The serum was moved into a new centrifuge tube and centrifuged at 2000×g for 30 min at 4 °C. The supernatant was carefully transferred to a new centrifuge tube and centrifuged at 10,000×g and 4 °C for 45 min to remove larger vesicles. The supernatant was removed and filtered through a 0.45 μm filter membrane (HAWG04700, Sigma-Aldrich.cn) to collect the filtrate, which was transferred to a new centrifuge tube and centrifuged at 100,000×g for 70 min at 4 °C with a superspeed rotor selected. After removal of the supernatant and resuspension with 10 mL of precooled Phosphate Buffered Saline (PBS), the rotor was selected and ultra-centrifugated at 100,000×g at 4 °C for 70 min. The precipitates were collected as exosomes, which were resuspended in 800 μL of precooled PBS for subsequent transmission electron microscopy (Olympus TM3030Plus, Japan), particle size identification, and miRNA high-throughput sequencing. This study has been approved by the Ethics Committee (Zhongshan Traditional Chinese Medicine Hospital Ethics Committee), and all patients have signed informed consent.

Exosomes were extracted from three types of serum: serum from 9 healthy subjects (NC’exo), serum from 9 Epstein–Barr Virus (EBV) infected patients (EBV’exo), and serum from 9 NPC patients (NPC’exo). Exosomes obtained by ultracentrifugation were added to the lysate to extract RNA, and a small RNA sequencing library was constructed. High-throughput sequencing of miRNA was performed on the Illumina platform. The differentially expressed miRNAs were screened and verified by quantitative Polymerase Chain Reaction (qPCR).

RNAs were extracted from cells using the Trizol Reagent (CW0580S, CWBIO, China), followed by obtaining the total miRNA using the miRNA extraction kit (CW0627S, CWBIO, China) and the total mRNA using the mRNA extraction kit (CWBIO, China). Total miRNA and mRNA were transcribed into micDNA and cDNA using the miRNA cDNA Synthesis Kit (CW2141S, CWBIO, China) and the HiScript II Q RT SuperMix for qPCR (+gDNA wiper) (R223-01, Vazyme, China), respectively. The PCR reaction was conducted with the miRNA Universal SYBR qPCR Master Mix (MQ101-02, Vazyme, China) and ChamQ Universal SYBR qPCR Master Mix Q711-02, Vazyme, China). Real-time fluorescent quantitative PCR instrument Gene amplification instrument system (ABI 7500, ThermoFisher Scientific) was used for PCR detection. The gene levels were checked utilizing the 2−ΔΔCt method. Sequences of primers were shown in Table 1. U6 and β-actin were utilized as the internal reference for miRNA and mRNA, respectively.

The BCA kit (Cwbio, Jiangsu, China) was utilized to quantify the protein isolated from cells, followed by being separated with the 12% SDS-PAGE. The separated protein was transferred from the gel to the Polyvinylidene Difluoride Membrane (PVDF) membrane, which was further introduced with 5% skim milk. Then, the membrane was introduced with the primary antibody against BAZ2A (1:1000, DF3858, Affinity, USA), Interleukin (IL)-1β (1:1000, AF5103, Affinity, USA), Monocyte Chemotactic Protein 1 (MCP1, 1:1000, DF7577, Affinity, USA), Nuclear transcription Factor-κB (NF-κB, 1:1000, 66535-1-Ig, Proteintech, USA), Vascular Endothelial Growth Factor A (VEGFA, 1:1000, AF5131, Affinity, USA), and GAPDH (1:2000, TA-08, ZSBIO, China). The second antibody (1:2000, ZB-2301, ZSBIO, China) was subsequently added to be incubated for 90 min. Finally, ECL reagent was added to expose the bands, which were further quantified with the Image J software.

Cells were implanted in 96-well plates for 24 h, followed by adding with 10 μL CCK8 solution. After incubating for 2 h, the OD value was detected using the microplate reader (WD-2012B, LIUYI, China).

