Glioblastoma cell-derived extracellular vesicle miR-27a-3p facilitates M2 macrophage polarization CURRENT STATUS: UNDER REVIEW

Background: M2 macrophage polarization has been found to be correlated with malignancy of glioblastoma. In this study, we investigated the potential role of microRNA (miRNA) derived from extracellular vesicles of glioblastoma (glioblastoma-EVs) in M2 macrophage polarization. Methods: After isolation of human glioblastoma-EVs, transmission electron microscopy (TEM) and Nano-particle tracking analysis (NTA) were performed to identify the EVs. Besides, the proliferation, migration and invasion of glioma cells were analyzed by CCK8 and Transwell assays, respectively. The target genes of miR-27a-3p were predicted through bioinformatics analysis and verified by dual-luciferase reporter gene assay. ChIP assay was applied to detect the binding of enhancer of zeste homologue 1 (EZH1) to lysine-specific demethylase 3A (KDM3A) promoter region and the interaction between KDM3A and connective tissue growth factor (CTGF). Glioblastoma mouse models were established followed by the implement of hematoxylin-eosin (HE) and ELISA staining on pathological changes of mouse brain tissues. Results: Human glioblastoma-EVs were successfully isolated. The high expression of miR-27a-3p was found in glioblastoma tissues as well as glioblastoma-EVs, which could induce M2 polarization, thus promoting glioblastoma cell proliferation, migration and invasion. It was also shown that miR-27a-3p targeted EZH1 and promoted KDM3A expression to elevate the expression of CTGF. Glioblastoma-EVs delivered miR-27a-3p to promote the KDM3A-upregulated CTGF by downregulating EZH1, thereby promoting M2 macrophage polarization and development of glioblastoma in vivo. Conclusion: These findings highlight that EV miR-27a-3p may promote M2 macrophage polarization, which is associated with the progression of tumors. performed on the expression of M1 macrophage marker IL-10 and M2 macrophage marker TNF-α. Also, the secretion levels were analyzed by ELISA. The results showed that compared with the control group, the expression of IL-10 under hypoxia was increased while the TNF-α expression was suppressed. Moreover, glioblastoma-EVs under hypoxic significantly increased the secretion of IL-10 in macrophages, but prominently reduced the secretion of TNF-α, confirming that it is more likely that glioblastoma-EVs induce M2 macrophage polarization under hypoxic condition than miR-27a-3p mimic/miR-27a-3p glioblastoma cells, which were co-cultured with CTGF overexpression vector (oe-CTGF)-treated macrophages. The results showed that the in EVs extracted from miR-27a-3p mimic-treated glioblastoma cells, upregulated CTGF prominently stimulated M2 macrophage polarization, in which the expression of miR-27a-3p, KDM3A and CTGF expression increased, but EZH1 and TNF-α expression reduced. In contrast, in the EVs extracted from miR-27a-3p inhibitor-treated cells,

modulated tumor microenvironment of glioblastoma may represent a prognostic biomarker for glioblastoma. It is reported that the extracellular vesicles of glioblastoma (glioblastoma-EVs) carry functional genomic and proteomic cargo and affect surrounding and distant recipient cells, which enables EVs to emerge as crucial mediators of tumor microenvironment in glioblastoma [6]. Previous work also highlights the promoting role of microRNAs (miRNAs) delivered by glioblastoma-EVs in M2 macrophage polarization [7]. However, the miRNAs derived from glioblastoma-EVs are not fully understood, and it is necessary to further explore the detailed function of glioblastoma-EVs carried miRNA in M2 macrophage polarization. miRNAs are small non-coding RNAs which are implicated in different physiological processes, and dysregulation of miRNAs can lead to various cancers [8]. miRNAs can be EVs' components playing signaling role in the progression of cancers [9]. miR-27a-3p has been revealed to be involved in the development of glioma [10]. Furthermore, hsa-miR-27a-3p has been found to be enriched in circulating EVs [11]. However, few studies analyzed the effects of miR-27a-3p derived from glioblastoma-EVs on glioblastoma and the relevant tumor microenvironment. Additionally, enhancer of zeste homologue 1 (EZH1) is strikingly downregulated in glioblastoma, which may play a role in the development of glioblastoma [12]. What's more, EZH1 is able to suppress the polarization of M2 macrophage [13]. This study aims to analyze the relationship between miR-27aa-3p carried in glioblastoma-EVs and the underlying regulatory mechanism to provide better understanding of glioblastoma and gain functional insights into the glioblastoma-EVs mediated miRNAs. To grasp EVsmodulated tumor microenvironment may unveil novel diagnostic and therapeutic targets for glioblastoma.

