Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

The miR155HG/miR-185/ANXA2 loop contributes to glioblastoma growth and progression

The miR155HG/miR-185/ANXA2 loop contributes to glioblastoma growth and progression Background: Glioblastoma multiforme (GBM) is the most common and aggressive form of astrocytoma among adult brain tumors. Multiple studies have shown that long non-coding RNAs (lncRNAs) play important roles in acting as molecular sponge for competing with microRNAs (miRNAs) to regulate downstream molecules in tumor progression. We previously reported that miR155 host gene (miR155HG), an lncRNA, and its derivative miR-155 promote epithelial-to-mesenchymal transition in glioma. However, the other biological functions and mechanisms of miR155HG sponging miRNAs have been unknown. Considering ANXA2 has been generally accepted as oncogene overexpressed in a vast of cancers correlated with tumorigenesis, which might be the target molecule of miR155HG sponging miRNA via bioinformatics analysis. We designed this study to explore the interaction of miR155HG and ANXA2 to reveal the malignancy of them in GBM development. Methods: The expression of miR155HG was analyzed in three independent databases and clinical GBM specimens. Bioinformatics analysis was performed to assess the potential tumor-related functions of miR155HG. The interaction of miR155HG and miR-185 and the inhibition of ANXA2 by miR-185 were analyzed by luciferase reporter experiments, and biological effects in GBM were explored by colony formation assays, EDU cell proliferation assays, flow cytometric analysis and intracranial GBM mouse model. Changes in protein expression were analyzed using western blot. We examined the regulatory mechanism of ANXA2 on miR155HG in GBM by gene expression profiling analysis, double immunofluorescence staining, chromatin immunoprecipitation and luciferase reporter assays. Results: We found that miR155HG was upregulated in GBM tissues and cell lines. Bioinformatic analyses of three GBM databases showed that miR155HG expression levels were closely associated with genes involved in cell proliferation and apoptosis. Knocking down miR155HG suppressed GBM cell proliferation in vitro, induced a G1/S-phase cell cycle arrest, and increased apoptosis. We also found that miR155HG functions as a competing endogenous RNA for miR-185. Moreover, miR-185 directly targets and inhibits ANXA2, which exhibits oncogenic functions in GBM. We also found that ANXA2 promoted miR155HG expression via STAT3 phosphorylation. Conclusion: Our results demonstrated that overexpressed miR155HG in GBM can sponge miR-185 to promote ANXA2 expression, and ANXA2 stimulates miR155HG level through phosphorylated STAT3 binding to the miR155HG promoter. We establish the miR155HG/miR185/ANXA2 loop as a mechanism that underlies the biological functions of miR155HG and ANXA2 in GBM and further suggest this loop may serve as a therapeutic target and/or prognostic biomarker for GBM. Keywords: GBM, miR155HG, ANXA2, P-STAT3, ceRNA * Correspondence: [email protected] Weining Wu, Tianfu Yu and Youzhi Wu these authors contributed equally to this work. Department of Neurosurgery, The First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210029, China Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 2 of 14 Background functions of miR155HG in glioma has not been Glioblastoma multiforme (GBM) is the most aggressive determined. form of astrocytoma and is characterized by rapid pro- Through bioinformatical analysis and experimental gression and poor prognosis [1]. Studies have indicated analyses, here we confirmed that miR155HG is overex- that GBM development is associated with disrupted gene pressed in GBM and acts as a ceRNA for the tumor sup- regulatory networks. In addition to aberrant gene ex- pressor miR-185 to upregulate ANXA2. We also showed pression in GBM, several reports have demonstrated a that ANXA2 promotes GBM growth and miR155HG ex- role for non-coding RNAs (ncRNAs), which lack protein pression by activating STAT3. Our results demonstrate coding capability due the lack of an open reading frame, that miR155HG, miR-185 and ANXA2 form a signaling in GBM progression [2]. Among these ncRNAs, long loop that promotes malignant phenotypes in GBM. ncRNAs (lncRNAs) are defined as RNAs > 200 nucleo- tides in length and have been demonstrated to partici- Methods pate in diverse cellular processes including epigenetics Public human astrocytoma databases, GBM specimens and transcriptional and posttranscriptional regulation and cell lines [3]. At the posttranscriptional level, lncRNAs can act as Three public human astrocytoma databases (TCGA, competing endogenous RNAs (ceRNAs) or molecular CGGA and Rembrandt) were described in our previous sponge that function by sponging microRNAs (miR- study [21]. A total of 24 GBM tissues and 15 NAs), short endogenous ncRNAs approximately 18–25 pair-matched adjacent normal brain edematous tissues nucleotides in length, resulting in inhibiting the effects (collected postoperatively between April 2016 and Feb- of miRNAs on target mRNAs [4]. ruary 2017) were collected from patients who underwent The lncRNA miR155HG is transcribed from a gene lo- surgical removal of GBM tumors at the First Affiliated cated on chromosome 21q21 and consists of three exons Hospital of Nanjing Medical University. Samples were that span 1.5 Kb. MiR155HG, also known as B-cell inte- frozen in liquid nitrogen immediately after isolation and gration cluster, includes an imperfectly base-paired stem stored at − 80 °C to avoid RNA deterioration. Tumor tis- loop in exon 3 that is conserved across species and is a sues were collected after participants signed written in- primary miRNA for miR-155 [5]. MiR155HG was ini- formed consent. The study protocol was approved by tially thought to be involved in the human immune re- The Institutional Review Board of the First Affiliated sponse. For example, Haasch et al. showed that Hospital of Nanjing Medical University. GBM diagnoses transcriptional activation of miR155HG is an early and were confirmed by two independent pathologists. Pa- sustained T cell activation event [6]. Several oncological tients recruited into this study received no preoperative studies have shown that miR155HG is highly expressed treatments. in diffuse large and primary mediastinal B cell lymph- The normal human astrocyte cell line NHA and the omas [7]. MiR155HG can be induced by abnormal B-cell human GBM cell lines U87, U251, Ln229, T98, and receptor in Hodgkin lymphoma [8], and miR-155 was A172 were purchased from the Chinese Academy of Sci- also shown to be upregulated by mutant p53 and facili- ences Cell Bank (Shanghai, China). A primary GBM cell tate invasion of breast cancer cells [9]. Our group previ- line GP1 was extracted in December 2016 from the ously reported that the miR155HG/miR-155 axis tumor of a patient with a temporal GBM; a second pri- exhibits an oncogenic function in glioma by promoting mary GBM cell line GP2 was extracted in January 2017 epithelial-to-mesenchymal transition [10]. However, from the tumor of a patient with a frontal GBM. All cell other biological mechanisms of miR155HG, such as po- lines were stored in liquid nitrogen before use. Cell cul- tential ceRNA functions, and the upstream regulation of ture was performed as described previously [22]. miR155HG in astrocytoma have remained unknown. Annexin A2 (ANXA2) is a calcium-dependent phos Quantitative real-time PCR (qRT-PCR) and western pholipid-binding protein with demonstrated roles in blotting stimulating fibrinolytic processes, degrading extracellular QRT-PCR and western blotting were performed as matrix, and promoting angiogenesis [11]. Recent reports described previously [22]. The primer for miR155HG is have shown that ANXA2 is aberrantly expressed in a F5′-GAGTGCTGAAGGCTTGCTGT-3′,R5′-TTGA wide variety of tumors [12–15] and has been implicated ACATCCCAGTGACCAG-3′, for β-actin is F 5′-TCAC in various processes of tumorigenesis, including cell in- CCACACTGTGCCCA-TCTACGA-3′,R5′- CAGCGG vasion [16], proliferation [17] and neovascularization AACCGCTCATTGCCAATGG-3′. The antibodies for [18]. Previous studies reveal invasion function for western blot analysis were: anti-ANXA2 (1:1000; Abcam, ANXA2 in glioma and suggest its role as a potential Cambridge, UK), anti-cell cycle-related proteins (cyclin diagnostic and prognostic marker for glioma [19, 20]. E, cyclin D, CDK4, CDK6) (1:1000; Cell Signaling However, whether ANXA2 is involved in the oncogenic Technology, Danvers, MA, USA), anti-Bax (1:500; Santa Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 3 of 14 Cruz Biotechnology, Dallas, TX, USA), anti-Bcl-2 (1:500; rabbit anti-IgG antibody (Abcam), and rabbit anti-RNA Santa Cruz Biotechnology), anti-caspase 3 (1:1000; Abcam) Polymerase II antibody (Abcam) was used for positive and anti-β-actin (1:1000, Cell Signaling Technology). Cells control. After immunoprecipitation, the beads were were treated with EGF Recombinant Human Protein Solu- washed sequentially with low-salt buffer, high-salt buffer, tion (Thermo Fisher Scientific, Waltham, MA, USA) or the LiCl buffer, and TE buffer each for 5 min at 4 °C. The SH-4-54 inhibitor of STAT3 phosphorylation (Selleck, immunoprecipitated DNA was then eluted by incubation Shanghai, China) according to the manufacturer’sprotocol. in 100 μl of elution buffer (0.1 M NaHCO and 1% SDS) containing 10 μg proteinase K (Millipore) at 62 °C for 2 Oligonucleotides, plasmid and transfection h with rotation. The eluted DNA was purified using the To construct a plasmid expressing miR155HG, the columns and buffers in the kit and then re-dissolved in full-length human miR155HG sequence or a mutated 50 μl of PCR-grade water. Immunoprecipitated chroma- miR155HG sequence for miR-185-5p (NCBI Reference tin was analyzed by qPCR using primers targeting the Sequence: NR_001458.3) was synthesized and inserted into phosphorylated (p)-STAT3 (p-STAT3) binding regions in the pCDNA3.1 vector to generate pCDNA3.1-miR155HG the human miR155HG promoter. The primer sequences WT or MUT, respectively (Genechem, Shanghai, China). used for ChIP-qPCR are binding region 1 (− 1982 to − The hsa-miR-185-5p mimic, hsa-miR-185-5p inhibitor, and 1972): F 5′-GAGACATCATTATTGTCATT-3′,R 5′-TA hsa-miR-scramble were chemically synthesized (Ribobio, GGAGTCAAATACACCTG-3′; binding region 2 (− Guangzhou, China). The sequence of ANXA2 and 1548 to − 1411): F 5′-ATGGGAAATTCAGAAAGGC miR155HG siRNAs were as follows: si-miR155HG: sense, -3′,R5′-TGATCATATGAGGGAGGAGC-3′; and bind- 5′-CUGGGAUGUU-CAACCUUAATT-3′,antisense,5′-U ing region 3 (− 275 to − 116): F 5′-TTAAGAACA UAAGGUUGAACAUCCCAGTT-3′; scramble: sense, 5′-U AAGGTTGGAGC-3′,R5′-TGTGACTCATAACCGAC UCUCCGAACGUGUCACGUTT-3′,antisense,5′-ACGU CAG-3′. PCR conditions were set according to the in- GACACGUU-CGGAGAATT-3′.ANXA2 siRNAs were structions provided in the SYBR Green Kit (Roche Ap- chemically synthesized (Invitrogen, Shanghai, China). plied Science, Upper Bavaria, Germany). Results were ANXA2 siRNA 1 target sequence is 5′-CUGGGAAGAAGG analyzed by agarose gel electrophoresis. CUUC-CUUTT-3′, ANXA2 siRNA 2 target sequence is 5′-AAGGAAGCCUUCUUCCC-AGTT-3′,ANXA2 siRNA RNA immunoprecipitation (RIP) 3target sequence is5′- GGGAAGAAGGCUUCCU-UCAT RIP assays were performed using U87 cell extracts with the T-3′. Cells were transfected with oligonucleotides or plasmid EZ-Magna RIP RNA-Binding Protein Immunoprecipitation using Lipofectamine 2000 Reagent (Invitrogen, Carlsbad, Kit (Millipore, Burlington, MA, USA) according to the man- CA, USA) following the manufacturer’s instructions. ufacturer’s instructions. U87 cells were rinsed with cold PBS and fixed with 1% formaldehyde for 10 min. After centrifuga- Lentiviral packaging and stable cell lines tion, cell pellets were collected and resuspended in NP-40 Lentiviruses carrying shRNA-miR155HG or shRNA- lysis buffer (Thermo Fisher Scientific, Waltham, MA, USA) ANXA2 and the negative control lentivirus (sh-miR supplemented with 1 mM PMSF, 1 mM DTT, 1% Protease 155HG sequence is 5′-CUGGGAUGUUCAACCUUAA Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO, USA) and TT-3′; sh-ANXA2 sequence is 5′- CGGGATGCTTTG 200 U/ml RNase Inhibitor (Life Technologies, Carlsbad, CA, AACATTGAA -3′; sh-NC sequence is 5′-UUCUCCG USA). Lysates were subjected to high-speed centrifugation, AACGUGUCACGUTT-3′) were assembled in the hu- and then 100 μl of the supernatant was incubated with RIP man embryonic kidney cell line 293 T, and the viruses buffer containing magnetic beads conjugated with human were collected according to the manufacturer’s manual anti-Ago2 antibody (Cell Signaling Technology). Mouse IgG (Genechem). Stably transfected cell lines were estab- (Cell Signaling Technology) was used as a negative control, lished by infecting U87 cells with lentiviruses using a while SNRNP70 (Cell Signaling Technology) was used as a lentiviral packaging kit purchased from Genechem, positive control. Co-precipitated RNAs were detected by re- followed by puromycin selection. verse transcription PCR. Total RNAs (input control), IgG and SNRNP70 were assayed simultaneously to evaluate the Chromatin immunoprecipitation (ChIP) efficiency of Ago2-specific RNAs. ChIP assays were performed as previously described [22]. The EZ-magna ChIP kit (Millipore, Bedford, MA, Colony formation assays USA) was used according to the manufacturer’s proto- Cells (4 × 10 ) were seeded into cell culture dishes and col. Crosslinked chromatin was sonicated into DNA cultured for 15 d. Cell colonies were fixed with 4% para- fragments in the range of 200–1000 bp and immunopre- formaldehyde for 20 min and stained with 0.2% crystal cipitated using rabbit anti-p-STAT3 antibodies (Abcam). violet. Images were captured and colonies (diameter > Negative control samples were prepared using control 0.5 mm) were counted using Image J software (National Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 4 of 14 Institutes of Health, Arlington, VA, USA). All assays times. Finally, tissue sections were mounted with a were repeated at least three times. medium containing DAPI for 8 min in the dark and im- ages were obtained with fluorescence microscope EDU cell proliferation assays (Nikon, Japan). EDU cell proliferation assays were conducted with the Molecular Probes EdU-Alexa imaging detection kit (Life Dual-luciferase reporter assay Technologies). Cells treated for 48 h were incubated The ANXA2 3′-untranslated region (UTR) and the full with 10 μM EdU for 2 h, fixed with 4% paraformalde- miR155HG sequence containing miR-185-5p seed hyde, permeabilized with 1% Triton X-100, and stained matching sites were amplified from human cDNA via with the Alexa-Fluor 594 reaction cocktail for EdU and PCR and cloned into the 3′ end of the pGL3-basic lucif- Hoechst 33342 (nuclei). Images were obtained using a erase vector (Genechem). Mutated versions of each con- fluorescence microscope (Olympus, Japan). All