The human laryngeal squamous carcinoma cell line AMC-HN-8 which was established by Kim et al in 1997 (18) from patients with head and neck cancer was preserved in our laboratory. Laryngeal squamous carcinoma cell lines Tu212 and Tu686 were obtained from the Central South University (Hunan, China). All 3 cell lines are representative in vitro models for studying the biology of head and neck carcinoma. All cells were cultured in RPMI-1640 (HyClone; GE Healthcare Life Sciences, Logan, UT, USA), 1% penicillin-streptomycin (Genom Biotechnology, Hangzhou, China) and 10% exosome-depleted fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) in humidified air with 5% CO₂ at 37°C.

The generation of exosome-depleted FBS was carried out by ultracentrifugation to reduce contamination from bovine exosomes. Briefly, centrifuge tubes were loaded with FBS and centrifuged at 120,000 × g for 6 h at 4°C (Beckman Coulter Optima L-100XP Ultracentrifuge, SW 32Ti; Beckman Coulter, Inc., Fullerton, CA, USA). The supernatant of FBS was then filtered using a 0.22-µm filter (Merck KGaA, Darmstadt, Germany).

Cells were cultured in conditioned medium containing 10% exosome-depleted FBS for 72 h at 90–100% density. The conditioned medium was harvested and centrifuged at 2,000 × g for 10 min followed by 10,000 × g for 30 min at 4°C. Then, the supernatant was filtered using a 0.22-µm filter to eliminate cellular debris thoroughly and concentrated using centrifugal ultrafiltration (Amicon^® Ultra-15 100 KDa; Merck KGaA) to minimize the potential contamination, respectively.

Exosomes were isolated from processed conditioned medium with Ribo™ Exosome Isolation Reagent (Guangzhou RiboBio Co., Ltd., Guangzhou, China), according to the manufacturer's instructions. Briefly, processed conditioned medium was mixed with Ribo™ Exosome Isolation Reagent at a ratio of 3:1 and incubated at 4°C overnight. After centrifugation at 1,500 × g for 30 min, exosomes pellets were resuspended in appropriate phosphate-buffered saline (PBS; HyClone; GE Healthcare Life Sciences) for TEM, NTA and further research. The exosome pellets were used immediately or stored at −80°C until use.

The morphology of exosome pellets was examined by TEM. Briefly, a 20-µl of exosome-PBS solution drop was loaded onto carbon-coated copper grids and permitted to stand for 1 min. The filter paper was utilized to remove the excess solution. Then the exosomes pellets were negatively stained with 20 µl uranyl acetate dihydrate (2%) and allowed to stand for 1 min. Subsequently, the filter paper was utilized to remove the excess fluid, again. The sample was allowed to dry under a lamp for 10 min before viewing on an FEI Tecnai G2 Spirit transmission electron microscope (FEI Company, Hillsboro, OR, USA) operated at 80 kV.

For the purpose of demonstrating the particle size distribution, NTA was performed using NanoSight NS300 (Malvern Instruments, Inc., Westborough, MA, USA), according to the operating instructions. In addition, FCM (BD Accuri C6 flow cytometer; BD Biosciences, Franklin Lakes, NJ, USA) was used to detect exosomal surface markers. Briefly, exosome pellets were resuspended in 100 µl filtered PBS, and then 20 µl immunofluorescence antibody CD63 (CD63-Antibody-FITC; cat. no. 557288) and CD81 (CD81-Antibody-FITC; cat. no. 551108; both from BD Biosciences) were added to the exosome-PBS solution. Following incubation, FCM was utilized to detect exosomal surface markers.

Total RNA in cells and exosomes was extracted using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) as previously described (14). The quality and concentration of RNA were assessed using NanoDrop 2000 Spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc.). RNA content was assessed using an Agilent 2200 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA). Total RNA was sequenced or reverse-transcribed to cDNA immediately for further research.

Small RNA library preparation and sample sequencing were accomplished with the assistance of Guangzhou RiboBio Co., Ltd., using an Illumina HiSeq™ 2500 device. Total RNA from LSCC AMC-HN-8 cells (n=3) and exosomes (n=3) were concatenated with 5′- and 3′-adaptors. After cDNA synthesis and PCR amplification, the cDNA library (18–40 nt) was obtained using an acrylamide gel purification method. Then, single end (SE) sequencing was performed: 1×50 bp.

To validate the results of sequencing, we selected 6 miRNAs (miR-1246, miR-122-5p, miR-320d, miR-4483, miR-30c-5p and let-7e-5p) for further research by RT-qPCR in all 3 cell lines and its exosomes. Total RNA (300 ng) was reversed-transcribed using PrimeScript™ RT reagent Kit (Takara Biotechnology Co., Ltd., Dalian, China) according to the manufacturer's instructions. The expression level of miRNA was measured using SYBR^® Premix Ex Taq™ (Takara Biotechnology Co., Ltd.) following the manufacturer's instructions. We used miR-93-5p as an internal control, which has a stable expression level between cells and exosomes, to normalize the relative expression ratio of miRNA, using the 2^−ΔΔCq method (19). An ABI 7500 PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.) was used to perform the RT-qPCR and analyze the data. The thermocycling conditions were as follows: holding stage: 95°C for 30 sec; cycling stage: 95°C for 5 sec and 60°C for 34 sec, 40 cycles; melt curve stage: 95°C for 15 sec, 60°C for 1 min and 95°C for 15 sec. The primers used are listed in Table I.