Methods And Materials
Ethical statement and study object

Transmission electron microscope (TEM)
The EV obtained by high-speed centrifugation of 400 mL medium was fixed in 2% glutaraldehyde overnight at 4℃, washed with PBS, fixed with 1% OsO 4 for 1 h, dehydrated in ethanol and embedded with epoxy resin. Embedded-sections were added with saturated sodium periodate and 0.1 N hydrochloric acid, the size and morphological characteristics of which were observed using TEM (JEM-1010; JEOL, Tokyo, Japan) after 10 min.

Nano-particle tracking analysis
The size distribution and concentration of EVs were determined following the instructions of NTA (Zetasizer Nano ZS90 instrument, Malvern, UK), which associated with light scattering and Brownian motion. PBS-diluted EVs were injected into Zetasizer Nano ZS90 instrument, followed by the measurement of particle size according to Brownian motion and diffusion coefficient. The filtered PBS was used as a control. All samples were measured using NP100 film with 44.5 mm and 0.64 V voltage parameters. Samples were calibrated by the dilution of CPC100 standard particles at 1,000 times under the same condition. With 5 videos (60-s-duration) taken, data were analyzed using Zetasizer software (Malvern) to identify and track each particle.

EV labeling and immunofluorescence
EVs with the concentration of 0.1-0.2 µg were resuspended in 400 µL PBS and stained with CellMask Deep Red (Thermo Fisher Scientific) at excitation/emission wavelengths of 649/666 nm. During the labeling, EVs were incubated with deep red staining solution (1:1000) for 20 min at 37℃. Then EVs were centrifuged at 100,000 g for 1 h and diluted in PBS, followed by the determination of protein concentration using bicinchoninic acid (BCA) protein detection kit.
Cells were stained with CellTrace™ Carboxyfluorescein succinimidyl ester (CFSE, Life Technologies, Carlsbad, CA, USA) with a maximum excitation/emission wavelength of 492/517 nm. The immunofluorescence staining was performed after the covalent binding of cells diffused by lactonedigested CFSE with intracellular amines. Glioblastoma cells (3-5 × 10 5 ) in serum-free medium were stained with CFSE (working concentration of 5 µM) at a dilution of 1:1000 and incubated at 37℃ for 20 min in the dark. After cells were seeded into 8-well slides (Millipore, Billerica, MA, USA), EVs were incubated with CFSE-stained cells at different time points. Cells were fixed with 3.7% (w/v) formaldehyde for 5 min at room temperature, imaged and observed under a fluorescence microscope.