When the cell density reached more than 90%, 200 μL of the gun tip was used to scratch in each well, followed by discarding the medium and replaced with DMEM incomplete medium. Then the scratch in each well was photographed. Cells were put into the incubator, and the scratch of each well was photographed again after 24 h. According to the scratch condition, the 24 h scratch data and the 0 h scratch data were determined. The corresponding scratch width and migration rate was calculated.

1 × 106 cells were collected and washed by PBS buffer, which were then re-suspended with 300 μL pre-cold 1× Annexin V-FITC binding buffer. Then, cells were introduced with 5 μL Annexin V-FITC reagent and 10 μL PI reagent, followed by 10 min incubation in the dark at room temperature. Lastly, cells were loaded onto the flow cytometry (NovoCyte 2060R, ACEA Biosciences, China) for the analysis of apoptosis.

The medium in the 6-well plate was discarded, and each well was rinsed twice with 1 mL PBS. Then 200 μL EDTA containing trypsin digestion solution was added and digested in 37 °C incubator. The cell suspension was centrifuged at 3000 rpm for 3 min, and the supernatant was discarded. 1 mL PBS was added to each tube, followed by centrifuged at 3000 rpm for 3 min. 1 mL of absolute ethanol precooled at 4 °C was added to each tube and placed in a 4 °C refrigerator for more than 2 h. Samples were centrifuged at 3000 rpm for 3 min and the supernatant was discarded. 1 mL of PBS was added, followed by centrifuged at 3000 rpm for 3 min and discarding the supernatant. 500 μL DNA staining solution was added to be incubated in the dark for 1 h, followed by loaded onto the flow cytometry (NovoCyte 2060R, ACEA Biosciences, China) for the analysis of cell cycle.

The full-length 3′-UTR fragment of the BAZ2A gene was amplified using the BEC genome as a template, and the PCR product was digested, followed by linking the amplified product to the pMIR plasmid. After transforming the amplified product into DH5α cells, positive clones were screened and sequenced for correction. The correct sequencing plasmids were used for transfection and named pMIR-BAZ2A-WT and pMIR-BAZ2A-MUT. Subsequently, pMIR-BAZ2A-WT or pMIR-BAZ2A-MUT was transfected into HEK293T cells, along with the miR-99a-5p mimic or mimic NC, using Lipofectamine 2000. After 2-day incubation, relative luciferase activity was measured according to the instructions of the Dual-Glo Luciferase Assay System Kit.

Four groups were divided to verify the miR-99a-5p function in NPC: Control, NPC’exo, NPC’exo+ mimic, and NPC’exo+ mimic NC. In the control group, NPC cells were treated with blank medium. In the NPC’exo group, NPC cells were treated with 20 μg/mL NPC’exo, while in the NPC’exo+ mimic group, NPC cells were treated with 20 μg/mL NPC’exo combined with 50 nM miR-99a-5p mimic. In the NPC’exo+ mimic NC group, NPC cells were treated with 20 μg/mL NPC’exo combined with 50 nM mimic NC.

Mean ± Standard Deviation (SD) was utilized to present data, which was analyzed using the one-way ANOVA method with the software of GraphPad prism software 6.0; p < 0.05 was considered to be a statistically significant difference.

Exosomes are derived from multivesicular bodies in cells, which are membrane vesicles with a diameter of about 30–150 nm secreted by living cells. The density of exosomes is 1.13–1.19 g/mL with a typical “cup and plate” shape.

As shown in Fig. 1A‒B, exosomes extracted from each group showed a typical “cup and plate” morphology, and the particle size was mainly distributed between 50 and 120 nm, which was consistent with the basic characteristics of exosomes.

Figure 1.

The extracted exosomes were identified. (A) The morphology of exosomes was imaged using the transmission electron microscopy. (B) The particle size of exosomes was detected by the Nanoparticle Tracking Analysis (NTA).

(0.44MB).