assays struct were generated by mutating the miR-185-5p seed were repeated at least three times. site sequences (pGL3-wt or mut). The miR155HG pro- moter region sequence (2000 bp to 1000 bp upstream of Flow cytometric analysis transcription starting point) were also amplified and Transfected GBM cells in logarithmic growth were col- cloned into the 5′ end of the pGL3-basic luciferase vec- lected and processed with the Cell Cycle Staining Kit tor. Mutated version was generated by deleting binding (Multi Sciences, Hangzhou, China) for cell cycle analysis. region sequences of p-STAT3 (wt or mut-pGL3). U87 After washing with PBS, cells were fixed with 70% cells seeded into 96-well plates were co-transfected with ice-cold ethanol, incubated with Cell Cycle Staining Kit wt or mut report gene, the pRL-TK control (Promega, for 30 min in the dark, and analyzed by flow cytometry. Madison, WI, USA) and miR-185-5p mimic or miRNA In other experiments, treated cells in logarithmic growth NC using Lipofectamine 2000 (Invitrogen). The wt or were harvested and stained with the Annexin V-FITC mut-pGL3 and the pRL-TK control were co-transfected Apoptosis Detection Kit (Multi Sciences). After washing into the cells, then treated with cell culture with or with- with PBS and incubating with Annexin V/propidium out SH-4-54 inhibitor of STAT3 phosphorylation. At 48 iodide for 30 min in the dark, cells were analyzed by flow h after transfection, luciferase activity was determined cytometry. using the Dual Luciferase Reporter Assay System (Pro- mega, WI, USA) according to the manufacturer’s proto- Immunohistochemistry (IHC) col. The relative luciferase activity was normalized to Fresh intracranial tumor tissues from nude mice were Renilla luciferase activity. All assays were performed in fixed with 4% paraformaldehyde and then embedded in triplicate. paraffin. Sections were incubated at 4 °C overnight with primary antibodies against ANXA2 (1:1000; Abcam) and p-STAT3 (1:500; Abcam). Sections were then incubated Intracranial GBM mouse model with secondary antibody (1:1000; Santa Cruz) for 2 h at The animal experiments were approved by the Animal room temperature and stained with diaminobenzidine Management Rule of the Chinese Ministry of Health until brown granules appeared. (document 55, 2001) and were performed conforming to the approved guidelines and experimental protocols of Fluorescence in situ hybridization (FISH) Nanjing Medical University. U87 cells (1 × 10 ) stably RNA FISH was performed as described previously [22]. expressing MCS-firefly luciferase for bioluminescence MiR155HG-Bio probe was synthesized from GoodBio imaging were transfected with lentivirus expressing con- (Wuhan, China); the sequence is 5′-CCTCCCACG trol shRNA, shRNA-ANXA2 or shRNA-miR155HG and GCAGCAATTTGTTCCA-3′. Frozen sections of fresh then were intracranially injected into the frontal lobe of tissues were fixed with 4% formaldehyde for 10 min, nude mice to generate GBM (n = 10 mice per group). washed with PBS, and then digested with Proteinase K Tumor volumes were measured by luciferase using a for 5 min. After eliminating auto-fluorescence and block- bioluminescence imaging system (Caliper IVIS ing endogenous biotin, the sections were hybridized with Spectrum, PerkinElmer, Waltham, MA, USA) on days 1, probes overnight. Sections were then washed with 11, and 21 after implantation. The integrated flux of pre-warmed 2× SSC at 37 °C for 10 min, 1 × SSC at 37 °C photons (photons/s) within each region was determined for 10 min, and 0.5 × SSC for 10 min. Tissue sections by the Living Images software package (Caliper Life were then blocked with bovine serum albumin for 30 Sciences). Mice were sacrificed when they were in deep min at room temperature, followed by staining with coma. Brains were extracted, fixed in 10% formalin and 488-avidin (1:400) at room temperature for 50 min. then embedded in paraffin for IHC or frozen at − 80 °C Stained sections were washed with PBS for 5 min four for western blotting or FISH. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 5 of 14 Statistical analysis Together these results suggested that miR155HG may Data are presented as the mean ± standard deviation be involved in the malignant phenotypes of GBM. (SD). Statistical analyses were performed using the To evaluate the role of miR155HG in promoting malig- Student t test to evaluate the significance of differences nant phenotypes of GBM, we generated an intracranial between groups, one-way ANOVA (Tukey’s post hoc) GBM mouse model by injecting U87 cells infected with was used to determine the difference among at least lentivirus expressing control shRNA or shRNA-miR155HG. three groups using SPSS v19.0 for Windows. (SPSS, IL, The tumor volumes in the group of mice treated with USA). Pearson’s correlations analysis and heat map shRNA-miR155HG were smaller than those in the control microarray analysis were performed using Multiple at 11 and 21 days after implantation. Furthermore, mice Array Viewer 4.9 software (MEV). Kaplan–Meier sur- treated with shRNA-miR155HG showed better survival than vival analysis was performed using GraphPad 5.0 soft- controls (Additional file 1:FigureS1A andB). Theseresults ware. P < 0.05 indicates a significant difference. indicated that miR155HG acted as oncogene in promoting GBM growth. Results MiR155HG is overexpressed in GBM and miR155HG- MiR155HG sponges and downregulates miR-185-5p related genes are enriched in cancer-associated processes Previous studies have shown that lncRNAs can act as a To explore miR155HG expression in human astrocy- sponge for miRNAs, therefore we speculated that toma tissues, we examined three public human astrocy- miR155HG may exhibit sponge activity. RNA hybrid toma databases (TCGA, CGGA and Rembrandt) and bioinformatics tools showed that miR155HG contains a found overexpression of miR155HG in GBM (Fig. 1a). putative binding site for miR-185-5p as tumor suppres- We also found that miR155HG expression was elevated sor in a wide range of tumors [23–26] (Fig. 2a). We in 24 GBM specimens compared with adjacent normal found that miR-185-5p levels were lower in GBM tissues brain tissue from patients histologically diagnosed with than in normal brain tissue, and that miR-185-5p was GBM, and miR155HG was generally overexpressed in significantly negatively correlated with miR155HG in the the GBM cells compared with NHA cells from normal same GBM samples (Fig. 2b; r = − 0.5970, p = 0.0021). brain tissue using fluorescent qPCR (Fig. 1b). FISH ana- RNA-induced silencing complex (RISC) is an essential lysis revealed that miR155HG was primarily concen- factor in the biological effect of miRNAs, and Ago2 is an trated in the cytoplasm of GBM cells (Fig. 1c). elementary catalytic constituent of RISC that is involved MEV software was used to perform a Pearson’s correl- in RNA cleavage [27]. To explore the possible interactions ation analysis to identify genes associated with between miR155HG and miR-185-5p, RIP was performed miR155HG expression in the TCGA, CGGA and Rem- in U87 cells. SNRNP70 protein, which interacts with U1 brandt GBM databases. The results identified 1037 spliceosomal RNA [28], was used as a positive control. up-regulated genes and 873 down-regulated genes from MiR155HG was predominantly enriched in beads contain- the overlap of the three databases, and these genes were ing anti-Ago2 antibody compared with those harboring examined using the DAVID Web tool (http://david.abcc. control IgG (Fig. 2c). This result suggests that miR155HG ncifcrf.gov/home.jsp) for Gene Oncology (GO) and is capable of sponging miRNA as ceRNA. KEGG Pathway enrichment analyses. GO analysis Then we constructed a luciferase reporter plasmid con- showed that the upregulated genes were primarily taining the putative miR-185-5p binding site from enriched in tumor progression processes, such as regu- miR155HG, as well as a mutant construct in which the lating apoptosis and cell proliferation. Moreover, KEGG binding site was mutated (Fig. 2a). Co-transfecting pathway analysis indicated that the upregulated genes miR-185-5p mimic decreased the relative luciferase activ- were closely associated with pathways activated in can- ity in U87 cells transfected with wild-type plasmid, but cers, such as the PI3K-Akt signaling pathway, a promin- had no impact on the mutant construct (Fig. 2d), which ent pathway in tumorigenesis and cancer progression suggests that miR155HG directly binds miR-185-5p. (Fig. 1d-e). Furthermore, gene set enrichment analysis (GSEA) (http://www.broadinstitute.org/gsea/index.jsp) ANXA2 is the target molecule of miR-185-5p was used to examine genes that were expressed in Previous studies demonstrated a potential role for TCGA GBM samples from patients with high ANXA2 in glioma [19]. Bioinformatics analytical tools miR155HG expression and those with low miR155HG TargetScan and miRNAWalk 2.0 showed that the expression. Genes expressed in patients with high 3′-UTR of ANXA2 mRNA contained a seed sequence of miR155HG expression were primarily associated with miR-185-5p (Fig. 2e). To determine whether ANXA2 reduced apoptosis and cell death, while genes with nega- may be involved in the miR155HG-miR-185-5p axis in tive expression were primarily associated with cell cycle GBM, we first examined the expression levels of ANXA2 progression and cell cycle phase transition (Fig. 1f-g). in frozen GBM tissue samples by western blot. We Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 6 of 14 Fig. 1 MiR155HG expression correlates positively with malignant degrees of glioma, and the miR155HG-associated genes were chiefly enriched in cancer related function. a Expression of miR155HG positively correlates with WHO grade in TCGA, CGGA and Rembrandt Public database. b Relative levels of miR155HG in the tumors and their adjacent normal brain tissues from 24 patients diagnosed as glioblastomas, and relative expressions of miR155HG in glioma cell lines NHA, A172, U87, U251, Ln229, T98, and primary glioma cells from two patients. c The distribution of miR155HG was evaluated via FISH in U87 and GP1 cells. d-e MiR155HG associated genes from overlapping CGGA, TCGA and Rembrandt databases were analyzed with KEEG pathway analysis, gene oncology analysis. f-g MiR155HG associated genes were analyzed with gene set enrichment analysis (GSEA) by TCGA genes data Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 7 of 14 Fig. 2 MiR155HG stimulated the ANXA2 expression by sponging endogenous miR-185-5p. a The binding site of MiR155HG and miR-185-5p was predicted by bioinformatics tools. b Expression levels of miR-185-5p in GBM tissues and adjacent normal brain tissues were analyzed by real-time PCR and normalized to U6. The correlation between MiR155HG and miR-185-5p in GBM tissues was applied with Pearson’s correlation coefficient (r = − 0.5970, P = 0.0021). c Luciferase assays was performed after transfection with miR-185-5p mimic and pGL3- miR155HG-wt or pGL3- miR155HG-mut into U87 cells as well as the internal control Renilla plasmid. Relative luciferase activity was analyzed after 48 h treatment. (*p < 0.05, **p < 0.01). d Amount of miR155HG bound to Ago2 or IgG measured by RT–qPCR after RIP assays with anti- IgG, anti-Ago2, anti-SNRNP70 and 10% input. e The binding site in 3’UTR of ANXA2 mRNA for miR-185-5p was predicted by bioinformatics tools. f Expression levels of ANXA2 in GBM tissues and adjacent normal brain tissues were analyzed by western blot and normalized to β- actin. The correlation between miR-185-5p and ANXA2 in GBM tissues was applied with Pearson’s correlation coefficient (r = − 0.4676, P = 0.0212). g Luciferase assays was performed after transfection with miR-185-5p mimic and pGL3- ANXA2-wt or pGL3- ANXA2-mut into U87 and GP1 cells as well as the internal control Renilla plasmid. Relative luciferase activity was analyzed after 48 h treatment. (*p < 0.05, **p < 0.01). h The protein expression levels of ANXA2 were analyzed by Western blotting after 48 h transfection in U87 cells with pcDNA3.1, pcDNA3.1-miR155HG wt, pcDNA3.1-miR155HG mut, miR-NC or different amount of miR-185-5p mimic, respectively found that ANXA2 was highly expressed in GBM tis- 0. − 4676, P = 0.0212; Fig. 2f). We generated luciferase sue but not in normal brain tissue (Fig. 2f and Additional constructs containing either the wild-type (WT) 3′-UTR file 1: Figure S1C). We next examined the correlation be- of ANXA2 mRNA or a mutated (MUT) sequence in which tween miR-185-5p and ANXA2 in GBM tissue and found the miR-185-5p seed sequences were mutated. Luciferase that miR-185-5p negatively correlated with ANXA2 (r = assays showed that expression of miR-185-5p decreased Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 8 of 14 the luciferase activity of the WT reporter but not the ac- revealed that tumor growth was significantly inhibited at tivity of the MUT reporter in U87 and GP1 cells (Fig. 2g). 