RNA isolation and quantification
According to the instructions of the Trizol reagent (Invitrogen, Carlsbad, CA, USA), RNA was extracted, which was reversely transcribed into complementary DNA (cDNA) using the first strand cDNA synthesis kit (Tiangen Biotech, Beijing, China) or Primer script™ one-step RT-PCR Kit (Takara, Shiga, Japan). Real-time polymerase chain reaction was conducted in ABI 7500 real-time PCR system (Applied Biosystems, Carlsbad, CA, USA) using SYBR Green IReal-time PCR kit (Cowin Bioscience, Beijng, China) with GAPDH or U6 used as an internal reference. Cel-miR-39 (No. miRB00000010-3-1 and MQPS0000071-1-100, Riobio, Guangzhou, China) was used as an external reference for the detection of EV miR-27a-3p. The relative expression of genes was analyzed by 2-ΔΔCT method, and primers involved in this experiment were shown in Table 1. Table 1 Primer sequences for RT-qPCR Human Mouse R 5-'AGATCAAGATCATTGCTCCTCCT − 3' Notes: F, Forward; R, Reverse; EZH1, enhancer of zeste 1; KDM3A, lysine demethylase 3A; CTGF, connective tissue growth factor; RT-qPCR, reverse transcription quantitative polymerase chain reaction The proliferative capacity of glioblastoma cells was detected using CCK-8 kit (Dojindo, Kumamoto, Japan) according to the instructions. Glioblastoma cells were seeded in 96-well plates for co-culture with different EVs at 0, 24, 48 and 72 h. CCK-8 solution was added to each well, which was incubated for 2 h. Then, the viable cells were measured at the optimal density (OD) of 450 nm using a microplate (Multiskan Sky Microplate Spectrophoto-meter, Cat. No. 51119570, Thermo Fisher Scientific). Each experiment was repeated 3 times.

Transwell assay
The migration and invasion ability of glioblastoma cells was evaluated using Transwell assay by following the manufacturer's protocol. Matrigel (BD Bioscience, Franklin Lakes, NJ, USA) coated apical Transwell chambers were placed at 37℃ for 30 min to polymerize Matrigel. Cells were cultured in serum-free medium for 12 h, harvested and resuspended in serum-free medium (1 × 10 5 /mL). While 100 µL cell suspension was incubated in the basolateral chamber containing 10% FBS at 37℃ for 24 h. Cells that did not invade the surface of Matrigel membrane were gently removed with cotton swabs. Cells were fixed with 100% methanol and stained with 1% toluidine blue (Sigma-Aldrich).
Stained cells were counted in five randomly selected areas under inverted light microscope (Carl Zeiss, German). Each experiment was repeated 3 times.
Animal experiment BALB/c female nude mice (n = 60) were purchased from SLAC Laboratory Animal (Shanghai, China). After 1-week adaptive feeding, mice were anesthetized with pentobarbital sodium and treated differently with 12 mice in each treatment. Sham-operated mice were taken as the sham group, while others were injected through the brain with murine glioblastoma cell line GL261 (with 10 6 cells each mouse) or macrophage (with 2 × 10 5 cells each mouse) with adenovirus-mediated CTGF knockdown.
Subsequently, mice in each group were intravenously injected with PBS or the equivalent volume of EV (8 mg/kg) extracted from miR-27a-3p mimic-treated GL261 cells via the caudal vein every 3 days.
Six mice were randomly selected from each group to record the survival time. The tumors of other mice were dissected 40 days after xenograft, frozen in liquid nitrogen or fixed in formalin, with serum samples collected.

Enzyme-linked immunosorbent assay (ELISA)
ELISA kits involving hsa-interlukin (IL)-10 (ab46034, Abcam), mmu-IL-10 (ab46103, Abcam), hsa-TNF-α (ab100654, Abcam) and mmu-TNF-α (ab208348, Abcam) were commercially obtained to detect the levels of IL-10 and TNF-α in the supernatant of U87-MG cells and peripheral blood of mice. Each well of the 96-well microtitration plates was added with 100 µL IL-10 or TNF-α (0.4 µg/mL) buffer solution and 50 mM sodium carbonate to adjust the pH value to 9.6. After overnight incubation at 4℃, the plate was blocked by PBS containing 1% BSA for 1 h at room temperature. IL-10 or TNF-a antibodies or samples to be tested were added to the wells, which were then incubated at room temperature for 2 h. Plates were incubated with 0.2 µg/mL biotinylated IL-10 or TNF-α antibodies for 2 h at room temperature. Diluted peroxidase-labeled antibiotic protein (1:1000; Sigma-Aldrich) was incubated with the plate for 1 h at room temperature. The chromogenic reaction was induced by the addition of 3,3',5,5'-tetramethylbenzidine (TMB)/H 2 O 2 substrate solution, which was terminated by 1 M phosphoric acid after 30 min. The OD was measured at 450 nm wavelength using an automatic microplate reader. The OD value was measured with normalization to diluted antibodies in the medium, ranging from 10 to 2,000 pg/mL with standard curves plotted.
Hematoxylin-Eosin (HE) staining Mouse brain tissues were fixed in 10% formalin for 24 h, routinely dehydrated and paraffinembedded. Brain tissue samples were cut into 3 µm-thick sections and HE staining was performed following previously described methods [14] to determine the lesion areas in mouse brain tissues. The images were photographed under a microscope at 100-fold magnification.