MiR-99b-5p and miR-8085 were selected as candidate genes by high-throughput sequencing analysis Fig. 2A, which were subsequently validated by qPCR. As shown in Fig. 2B‒C, compared to control, miR-99a-5p and miR-8085 were significantly upregulated in EBV’exo (miR-99a-5p expression, 18.02 ± 4.82; miR-8085 expression, 5.63 ± 2.33; n = 3), and NPC’exo (miR-99a-5p expression, 259.72 ± 23.80; miR-8085 expression, 5.14 ± 1.64; n = 3), in which miR-99a-5p showed the most significant differential expression. Therefore, miR-99a-5p was selected for subsequent assays.

Figure 2.

Mir-99a-5p and miR-8085 were significantly highly expressed in EBV’exo and NPC’exo. (A) Screening of differentially expressed miRNAs in exosomes from healthy subjects’ sera (NC’exo), EBV-infected patients’ sera (EBV’exo) and NPC patients’ sera (NPC’exo) by high-throughput sequencing of miRNAs. (B‒C) RT-qPCR was used to detect miR-99a-5p (B) and miR-8085 (C) expression levels in NC’exo, EBV’exo and NPC’exo (*p < 0.05 vs. control).

(0.3MB).

As shown in Fig. 3A‒B, in both 5‒8F and NPC/HK1 cells, compared to control, the cell viability and migration rate were extremely elevated by the introduction of EBV’exo and NPC’exo. Furthermore, higher cell viability (EBV’exo group 107.59 ± 2.87 and NPC’exo group 113.21 ± 2.08 in 5‒8F cell, compared to the control group 100 ± 3.1; EBV’exo group 108.41 ± 1.69 and NPC’exo group 112.41 ± 3.12 in NPC/HK1 cell, compared to the control group 100 ± 3.12; n = 5) and migration rate (EBV’exo group 23.62 ± 2.42 and NPC’exo group 24.29 ± 4.21 in 5‒8F cell, compared to the control group 18.78 ± 1.64; EBV’exo group 20.39 ± 3.34 and NPC’exo group 23.00 ± 3.94 in NPC/HK1 cell, compared to the control group 10.92 ± 3.21; n = 3) were observed in the NPC’exo compared to the EBV’exo group. As shown in Fig. 3C, in NPC/HK1 cells, compared to control, the apoptotic rate was greatly reduced by the introduction of EBV’exo (4.92 ± 0.36; n = 3) and NPC’exo (5.96 ± 0.34; n = 3). Moreover, no significant difference was observed among different groups regarding the cell cycle.

Figure 3.

The impact of different exosomes on the biofunction of NPC cells. (A) The cell viability was detected by the CCK-8 assay. (B) The migration ability of NPC cells was evaluated by the wound healing assay. (C) The apoptosis and cell cycle were detected by the flow cytometry (*p < 0.05 vs. control).

(0.88MB).

As shown in Fig. 4, compared with the control group, the expression levels of BAZ2A (EBV’exo group 0.97 ± 0.1 and NPC’exo group 1.29 ± 0.07 in 5‒8F cell; EBV’exo group 0.80 ± 0.02 and NPC’exo group 1.09 ± 0.03 in NPC/HK1 cell; n = 3), MCP1 (EBV’exo group 1.10 ± 0.09 and NPC’exo group 1.16 ± 0.05 in 5‒8F cell; EBV’exo group 0.73 ± 0.04 and NPC’exo group 1.02 ± 0.01 in NPC/HK1 cell; n = 3) and VEGFA (EBV’exo group 0.55 ± 0.04 and NPC’exo group 0.68 ± 0.06 in 5‒8F cell; EBV’exo group 0.79 ± 0.04 and NPC’exo group 1.02 ± 0.07 in NPC/HK1 cell; n = 3) protein were significantly up-regulated, and the expression levels of IL-1β (EBV’exo group 0.67 ± 0.05 and NPC’exo group 0.41 ± 0.01 in 5‒8F cell; EBV’exo group 0.56 ± 0.03 and NPC’exo group 0.53 ± 0.02 in NPC/HK1 cell; n = 3) and NF-κB (EBV’exo group 0.79 ± 0.06 and NPC’exo group 0.55 ± 0.03 in 5–8F cell; EBV’exo group 0.86 ± 0.05 and NPC’exo group 0.69 ± 0.03 in NPC/HK1 cell; n = 3) protein were significantly down-regulated in NPC cells treated with EBV’exo and NPC’exo. In addition, the expression levels of BAZ2A, MCP1 and VEGFA protein in NPC’exo group were significantly higher than those in NC’exo group, while the expression levels of IL-1β and NF-κB protein were significantly lower than those in NC’exo group.