The target genes of selective exosomal miRNAs were predicted using TargetScan (http://www.targetscan.org/mamm_31/), miRDB (http://www.mirdb.org/), miRanda (http://www.microrna.org/microrna/home.do) and StarBase (http://starBase.sysu.edu.cn/). Only target genes that were found by 3 of the 4 tools were identified to be the target genes of exosomal miRNAs. The comprehensive function annotations of the potential targets of selective exosomal miRNAs were performed with Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis based on the DAVID 6.7 software (http://david.abcc.ncifcrf.gov/home.jsp). All GO terms and signaling pathways were analyzed with a threshold of significance that was defined by a P-value of <0.05.

Our first step was to capture and verify the exosomes derived from LSCC cell line AMC-HN-8. Exosomes were isolated from conditioned medium of AMC-HN-8 using Ribo™ Exosome Isolation Reagent. After isolation, the morphology of exosome pellets was examined by TEM revealing membrane-bound particles that were homogeneous in appearance (Fig. 1A). In addition, the size distribution of exosome pellets was analyzed using NTA and the results revealed prospective diameters which ranged from 30–150 nm (Fig. 1B). Furthermore, we assessed two well-established surface markers for exosomes, CD63 and CD81, using flow cytometry (Fig. 1C). The results revealed a significant difference after incubation with CD63 and CD81 antibodies. These results indicated that these pellets which we isolated had the size (30–150 nm), specific surface markers and morphology typical of that of exosomes.

To identify the miRNA expression profiles of parental cells and exosomes to investigate which miRNAs are selectively enclosed in exosomes, we performed small RNA sequencing. First, total RNA was extracted from parental cells (n=3) and exosomes (n=3). A small RNA library preparation and sample sequencing were conducted under the assistance of Guangzhou RiboBio Co., Ltd. The miRNAs were the most common among the known sequences, followed by tRNAs, snRNAs, Y RNAs, piRNAs and other RNAs which can not be classified. Of all the RNA examined, an average of 72.99% miRNAs were detected in parental cells compared with an average of 69.76% miRNAs in exosomes derived from parental cells, which did not exhibit an obvious discrepancy in composition of small RNAs (Fig. 2A). However, only 5.63% miRNAs were common for both cellular miRNAs and exosomal miRNAs, which meant that the process of miRNA packing into exosomes is selective (Fig. 2B). The most abundant miRNA in cells and exosomes was miR-21-5p and miR-1246, respectively (Fig. 2C). miR-21-5p has been previously reported to act as a predictor of recurrence and novel biomarker of urothelial (20) and gastric carcinoma (21). In addition, exosomal miR-1246 induced cell motility and invasion in oral squamous cell carcinoma (22), and functioned as a promising biomarker in peripheral circulation of patients who have breast and colon cancer (23,24). To explore the functions of selective miRNAs in exosomes, we then detected 15 miRNAs downregulated in exosomes and 14 miRNAs upregulated in exosomes as compared to parental cells, with both a P-value of ≤0.05 and >5-fold-changes after filtering (Fig. 3). Detailed data are listed in Table II.

All of these data revealed altered miRNA expression profiles between parental cells and exosomes and indicated that miRNAs are selectively enriched or encapsulated in exosome pellets secreted from AMC-HN-8 cells rather than indiscriminate. These selective miRNAs in exosomes may promote malignant biological properties of tumors and play a crucial role in the cell-cell communication, including tumor cells-tumor cells or tumor cells-stromal cells crosstalk.

To validate the results of small RNA sequencing, RT-qPCR was carried out to detect 3 upregulated miRNAs (miR-320d, miR-1246 and miR-122-5p) and 3 downregulated miRNAs (miR-4483, miR-30c-5p and let-7e-5p) in exosomes derived from 3 cell lines. Since no internal controls for exosomal microRNA analysis have been established, we used miR-93-5p due to its stable expression level between cells and exosomes. As anticipated, the results revealed that the expression level of miR-320d, miR-1246 and miR-122-5p was significantly upregulated, and miR-4483, miR-30c-5p and let-7e-5p was significantly downregulated in exosomes derived from all 3 cell lines, respectively (Fig. 4). These results were consistent with the small RNA sequencing data, which signified that the results of the small RNA sequencing were credible.

Furthermore, GO annotation and KEGG pathway enrichment analysis were performed to explore the potential targets of these 29 selective exosomal miRNAs. The results revealed that a total of 485 GO terms were involved in biological processes, with P<0.05 (Fig. 5A). The top 3 biological processes with the lowest P-value were related to vasculature development, regulation of transcription from RNA polymerase II promoter and negative regulation of macromolecule biosynthetic process. Up to 51 GO terms were in relation to cellular components, with P<0.05. The top 3 cellular components with the lowest P-value were related to nucleoplasm, complex of collagen trimers and extracellular matrix component. There were 88 GO terms related to molecular function in all, with P<0.05. The top 3 molecular functions with the lowest P-value were related to chromatin binding, transcription factor activity, transcription factor binding and SMAD binding.

The KEGG pathway enrichment analysis indicated that 279 pathways were involved in the current small RNA sequencing data, in which many of these pathways were related to organismal systems, cellular processes, environmental information processing and human diseases (Fig. 5B). The top 5 pathways, with the lowest P-value, were the AGE-RAGE signaling pathway in diabetic complications, pathways in cancer, colorectal cancer, the p53 signaling pathway and signaling pathways regulating pluripotency of stem cells.

These results indicated that selective exosomal miRNAs play a distinctly important role in the process of cell metabolism, cell cycle and tumorigenesis.

We then constructed the co-expression network of selective exosomal miRNAs and genes (Fig. 6A). The most common overlay was the gene ‘TNRC6A’ (Fig. 6B), which was previously reported that along with Ago2, was overexpressed in the tissue of prostate and esophageal cancers (25). This network indicated that one single miRNA was bound up with a couple of genes and vice versa.