Statistical analysis
All data were analyzed using SPSS 21.0 software (SPSS, Inc, Armonk, NY, USA). Measurement data were expressed by mean ± standard deviation. Unpaired t-test was used for comparison between two groups, while one-way analysis of variance (ANOVA) was used for comparison among multiple groups and Tukey's post hoc test. Comparison among groups at different time points was conducted using two-way ANOVA and Bonferroni's post hoc test. A p < 0.05 was considered statistically significant.

Glioblastoma-EVs induce M2 macrophage polarization under hypoxia condition
To explore whether glioblastoma-EVs promote M2 macrophage polarization under hypoxia condition, glioblastoma-EVs were extracted from the supernatant of cultured human glioblastoma cell line U87MG, which were analyzed by TEM, NTA and Western blotting. The results showed that glioblastoma-EVs were typical round particles expressing CD63 and TSG101 proteins instead of calnexin proteins, which were with an average diameter of 105 ± 5.8 nm (Fig. 1A-1C). Meanwhile, human monocyte cell line U937 was treated with PMA to induce the differentiation into macrophages.
Following that, macrophages were stained with CFSE while glioblastoma-EVs were stained with deep red staining solution. Subsequently, the stained glioblastoma-EVs were co-cultured with CFSE-stained macrophages under normoxia or hypoxia conditions, respectively. After 24-h co-culture, it was found through the fluorescence microscope that macrophages would adhere to and took in glioblastoma-EVs, and under hypoxia condition macrophages could take in more glioblastoma-EVs (Fig. 1D).
To further determine that glioblastoma-EVs can induce M2 macrophage polarization under hypoxia, RT-qPCR was performed on the expression of M1 macrophage marker IL-10 and M2 macrophage marker TNF-α. Also, the secretion levels were analyzed by ELISA. The results showed that compared with the control group, the expression of IL-10 under hypoxia was increased while the TNF-α expression was suppressed. Moreover, glioblastoma-EVs under hypoxic significantly increased the secretion of IL-10 in macrophages, but prominently reduced the secretion of TNF-α, confirming that it is more likely that glioblastoma-EVs induce M2 macrophage polarization under hypoxic condition than normoxic condition (Fig. 1E-1F).
Glioblastoma-EVs transfer miR-27a-3p to promote M2 macrophage polarization Furthermore, we found that the expression of miR-27a-3p in glioblastoma was significantly increased in glioblastoma-related microarray GSE65626 which was obtained from GEO database (https://www.ncbi.nlm.nih.gov/geo/) ( Fig. 2A), and the expression of miR-27a-3p was also found to be increased in clinical specimens (Fig. 2B). Therefore, we explored the role of miR-27a-3p in glioblastoma-EVs-induced M2 macrophage polarization under hypoxia. The expression of miR-27a-3p was determined in macrophages incubated with glioblastoma-EVs under hypoxia or normoxia.
Compared with macrophages under normoxia, miR-27a-3p showed higher expression in macrophages under hypoxia (Fig. 2C). Together with the results shown in Fig. 1D, these results suggested that miR-27a-3p could be delivered to macrophages by glioblastoma-EVs. Subsequently, macrophages were transfected with miR-27a-3p mimic or miR-27a-3p inhibitor and then cultured under normoxia or hypoxia. The expression of miR-27a-3p, IL-10 and TNF-α in each group was detected by RT-qPCR, the results of which showed that IL-10 expression was increased and TNF-α expression was decreased after treatment of miR-27a-3p mimic under hypoxia/normoxia, while the plasmid miR-27a-3p inhibitor induced the opposite effect. (Fig. 2D-2F). In addition, the expression of macrophage polarization markers iNOS and Arg-1 was evaluated, and the results revealed that iNOS was downregulated and Arg-1 was upregulated in macrophages transfected with miR-27a-3p mimic under hypoxia/normoxia, while the opposite effects were observed in macrophages transfected with miR-27a-3p