound Healing, Cell Invasion Assay, and Cytoskeleton
When cells grew in six-well plates to 80-90% confluence, scratches were made with a 200 µL tip, they were washed with PBS (Phosphate Buffered Saline), and then the cells were further incubated in FBS (fetal bovine serum)-free medium. The width of wound closure (original width at 0 h-width after cell migration at 24 h) was counted in the same wound point with five replicates. The cell invasion assays were performed using 24-well Transwell chambers with a pore size of 8 µm (BD BioCoat, Bedford, MA, USA). 5 × 10 5 cells in serum-free RPMI1640 were added to the upper chamber coated with matrigel (BD Biosciences, Franklin Lakes, NJ, USA), and the bottom chamber contained 500 µL of culture medium with 10% FBS. After 24 h, noninvading cells on the upper chamber were rubbed away with a cotton bud and invading cells in the lower chamber were stained by crystal violet, photographed, and counted with an Olympus multifunction microscope (Tokyo, Japan). Assays were repeated three times. The cells were seeded into 24-well plates after 48 h, rinsed with PBS, fixed in 4% paraformaldehyde, permeabilized by 0.3% Triton X-100 at 37 • C for 6 min, then rinsed three times with PBS, and blocked with 3% BSA (bovine serum albumin) at 37 • C for 30 min. The cells were incubated with Paxillin antibody (1:200 dilution) overnight at 4 • C, then incubated with IgG-TRITC for 1 h, then incubated with 1% FITC-phalloidin (Sigma Chemical Corp., St. Louis, MO, USA) at 37 • C for 30 min. The cells were stained by DAPI (4 ,6-diamidino-2-phenylindole, SouthernBiotech, Birmingham, AL, USA) for 10 min, sealed with antifluorescent solution, and were then were observed under a confocal laser scanning microscope (Leica, Wetzlar, Germany).

Cell Apoptosis and Cell Cycle Assay
The cells were fixed with 4% paraform for 20 min, then dyed with Hoechst 33342 for 15 min and observed under a Leica TCS-SP2 Laser Confocal Microscopy. Apoptosis was assessed by TUNEL assay using an apoptosis detection kit (Keygen Biotech, Nanjing, China). The FITC-labeled apoptotic cells were photographed under a fluorescent microscope. The cells were harvested and apoptosis was determined with flow cytometry (Becon Dickinson FACSCalibur, New York, NY, USA) after annexin V-FITC/PI double staining. 2 × 10 6 cells were fixed for 12 h in 70% ethanol, then stained with PI. The cell cycle was observed with flow cytometry (Becon Dickinson FACSCalibur, New York, NY, USA).

Fluorescence In Situ Hybridization (FISH)
The subcellular localizations of LncRNA RP11-79H23.3 and miR-107 were examined by FISH kit (Roche Applied Science, Penzberg, Germany). Cells grew to 70% confluence, were fixed with 4% formaldehyde for 15 min, then permeabilized with 0.5% TritonX-100 for 15 min and rinsed with PBS. Cells were incubated in a mixture of RP11-79H23.3 probes labeled with Cy3 at 37 • C overnight and washed with prewarmed 2× saline-sodium citrate (SSC). The FITC-labeled miR-107 probes were incubated in prehybridization buffer (1:100) at 88 • C for 5 min and at 4 • C for 3 min with the PCR instrument (Bio-Rad, Hercules, CA, USA). Next, the FITC-labeled miR-107 probes were added into cells at 37 • C overnight, washed with 2× SSC, and stained by DAPI. Observations were undertaken with a Leica TCS-SP2 Laser Scanning Confocal Microscope.

RIP (RNA Immunoprecipitation) Assay
RIP is an antibody-based technique used to map in vivo RNA-protein interactions. The RNA binding protein (RBP) of interest is immunoprecipitated together with its associated RNA for identification of bound transcripts (mRNAs, noncoding RNAs, or viral RNAs). Transcripts are detected by real-time PCR. RIP detection was implemented according to the manufacturer's guide with the Magna RNA binding protein immunoprecipitation kit (Millipore, Billerica, MA, USA). EJ cells were transfected with miR-107 mimics and miR-107 NC, respectively. Then, after cells were lysed by RIP lysis buffer, whole cell lysate was incubated in the buffer containing anti-AGO2 antibody coupled to magnetic beads (Millipore). RNA was separated and deposited by immunoprecipitation, then the products were purified and analyzed by qRT-PCR. In addition, EJ cells were co-transfected with pcDNA3.1-MS2, pcDNA3.1-MS2-RB-RP11-79H23.3, and pcDNA3.1-MS2-RB-RP11-79H23.3-mut (no miR-107 binding site), respectively, and pMS2-GFP (Addgene plasmids). After 48 h, RIP experiments were executed using an anti-GFP antibody (Abcam, Burlingame, CA, USA) with IgG antibody as a negative antibody. The products were purified and enriched to detect the target miRNAs by qRT-PCR.

Immunohistochemistry (IHC) and Immunofluorescence (IF)
Cells were cultured on the slides in a 24-well plate and fixed. The frozen and paraffin section were obtained from nude mouse tumor tissues and human tissues for IF and IHC, respectively. For IHC assay, after dewaxing, rehydration, and antigen retrieval, the slides were then incubated with antibodies specific for PTEN and PI3K/AKT signaling pathway molecules (Abcam) and secondary HRP labelled goat anti-rabbit, stained with DAB, then counterstained with hematoxylin. For IF assays, the cells and frozen section were incubated with indicated primary antibodies overnight at 4 • C. Then, samples were incubated with fluorescein Alexa-Fluor 488-or 594-conjugated secondary antibodies. The slides were dyed by DAPI, then observed under a confocal laser microscope and Olympus multifunction microscope.

Tumor Xenograft Model
BALB/c nude mice (4-6 weeks old, female) were purchased from Beijing HFK Bioscience Co., Ltd. (Beijing, China). The stably transfected EJ cells (2 × 10 6 ) with indicated vector or 3 -cholesterol and 2 -OMe-conjugated siRNA and NC random fragments (RIBOBIO) were subcutaneously injected into BALB/c mice, respectively. The volume of tumor was measured per week. The mice were sacrificed after 35 days, and tumor xenografts and lungs were harvested and subjected to pathologic, immunohistochemical, and immunofluorescent examination. The microvessels were counted from six different fields in the tumor HE sections under microscope. Metastatic nodules of the lung were recorded under the microscope. All animal care and experiment procedures were implemented according to the guidelines of the National Institutes of Health. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of Chongqing Medical University (Date: 16 February 2016, Identification code: 2016021606).

Statistical Analysis
Statistical analyses were executed by Student's t-test and One-way ANOVA using GraphPad Prism 6.0 and SPSS 19.0 statistical software. The significance of differences between groups was assessed by Student's t test and Fisher's exact test. The data were expressed as mean ± SD. p < 0.05 was considered as statistical significance.