21 d in the group with decreased ANXA2 expression We speculated that ANXA2 levels in GBM cells may (Fig. 3f). Mice injected with cells expressing sh-ANXA2 be regulated by miR-185-5p and affected by its inter- also showed better survival than control mice (Fig. 3g). action with miR155HG. Indeed, transfection of a IHC of tumor sections from the sh-ANXA2 and control miR155HG expression vector increased ANXA2 levels groups confirmed decreased ANXA2 levels lead to in- in U87 and GP1 cells; however, the vector expressing hibition of tumor growth in vivo (Fig. 3h). Taken to- miR155HG with mutated binding sites for miR-185-5p gether, our in vitro and in vivo results demonstrate that had no effect on ANXA2 levels. In addition, ANXA2 exhibits oncogenic functions in GBM cells and miR155HG-mediated elevation of ANXA2 was blocked enhances the malignant phenotypes of tumors. by co-transfection with miR-185-5p mimic in a dose-dependent manner (Fig. 2h). Furthermore, inhibit- MiR155HG and miR-185-5p participate in GBM growth by ing miR155HG by siRNA downregulated ANXA2 levels regulating the proliferation and apoptosis of GBM cells in U87 and GP1 cells, which could be reversed by treat- As ANXA2 plays a crucial tumor-promoting role in ment with miR-185-5p inhibitor (Additional file 1: Fig- GBM and since ANXA2 levels are modulated by ure S1D). Together these results demonstrated that miR155HG and miR-185-5p, we hypothesized that miR155HG may promote ANXA2 expression by modu- miR155HG and miR-185-5p could interfere with the lating the capacity of miR-185-5p to bind the 3′-UTR of proliferation and apoptosis of GBM cells. Downregulat- ANXA2 mRNA. ing miR155HG levels in U87 and GP1 cells by siRNA re- duced proliferation, blocked cell cycle progression, and ANXA2 enhances the malignant phenotypes of GBM cells stimulated apoptosis in GBM cells, and these effects As ANXA2 was the downstream molecule positively were reversed by miR-185-5p inhibitor (Fig. 4a–d). modulated by miR155HG via the ceRNA mechanism, we Western blot analysis showed altered levels of proteins needed to investigate the function of ANXA2 to explain associated with proliferation and apoptosis, which was the oncological role of miR155HG in GBM. Bioinformat- consistent with the above results (Fig. 4e). These results ics analysis showed that ANXA2 expression was mostly indicate that the biological role of miR155HG and expressed in GBM samples from TCGA, CGGA and miR-185-5p in GBM cells was due to regulating proteins Rembrandt database (Additional file 2: Figure S2A). GO associated with proliferation and apoptosis, and this may analysis showed ANXA2 was closely associated with be partially by regulating ANXA2 expression. genes involved in cell apoptosis and proliferation (Add- itional file 2: Figure S2B). ANXA2 affected miR155HG expression in GBM cells via p- We next performed a series of experiments to evaluate STAT3 levels the possible oncogenic function of ANXA2 in GBM. Our results show that miR155HG can interfere with Colony formation and EDU assays were performed in ANXA2 expression by sponging miR-185-5p in GBM cells transfected with ANXA2 siRNA to evaluate the ef- cells. Previous studies show that ANXA2 can act with fect of ANXA2 on proliferation. The effect of ANXA2 AKT, STAT3 to promote downstream oncogenes [29, siRNA was shown in Additional file 3: Figure S3B. A signifi- 30], we speculate if ANXA2 can contribute to cant reduction of proliferation was observed in U87 and miR155HG aberrantly overexpression in GBM cells in GP1 cells transfected with ANXA2 siRNA compared with this way. Pearson’s correlation analysis showed that controls (Fig. 3a–b). Flow cytometry revealed that knocking ANXA2 was significantly positively correlated with down ANXA2 induced a G1/S arrest and decreased the miR155HG in the TCGA, CGGA and Rembrandt GBM percentage of cells in S phase (Fig. 3c). Flow cytometry also databases (Fig. 5a). We also found a positive correlation showed that knocking down ANXA2 remarkably increased between ANXA2 and miR155HG in WHO II and WHO apoptosis rates in GBM cells (Fig. 3d). Western blot results III astrocytoma patients in these independent public da- of proliferation- and apoptosis-associated proteins were tabases (Additional file 2: Figure S2C). Next, one repre- consistent with the above results (Fig. 3e). sentative GBM tissue exacted from a GBM patient To examine the oncogenic ability of ANXA2 in GBM during surgery was examined by FISH double staining in vivo, nude mouse tumorigenicity assays were per- and the results showed that ANXA2 and miR155HG formed using a U87 xenograft model. U87 cells infected were both significantly expressed in dense tumor tissues, with fluorescent lentiviruses expressing sh-ANXA2 or but not in loose normal brain tissue (Fig. 5b). Then controls were injected into nude mouse brains. The ef- double-staining was performed to investigate to show fect of ANXA2 knockdown in cells and GBM tissue was that inhibiting ANXA2 caused a miR155HG downregula- shown in Additional file 1: Figure S1E. In vivo imaging tion in the brain of nude mice from Fig. 3g and interfered of the nude mice at 1, 11 and 21 d after implantation with tumor growth (Fig. 5c). We thus concluded that Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 9 of 14 Fig. 3 Downregulating ANXA2 expression inhibited GBM cells proliferation, cell cycle and apoptosis in vitro and vivo. a Colony formation assays in U87 and Primary glioblastoma cells (GP1) transfected with scramble, ANXA2 siRNA 1 and siRNA 2. Scale bar> 500 μm. b U87 and GP1 in EDU transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. Representative merged images were shown (original magnification, 200×). c The cell cycle was detected in U87 and GP1 transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. d The apoptotic cells were measured with flow cytometry in U87 and GP1 transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. e The protein related with cell proliferation, cell cycle and apoptosis was measured with immunoblotting in U87 and GP1 cells transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. All experiments above were performed 3 times, and average scores are indicated with error bars on the histogram.*P < 0.05, **P < 0.01. f U87 cells transfected with a lentivirus with sh-ANXA2 or sh-NC and a lentivirus containing luciferase were implanted in the brain of 10 nude mice, respectively. The tumor formation was assessed by bioluminescence imaging. The bioluminescent images were measured at days 1, 11 and 21 after implantation. g Overall survival was determined by Kaplan-Meier survival curves between sh-ANXA2 or sh-NC group, and a log- rank test was used to assess the statistical differences. h Two representative immunohistochemical image of tumors from groups of nude mice implanted with U87 cells, transfected with a lentivirus with sh-ANXA2 or sh-NC, were shown to compare the volume size of tumors ANXA2 expression positively correlated with miR155HG resulted in reduced levels of p-STAT3 without impacting levels in GBM. STAT3levelsinGBM cells(Additional file 3:FigureS3B). Previous studies showed that ANXA2 modulates STAT3 Inhibiting miR155HG also resulted in decreased ANXA2 phosphorylation (p-STAT3) levels to stimulate proliferation, and p-STAT3 in vivo (Additional file 3:FigureS3E). Knock- angiogenesis, metastasis and epithelial to mesenchymal ing down ANXA2 resulted in reduced miR155HG expres- transition in breast cancer cells [30–32]. We found higher sion, but this effect was not observed in GBM cells treated levels of p-STAT3 in GBM cells and tissues compared with with a STAT3 phosphorylation inhibitor (Fig. 5e). Moreover, normal brain cell and tissue(Additional file 3:FigureS3A), induction of STAT3 phosphorylation by Epidermal Growth with distribution in nuclei in GBM cells (Fig. 5d). We thus Factor (EGF) [33, 34] could rescue the inhibitory effect of wondered if ANXA2-mediated induction of miR155HG in- ANXA2 depletion on miR155HG expression (Fig. 5fand volved p-STAT3. We first found that decreasing ANXA2 Additional file 3: Figure S3C). These results suggested that Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 10 of 14 Fig. 4 The impact of miR155HG and miR-185-5p on cell proliferation, cell cycle and apoptosis in GBM cells. a Colony formation assays in U87 and GP1 cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor. Scale bar> 500 μm. b U87 and Primary glioblastoma cells in EDU transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor after 48 h. Representative merged images were shown (original magnification, 200×). c The cell cycle was detected in U87 and Primary glioblastoma cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor after 48 h. d The apoptotic cells were measured with flow cytometry in U87 and Primary glioblastoma cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor after 48 h. e The protein related with cell proliferation, cell cycle and apoptosis was measured with immunoblotting in U87 and Primary glioblastoma cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR- 185-5P inhibitor after 48 h. All experiments were performed 3 times, and average scores are indicated with error bars on the histogram.*P < 0.05, **P < 0.01 p-STAT3 might directly promote miR155HG expression in Together these results indicated that the activated GBM. transcription factor p-STAT3 plays a key role in Bioinformatics tools identified three putative binding re- ANXA2-driven miR155HG expression and promotes gions for p-STAT3 in the miR155HG promoter. ChIP assays GBM cell growth, through the DNA binding activity of confirmed that p-STAT3 could bind putative binding region the p-STAT3 transcription factor. 2(− 1548 bp to − 1411 bp) but not binding region 1 (− 275 bp to − 161 bp) and binding region 3 (− 1982 bp to − 1972 Discussion bp) in the miR155HG promoter (Fig. 5g). Two luciferase We previously revealed that miR-155, which is derived reporter plasmids containing either the entire miR155HG from miR155HG, functions in epithelial-mesenchymal promoter region 2000 bp to 1000 bp upstream of the tran- transition in glioma [10]. In this report, we further scription start site (wt-pGL3) or the promoter deleted for showed that the lncRNA miR155HG is highly expressed binding region 2 (mut-pGL3) were transfected into GBM in GBM, where it acts as a ceRNA to sponge miR- cells along with or without STAT3 phosphorylation inhibitor. 185-5p, thus promoting miR155HG downstream mole- While inhibiting STAT3 phosphorylation resulted in down- cules level such as ANXA2. Our study clarifies the regulation of luciferase activity driven by the wild-type pro- mechanism by which miR155HG positively regulates moter, the p-STAT3-driven luciferase activity on the mutated AXNA2 to sustain the malignant phenotypes of astrocy- reporter remained unchanged (Additional file 3:Figure S3D). toma, particularly GBM. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 11 of 14 Fig. 5 ANXA2 promote miR155HG expression via P-STAT3 in GBM cells a ANXA2 positively correlates with miR155HG in GBM specimens of three independent public database. b The level of ANXA2 and miR155HG was determined by immunofluorescence in boundary between invasive GBM and normal brain tissue from a GBM patient. c Two representative image of tumors from groups of nude mice implanted with U87 cells transfected with a lentivirus of sh-ANXA2 or sh-NC, were shown to compare the volume size of tumors and the expression of miR155HG via double immunofluorescent staining of the same frozen slice of tumor. d The specific localization of p-STAT3 and ANXA2 in U87 and GP1 cells was detected by immunofluorescence. e Q-PCR was performed to explore the level of miR155HG after downregulating ANXA2 in U87 and GP1 cells treated or untreated with inhibitor of STAT3 phosphorylation SH-4-54. f EGF-induced p-STAT3 rescued the ANXA2 knockdown-mediated downregulation of miR155HG in U87 and GP1 cells. g CHIP assays was performed to confirm the three putative binding region for p-STAT3 in promoter region of miR155HG Through bioinformatics analyses, we found that glioma associated with inhibiting glioma cell invasion [40]. miR155HG was closely associated with the proliferative Another report showed that the lncRNA Linc00176 regu- activity and apoptosis resistance of GBM in three inde- lates the cell cycle by sponging miR-185 in hepatocellular pendent GBM gene expression arrays. Knocking down carcinoma [41]. Here, we confirmed that lncRNA miR155HG in GBM cells resulted in cell cycle arrest, de- miR155HG binds to miR-185 to impact proliferation, cell creased cell growth, and apoptosis. As lncRNAs have cycle progression and apoptosis in GBM cell lines. been proven to function as miRNA sponges [35], we hy- ANXA2, a 36-kDa protein that belongs to the family pothesized that miR155HG might also regulate gene ex- of calcium-dependent phospholipid binding proteins pression by competing for shared miRNA response [11], is a DNA-binding protein that modulates DNA elements in GBM. Several studies have shown that synthesis. Several studies showed that ANXA is involved miR-185 is involved in suppressing non-small cell lung in cell proliferation and cell cycle progression [42–45]in cancer [36], gastric cancer [37], hepatocellular carcinoma a variety of cancer cell types, such as breast cancer [18], [38] and prostate cancer [39] and is downregulated in hepatocellular carcinoma [46], colorectal cancer [47] and Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 12 of 14 Fig. 6 Schematic diagram of the relationship among miR155HG, miR-185-5p, p-STAT3 and ANXA2 in GBM. The binding of miR-185-5p to miR155HG leads to abolish miR-185-5p targeting ANXA2 3’UTR, which stimulated ANXA2-elevated level in cell. In turn ANXA2 then activates STAT3 phosphorylation, and STAT3 phosphorylation shifts from the cytosol to the nucleus, forms dimers and binds with promoter region to upregulate the expression of miR155HG pancreatic cancer [48]. We found that ANXA2 was over- that p-STAT3 could bind to the miR155HG promoter expressed in GBM. Based on TCGA, CGGA and Rem- region from − 1548 bp to − 1411 bp upstream of the brandt GBM gene expression profiles, we found that transcription start site to stimulate miR155HG expres- ANXA2-associated genes were primarily enriched in cell sion. This suggests that STAT3 phosphorylation is the proliferation and apoptosis. Knocking down ANXA2 critical role in driving lncRNAs regulation similar to inhibited proliferation and increased the G1/S cell cycle miR155HG in malignant brain tumor. arrest apoptosis of U87 and GP1 cells. We also showed Here we have established miR155HG/miR-185-5p/ that ANXA2 is a key determining factor of survival that ANXA2 loop in GBM formation and progression. How- promotes the growth of intracranial GBM tumors in ever, other interactions as epigenetic regulation between nude mice. We further found ANXA2 is a direct target miR155HG and its downstream effector molecules re- of miR-185-5p, and its expression was perturbed by mains obscure since lncRNA might recruit chromatin re- miR-185-5p. Taken together, these results established modeling complex. More biological studies and clinical the miR155HG/miR-185-5p/ANXA2 axis, which under- trials are needed to evaluate the practicality of targeting lies the biological mechanisms of miR155HG in GBM. miR155HG for the treatment of GBM. Silencing ANXA2 was previously reported to inhibit activation of STAT3 (p-STAT3) [31, 32, 49], and we con- Conclusions firmed that ANXA2 knockdown decreased p-STAT3 in Our study suggested that the lncRNA miR155HG in- GBM cells. Activated p-STAT3 forms homologous di- creases ANXA2 expression by sponging miR-185-5p to mers and enters into the nucleus to function as a tran- exert tumorigenic effects and that ANXA2 stimulates scription factor to promote target gene expression [50]. miR155HG level via ANXA2-driven p-STAT3 in GBM Since constitutively activated STAT3 is closely associated (Fig. 6). Thus, we have identified the miR155HG/ with GBM [51, 52], we find that ANXA2-mediated ele- miR-185-5p/ANXA2 loop and its mechanisms and bio- vated miR155HG levels is mainly due to p-STAT3 level. logical effects in malignant brain tumors. This loop ChIP assay and luciferase reporter gene assay showed could serve as a novel therapeutic biomarker for GBM. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 13 of 14 Additional files manuscript; and WWN, ZJX and WYY designed this study. All authors have read and approved the final manuscript. Additional file 1: Figure S1. (A) U87 cells pretreated with a lentivirus Ethics approval and consent to participate with sh-miR155HG or sh-NC and a lentivirus containing luciferase were The procedures of this study were approved by the Institutional Review implanted in the brains of nude mice, and tumor formation was assessed Board of Nanjing Medical University. by bioluminescence imaging. The bioluminescent images were measured at days 1, 11 and 21 after implantation. (B) Overall survival from two Consent for publication groups of nude mice injected with U87 cells, transfected with sh-NC or Not applicable. sh-miR155HG lentivirus, was determined by Kaplan-Meier survival curves, and a log-rank test was used to assess the statistical significance of the Competing interests differences. (C) Expression levels of ANXA2 in GBM tissues and adjacent The authors declare that they have no competing interests. normal brain tissues were analyzed by western blotting and normalized to β-catenin. (D) The protein expression levels of ANXA2 were analyzed by western blotting after 48 h transfection in U87 and GP1 cells with Publisher’sNote scramble, miR155HG siRNA 1, miR155HG siRNA 2, miR-NC or inhibitor, Springer Nature remains neutral with regard to jurisdictional claims in respectively. (E) The effect of sh-ANXA2 in U87 cell and tumor tissue of published maps and institutional affiliations. nude mice after implantation were analyzed by western blotting. (E) Downregulating ANXA2 contributed to the reduction of p-STAT3 level in Author details GBM cells. (TIF 9482 kb) Department of Neurosurgery, The First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210029, China. Department of Neurosurgery, Additional file 2: Figure S2. (A) Expression of ANXA2 in TCGA, CGGA Nanjing First Hospital of Nanjing Medical University, Nanjing, Jiangsu, China. and Rembrandt astrocytoma database. (B) ANXA2 associated genes from overlapping CGGA, TCGA and Rembrandt databases were analyzed with Received: 14 November 2018 Accepted: 8 March 2019 gene oncology analysis. (C) ANXA2 positively correlates with miR155HG in WHOII/III astrocytoma specimens of three independent public database. (TIF 3895 kb) References Additional file 3: Figure S3. (A) Expression levels of p-STAT3 in cell lines, 1. Abou-Antoun TJ, Hale JS, Lathia JD, Dombrowski SM. Brain Cancer Stem GBM tissues and normal brain tissues were analyzed by western blot. (B) Cells in Adults and Children: Cell Biology and Therapeutic Downregulating ANXA2 contributed to the reduction of p-STAT3 level in GBM ImplicationsNeurotherapeutics : the journal of the American Society for cells. (C) Overexpression STAT3 was constitutively activated by EGF in ANXA2- Experimental NeuroTherapeutics. 2017. depleted GBM cells in U87 and GP1 cells. (D) Luciferase assays was performed 2. Patel AP, Tirosh I, Trombetta JJ, Shalek AK, Gillespie SM, Wakimoto H, et al. after transfection with miR155HG promoter wt-pGL3 or miR155HG promoer Single-cell RNA-seq highlights intratumoral heterogeneity in primary mut-pGL3 as well as the internal control Renilla plasmid into U87 and GP1 glioblastoma. Science (New York, NY). 2014;344:1396–401. cells. The cells then were treated with or without SH-4-54. Relative luciferase 3. Bartonicek N, Maag JL, Dinger ME. Long noncoding RNAs in cancer: activity was analyzed after 48 h treatment. (*p < 0.05, **p < 0.01). (E) Two mechanisms of action and technological advancements. Mol Cancer. groups of representative immunohistochemical image of tumors from groups 2016;15:43. of nude mice implanted with U87 cells, transfected with a lentivirus with sh- 4. Kallen AN, Zhou XB, Xu J, Qiao C, Ma J, Yan L, et al. The imprinted H19 miR155HG or sh-NC, were shown to compare the volume size of tumors and lncRNA antagonizes let-7 microRNAs. Mol Cell. 2013;52:101–12. the expression of ANXA2 and p-STAT3 through serial slices of same section of 5. Elton TS, Selemon H, Elton SM, Parinandi NL. Regulation of the MIR155 host tumor. (TIF 13681 kb) gene in physiological and pathological processes. Gene. 2013;532:1–12. 6. Haasch D, Chen YW, Reilly RM, Chiou XG, Koterski S, Smith ML, et al. T cell activation induces a noncoding RNA transcript sensitive to inhibition by Abbreviations immunosuppressant drugs and encoded by the proto-oncogene, BIC. Cell EGF: Epidermal Growth Factor; CeRNAs: Competing endogenous RNAs; Immunol. 2002;217:78–86. ChIP: Chromatin immunoprecipitation; FISH: Fluorescence in situ 7. Kluiver J, Poppema S, de Jong D, Blokzijl T, Harms G, Jacobs S, et al. BIC and hybridization; GBM: Glioblastoma; GP1: Glioblastoma cell from patient 1; miR-155 are highly expressed in Hodgkin, primary mediastinal and diffuse IHC: Immunoprecipitation; RIP: RNA immunoprecipitation large B cell lymphomas. J Pathol. 2005;207:243–9. 8. van den Berg A, Kroesen BJ, Kooistra K, de Jong D, Briggs J, Blokzijl T, et al. Acknowledgments High expression of B-cell receptor inducible gene BIC in all subtypes of We thank Department of Neurosurgery, The First Affiliated Hospital of Hodgkin lymphoma. Genes, chromosomes & cancer. 2003;37:20–8. Nanjing Medical University (Nanjing, China) and Department of 9. Neilsen PM, Noll JE, Mattiske S, Bracken CP, Gregory PA, Schulz RB, et al. Neurosurgery, Nanjing First Hospital of Nanjing Medical University for their Mutant p53 drives invasion in breast tumors through up-regulation of miR- generous help. We thank Edanz Group (www.edanzediting.com/ac) for 155. Oncogene. 2013;32:2992–3000. editing a draft of this manuscript. 10. Wu X, Wang Y, Yu T, Nie E, Hu Q, Wu W, et al. Blocking MIR155HG/miR-155 axis inhibits mesenchymal transition in glioma. Neuro-oncology. 2017;19: 1195–205. Funding 11. Rescher U, Gerke V. Annexins--unique membrane binding proteins with This work was supported by grants from National Natural Science diverse functions. J Cell Sci. 2004;117:2631–9. Foundation of China (81672501), Jiangsu Province’s Science and Technology 12. Yang T, Peng H, Wang J, Yang J, Nice EC, Xie K, et al. Prognostic and Foundation (BE2018724), Jiangsu Province’s Natural Science Foundation diagnostic significance of annexin A2 in colorectal cancer. Colorectal (20170108), the Program for Advanced Talents within Six Industries of disease : the official journal of the Association of Coloproctology of great. Jiangsu Province (2016-WSW-013), Jiangsu Province’s Key Discipline of Britain and Ireland. 2013;15:e373–81. Medicine (ZDXKA2016001), and the Priority Academic Program Development 13. Shetty PK, Thamake SI, Biswas S, Johansson SL, Vishwanatha JK. Reciprocal of Jiangsu Higher Education Institutions (PAPD). regulation of annexin A2 and EGFR with Her-2 in Her-2 negative and herceptin-resistant breast cancer. PLoS One. 2012;7:e44299. Availability of data and materials 14. Sharma MR, Koltowski L, Ownbey RT, Tuszynski GP, Sharma MC. The dataset(s) supporting the findings of this study are included within the article. Angiogenesis-associated protein annexin II in breast cancer: selective expression in invasive breast cancer and contribution to tumor invasion and Authors’ contributions progression. Exp Mol Pathol. 2006;81:146–56. WWN, YTF and WYZ performed in vitro assays, collected clinical samples and 15. Takano S, Togawa A, Yoshitomi H, Shida T, Kimura F, Shimizu H, et al. analyzed the data; TW performed in vivo experiments; WWN wrote the Annexin II overexpression predicts rapid recurrence after surgery in Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 14 of 14 pancreatic cancer patients undergoing gemcitabine-adjuvant 38. Qadir XV, Han C, Lu D, Zhang J, Wu T. miR-185 inhibits hepatocellular chemotherapy. Ann Surg Oncol. 2008;15:3157–68. carcinoma growth by targeting the DNMT1/PTEN/Akt pathway. Am J 16. Zhuang H, Tan M, Liu J, Hu Z, Liu D, Gao J, et al. Human epididymis protein Pathol. 2014;184:2355–64. 4 in association with Annexin II promotes invasion and metastasis of 39. Liu C, Chen Z, Hu X, Wang L, Li C, Xue J, et al. MicroRNA-185 downregulates ovarian cancer cells. Mol Cancer. 2014;13:243. androgen receptor expression in the LNCaP prostate carcinoma cell line. Mol Med Rep. 2015;11:4625–32. 17. Shetty P, Bargale A, Patil BR, Mohan R, Dinesh US, Vishwanatha JK, et al. Cell 40. Tang H, Wang Z, Liu X, Liu Q, Xu G, Li G, et al. LRRC4 inhibits glioma cell surface interaction of annexin A2 and galectin-3 modulates epidermal growth factor receptor signaling in Her-2 negative breast cancer cells. Mol growth and invasion through a miR-185-dependent pathway. Curr Cancer Drug Targets. 2012;12:1032–42. Cell Biochem. 2016;411:221–33. 41. Tran DDH, Kessler C, Niehus SE, Mahnkopf M, Koch A, Tamura T. Myc target 18. Onishi M, Ichikawa T, Kurozumi K, Inoue S, Maruo T, Otani Y, et al. Annexin gene, long intergenic noncoding RNA, Linc00176 in hepatocellular A2 regulates angiogenesis and invasion phenotypes of malignant glioma. carcinoma regulates cell cycle and cell survival by titrating tumor Brain tumor pathology. 2015;32:184–94. suppressor microRNAs. Oncogene. 2018;37:75–85. 19. Zhai H, Acharya S, Gravanis I, Mehmood S, Seidman RJ, Shroyer KR, et al. 42. Boyko V, Mudrak O, Svetlova M, Negishi Y, Ariga H, Tomilin N. A major Annexin A2 promotes glioma cell invasion and tumor progression. J cellular substrate for protein kinases, annexin II, is a DNA-binding protein. Neurosci. 2011;31:14346–60. FEBS Lett. 1994;345:139–42. 20. Maruo T, Ichikawa T, Kanzaki H, Inoue S, Kurozumi K, Onishi M, et al. 43. Krutilina RI, Babich VS, Kropotov AV, Mudrak OS, Chesnokov IN, Turoverova Proteomics-based analysis of invasion-related proteins in malignant gliomas. LV, et al. The DNA-binding activity of the membrane protein annexin II and Neuropathology : official journal of the Japanese Society of its interaction with antibodies to the chromatin nuclear ribonucleoprotein Neuropathology. 2013;33:264–75. (alpha-RNP). Tsitologiia. 1996;38:1106–14. 21. Wu W, Zhou X, Yu T, Bao Z, Zhi T, Jiang K, et al. The malignancy of 44. Bao H, Jiang M, Zhu M, Sheng F, Ruan J, Ruan C. Overexpression of Annexin miR-18a in human glioblastoma via directly targeting CBX7. Am J II affects the proliferation, apoptosis, invasion and production of Cancer Res. 2017;7:64–76. proangiogenic factors in multiple myeloma. Int J Hematol. 2009;90:177–85. 22. Wu W, Hu Q, Nie E, Yu T, Wu Y, Zhi T, et al. Hypoxia induces H19 expression 45. Lokman NA, Ween MP, Oehler MK, Ricciardelli C. The role of annexin A2 in through direct and indirect Hif-1alpha activity, promoting oncogenic effects tumorigenesis and cancer progression. Cancer microenvironment : official in glioblastoma. Sci Rep. 2017;7:45029. journal of the International Cancer Microenvironment Society. 2011;4:199–208. 23. Tian C, Deng Y, Jin Y, Shi S, Bi H. Long non-coding RNA RNCR3 promotes 46. Kohli S, Bhardwaj A, Kumari R, Das S. SIRT6 is a target of regulation by prostate cancer progression through targeting miR-185-5p. Am J Transl Res. UBE3A that contributes to liver tumorigenesis in an ANXA2-dependent 2018;10:1562–70. manner. Cancer Res. 2018;78:645–58. 24. Afshar S, Najafi R, Sedighi Pashaki A, Sharifi M, Nikzad S, Gholami MH, et al. 47. Rocha MR, Barcellos-de-Souza P, Sousa-Squiavinato ACM, Fernandes PV, de MiR-185 enhances radiosensitivity of colorectal cancer cells by targeting Oliveira IM, Boroni M. Annexin A2 overexpression associates with colorectal IGF1R and IGF2. Biomedicine & pharmacotherapy = Biomedecine & cancer invasiveness and TGF-ss induced epithelial mesenchymal transition pharmacotherapie. 2018;106:763–9. via Src/ANXA2/STAT3. 2018;8:11285. 25. Zhang Q, Chen Y, Liu K. miR-185 inhibits cell migration and invasion of 48. Keklikoglou I, Hosaka K, Bender C, Bott A, Koerner C, Mitra D, et al. hepatocellular carcinoma through CDC42. Oncol Lett. 2018;16:3101–7. MicroRNA-206 functions as a pleiotropic modulator of cell proliferation, 26. Zhao L, Zhang Y, Liu J, Yin W, Jin D, Wang D, et al. MiR-185 inhibits cell invasion and lymphangiogenesis in pancreatic adenocarcinoma by proliferation and invasion of non-small cell lung cancer by targeting KLF7. targeting ANXA2 and KRAS genes. Oncogene. 2015;34:4867–78. Oncol Res. 2018. 49. Yuan J, Yang Y, Gao Z, Wang Z, Ji W, Song W, et al. Tyr23 phosphorylation 27. Filipowicz W, Bhattacharyya SN, Sonenberg N. Mechanisms of post- of Anxa2 enhances STAT3 activation and promotes proliferation and transcriptional regulation by microRNAs: are the answers in sight? Nat Rev invasion of breast cancer cells. Breast Cancer Res Treat. 2017;164:327–40. Genet. 2008;9:102–14. 50. Aaronson DS, Horvath CM. A road map for those who don't know JAK- 28. Schumperli D, Pillai RS. The special Sm core structure of the U7 snRNP: far- STAT, vol. 296. New York, NY: Science; 2002. p. 1653–5. reaching significance of a small nuclear ribonucleoprotein. Cellular and 51. Villalva C, Martin-Lanneree S, Cortes U, Dkhissi F, Wager M, Le Corf A, et al. molecular life sciences : CMLS. 2004;61:2560–70. STAT3 is essential for the maintenance of neurosphere-initiating tumor cells 29. Su SC, Maxwell SA, Bayless KJ. Annexin 2 regulates endothelial in patients with glioblastomas: a potential for targeted therapy? Int J morphogenesis by controlling AKT activation and junctional integrity. J Biol Cancer. 2011;128:826–38. Chem. 2010;285:40624–34. 52. McFarland BC, Gray GK, Nozell SE, Hong SW, Benveniste EN. Activation of 30. Zhang F, Wang Z, Yuan J, Wei X, Tian R, Niu R. RNAi-mediated silencing of the NF-kappaB pathway by the STAT3 inhibitor JSI-124 in human Anxa2 inhibits breast cancer cell proliferation by downregulating cyclin D1 glioblastoma cells. Mol Cancer Res. 2013;11:494–505. in STAT3-dependent pathway. Breast Cancer Res Treat. 2015;153:263–75. 31. Maji S, Chaudhary P, Akopova I, Nguyen PM, Hare RJ, Gryczynski I, et al. Exosomal Annexin II promotes angiogenesis and breast Cancer metastasis. Molecular cancer research : MCR. 2017;15:93–105. 32. Wang T, Yuan J, Zhang J, Tian R, Ji W, Zhou Y, et al. Anxa2 binds to STAT3 and promotes epithelial to mesenchymal transition in breast cancer cells. Oncotarget. 2015;6:30975–92. 33. Lo HW, Hsu SC, Xia W, Cao X, Shih JY, Wei Y, et al. Epidermal growth factor receptor cooperates with signal transducer and activator of transcription 3 to induce epithelial-mesenchymal transition in cancer cells via up-regulation of TWIST gene expression. Cancer Res. 2007;67:9066–76. 34. Davis FM, Azimi I, Faville RA, Peters AA, Jalink K, Putney JW Jr, et al. Induction of epithelial-mesenchymal transition (EMT) in breast cancer cells is calcium signal dependent. Oncogene. 2014;33:2307–16. 35. Ebert MS, Sharp PA. Emerging roles for natural microRNA sponges. Curr Biol. 2010;20:R858–61. 36. Li S, Ma Y, Hou X, Liu Y, Li K, Xu S, et al. MiR-185 acts as a tumor suppressor by targeting AKT1 in non-small cell lung cancer cells. Int J Clin Exp Pathol. 2015;8:11854–62. 37. Li Q, Wang JX, He YQ, Feng C, Zhang XJ, Sheng JQ, et al. MicroRNA-185 regulates chemotherapeutic sensitivity in gastric cancer by targeting apoptosis repressor with caspase recruitment domain. Cell Death Dis. 2014; 5:e1197. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Experimental & Clinical Cancer Research Springer Journals

The miR155HG/miR-185/ANXA2 loop contributes to glioblastoma growth and progression

Download PDF
14 pages

Loading next page...
 
/lp/springer-journals/the-mir155hg-mir-185-anxa2-loop-contributes-to-glioblastoma-growth-and-8Zw2ueMMod

References (55)

Publisher
Springer Journals
Copyright
Copyright © 2019 by The Author(s).