inhibitor ( Fig. 2G-H). These results indicate that upregulation of miR-27a-3p can induce M2 macrophage polarization under hypoxia/normoxia. Next, the migration, invasion and proliferation of glioblastoma cells were evaluated after co-culture of macrophages with glioblastoma cells. The results of CCK-8 assay showed that the proliferation ability of glioblastoma cells co-cultured with macrophages was strengthened after miR-27a-3p mimic treatment relative to mimic NC, but weakened in glioblastoma cells co-cultured with miR-27a-3p inhibitor-treated macrophages in contrast to inhibitor NC (Fig. 2I). In addition, the results of Transwell assay showed higher motility of glioblastoma cells co-cultured with miR-27a-3p mimic-treated macrophages, while cell motility was curtailed in glioblastoma cells co-cultured with miR-27a-3p inhibitor-treated macrophages (Fig. 2J-2K). These results indicate that miR-27a-3p can promote the polarization of M2 macrophages, thereby facilitating the proliferation, migration and invasion of glioblastoma cells in vitro.
Meanwhile, differential analysis was performed on the microarrays GSE12657, GSE104291 and GSE50161 using limma package of R language with |logFoldChange| > 1 and p-value < 0.05 as threshold. The prediction results of differentially expressed genes and target genes in the differential analysis results were intersected (Fig. 3A). Subsequently, GO functional analysis of intersected genes revealed that those genes were mainly enriched in the pathways of "modulation of chemical synaptic transmission", "neuron to neuron synapase" and "protein serine/threonine kinase activity" (Fig. 3B).
The expression of EZH1 was analyzed in the three glioblastoma-related microarrays, showing that EZH1 was reduced. The expression of EZH1 in microarray GSE50161 was shown in Fig. 3C. We further detected the mRNA expression of EZH1 in clinical specimens and found that EZH1 expression was also decreased in glioblastoma patients (Fig. 3D).
To further validate the regulation of EZH1 by miR-27a-3p at the cellular level, U87MG cells were transfected with miR-27a-3p mimic, followed by the extraction of EVs. Then, EVs were co-cultured with macrophages overexpressing EZH1 under hypoxia. Results of RT-qPCR showed that the expression of miR-27a-3p elevated while EZH1 expression was reduced in the miR-27a-3p mimictreated cells, along with increased IL-10 and decreased TNF-α expression. However, the expression of miR-27a-3p did not significant change in the cells co-transfected with miR-27a-3p mimic and EZH1 overexpression vector (oe-EZH1) and cells co-transfected with miR-27a-3p and oe-NC, while among miR-27a-3p mimic-and oe-EZH1-treated cells the expression of EZH1 and TNF-α increased and IL-10 expression was decreased, indicating that the overexpression of EZH1 reversed the effects of miR-27a-3p on M2 macrophage polarization (Fig. 3G-3I). In addition, the expression of iNOS and Arg-1, the polarization markers of macrophages, was further determined in each group. Results showed that the iNOS expression was decreased and Arg-1 expression was increased after the treatment of miR-27a-3p mimic of miR-27a-3p, while the overexpression of EZH 1 could reverse the effects of the miR-27a-3p mimic (Fig. 3J). The findings suggest that the overexpression of EZH 1 could reverse the expression of miR-27a-3p effects on M2 macrophage polarization.
CCK-8 and Transwell assays were conducted to detect the effects of miR-27a-3p mimic and/or oe-EZH1 on glioblastoma cell proliferation, migration and invasion. The proliferation, migration and invasion of glioblastoma cells were increased after treatment of miR-27a-3p mimic, but were distinctly suppressed after the addition of oe-EZH1, indicating that the overexpression of EZH1 offset the effects of overexpressed miR-27a-3p on the proliferation (Fig. 3K), migration and invasion (Fig. 3L-3M) of glioblastoma cells. These results indicate that miR-27a-3p induces M2 macrophage polarization by targeting EZH1.

