A retrospective study was designed and included 60 patients with glioma (21 female patients, 39 male patients; age, 40.78±11.61 years, 18-64 years), who were admitted to The Shenzhen Second People's Hospital (Shenzhen, China) between January 2014 and December 2017. The following patients were excluded: i) Aged <18 years or ≥65 years; ii) those with other tumors or neurological diseases. After obtaining the written consent from all participants at The Second People's Hospital of Shenzhen (Shenzhen, China), formalin-fixed paraffin-embedded (FFPE) samples were collected from the Department of Pathology after diagnosis and treatment. The tissue was fixed in 10% formalin at room temperature for 12-24 h before embedding with paraffin. FFPE samples sections with a thickness of 5-µm were cut for hematoxylin and eosin (H&E) staining and miR-137 in situ hybridization (ISH). The pathological diagnosis was made independently by two neuropathologists according to the WHO classification of central nervous system tumors in 2016(24). In a previous study, the authors summarized the WHO classification and histopathological subtypes of these gliomas (22). The present study followed the principles of The Declaration of Helsinki and was approved by the Ethics Committee of Shenzhen Second People's Hospital (approval no. XZ2019103101).

The deparaffinized tissue was partially hybridized with 50 nm LNA-modified and digoxin-labeled miR-137 oligonucleotide probes (Exiqon A/S) for 1 h at 55˚C, incubated with 5 µg/ml anti-digoxin-Rhodamine antibody (cat. no. 11093274910; Roche Applied Science) overnight at 4˚C, and stained with DAPI (cat. no. ZLI-9557; OriGene Technologies, Ins.) in the dark at room temperature for 15 min. The labeling index (LI) was expressed as the percentage of positive cells to the total number of cells.

In the present study, the human GBM cell lines U251, U373, SK-MG3, U-343, A172, LNZ-308, U118 and U138, and normal human astrocytes (HAs) were used. These cell lines were purchased from The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences and cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) (25,26). Authentications of U118MG and U373MG cell lines were detected with STR profiling. According to the authentication, U118MG and U373MG cell lines are from American Type Culture Collection. All cell lines were cultured in a humidified incubator at 37˚C in 5% CO₂/95% air.

The 2nd generation system was used for lentivirus transduction. 293T cells (The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences) were used as the interim cell line. Lentivirus packaging was performed by Hanbio Biotechnology Co., Ltd. in 293T cells using pSPAX2 (10 µg), pMD2G (5 µg), shuttle plasmids (10 µg) and Lipofiter (75 µl; cat. no. HB-LF-1000; Hanbio Biotechnology Co., Ltd.) for 16 h. A total of 48 and 72 h after transfection, lentiviruses were collected because more viruses are collected at two time points to ensure enough for the later experiments. Lentiviruses collected 48 and 72 h after transfection were both used in the subsequent experiments. Hblv-cmv-miR-137-GFP-puro, hblv-GFP-puro (control), hblv-cmv-SPHK2-GFP-puro [short hairpin RNA (sh)-SPHK2] and hblv-GFP-puro (NC) vectors (Hanbio Biotechnology Co., Ltd.) were used to transduce U251 and U373 cell lines at a multiplicity of infection of 3 at 37˚C for 24 h. A total of 72 h after transduction, puromycin (3 µg/ml; Gibco; Thermo Fisher Scientific, Inc.) was added to the culture medium of transduced cells to create stable cell lines for 7 days. The overexpression of miR-137 and knockdown of SPHK2 following transduction with lentiviral vector was detected according to the manufacturer's protocol.

The culture supernatant was collected and the cytokines were analyzed. The quantity of cytokine secretion was quantified using the Quantibody^® Human Inflammation Array 1 (RayBiotech, Inc.) to measure the levels of 10 types of human cytokines, including IFN-γ, IL-1α, IL-1β, IL-10, IL-13, IL-4, IL-6, IL-8, monocyte chemoattractant protein-1 (MCP-1) and TNFα. IL-13 was examined using an ELISA kit (cat. no. CHE0004; Beijing 4A Biotech Co., Ltd.). TNF-α and IFN-γ were detected with ELISA Kits purchased from PeproTech, Inc. (cat. nos. 900-M25 and 900-M27, respectively).