Subject
Biomedicine; Cancer Research; Immunology; Apoptosis; Oncology
eISSN
1756-9966
DOI
10.1186/s13046-019-1132-0
Publisher site
See Article on Publisher Site

Abstract

Background: Glioblastoma multiforme (GBM) is the most common and aggressive form of astrocytoma among adult brain tumors. Multiple studies have shown that long non-coding RNAs (lncRNAs) play important roles in acting as molecular sponge for competing with microRNAs (miRNAs) to regulate downstream molecules in tumor progression. We previously reported that miR155 host gene (miR155HG), an lncRNA, and its derivative miR-155 promote epithelial-to-mesenchymal transition in glioma. However, the other biological functions and mechanisms of miR155HG sponging miRNAs have been unknown. Considering ANXA2 has been generally accepted as oncogene overexpressed in a vast of cancers correlated with tumorigenesis, which might be the target molecule of miR155HG sponging miRNA via bioinformatics analysis. We designed this study to explore the interaction of miR155HG and ANXA2 to reveal the malignancy of them in GBM development. Methods: The expression of miR155HG was analyzed in three independent databases and clinical GBM specimens. Bioinformatics analysis was performed to assess the potential tumor-related functions of miR155HG. The interaction of miR155HG and miR-185 and the inhibition of ANXA2 by miR-185 were analyzed by luciferase reporter experiments, and biological effects in GBM were explored by colony formation assays, EDU cell proliferation assays, flow cytometric analysis and intracranial GBM mouse model. Changes in protein expression were analyzed using western blot. We examined the regulatory mechanism of ANXA2 on miR155HG in GBM by gene expression profiling analysis, double immunofluorescence staining, chromatin immunoprecipitation and luciferase reporter assays. Results: We found that miR155HG was upregulated in GBM tissues and cell lines. Bioinformatic analyses of three GBM databases showed that miR155HG expression levels were closely associated with genes involved in cell proliferation and apoptosis. Knocking down miR155HG suppressed GBM cell proliferation in vitro, induced a G1/S-phase cell cycle arrest, and increased apoptosis. We also found that miR155HG functions as a competing endogenous RNA for miR-185. Moreover, miR-185 directly targets and inhibits ANXA2, which exhibits oncogenic functions in GBM. We also found that ANXA2 promoted miR155HG expression via STAT3 phosphorylation. Conclusion: Our results demonstrated that overexpressed miR155HG in GBM can sponge miR-185 to promote ANXA2 expression, and ANXA2 stimulates miR155HG level through phosphorylated STAT3 binding to the miR155HG promoter. We establish the miR155HG/miR185/ANXA2 loop as a mechanism that underlies the biological functions of miR155HG and ANXA2 in GBM and further suggest this loop may serve as a therapeutic target and/or prognostic biomarker for GBM. Keywords: GBM, miR155HG, ANXA2, P-STAT3, ceRNA * Correspondence: [email protected] Weining Wu, Tianfu Yu and Youzhi Wu these authors contributed equally to this work. Department of Neurosurgery, The First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210029, China Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 2 of 14 Background functions of miR155HG in glioma has not been Glioblastoma multiforme (GBM) is the most aggressive determined. form of astrocytoma and is characterized by rapid pro- Through bioinformatical analysis and experimental gression and poor prognosis [1]. Studies have indicated analyses, here we confirmed that miR155HG is overex- that GBM development is associated with disrupted gene pressed in GBM and acts as a ceRNA for the tumor sup- regulatory networks. In addition to aberrant gene ex- pressor miR-185 to upregulate ANXA2. We also showed pression in GBM, several reports have demonstrated a that ANXA2 promotes GBM growth and miR155HG ex- role for non-coding RNAs (ncRNAs), which lack protein pression by activating STAT3. Our results demonstrate coding capability due the lack of an open reading frame, that miR155HG, miR-185 and ANXA2 form a signaling in GBM progression [2]. Among these ncRNAs, long loop that promotes malignant phenotypes in GBM. ncRNAs (lncRNAs) are defined as RNAs > 200 nucleo- tides in length and have been demonstrated to partici- Methods pate in diverse cellular processes including epigenetics Public human astrocytoma databases, GBM specimens and transcriptional and posttranscriptional regulation and cell lines [3]. At the posttranscriptional level, lncRNAs can act as Three public human astrocytoma databases (TCGA, competing endogenous RNAs (ceRNAs) or molecular CGGA and Rembrandt) were described in our previous sponge that function by sponging microRNAs (miR- study [21]. A total of 24 GBM tissues and 15 NAs), short endogenous ncRNAs approximately 18–25 pair-matched adjacent normal brain edematous tissues nucleotides in length, resulting in inhibiting the effects (collected postoperatively between April 2016 and Feb- of miRNAs on target mRNAs [4]. ruary 2017) were collected from patients who underwent The lncRNA miR155HG is transcribed from a gene lo- surgical removal of GBM tumors at the First Affiliated cated on chromosome 21q21 and consists of three exons Hospital of Nanjing Medical University. Samples were that span 1.5 Kb. MiR155HG, also known as B-cell inte- frozen in liquid nitrogen immediately after isolation and gration cluster, includes an imperfectly base-paired stem stored at − 80 °C to avoid RNA deterioration. Tumor tis- loop in exon 3 that is conserved across species and is a sues were collected after participants signed written in- primary miRNA for miR-155 [5]. MiR155HG was ini- formed consent. The study protocol was approved by tially thought to be involved in the human immune re- The Institutional Review Board of the First Affiliated sponse. For example, Haasch et al. showed that Hospital of Nanjing Medical University. GBM diagnoses transcriptional activation of miR155HG is an early and were confirmed by two independent pathologists. Pa- sustained T cell activation event [6]. Several oncological tients recruited into this study received no preoperative studies have shown that miR155HG is highly expressed treatments. in diffuse large and primary mediastinal B cell lymph- The normal human astrocyte cell line NHA and the omas [7]. MiR155HG can be induced by abnormal B-cell human GBM cell lines U87, U251, Ln229, T98, and receptor in Hodgkin lymphoma [8], and miR-155 was A172 were purchased from the Chinese Academy of Sci- also shown to be upregulated by mutant p53 and facili- ences Cell Bank (Shanghai, China). A primary GBM cell tate invasion of breast cancer cells [9]. Our group previ- line GP1 was extracted in December 2016 from the ously reported that the miR155HG/miR-155 axis tumor of a patient with a temporal GBM; a second pri- exhibits an oncogenic function in glioma by promoting mary GBM cell line GP2 was extracted in January 2017 epithelial-to-mesenchymal transition [10]. However, from the tumor of a patient with a frontal GBM. All cell other biological mechanisms of miR155HG, such as po- lines were stored in liquid nitrogen before use. Cell cul- tential ceRNA functions, and the upstream regulation of ture was performed as described previously [22]. miR155HG in astrocytoma have remained unknown. Annexin A2 (ANXA2) is a calcium-dependent phos Quantitative real-time PCR (qRT-PCR) and western pholipid-binding protein with demonstrated roles in blotting stimulating fibrinolytic processes, degrading extracellular QRT-PCR and western blotting were performed as matrix, and promoting angiogenesis [11]. Recent reports described previously [22]. The primer for miR155HG is have shown that ANXA2 is aberrantly expressed in a F5′-GAGTGCTGAAGGCTTGCTGT-3′,R5′-TTGA wide variety of tumors [12–15] and has been implicated ACATCCCAGTGACCAG-3′, for β-actin is F 5′-TCAC in various processes of tumorigenesis, including cell in- CCACACTGTGCCCA-TCTACGA-3′,R5′- CAGCGG vasion [16], proliferation [17] and neovascularization AACCGCTCATTGCCAATGG-3′. The antibodies for [18]. Previous studies reveal invasion function for western blot analysis were: anti-ANXA2 (1:1000; Abcam, ANXA2 in glioma and suggest its role as a potential Cambridge, UK), anti-cell cycle-related proteins (cyclin diagnostic and prognostic marker for glioma [19, 20]. E, cyclin D, CDK4, CDK6) (1:1000; Cell Signaling However, whether ANXA2 is involved in the oncogenic Technology, Danvers, MA, USA), anti-Bax (1:500; Santa Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 3 of 14 Cruz Biotechnology, Dallas, TX, USA), anti-Bcl-2 (1:500; rabbit anti-IgG antibody (Abcam), and rabbit anti-RNA Santa Cruz Biotechnology), anti-caspase 3 (1:1000; Abcam) Polymerase II antibody (Abcam) was used for positive and anti-β-actin (1:1000, Cell Signaling Technology). Cells control. After immunoprecipitation, the beads were were treated with EGF Recombinant Human Protein Solu- washed sequentially with low-salt buffer, high-salt buffer, tion (Thermo Fisher Scientific, Waltham, MA, USA) or the LiCl buffer, and TE buffer each for 5 min at 4 °C. The SH-4-54 inhibitor of STAT3 phosphorylation (Selleck, immunoprecipitated DNA was then eluted by incubation Shanghai, China) according to the manufacturer’sprotocol. in 100 μl of elution buffer (0.1 M NaHCO and 1% SDS) containing 10 μg proteinase K (Millipore) at 62 °C for 2 Oligonucleotides, plasmid and transfection h with rotation. The eluted DNA was purified using the To construct a plasmid expressing miR155HG, the columns and buffers in the kit and then re-dissolved in full-length human miR155HG sequence or a mutated 50 μl of PCR-grade water. Immunoprecipitated chroma- miR155HG sequence for miR-185-5p (NCBI Reference tin was analyzed by qPCR using primers targeting the Sequence: NR_001458.3) was synthesized and inserted into phosphorylated (p)-STAT3 (p-STAT3) binding regions in the pCDNA3.1 vector to generate pCDNA3.1-miR155HG the human miR155HG promoter. The primer sequences WT or MUT, respectively (Genechem, Shanghai, China). used for ChIP-qPCR are binding region 1 (− 1982 to − The hsa-miR-185-5p mimic, hsa-miR-185-5p inhibitor, and 1972): F 5′-GAGACATCATTATTGTCATT-3′,R 5′-TA hsa-miR-scramble were chemically synthesized (Ribobio, GGAGTCAAATACACCTG-3′; binding region 2 (− Guangzhou, China). The sequence of ANXA2 and 1548 to − 1411): F 5′-ATGGGAAATTCAGAAAGGC miR155HG siRNAs were as follows: si-miR155HG: sense, -3′,R5′-TGATCATATGAGGGAGGAGC-3′; and bind- 5′-CUGGGAUGUU-CAACCUUAATT-3′,antisense,5′-U ing region 3 (− 275 to − 116): F 5′-TTAAGAACA UAAGGUUGAACAUCCCAGTT-3′; scramble: sense, 5′-U AAGGTTGGAGC-3′,R5′-TGTGACTCATAACCGAC UCUCCGAACGUGUCACGUTT-3′,antisense,5′-ACGU CAG-3′. PCR conditions were set according to the in- GACACGUU-CGGAGAATT-3′.ANXA2 siRNAs were structions provided in the SYBR Green Kit (Roche Ap- chemically synthesized (Invitrogen, Shanghai, China). plied Science, Upper Bavaria, Germany). Results were ANXA2 siRNA 1 target sequence is 5′-CUGGGAAGAAGG analyzed by agarose gel electrophoresis. CUUC-CUUTT-3′, ANXA2 siRNA 2 target sequence is 5′-AAGGAAGCCUUCUUCCC-AGTT-3′,ANXA2 siRNA RNA immunoprecipitation (RIP) 3target sequence is5′- GGGAAGAAGGCUUCCU-UCAT RIP assays were performed using U87 cell extracts with the T-3′. Cells were transfected with oligonucleotides or plasmid EZ-Magna RIP RNA-Binding Protein Immunoprecipitation using Lipofectamine 2000 Reagent (Invitrogen, Carlsbad, Kit (Millipore, Burlington, MA, USA) according to the man- CA, USA) following the manufacturer’s instructions. ufacturer’s instructions. U87 cells were rinsed with cold PBS and fixed with 1% formaldehyde for 10 min. After centrifuga- Lentiviral packaging and stable cell lines tion, cell pellets were collected and resuspended in NP-40 Lentiviruses carrying shRNA-miR155HG or shRNA- lysis buffer (Thermo Fisher Scientific, Waltham, MA, USA) ANXA2 and the negative control lentivirus (sh-miR supplemented with 1 mM PMSF, 1 mM DTT, 1% Protease 155HG sequence is 5′-CUGGGAUGUUCAACCUUAA Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO, USA) and TT-3′; sh-ANXA2 sequence is 5′- CGGGATGCTTTG 200 U/ml RNase Inhibitor (Life Technologies, Carlsbad, CA, AACATTGAA -3′; sh-NC sequence is 5′-UUCUCCG USA). Lysates were subjected to high-speed centrifugation, AACGUGUCACGUTT-3′) were assembled in the hu- and then 100 μl of the supernatant was incubated with RIP man embryonic kidney cell line 293 T, and the viruses buffer containing magnetic beads conjugated with human were collected according to the manufacturer’s manual anti-Ago2 antibody (Cell Signaling Technology). Mouse IgG (Genechem). Stably transfected cell lines were estab- (Cell Signaling Technology) was used as a negative control, lished by infecting U87 cells with lentiviruses using a while SNRNP70 (Cell Signaling Technology) was used as a lentiviral packaging kit purchased from Genechem, positive control. Co-precipitated RNAs were detected by re- followed by puromycin selection. verse transcription PCR. Total RNAs (input control), IgG and SNRNP70 were assayed simultaneously to evaluate the Chromatin immunoprecipitation (ChIP) efficiency of Ago2-specific RNAs. ChIP assays were performed as previously described [22]. The EZ-magna ChIP kit (Millipore, Bedford, MA, Colony formation assays USA) was used according to the manufacturer’s proto- Cells (4 × 10 ) were seeded into cell culture dishes and col. Crosslinked chromatin was sonicated into DNA cultured for 15 d. Cell colonies were fixed with 4% para- fragments in the range of 200–1000 bp and immunopre- formaldehyde for 20 min and stained with 0.2% crystal cipitated using rabbit anti-p-STAT3 antibodies (Abcam). violet. Images were captured and colonies (diameter > Negative control samples were prepared using control 0.5 mm) were counted using Image J software (National Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 4 of 14 Institutes of Health, Arlington, VA, USA). All assays times. Finally, tissue sections were mounted with a were repeated at least three times. medium containing DAPI for 8 min in the dark and im- ages were obtained with fluorescence microscope EDU cell proliferation assays (Nikon, Japan). EDU cell proliferation assays were conducted with the Molecular Probes EdU-Alexa imaging detection kit (Life Dual-luciferase reporter assay Technologies). Cells treated for 48 h were incubated The ANXA2 3′-untranslated region (UTR) and the full with 10 μM EdU for 2 h, fixed with 4% paraformalde- miR155HG sequence containing miR-185-5p seed hyde, permeabilized with 1% Triton X-100, and stained matching sites were amplified from human cDNA via with the Alexa-Fluor 594 reaction cocktail for EdU and PCR and cloned into the 3′ end of the pGL3-basic lucif- Hoechst 33342 (nuclei). Images were obtained using a erase vector (Genechem). Mutated versions of each con- fluorescence microscope (Olympus, Japan). All assays struct were generated by mutating the miR-185-5p seed were repeated at least three times. site sequences (pGL3-wt or mut). The miR155HG pro- moter region sequence (2000 bp to 1000 bp upstream of Flow cytometric analysis transcription starting point) were also amplified and Transfected GBM cells in logarithmic growth were col- cloned into the 5′ end of the pGL3-basic luciferase vec- lected and processed with the Cell Cycle Staining Kit tor. Mutated version was generated by deleting binding (Multi Sciences, Hangzhou, China) for cell cycle analysis. region sequences of p-STAT3 (wt or mut-pGL3). U87 After washing with PBS, cells were fixed with 70% cells seeded into 96-well plates were co-transfected with ice-cold ethanol, incubated with Cell Cycle Staining Kit wt or mut report gene, the pRL-TK control (Promega, for 30 min in the dark, and analyzed by flow cytometry. Madison, WI, USA) and miR-185-5p mimic or miRNA In other experiments, treated cells in logarithmic growth NC using Lipofectamine 2000 (Invitrogen). The wt or were harvested and stained with the Annexin V-FITC mut-pGL3 and the pRL-TK control were co-transfected Apoptosis Detection Kit (Multi Sciences). After washing into the cells, then treated with cell culture with or with- with PBS and incubating with Annexin V/propidium out SH-4-54 inhibitor of STAT3 phosphorylation. At 48 iodide for 30 min in the dark, cells were analyzed by flow h after transfection, luciferase activity was determined cytometry. using the Dual Luciferase Reporter Assay System (Pro- mega, WI, USA) according to the manufacturer’s proto- Immunohistochemistry (IHC) col. The relative luciferase activity was normalized to Fresh intracranial tumor tissues from nude mice were Renilla luciferase activity. All assays were performed in fixed with 4% paraformaldehyde and then embedded in triplicate. paraffin. Sections were incubated at 4 °C overnight with primary antibodies against ANXA2 (1:1000; Abcam) and p-STAT3 (1:500; Abcam). Sections were then incubated Intracranial GBM mouse model with secondary antibody (1:1000; Santa Cruz) for 2 h at The animal experiments were approved by the Animal room temperature and stained with diaminobenzidine Management Rule of the Chinese Ministry of Health until brown granules appeared. (document 55, 2001) and were performed conforming to the approved guidelines and experimental protocols of Fluorescence in situ hybridization (FISH) Nanjing Medical University. U87 cells (1 × 10 ) stably RNA FISH was performed as described previously [22]. expressing MCS-firefly luciferase for bioluminescence MiR155HG-Bio probe was synthesized from GoodBio imaging were transfected with lentivirus expressing con- (Wuhan, China); the sequence is 5′-CCTCCCACG trol shRNA, shRNA-ANXA2 or shRNA-miR155HG and GCAGCAATTTGTTCCA-3′. Frozen sections of fresh then were intracranially injected into the frontal lobe of tissues were fixed with 4% formaldehyde for 10 min, nude mice to generate GBM (n = 10 mice per group). washed with PBS, and then digested with Proteinase K Tumor volumes were measured by luciferase using a for 5 min. After eliminating auto-fluorescence and block- bioluminescence imaging system (Caliper IVIS ing endogenous biotin, the sections were hybridized with Spectrum, PerkinElmer, Waltham, MA, USA) on days 1, probes overnight. Sections were then washed with 11, and 21 after implantation. The integrated flux of pre-warmed 2× SSC at 37 °C for 10 min, 1 × SSC at 37 °C photons (photons/s) within each region was determined for 10 min, and 0.5 × SSC for 10 min. Tissue sections by the Living Images software package (Caliper Life were then blocked with bovine serum albumin for 30 Sciences). Mice were sacrificed when they were in deep min at room temperature, followed by staining with coma. Brains were extracted, fixed in 10% formalin and 488-avidin (1:400) at room temperature for 50 min. then embedded in paraffin for IHC or frozen at − 80 °C Stained sections were washed with PBS for 5 min four for western blotting or FISH. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 5 of 14 Statistical analysis Together these results suggested that miR155HG may Data are presented as the mean ± standard deviation be involved in the malignant phenotypes of GBM. (SD). Statistical analyses were performed using the To evaluate the role of miR155HG in promoting malig- Student t test to evaluate the significance of differences nant phenotypes of GBM, we generated an intracranial between groups, one-way ANOVA (Tukey’s post hoc) GBM mouse model by injecting U87 cells infected with was used to determine the difference among at least lentivirus expressing control shRNA or shRNA-miR155HG. three groups using SPSS v19.0 for Windows. (SPSS, IL, The tumor volumes in the group of mice treated with USA). Pearson’s correlations analysis and heat map shRNA-miR155HG were smaller than those in the control microarray analysis were performed using Multiple at 11 and 21 days after implantation. Furthermore, mice Array Viewer 4.9 software (MEV). Kaplan–Meier sur- treated with shRNA-miR155HG showed better survival than vival analysis was performed using GraphPad 5.0 soft- controls (Additional file 1:FigureS1A andB). Theseresults ware. P < 0.05 indicates a significant difference. indicated that miR155HG acted as oncogene in promoting GBM growth. Results MiR155HG is overexpressed in GBM and miR155HG- MiR155HG sponges and downregulates miR-185-5p related genes are enriched in cancer-associated processes Previous studies have shown that lncRNAs can act as a To explore miR155HG expression in human astrocy- sponge for miRNAs, therefore we speculated that toma tissues, we examined three public human astrocy- miR155HG may exhibit sponge activity. RNA hybrid toma databases (TCGA, CGGA and Rembrandt) and bioinformatics tools showed that miR155HG contains a found overexpression of miR155HG in GBM (Fig. 1a). putative binding site for miR-185-5p as tumor suppres- We also found that miR155HG expression was elevated sor in a wide range of tumors [23–26] (Fig. 2a). We in 24 GBM specimens compared with adjacent normal found that miR-185-5p levels were lower in GBM tissues brain tissue from patients histologically diagnosed with than in normal brain tissue, and that miR-185-5p was GBM, and miR155HG was generally overexpressed in significantly negatively correlated with miR155HG in the the GBM cells compared with NHA cells from normal same GBM samples (Fig. 2b; r = − 0.5970, p = 0.0021). brain tissue using fluorescent qPCR (Fig. 1b). FISH ana- RNA-induced silencing complex (RISC) is an essential lysis revealed that miR155HG was primarily concen- factor in the biological effect of miRNAs, and Ago2 is an trated in the cytoplasm of GBM cells (Fig. 1c). elementary catalytic constituent of RISC that is involved MEV software was used to perform a Pearson’s correl- in RNA cleavage [27]. To explore the possible interactions ation analysis to identify genes associated with between miR155HG and miR-185-5p, RIP was performed miR155HG expression in the TCGA, CGGA and Rem- in U87 cells. SNRNP70 protein, which interacts with U1 brandt GBM databases. The results identified 1037 spliceosomal RNA [28], was used as a positive control. up-regulated genes and 873 down-regulated genes from MiR155HG was predominantly enriched in beads contain- the overlap of the three databases, and these genes were ing anti-Ago2 antibody compared with those harboring examined using the DAVID Web tool (http://david.abcc. control IgG (Fig. 2c). This result suggests that miR155HG ncifcrf.gov/home.jsp) for Gene Oncology (GO) and is capable of sponging miRNA as ceRNA. KEGG Pathway enrichment analyses. GO analysis Then we constructed a luciferase reporter plasmid con- showed that the upregulated genes were primarily taining the putative miR-185-5p binding site from enriched in tumor progression processes, such as regu- miR155HG, as well as a mutant construct in which the lating apoptosis and cell proliferation. Moreover, KEGG binding site was mutated (Fig. 2a). Co-transfecting pathway analysis indicated that the upregulated genes miR-185-5p mimic decreased the relative luciferase activ- were closely associated with pathways activated in can- ity in U87 cells transfected with wild-type plasmid, but cers, such as the PI3K-Akt signaling pathway, a promin- had no impact on the mutant construct (Fig. 2d), which ent pathway in tumorigenesis and cancer progression suggests that miR155HG directly binds miR-185-5p. (Fig. 1d-e). Furthermore, gene set enrichment analysis (GSEA) (http://www.broadinstitute.org/gsea/index.jsp) ANXA2 is the target molecule of miR-185-5p was used to examine genes that were expressed in Previous studies demonstrated a potential role for TCGA GBM samples from patients with high ANXA2 in glioma [19]. Bioinformatics analytical tools miR155HG expression and those with low miR155HG TargetScan and miRNAWalk 2.0 showed that the expression. Genes expressed in patients with high 3′-UTR of ANXA2 mRNA contained a seed sequence of miR155HG expression were primarily associated with miR-185-5p (Fig. 2e). To determine whether ANXA2 reduced apoptosis and cell death, while genes with nega- may be involved in the miR155HG-miR-185-5p axis in tive expression were primarily associated with cell cycle GBM, we first examined the expression levels of ANXA2 progression and cell cycle phase transition (Fig. 1f-g). in frozen GBM tissue samples by western blot. We Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 6 of 14 Fig. 1 MiR155HG expression correlates positively with malignant degrees of glioma, and the miR155HG-associated genes were chiefly enriched in cancer related function. a Expression of miR155HG positively correlates with WHO grade in TCGA, CGGA and Rembrandt Public database. b Relative levels of miR155HG in the tumors and their adjacent normal brain tissues from 24 patients diagnosed as glioblastomas, and relative expressions of miR155HG in glioma cell lines NHA, A172, U87, U251, Ln229, T98, and primary glioma cells from two patients. c The distribution of miR155HG was evaluated via FISH in U87 and GP1 cells. d-e MiR155HG associated genes from overlapping CGGA, TCGA and Rembrandt databases were analyzed with KEEG pathway analysis, gene oncology analysis. f-g MiR155HG associated genes were analyzed with gene set enrichment analysis (GSEA) by TCGA genes data Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 7 of 14 Fig. 2 MiR155HG stimulated the ANXA2 expression by sponging endogenous miR-185-5p. a The binding site of MiR155HG and miR-185-5p was predicted by bioinformatics tools. b Expression levels of miR-185-5p in GBM tissues and adjacent normal brain tissues were analyzed by real-time PCR and normalized to U6. The correlation between MiR155HG and miR-185-5p in GBM tissues was applied with Pearson’s correlation coefficient (r = − 0.5970, P = 0.0021). c Luciferase assays was performed after transfection with miR-185-5p mimic and pGL3- miR155HG-wt or pGL3- miR155HG-mut into U87 cells as well as the internal control Renilla plasmid. Relative luciferase activity was analyzed after 48 h treatment. (*p < 0.05, **p < 0.01). d Amount of miR155HG bound to Ago2 or IgG measured by RT–qPCR after RIP assays with anti- IgG, anti-Ago2, anti-SNRNP70 and 10% input. e The binding site in 3’UTR of ANXA2 mRNA for miR-185-5p was predicted by bioinformatics tools. f Expression levels of ANXA2 in GBM tissues and adjacent normal brain tissues were analyzed by western blot and normalized to β- actin. The correlation between miR-185-5p and ANXA2 in GBM tissues was applied with Pearson’s correlation coefficient (r = − 0.4676, P = 0.0212). g Luciferase assays was performed after transfection with miR-185-5p mimic and pGL3- ANXA2-wt or pGL3- ANXA2-mut into U87 and GP1 cells as well as the internal control Renilla plasmid. Relative luciferase activity was analyzed after 48 h treatment. (*p < 0.05, **p < 0.01). h The protein expression levels of ANXA2 were analyzed by Western blotting after 48 h transfection in U87 cells with pcDNA3.1, pcDNA3.1-miR155HG wt, pcDNA3.1-miR155HG mut, miR-NC or different amount of miR-185-5p mimic, respectively found that ANXA2 was highly expressed in GBM tis- 0. − 4676, P = 0.0212; Fig. 2f). We generated luciferase sue but not in normal brain tissue (Fig. 2f and Additional constructs containing either the wild-type (WT) 3′-UTR file 1: Figure S1C). We next examined the correlation be- of ANXA2 mRNA or a mutated (MUT) sequence in which tween miR-185-5p and ANXA2 in GBM tissue and found the miR-185-5p seed sequences were mutated. Luciferase that miR-185-5p negatively correlated with ANXA2 (r = assays showed that expression of miR-185-5p decreased Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 8 of 14 the luciferase activity of the WT reporter but not the ac- revealed that tumor growth was significantly inhibited at tivity of the MUT reporter in U87 and GP1 cells (Fig. 2g). 