EZH1 inhibits M2 macrophage polarization by suppressing KDM3A expression
As ChIP-Seq data analysis results unraveled, EZH1 was found to bind to the promoter of KDM3A ( Fig. 4A), suggesting that EZH1 inhibited KDM3A at a transcriptional level. Next, we performed ChIP analysis using anti-H3K27me3 antibody after si-EZH1 treatment to detect changes of histone modifications. It was found that EZH1 protein was highly enriched in the promoter region of KDM3A after macrophages were treated with 3xFLAG-EZH1 vector (Fig. 4B). Moreover, si-EZH1 significantly enhanced H3K27me3 methylation in the promoter region of KDM3A gene (Fig. 4C). Therefore, endogenous EZH1 protein can bind to the promoter region of KDM3A and inhibit the expression of KDM3A gene by removing histone H3K27me3. It can be inferred that H3K27me3 plays an important role in the activation of the promoter region of KDM3A gene.
To further explore the regulatory relationship between EZH1 and KDM3A, EZH1 or/and KDM3A was knocked down in macrophages. Results of RT-qPCR displayed that silencing EZH1 could upregulate KDM3A, increase IL-10 and decrease TNF-α expression to induce M2 macrophage polarization. While knockdown of KDM3A reversed the effects of silenced EZH1 on M2 macrophage polarization when cotransfected with si-EZH1 (Fig. 4D-4F). Additionally, the results of Western blotting exhibited that silencing EZH1 could downregulate the expression of iNOS and upregulate the expression of Arg-1, while KDM3A knockdown could reverse the effects of silenced EZH1 on M2 polarization (Fig. 4G).
These results suggest that EZH1 regulates M2 macrophage polarization by inhibiting KDM3A expression through H3K27me3 methylation.

KDM3A facilitates macrophage polarization by upregulating CTGF
To examine the regulatory function of endogenous KDM3A protein in macrophages, we performed ChIP analysis on cells using antibodies to CTGF, H3K27ac and H3K4me1 after treatment of macrophages with siRNA targeting KDM3A (si-KDM3A). It was found that KDM3A protein was highly enriched in the enhancer region of CTGF gene. Meanwhile, the changes of histone modification in the enhancer region of the gene were also detected. The results displayed that H3K4me1 and H3K27ac were simultaneously enriched in the enhancer region, and activation of enhancers promotes CTGF gene expression. The data also revealed that the degree of H3K4me1 and H3K27ac modification was remarkably elevated after siRNA treatment (Fig. 5A-5B). Therefore, endogenous KDM3A protein can bind to the enhancer region of CTGF and activate enhancers to promote CTGF gene expression by removing histone H3K4me1 and H3K27ac modifications.
To further explore the regulatory relationship between KDM3A and CTGF, KDM3A was overexpressed or/and CTGF was knocked down in macrophages. Results of RT-qPCR exhibited that overexpression of KDM3A upregulated CTGF and IL-10 expression but decreased TNF-α expression to induce M2 macrophage polarization. However, after overexpressing KDM3A or knocking CTGF down in cells, it was found that depleted CTGF could inhibit M2 macrophage polarization which was induced by overexpressed KDM3A (Fig. 5C-5E). The protein expression of macrophage polarization markers iNOS and Arg-1 was further examined in each group. The results showed that overexpression of KDM3A could downregulate iNOS protein expression and upregulate Arg-1 protein expression, while knockdown of CTGF could reverse the M2 macrophage polarization induced by overexpression of KDM3A (Fig. 5F), which was consistent with the above results. These results suggest that KDM3A promotes M2 macrophage polarization by binding to CTGF enhancer regions and promoting CTGF gene expression.
M2 macrophage polarization can be accelerated by glioma-EVs-released miR-27a- To determine whether EZH1 can regulate CTGF expression through KDM3A, EZH1 or CTGF was knocked down in macrophages, followed by the detection of EZH1, KDM3A and CTGF expression. It was found that the expression of EZH1 and TNF-α decreased while the expression of KDM3A, CTGF and IL-10 increased after EZH1 was silenced, accompanied with accelerated M2 macrophage polarization. However, the expression of EZH1 and KDM3A did not change significantly in the si-EZH1 and sh-CTGF group compared with the si-EZH1 and sh-NC group, while the expression of CTGF and IL-10 decreased after EZH1 and CTGF were both silenced, along with increased TNF-α expression, indicating that knockdown of CTGF could reverse the M2 macrophage polarization induced by overexpressed KDM3A and silenced EZH1 (Fig. 6A-6D).
Besides, expression of iNOS and Arg-1 was measured using Western blotting, he results of which showed that the expression of iNOS decreased and Arg-1 increased in the glioblastoma-EVs containing miR-27a-3p mimic and macrophages overexpressing CTGF. However, the results were reversed in the glioblastoma-EVs containing miR-27a-3p inhibitor and macrophages overexpressing CTGF (Fig. 6K).
The role of conditioned macrophages in the migration, invasion and proliferation of glioblastoma cells was further explored in vitro. The results of CCK-8 and Transwell assays showed that in glioblastoma-EVs containing miR-27a-3p mimic and macrophages overexpressing CTGF, strengthened proliferation, invasion and migration ability was observed, which was abolished in glioblastoma-EVs containing miR-27a-3p inhibitor and macrophages overexpressing CTGF (Fig. 6L-6N). These results suggest that glioblastoma-EVs carrying miR-27a-3p promote the activation of CTGF by inhibiting EZH1 expression and upregulating KDM3A, which in turn boost M2 macrophage polarization and ultimately facilitate the proliferation, migration and invasion of glioblastoma cells in vitro.