Figure 4.

The expression level of BAZ2A, IL-1β, MCP1, NF-κB, and VEGFA in NPC cells treated by different exosomes was detected by the Western blotting assay (*p < 0.05 vs. control).

(0.57MB).

As shown in Fig. 5, compared to the BAZ2A-WT + NC group (590.00 ± 41.59; n = 9), dramatically declined fluorescence intensity was observed in the BAZ2A-WT + miR-99a-5p mimic group (398.56 ± 75.50; n = 9). However, no significant difference on the fluorescence intensity was observed between the BAZ2A-WUT + NC group (520.11 ± 38.61; n = 9) and BAZ2A-WUT miR-99a-5p mimic groups (543.56 ± 64.33; n = 9). These data suggested that miR-99a-5p specifically targeted BAZ2A.

Figure 5.

The binding between miR-99a-5p and BAZ2A was identified by the dual-luciferase assay (*p < 0.05 vs. BAZ2A-WT).

(0.23MB).

To overexpress miR-99a-5p in NPC cells, cells were transfected with miR-99a-5p mimic, with mimic NC as a negative control. As shown in Fig. 6, in both 5‒8F and NPC/HK1 cells, compared to mimic NC (1.00 ± 0.03 in 5‒8F cell, 1.00 ± 0.09 in NPC/HK1 cell; n = 3), the level of miR-99a-5p (4610.04 ± 669.21 in 5‒8F cell, 7054.10 ± 501.77 in NPC/HK1 cell; n = 3) was extremely elevated by the transfection of miR-99a-5p mimic, suggesting a successful overexpression of miR-99a-5p in NPC cells.

Figure 6.

The transfection efficacy of miR-99a-5p in NPC cells was identified by the RT-PCR assay (*p < 0.05 vs. mimic NC).

(0.22MB).

As shown in Fig. 7A, in both 5–8F and NPC/HK1 cells, compared to control (100.00 ± 2.51 in 5‒8F cell, 100.00 ± 1.43 in NPC/HK1 cell; n = 5), the cell viability was extremely elevated in the NPC’exo (108.54 ± 3.59 in 5‒8F cell, 112.69 ± 3.08 in NPC/HK1 cell; n = 5), which was greatly repressed in the NPC’exo+ mimic group (98.66 ± 2.82 in 5‒8F cell, 89.83 ± 3.97 in NPC/HK1 cell; n = 5), suggesting that the effect of miR-99a-5p on the proliferation of NPC cells was partly abolished by miR-99a-5p.

Figure 7.

The verification of miR-99a-5p function in NPC cells. (A) The cell viability was detected by the CCK-8 assay. (B) The level of BAZ2A in NPC cells was measured by the RT-PCR assay (*p < 0.05 vs. control, #p < 0.05 vs. NPC’exo).

(0.47MB).

As shown in Fig. 7B, in both 5–8F and NPC/HK1 cells, compared to NPC’exo (1.12 ± 0.05 in 5‒8F cell, 1.14 ± 0.21 in NPC/HK1 cell; n = 3), the BAZ2A level was dramatically reduced in the NPC’exo+ mimic group (0.92 ± 0.01 in 5‒8F cell, 0.84 ± 0.09 in NPC/HK1 cell; n = 3), suggesting that BAZ2A was downregulated by the overexpression of miR-99a-5p.