The candidate targets of miR-137 were predicted using TargetScan (http://www.targetscan.org/). The wild-type (p-WT) and mutant-type (p-MT) reporter vectors of the SPHK2 3'-UTR were constructed using the pEZX-MT01 Luciferase miRNA Expression Reporter Vector (GeneCopoeia, Inc.). The predicted target coding sequence of miR-137 was deleted from SPHK2 3'-UTR cDNA for p-MT construction by site-directed mutagenesis. The sequence and direction of the two vectors were verified with Sanger DNA sequencing by Genscript, Inc.

dsRNA oligonucleotides of miR-137 mimics (cat. no. miR10000429) and scramble sequence (Scr; cat. no. miR1N0000001-1-5) were purchased from Guangzhou RiboBio Co., Ltd. dsRNA oligonucleotides (2,500 ng; 50 nM) were transfected either alone or with 0.75 µg/ml luciferase plasmids (p-WT and p-MT; GeneCopoeia, Inc.) using X-tremeGENE small interfering RNA (siRNA) Transfection reagent (cat. no. 4476093001; MilliporeSigma) for 48 h at 37˚C. A total of 48 h after transfection, subsequent experimentation was performed.

U373 and U251 cells (5,000 cells/well) inoculated into 96-well plates were transfected with 0.15 µg p-WT or p-MT and 0.08 µg miR-137 mimics or scrambled sequences using X-tremeGENE siRNA Transfection reagent for 48 h. Subsequently, the activities of Renilla and Firefly luciferase were detected on a Synergy 2 Microplate Reader Fluorometer (BioTek Instruments, Inc.) using the Dual-Luciferase Reporter Assay system (Promega Corporation). The results were normalized to the measured value of Firefly luciferase.

TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) was used to extract total RNA from the two groups of cell lines. miR-137 was quantified using the Bulge-loop miRNA RT-qPCR Detection kit (Guangzhou RiboBio Co., Ltd.). The qPCR thermocycling conditions were as follows: Initial denaturation at 95˚C for 10 min, followed by 40 cycles of denaturation at 95˚C for 2 sec, annealing at 60˚C for 20 sec and extension at 70˚C for 10 sec; final dissociation was performed at 95˚C for 15 sec, 60˚C for 1 min and 95˚C for 15 sec. The expression of SPHK2 mRNA was detected using the Reverse Transcription System and GoTaq qPCR Master Mix kit (Promega Corporation) according to the manufacturer's instructions. U6 and GAPDH were used as internal controls for miR-137 and SPHK2 mRNA, respectively. All reactions were performed on the CFX Connect^™ Real-Time PCR Detection system (Bio-Rad Laboratories, Inc.). The expression of miR-137 and SPHK2 were calculated using the 2^-ΔΔCq method (27). The miR-137-3P Primer Set (cat. no. MQPS0000619-1-100; Guangzhou RiboBio Co., Ltd.) and the U6 qPCR Primer Set (cat. no. MQPS0000002-1-100; Guangzhou RiboBio Co., Ltd.) were used in these experiments.

Western blot analysis was performed as previously described (28). The primary antibodies used were rabbit anti-SPHK2 (1:1,000; cat. no. ab264042; Abcam) and mouse anti-β-actin (1:1,000; cat. no. sc-81178; Santa Cruz Biotechnology, Inc.), used for incubation at 4˚C overnight. The secondary HRP-conjugated antibodies (used for incubation at room temperature for 1 h) were as follows: Anti-mouse IgG (1:2,000; cat. no. 7076) and Anti-rabbit IgG (1:2,000; cat. no. 7074) both purchased from Cell Signaling Technology, Inc.

All statistical analysis was performed using SPSS 24.0 software (IBM Corp.). The data are expressed as the mean ± SD. The normality of the distribution was estimated using the Kolmogorov-Smirnov test. The unpaired double-tailed Student's t-test was used to analyze the statistical difference between the two groups. The differences among the sample groups were analyzed by one-way ANOVA followed by multiple comparison Tukey's honestly significant difference post hoc test. In order to determine the optimal cut-off level of miR-137 expression for the diagnosis of gliomas, receiver operating characteristic (ROC) curve analysis was performed. The Youden index (J=sensitivity + specificity-1) was used to determine the optimal cutoff level. The optimal cutoff value was the cutoff value used to obtain the highest Youden index. P<0.05 was considered to indicate a statistically significant difference. All cell lines were tested at least three times and the samples were repeated three times.