M2 macrophage polarization aggravates glioma progression in vivo
To validate the above-mentioned results in vivo, nude mice were injected with glioblastoma cells overexpressing miR-27a-3p or/and macrophages transfected with lentivirus-mediated CTGF to develop xenograft tumors. The results of RT-qPCR showed higher transfection efficiency of miR-27a-3p with CTGF (Fig. 7A). Results of HE staining exhibited that the boundary of mouse brain tissues was blurred after treatment of miR-27a-3p overexpression, accompanied with elevated invasive ability of glioblastoma cells. While the invasive ability of tumors to brain tissues was weaker in mice after CTGF expression was silenced (Fig. 7B). In addition, the survival time of mice was shortened after in glioblastoma cells overexpressing miR-27a-3p was injected into mice, while the survival time of mice was prolonged after interfering with CTGF (Fig. 7C). Results of Western blotting displayed that overexpression of miR-27a-3p led to downregulated EZH1 and upregulated KDM3A and CTGF.
However, after silencing of CTGF, the expression of EZH1 and KDM3A did not change significantly, while the expression of CTGF was downregulated (Fig. 7D). The secretion levels of IL-10 and TNF-α in peripheral serum of mice were detected by ELISA. The results showed that the overexpression of miR-27a-3p resulted in the increase of IL-10 and the decrease of TNF-α, indicating that M2 polarization was induced in mice. However, the decrease of IL-10 and the increase of TNF-α were observed after silencing of CTGF, suggesting that loss of CTGF could prevent M2 macrophages polarization (Fig. 7E).
These results suggest that glioblastoma-EVs promote the development of glioblastoma in vivo by reprogramming M2 macrophages.