21 d in the group with decreased ANXA2 expression We speculated that ANXA2 levels in GBM cells may (Fig. 3f). Mice injected with cells expressing sh-ANXA2 be regulated by miR-185-5p and affected by its inter- also showed better survival than control mice (Fig. 3g). action with miR155HG. Indeed, transfection of a IHC of tumor sections from the sh-ANXA2 and control miR155HG expression vector increased ANXA2 levels groups confirmed decreased ANXA2 levels lead to in- in U87 and GP1 cells; however, the vector expressing hibition of tumor growth in vivo (Fig. 3h). Taken to- miR155HG with mutated binding sites for miR-185-5p gether, our in vitro and in vivo results demonstrate that had no effect on ANXA2 levels. In addition, ANXA2 exhibits oncogenic functions in GBM cells and miR155HG-mediated elevation of ANXA2 was blocked enhances the malignant phenotypes of tumors. by co-transfection with miR-185-5p mimic in a dose-dependent manner (Fig. 2h). Furthermore, inhibit- MiR155HG and miR-185-5p participate in GBM growth by ing miR155HG by siRNA downregulated ANXA2 levels regulating the proliferation and apoptosis of GBM cells in U87 and GP1 cells, which could be reversed by treat- As ANXA2 plays a crucial tumor-promoting role in ment with miR-185-5p inhibitor (Additional file 1: Fig- GBM and since ANXA2 levels are modulated by ure S1D). Together these results demonstrated that miR155HG and miR-185-5p, we hypothesized that miR155HG may promote ANXA2 expression by modu- miR155HG and miR-185-5p could interfere with the lating the capacity of miR-185-5p to bind the 3′-UTR of proliferation and apoptosis of GBM cells. Downregulat- ANXA2 mRNA. ing miR155HG levels in U87 and GP1 cells by siRNA re- duced proliferation, blocked cell cycle progression, and ANXA2 enhances the malignant phenotypes of GBM cells stimulated apoptosis in GBM cells, and these effects As ANXA2 was the downstream molecule positively were reversed by miR-185-5p inhibitor (Fig. 4a–d). modulated by miR155HG via the ceRNA mechanism, we Western blot analysis showed altered levels of proteins needed to investigate the function of ANXA2 to explain associated with proliferation and apoptosis, which was the oncological role of miR155HG in GBM. Bioinformat- consistent with the above results (Fig. 4e). These results ics analysis showed that ANXA2 expression was mostly indicate that the biological role of miR155HG and expressed in GBM samples from TCGA, CGGA and miR-185-5p in GBM cells was due to regulating proteins Rembrandt database (Additional file 2: Figure S2A). GO associated with proliferation and apoptosis, and this may analysis showed ANXA2 was closely associated with be partially by regulating ANXA2 expression. genes involved in cell apoptosis and proliferation (Add- itional file 2: Figure S2B). ANXA2 affected miR155HG expression in GBM cells via p- We next performed a series of experiments to evaluate STAT3 levels the possible oncogenic function of ANXA2 in GBM. Our results show that miR155HG can interfere with Colony formation and EDU assays were performed in ANXA2 expression by sponging miR-185-5p in GBM cells transfected with ANXA2 siRNA to evaluate the ef- cells. Previous studies show that ANXA2 can act with fect of ANXA2 on proliferation. The effect of ANXA2 AKT, STAT3 to promote downstream oncogenes [29, siRNA was shown in Additional file 3: Figure S3B. A signifi- 30], we speculate if ANXA2 can contribute to cant reduction of proliferation was observed in U87 and miR155HG aberrantly overexpression in GBM cells in GP1 cells transfected with ANXA2 siRNA compared with this way. Pearson’s correlation analysis showed that controls (Fig. 3a–b). Flow cytometry revealed that knocking ANXA2 was significantly positively correlated with down ANXA2 induced a G1/S arrest and decreased the miR155HG in the TCGA, CGGA and Rembrandt GBM percentage of cells in S phase (Fig. 3c). Flow cytometry also databases (Fig. 5a). We also found a positive correlation showed that knocking down ANXA2 remarkably increased between ANXA2 and miR155HG in WHO II and WHO apoptosis rates in GBM cells (Fig. 3d). Western blot results III astrocytoma patients in these independent public da- of proliferation- and apoptosis-associated proteins were tabases (Additional file 2: Figure S2C). Next, one repre- consistent with the above results (Fig. 3e). sentative GBM tissue exacted from a GBM patient To examine the oncogenic ability of ANXA2 in GBM during surgery was examined by FISH double staining in vivo, nude mouse tumorigenicity assays were per- and the results showed that ANXA2 and miR155HG formed using a U87 xenograft model. U87 cells infected were both significantly expressed in dense tumor tissues, with fluorescent lentiviruses expressing sh-ANXA2 or but not in loose normal brain tissue (Fig. 5b). Then controls were injected into nude mouse brains. The ef- double-staining was performed to investigate to show fect of ANXA2 knockdown in cells and GBM tissue was that inhibiting ANXA2 caused a miR155HG downregula- shown in Additional file 1: Figure S1E. In vivo imaging tion in the brain of nude mice from Fig. 3g and interfered of the nude mice at 1, 11 and 21 d after implantation with tumor growth (Fig. 5c). We thus concluded that Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 9 of 14 Fig. 3 Downregulating ANXA2 expression inhibited GBM cells proliferation, cell cycle and apoptosis in vitro and vivo. a Colony formation assays in U87 and Primary glioblastoma cells (GP1) transfected with scramble, ANXA2 siRNA 1 and siRNA 2. Scale bar> 500 μm. b U87 and GP1 in EDU transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. Representative merged images were shown (original magnification, 200×). c The cell cycle was detected in U87 and GP1 transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. d The apoptotic cells were measured with flow cytometry in U87 and GP1 transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. e The protein related with cell proliferation, cell cycle and apoptosis was measured with immunoblotting in U87 and GP1 cells transfected with scramble, ANXA2 siRNA 1 and siRNA 2 after 48 h. All experiments above were performed 3 times, and average scores are indicated with error bars on the histogram.*P < 0.05, **P < 0.01. f U87 cells transfected with a lentivirus with sh-ANXA2 or sh-NC and a lentivirus containing luciferase were implanted in the brain of 10 nude mice, respectively. The tumor formation was assessed by bioluminescence imaging. The bioluminescent images were measured at days 1, 11 and 21 after implantation. g Overall survival was determined by Kaplan-Meier survival curves between sh-ANXA2 or sh-NC group, and a log- rank test was used to assess the statistical differences. h Two representative immunohistochemical image of tumors from groups of nude mice implanted with U87 cells, transfected with a lentivirus with sh-ANXA2 or sh-NC, were shown to compare the volume size of tumors ANXA2 expression positively correlated with miR155HG resulted in reduced levels of p-STAT3 without impacting levels in GBM. STAT3levelsinGBM cells(Additional file 3:FigureS3B). Previous studies showed that ANXA2 modulates STAT3 Inhibiting miR155HG also resulted in decreased ANXA2 phosphorylation (p-STAT3) levels to stimulate proliferation, and p-STAT3 in vivo (Additional file 3:FigureS3E). Knock- angiogenesis, metastasis and epithelial to mesenchymal ing down ANXA2 resulted in reduced miR155HG expres- transition in breast cancer cells [30–32]. We found higher sion, but this effect was not observed in GBM cells treated levels of p-STAT3 in GBM cells and tissues compared with with a STAT3 phosphorylation inhibitor (Fig. 5e). Moreover, normal brain cell and tissue(Additional file 3:FigureS3A), induction of STAT3 phosphorylation by Epidermal Growth with distribution in nuclei in GBM cells (Fig. 5d). We thus Factor (EGF) [33, 34] could rescue the inhibitory effect of wondered if ANXA2-mediated induction of miR155HG in- ANXA2 depletion on miR155HG expression (Fig. 5fand volved p-STAT3. We first found that decreasing ANXA2 Additional file 3: Figure S3C). These results suggested that Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 10 of 14 Fig. 4 The impact of miR155HG and miR-185-5p on cell proliferation, cell cycle and apoptosis in GBM cells. a Colony formation assays in U87 and GP1 cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor. Scale bar> 500 μm. b U87 and Primary glioblastoma cells in EDU transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor after 48 h. Representative merged images were shown (original magnification, 200×). c The cell cycle was detected in U87 and Primary glioblastoma cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor after 48 h. d The apoptotic cells were measured with flow cytometry in U87 and Primary glioblastoma cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR-185-5P inhibitor after 48 h. e The protein related with cell proliferation, cell cycle and apoptosis was measured with immunoblotting in U87 and Primary glioblastoma cells transfected with scramble, si-miR155HG, miR-185-5P inhibitor and si- miR155HG + miR- 185-5P inhibitor after 48 h. All experiments were performed 3 times, and average scores are indicated with error bars on the histogram.*P < 0.05, **P < 0.01 p-STAT3 might directly promote miR155HG expression in Together these results indicated that the activated GBM. transcription factor p-STAT3 plays a key role in Bioinformatics tools identified three putative binding re- ANXA2-driven miR155HG expression and promotes gions for p-STAT3 in the miR155HG promoter. ChIP assays GBM cell growth, through the DNA binding activity of confirmed that p-STAT3 could bind putative binding region the p-STAT3 transcription factor. 2(− 1548 bp to − 1411 bp) but not binding region 1 (− 275 bp to − 161 bp) and binding region 3 (− 1982 bp to − 1972 Discussion bp) in the miR155HG promoter (Fig. 5g). Two luciferase We previously revealed that miR-155, which is derived reporter plasmids containing either the entire miR155HG from miR155HG, functions in epithelial-mesenchymal promoter region 2000 bp to 1000 bp upstream of the tran- transition in glioma [10]. In this report, we further scription start site (wt-pGL3) or the promoter deleted for showed that the lncRNA miR155HG is highly expressed binding region 2 (mut-pGL3) were transfected into GBM in GBM, where it acts as a ceRNA to sponge miR- cells along with or without STAT3 phosphorylation inhibitor. 185-5p, thus promoting miR155HG downstream mole- While inhibiting STAT3 phosphorylation resulted in down- cules level such as ANXA2. Our study clarifies the regulation of luciferase activity driven by the wild-type pro- mechanism by which miR155HG positively regulates moter, the p-STAT3-driven luciferase activity on the mutated AXNA2 to sustain the malignant phenotypes of astrocy- reporter remained unchanged (Additional file 3:Figure S3D). toma, particularly GBM. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 11 of 14 Fig. 5 ANXA2 promote miR155HG expression via P-STAT3 in GBM cells a ANXA2 positively correlates with miR155HG in GBM specimens of three independent public database. b The level of ANXA2 and miR155HG was determined by immunofluorescence in boundary between invasive GBM and normal brain tissue from a GBM patient. c Two representative image of tumors from groups of nude mice implanted with U87 cells transfected with a lentivirus of sh-ANXA2 or sh-NC, were shown to compare the volume size of tumors and the expression of miR155HG via double immunofluorescent staining of the same frozen slice of tumor. d The specific localization of p-STAT3 and ANXA2 in U87 and GP1 cells was detected by immunofluorescence. e Q-PCR was performed to explore the level of miR155HG after downregulating ANXA2 in U87 and GP1 cells treated or untreated with inhibitor of STAT3 phosphorylation SH-4-54. f EGF-induced p-STAT3 rescued the ANXA2 knockdown-mediated downregulation of miR155HG in U87 and GP1 cells. g CHIP assays was performed to confirm the three putative binding region for p-STAT3 in promoter region of miR155HG Through bioinformatics analyses, we found that glioma associated with inhibiting glioma cell invasion [40]. miR155HG was closely associated with the proliferative Another report showed that the lncRNA Linc00176 regu- activity and apoptosis resistance of GBM in three inde- lates the cell cycle by sponging miR-185 in hepatocellular pendent GBM gene expression arrays. Knocking down carcinoma [41]. Here, we confirmed that lncRNA miR155HG in GBM cells resulted in cell cycle arrest, de- miR155HG binds to miR-185 to impact proliferation, cell creased cell growth, and apoptosis. As lncRNAs have cycle progression and apoptosis in GBM cell lines. been proven to function as miRNA sponges [35], we hy- ANXA2, a 36-kDa protein that belongs to the family pothesized that miR155HG might also regulate gene ex- of calcium-dependent phospholipid binding proteins pression by competing for shared miRNA response [11], is a DNA-binding protein that modulates DNA elements in GBM. Several studies have shown that synthesis. Several studies showed that ANXA is involved miR-185 is involved in suppressing non-small cell lung in cell proliferation and cell cycle progression [42–45]in cancer [36], gastric cancer [37], hepatocellular carcinoma a variety of cancer cell types, such as breast cancer [18], [38] and prostate cancer [39] and is downregulated in hepatocellular carcinoma [46], colorectal cancer [47] and Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 12 of 14 Fig. 6 Schematic diagram of the relationship among miR155HG, miR-185-5p, p-STAT3 and ANXA2 in GBM. The binding of miR-185-5p to miR155HG leads to abolish miR-185-5p targeting ANXA2 3’UTR, which stimulated ANXA2-elevated level in cell. In turn ANXA2 then activates STAT3 phosphorylation, and STAT3 phosphorylation shifts from the cytosol to the nucleus, forms dimers and binds with promoter region to upregulate the expression of miR155HG pancreatic cancer [48]. We found that ANXA2 was over- that p-STAT3 could bind to the miR155HG promoter expressed in GBM. Based on TCGA, CGGA and Rem- region from − 1548 bp to − 1411 bp upstream of the brandt GBM gene expression profiles, we found that transcription start site to stimulate miR155HG expres- ANXA2-associated genes were primarily enriched in cell sion. This suggests that STAT3 phosphorylation is the proliferation and apoptosis. Knocking down ANXA2 critical role in driving lncRNAs regulation similar to inhibited proliferation and increased the G1/S cell cycle miR155HG in malignant brain tumor. arrest apoptosis of U87 and GP1 cells. We also showed Here we have established miR155HG/miR-185-5p/ that ANXA2 is a key determining factor of survival that ANXA2 loop in GBM formation and progression. How- promotes the growth of intracranial GBM tumors in ever, other interactions as epigenetic regulation between nude mice. We further found ANXA2 is a direct target miR155HG and its downstream effector molecules re- of miR-185-5p, and its expression was perturbed by mains obscure since lncRNA might recruit chromatin re- miR-185-5p. Taken together, these results established modeling complex. More biological studies and clinical the miR155HG/miR-185-5p/ANXA2 axis, which under- trials are needed to evaluate the practicality of targeting lies the biological mechanisms of miR155HG in GBM. miR155HG for the treatment of GBM. Silencing ANXA2 was previously reported to inhibit activation of STAT3 (p-STAT3) [31, 32, 49], and we con- Conclusions firmed that ANXA2 knockdown decreased p-STAT3 in Our study suggested that the lncRNA miR155HG in- GBM cells. Activated p-STAT3 forms homologous di- creases ANXA2 expression by sponging miR-185-5p to mers and enters into the nucleus to function as a tran- exert tumorigenic effects and that ANXA2 stimulates scription factor to promote target gene expression [50]. miR155HG level via ANXA2-driven p-STAT3 in GBM Since constitutively activated STAT3 is closely associated (Fig. 6). Thus, we have identified the miR155HG/ with GBM [51, 52], we find that ANXA2-mediated ele- miR-185-5p/ANXA2 loop and its mechanisms and bio- vated miR155HG levels is mainly due to p-STAT3 level. logical effects in malignant brain tumors. This loop ChIP assay and luciferase reporter gene assay showed could serve as a novel therapeutic biomarker for GBM. Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 13 of 14 Additional files manuscript; and WWN, ZJX and WYY designed this study. All authors have read and approved the final manuscript. Additional file 1: Figure S1. (A) U87 cells pretreated with a lentivirus Ethics approval and consent to participate with sh-miR155HG or sh-NC and a lentivirus containing luciferase were The procedures of this study were approved by the Institutional Review implanted in the brains of nude mice, and tumor formation was assessed Board of Nanjing Medical University. by bioluminescence imaging. The bioluminescent images were measured at days 1, 11 and 21 after implantation. (B) Overall survival from two Consent for publication groups of nude mice injected with U87 cells, transfected with sh-NC or Not applicable. sh-miR155HG lentivirus, was determined by Kaplan-Meier survival curves, and a log-rank test was used to assess the statistical significance of the Competing interests differences. (C) Expression levels of ANXA2 in GBM tissues and adjacent The authors declare that they have no competing interests. normal brain tissues were analyzed by western blotting and normalized to β-catenin. (D) The protein expression levels of ANXA2 were analyzed by western blotting after 48 h transfection in U87 and GP1 cells with Publisher’sNote scramble, miR155HG siRNA 1, miR155HG siRNA 2, miR-NC or inhibitor, Springer Nature remains neutral with regard to jurisdictional claims in respectively. (E) The effect of sh-ANXA2 in U87 cell and tumor tissue of published maps and institutional affiliations. nude mice after implantation were analyzed by western blotting. (E) Downregulating ANXA2 contributed to the reduction of p-STAT3 level in Author details GBM cells. (TIF 9482 kb) Department of Neurosurgery, The First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210029, China. Department of Neurosurgery, Additional file 2: Figure S2. (A) Expression of ANXA2 in TCGA, CGGA Nanjing First Hospital of Nanjing Medical University, Nanjing, Jiangsu, China. and Rembrandt astrocytoma database. (B) ANXA2 associated genes from overlapping CGGA, TCGA and Rembrandt databases were analyzed with Received: 14 November 2018 Accepted: 8 March 2019 gene oncology analysis. (C) ANXA2 positively correlates with miR155HG in WHOII/III astrocytoma specimens of three independent public database. (TIF 3895 kb) References Additional file 3: Figure S3. (A) Expression levels of p-STAT3 in cell lines, 1. Abou-Antoun TJ, Hale JS, Lathia JD, Dombrowski SM. Brain Cancer Stem GBM tissues and normal brain tissues were analyzed by western blot. (B) Cells in Adults and Children: Cell Biology and Therapeutic Downregulating ANXA2 contributed to the reduction of p-STAT3 level in GBM ImplicationsNeurotherapeutics : the journal of the American Society for cells. (C) Overexpression STAT3 was constitutively activated by EGF in ANXA2- Experimental NeuroTherapeutics. 2017. depleted GBM cells in U87 and GP1 cells. (D) Luciferase assays was performed 2. Patel AP, Tirosh I, Trombetta JJ, Shalek AK, Gillespie SM, Wakimoto H, et al. after transfection with miR155HG promoter wt-pGL3 or miR155HG promoer Single-cell RNA-seq highlights intratumoral heterogeneity in primary mut-pGL3 as well as the internal control Renilla plasmid into U87 and GP1 glioblastoma. Science (New York, NY). 2014;344:1396–401. cells. The cells then were treated with or without SH-4-54. Relative luciferase 3. Bartonicek N, Maag JL, Dinger ME. Long noncoding RNAs in cancer: activity was analyzed after 48 h treatment. (*p < 0.05, **p < 0.01). (E) Two mechanisms of action and technological advancements. Mol Cancer. groups of representative immunohistochemical image of tumors from groups 2016;15:43. of nude mice implanted with U87 cells, transfected with a lentivirus with sh- 4. Kallen AN, Zhou XB, Xu J, Qiao C, Ma J, Yan L, et al. The imprinted H19 miR155HG or sh-NC, were shown to compare the volume size of tumors and lncRNA antagonizes let-7 microRNAs. Mol Cell. 2013;52:101–12. the expression of ANXA2 and p-STAT3 through serial slices of same section of 5. Elton TS, Selemon H, Elton SM, Parinandi NL. Regulation of the MIR155 host tumor. (TIF 13681 kb) gene in physiological and pathological processes. Gene. 2013;532:1–12. 6. Haasch D, Chen YW, Reilly RM, Chiou XG, Koterski S, Smith ML, et al. T cell activation induces a noncoding RNA transcript sensitive to inhibition by Abbreviations immunosuppressant drugs and encoded by the proto-oncogene, BIC. Cell EGF: Epidermal Growth Factor; CeRNAs: Competing endogenous RNAs; Immunol. 2002;217:78–86. ChIP: Chromatin immunoprecipitation; FISH: Fluorescence in situ 7. Kluiver J, Poppema S, de Jong D, Blokzijl T, Harms G, Jacobs S, et al. BIC and hybridization; GBM: Glioblastoma; GP1: Glioblastoma cell from patient 1; miR-155 are highly expressed in Hodgkin, primary mediastinal and diffuse IHC: Immunoprecipitation; RIP: RNA immunoprecipitation large B cell lymphomas. J Pathol. 2005;207:243–9. 8. van den Berg A, Kroesen BJ, Kooistra K, de Jong D, Briggs J, Blokzijl T, et al. Acknowledgments High expression of B-cell receptor inducible gene BIC in all subtypes of We thank Department of Neurosurgery, The First Affiliated Hospital of Hodgkin lymphoma. Genes, chromosomes & cancer. 2003;37:20–8. Nanjing Medical University (Nanjing, China) and Department of 9. Neilsen PM, Noll JE, Mattiske S, Bracken CP, Gregory PA, Schulz RB, et al. Neurosurgery, Nanjing First Hospital of Nanjing Medical University for their Mutant p53 drives invasion in breast tumors through up-regulation of miR- generous help. We thank Edanz Group (www.edanzediting.com/ac) for 155. Oncogene. 2013;32:2992–3000. editing a draft of this manuscript. 10. Wu X, Wang Y, Yu T, Nie E, Hu Q, Wu W, et al. Blocking MIR155HG/miR-155 axis inhibits mesenchymal transition in glioma. Neuro-oncology. 2017;19: 1195–205. Funding 11. Rescher U, Gerke V. Annexins--unique membrane binding proteins with This work was supported by grants from National Natural Science diverse functions. J Cell Sci. 2004;117:2631–9. Foundation of China (81672501), Jiangsu Province’s Science and Technology 12. Yang T, Peng H, Wang J, Yang J, Nice EC, Xie K, et al. Prognostic and Foundation (BE2018724), Jiangsu Province’s Natural Science Foundation diagnostic significance of annexin A2 in colorectal cancer. Colorectal (20170108), the Program for Advanced Talents within Six Industries of disease : the official journal of the Association of Coloproctology of great. Jiangsu Province (2016-WSW-013), Jiangsu Province’s Key Discipline of Britain and Ireland. 2013;15:e373–81. Medicine (ZDXKA2016001), and the Priority Academic Program Development 13. Shetty PK, Thamake SI, Biswas S, Johansson SL, Vishwanatha JK. Reciprocal of Jiangsu Higher Education Institutions (PAPD). regulation of annexin A2 and EGFR with Her-2 in Her-2 negative and herceptin-resistant breast cancer. PLoS One. 2012;7:e44299. Availability of data and materials 14. Sharma MR, Koltowski L, Ownbey RT, Tuszynski GP, Sharma MC. The dataset(s) supporting the findings of this study are included within the article. Angiogenesis-associated protein annexin II in breast cancer: selective expression in invasive breast cancer and contribution to tumor invasion and Authors’ contributions progression. Exp Mol Pathol. 2006;81:146–56. WWN, YTF and WYZ performed in vitro assays, collected clinical samples and 15. Takano S, Togawa A, Yoshitomi H, Shida T, Kimura F, Shimizu H, et al. analyzed the data; TW performed in vivo experiments; WWN wrote the Annexin II overexpression predicts rapid recurrence after surgery in Wu et al. Journal of Experimental & Clinical Cancer Research (2019) 38:133 Page 14 of 14 pancreatic cancer patients undergoing gemcitabine-adjuvant 38. Qadir XV, Han C, Lu D, Zhang J, Wu T. miR-185 inhibits hepatocellular chemotherapy. Ann Surg Oncol. 2008;15:3157–68. carcinoma growth by targeting the DNMT1/PTEN/Akt pathway. Am J 16. Zhuang H, Tan M, Liu J, Hu Z, Liu D, Gao J, et al. Human epididymis protein Pathol. 2014;184:2355–64. 4 in association with Annexin II promotes invasion and metastasis of 39. Liu C, Chen Z, Hu X, Wang L, Li C, Xue J, et al. MicroRNA-185 downregulates ovarian cancer cells. Mol Cancer. 2014;13:243. androgen receptor expression in the LNCaP prostate carcinoma cell line. Mol Med Rep. 2015;11:4625–32. 17. Shetty P, Bargale A, Patil BR, Mohan R, Dinesh US, Vishwanatha JK, et al. Cell 40. Tang H, Wang Z, Liu X, Liu Q, Xu G, Li G, et al. LRRC4 inhibits glioma cell surface interaction of annexin A2 and galectin-3 modulates epidermal growth factor receptor signaling in Her-2 negative breast cancer cells. Mol growth and invasion through a miR-185-dependent pathway. Curr Cancer Drug Targets. 2012;12:1032–42. Cell Biochem. 2016;411:221–33. 41. Tran DDH, Kessler C, Niehus SE, Mahnkopf M, Koch A, Tamura T. Myc target 18. Onishi M, Ichikawa T, Kurozumi K, Inoue S, Maruo T, Otani Y, et al. Annexin gene, long intergenic noncoding RNA, Linc00176 in hepatocellular A2 regulates angiogenesis and invasion phenotypes of malignant glioma. carcinoma regulates cell cycle and cell survival by titrating tumor Brain tumor pathology. 2015;32:184–94. suppressor microRNAs. Oncogene. 2018;37:75–85. 19. Zhai H, Acharya S, Gravanis I, Mehmood S, Seidman RJ, Shroyer KR, et al. 42. Boyko V, Mudrak O, Svetlova M, Negishi Y, Ariga H, Tomilin N. A major Annexin A2 promotes glioma cell invasion and tumor progression. J cellular substrate for protein kinases, annexin II, is a DNA-binding protein. Neurosci. 2011;31:14346–60. FEBS Lett. 1994;345:139–42. 20. Maruo T, Ichikawa T, Kanzaki H, Inoue S, Kurozumi K, Onishi M, et al. 43. Krutilina RI, Babich VS, Kropotov AV, Mudrak OS, Chesnokov IN, Turoverova Proteomics-based analysis of invasion-related proteins in malignant gliomas. LV, et al. The DNA-binding activity of the membrane protein annexin II and Neuropathology : official journal of the Japanese Society of its interaction with antibodies to the chromatin nuclear ribonucleoprotein Neuropathology. 2013;33:264–75. (alpha-RNP). Tsitologiia. 1996;38:1106–14. 21. Wu W, Zhou X, Yu T, Bao Z, Zhi T, Jiang K, et al. The malignancy of 44. Bao H, Jiang M, Zhu M, Sheng F, Ruan J, Ruan C. Overexpression of Annexin miR-18a in human glioblastoma via directly targeting CBX7. Am J II affects the proliferation, apoptosis, invasion and production of Cancer Res. 2017;7:64–76. proangiogenic factors in multiple myeloma. Int J Hematol. 2009;90:177–85. 22. Wu W, Hu Q, Nie E, Yu T, Wu Y, Zhi T, et al. Hypoxia induces H19 expression 45. Lokman NA, Ween MP, Oehler MK, Ricciardelli C. The role of annexin A2 in through direct and indirect Hif-1alpha activity, promoting oncogenic effects tumorigenesis and cancer progression. Cancer microenvironment : official in glioblastoma. Sci Rep. 2017;7:45029. journal of the International Cancer Microenvironment Society. 2011;4:199–208. 23. Tian C, Deng Y, Jin Y, Shi S, Bi H. Long non-coding RNA RNCR3 promotes 46. Kohli S, Bhardwaj A, Kumari R, Das S. SIRT6 is a target of regulation by prostate cancer progression through targeting miR-185-5p. Am J Transl Res. UBE3A that contributes to liver tumorigenesis in an ANXA2-dependent 2018;10:1562–70. manner. Cancer Res. 2018;78:645–58. 24. Afshar S, Najafi R, Sedighi Pashaki A, Sharifi M, Nikzad S, Gholami MH, et al. 47. Rocha MR, Barcellos-de-Souza P, Sousa-Squiavinato ACM, Fernandes PV, de MiR-185 enhances radiosensitivity of colorectal cancer cells by targeting Oliveira IM, Boroni M. Annexin A2 overexpression associates with colorectal IGF1R and IGF2. Biomedicine & pharmacotherapy = Biomedecine & cancer invasiveness and TGF-ss induced epithelial mesenchymal transition pharmacotherapie. 2018;106:763–9. via Src/ANXA2/STAT3. 2018;8:11285. 25. Zhang Q, Chen Y, Liu K. miR-185 inhibits cell migration and invasion of 48. Keklikoglou I, Hosaka K, Bender C, Bott A, Koerner C, Mitra D, et al. hepatocellular carcinoma through CDC42. Oncol Lett. 2018;16:3101–7. MicroRNA-206 functions as a pleiotropic modulator of cell proliferation, 26. Zhao L, Zhang Y, Liu J, Yin W, Jin D, Wang D, et al. MiR-185 inhibits cell invasion and lymphangiogenesis in pancreatic adenocarcinoma by proliferation and invasion of non-small cell lung cancer by targeting KLF7. targeting ANXA2 and KRAS genes. Oncogene. 2015;34:4867–78. Oncol Res. 2018. 49. Yuan J, Yang Y, Gao Z, Wang Z, Ji W, Song W, et al. Tyr23 phosphorylation 27. Filipowicz W, Bhattacharyya SN, Sonenberg N. Mechanisms of post- of Anxa2 enhances STAT3 activation and promotes proliferation and transcriptional regulation by microRNAs: are the answers in sight? Nat Rev invasion of breast cancer cells. Breast Cancer Res Treat. 2017;164:327–40. Genet. 2008;9:102–14. 50. Aaronson DS, Horvath CM. A road map for those who don't know JAK- 28. Schumperli D, Pillai RS. The special Sm core structure of the U7 snRNP: far- STAT, vol. 296. New York, NY: Science; 2002. p. 1653–5. reaching significance of a small nuclear ribonucleoprotein. Cellular and 51. Villalva C, Martin-Lanneree S, Cortes U, Dkhissi F, Wager M, Le Corf A, et al. molecular life sciences : CMLS. 2004;61:2560–70. STAT3 is essential for the maintenance of neurosphere-initiating tumor cells 29. Su SC, Maxwell SA, Bayless KJ. Annexin 2 regulates endothelial in patients with glioblastomas: a potential for targeted therapy? Int J morphogenesis by controlling AKT activation and junctional integrity. J Biol Cancer. 2011;128:826–38. Chem. 2010;285:40624–34. 52. McFarland BC, Gray GK, Nozell SE, Hong SW, Benveniste EN. Activation of 30. Zhang F, Wang Z, Yuan J, Wei X, Tian R, Niu R. RNAi-mediated silencing of the NF-kappaB pathway by the STAT3 inhibitor JSI-124 in human Anxa2 inhibits breast cancer cell proliferation by downregulating cyclin D1 glioblastoma cells. Mol Cancer Res. 2013;11:494–505. in STAT3-dependent pathway. Breast Cancer Res Treat. 2015;153:263–75. 31. Maji S, Chaudhary P, Akopova I, Nguyen PM, Hare RJ, Gryczynski I, et al. Exosomal Annexin II promotes angiogenesis and breast Cancer metastasis. Molecular cancer research : MCR. 2017;15:93–105. 32. Wang T, Yuan J, Zhang J, Tian R, Ji W, Zhou Y, et al. Anxa2 binds to STAT3 and promotes epithelial to mesenchymal transition in breast cancer cells. Oncotarget. 2015;6:30975–92. 33. Lo HW, Hsu SC, Xia W, Cao X, Shih JY, Wei Y, et al. Epidermal growth factor receptor cooperates with signal transducer and activator of transcription 3 to induce epithelial-mesenchymal transition in cancer cells via up-regulation of TWIST gene expression. Cancer Res. 2007;67:9066–76. 34. Davis FM, Azimi I, Faville RA, Peters AA, Jalink K, Putney JW Jr, et al. Induction of epithelial-mesenchymal transition (EMT) in breast cancer cells is calcium signal dependent. Oncogene. 2014;33:2307–16. 35. Ebert MS, Sharp PA. Emerging roles for natural microRNA sponges. Curr Biol. 2010;20:R858–61. 36. Li S, Ma Y, Hou X, Liu Y, Li K, Xu S, et al. MiR-185 acts as a tumor suppressor by targeting AKT1 in non-small cell lung cancer cells. Int J Clin Exp Pathol. 2015;8:11854–62. 37. Li Q, Wang JX, He YQ, Feng C, Zhang XJ, Sheng JQ, et al. MicroRNA-185 regulates chemotherapeutic sensitivity in gastric cancer by targeting apoptosis repressor with caspase recruitment domain. Cell Death Dis. 2014; 5:e1197.

Journal

Journal of Experimental & Clinical Cancer Research – Springer Journals

Published: Mar 21, 2019

There are no references for this article.