ISH with LNA-modified probes was applied to detect endogenous miR-137 expression in FFPE specimens of 60 gliomas and 11 non-tumorous brain tissues. It was found that the expression of miR-137 in gliomas was decreased compared with non-tumorous brain tissues, and that its expression was significantly decreased as the glioma grade increased; however, there was no significant difference between two high-grade gliomas (WHO III grade and WHO IV grade) (Fig. 1A and B). miR-137 expression was also detected in 8 glioma cell lines and normal HAs. The results revealed that the expression of miR-137 in all glioma cell lines was significantly lower compared with normal HAs (Fig. 1C). The expression curve of miR-137 between non-tumorous brain tissues (control group) and glioma tissues revealed that the area under the ROC curve was 0.7924 [95% Confidence Interval (CI), 0.6425-0.9423; P<0.05]. As demonstrated in Fig. 2A and B, the expression of miR-137 may prove helpful for the diagnosis of glioma. The optimal cutoff value of the miR-137 LI was 0.37% (sensitivity, 55%; specificity, 90.91%). The ROC curve of miR-137 between low-grade (WHO II) and high-grade (WHO III and V) gliomas revealed an area under the ROC curve of 0.6998 (95% CI, 0.5584-0.8411; P<0.05; Fig. 2C and D), indicating that the expression of miR-137 in gliomas contributes to the differentiation of glioma grades. The optimal cutoff value of the miR-137 LI in the diagnosis of glioma was 0.90% (sensitivity, 60%; specificity, 81.82%). These data indicated the inverse association of miR-137 expression with glioma and revealed that miR-137 may be a potential biomarker for the diagnosis of patients with glioma.

To examine the effects of miR-137 on TAM-associated release of cytokines, miR-137-overexpression cells were constructed by transducing miR-137 overexpression lentiviral vector (empty lentiviral vector transduced cells as control). The Quantibody^® Human Inflammation Array 1 was used to detect the 10 human cytokines, including IFN-γ, IL-1α, IL-1β, IL-10, IL-13, IL-4, IL-6, IL-8, MCP-1 and TNFα. The results revealed that miR-137 overexpression reduced the IL-13 levels, whereas it promoted TNFα and IFNγ production (Fig. 3A-C). As previously reported, IL-13 can promote M2 polarization, and TNFα and IFNγ can promote M1 polarization (29,30). Hence, it was hypothesized that miR-137 may promote M1 TAM polarization by inhibiting M2-associated cytokine and promoting M1-associated cytokine release. There were no differences in the levels of IL-1α, IL-1β, IL-10, IL-4, IL-6, IL-8 and MCP-1 between NC and miR-137 overexpression (data not shown). To analyze the effect of SPHK2 on TAMs associated release of cytokines, knockdown of SPHK2 was carried out in U251 and U373 cells at mRNA and protein levels (Fig. 3D-F). Two cell lines were transduced with lentivirus containing knockdown shRNA sequence (HK as control group, and SH1, SH2, SH3 as knockdown groups). After confirming SPHK2 knockdown, IL-13, TNFα and IFNγ released by NC and sh-SPHK2 cells were examined using ELISA. It was found that SPHK2 knockdown reduced IL-13 concentration in U373 and U251 cells, promoted TNFα release in U373 cells, and increased IFNγ release both in U373 and U251 cells (Fig. 3G-I).

Using TargetScan, it was found that SPHK2 was one of the potential targets of miR-137. miR-137 only had one putative binding site in the SPHK2 3'-UTR (Fig. 4A). In order to obtain direct evidence of miR-137 targeting SPHK2, two recombinant luciferase reporter gene vectors of SPHK2 3'-UTR were constructed: p-WT and p-MT. The recombinant luciferase mRNA transcribed by p-WT carried the predicted miR-137 target (SPHK2-3'-UTR-WT) in SPHK2-3'-UTR, while the mRNA transcribed by p-MT lacked the predicted target (SPHK2-3'-UTR-MT) (Fig. 4A). The results of the dual-luciferase reporter assay revealed that the relative luciferase activity significantly decreased following p-WT co-transfection with the miR-137 mimic (the scramble sequence was used in the control group); however, no change was observed following p-MT transfection (P<0.05; Fig. 4D and E). To further verify whether miR-137 directly targeted SPHK2, RT-qPCR and western blot analysis were performed to detect the changes in SPHK2 expression in miR-137-transfected cell lines. As revealed in Fig. 4B and C, SPHK2 protein expression in the miR-137-transfected cell lines was significantly decreased compared with the scramble control-transfected cell lines (Fig. 4C); however, the SPHK2 mRNA level was not altered (Fig. 4B). These results suggested that miR-137 may directly target SPHK2 in glioma cells and downregulate SPHK2 protein expression by inhibiting its mRNA translation.