Discussion
EVs activating macrophages is now considered as key players in cancer progression, as macrophages are able to promote the growth of tumors [15]. Moreover, hypoxia condition is likely to stimulate glioblastoma cells to secrete EVs, hence promoting the glioblastoma cell motility [16]. This study investigated the effects and mechanism of miR-27a-3p derived from glioblastoma-EVs on M2 macrophage polarization. We found that glioblastoma-EVs could deliver miR-27a-3p to participate in the polarization of M2 macrophage, thereby playing a key role in tumor microenvironment.
In the present study, we initially revealed that glioblastoma-EVs could deliver miR-27a-3p to polarize M2 macrophage. Coincidentally, EVs derived from monocytes containing miR-27a also play promoting role in the polarization of M2 macrophages [17]. Moreover, miR-27a that promotes proliferation and migration of glioblastoma cells is abundant in glioblastoma-EVs [18]. These works further support our statement that miR-27a-3p contained in glioblastoma-EVs could polarize M2 macrophage.
Furthermore, miR-27a-3p expression was silenced by using plasmid of miR-27a-3p inhibitor. We found that loss of miR-27a-3p could upregulate the protein level of iNOS but downregulate Arg-1. It is reported that iNOS is one of the phenotypes of M1 macrophage while Arg-1 was a marker for M2 macrophage [19]. Therefore, it can be inferred that inhibited miR-27a-3p could curtail the M2 macrophage polarization.
We next found that miR-27a-3p could suppress the expression of EZH1, the target of miR-27a-3p.
EZH1 is reported to participate in the macrophage phenotype shifting and downregulation of EZH1 is able to promote M2 macrophage [13]. To our best knowledge, the relationship between miR-27a-3p and EZH1 is barely explored in previous literature. Furthermore, we found that EZH1 could participate in the regulation of M2 polarization in association with KDM3A. Prior work also demonstrates that EZH1, as a H3K27me3 methylase, binds to the downstream gene promoter and promotes H3K27me3 to suppress gene expression [20], which is in line with our study proposing that EZH1 could inhibit the expression of KDM3A. As prior work verified, M2 macrophage can be induced by KDM3A [21].
Additionally, it is addressed that KDM3A, a demethylase of histone H3K9me1/2, can promote the expression of CTGF by facilitating H3K27ac on the enhancers of CTGF [22]. While CTGF promotes drug-resistance in glioblastoma cells and facilitates the progression of glioblastoma [23]. Based on what has been discussed above, it was indicated that miR-27a-3p secreted from glioblastoma-EVs could downregulate EZH1 expression, elevate expression of KDM3A and further upregulate CTGF to polarize M2 macrophage.
In vitro analysis further unraveled that proliferation, migration and invasion of glioblastoma cells were expedited by overexpressed miR-27a-3p and the underlying mechanism. Promoted M2 macrophage is associated with the enhanced proliferation and migration abilities of glioblastoma cells [24]. To confirm our results in vivo, we also employed murine glioblastoma cell line GL261 for generating xenograft rat models to study the effects of miR-27a-3p/EZH1/KDM3A/CTGF on development of glioblastoma in vivo. The data of our study displayed that overexpressed miR-27a-3p induced upregulated IL-10 and downregulated TNF-α in peripheral serum of mice. Promoted M2 macrophage polarization is related to increased expression of IL-10 and decreased expression of the proinflammatory cytokine TNF-α [25]. Therefore, the results derived from in vitro analysis were consistent with what we had concluded from in vivo experiment.

Conclusions
In summary, the evidence collected from our study collaboratively suggests that miR-27a-3p contained in glioblastoma-EVs inhibits the expression of target gene EZH1 to upregulate KDM3Amediated CTGF expression, which induces M2 macrophage polarization and further facilitates proliferation, migration and invasion of glioblastoma cells (Figure 8). Our findings show that the glioblastoma-EVs miR-27a-3p may be a biomarker for diagnosis of glioblastoma. It is likely that inhibition of miR-27a-3p in glioblastoma-EVs could serve as a tool to combat glioblastoma. Further studies of the molecular mechanisms underlying tumor-associated macrophage via the EVs will facilitate better understanding of the effects of glioblastoma-EVs on the progression of glioblastoma.
EV miR-27a-3p inhibits the expression of target gene EZH1 to upregulate KDM3A-mediated CTGF expression, which induces M2 macrophage polarization and further facilitates proliferation, migration and invasion of glioblastoma cells.