Infection of Epstein–Barr Virus in Type III Latency Modulates Biogenesis of Exosomes and the Expression Profile of Exosomal miRNAs in the Burkitt Lymphoma Mutu Cell Lines

Infection of Epstein–Barr virus (EBV), a ubiquitous human gamma herpesvirus, is associated with various malignancies in B lymphocytes and epithelial cells. EBV encodes 49 microRNAs in two separated regions, termed the BART and BHRF1 loci. Although accumulating evidence demonstrates that EBV infection regulates the profile of microRNAs in the cells, little is known about the microRNAs in exosomes released from infected cells. Here, we characterized the expression profile of intracellular and exosomal microRNAs in EBV-negative, and two related EBV-infected Burkitt lymphoma cell lines having type I and type III latency by next-generation sequencing. We found that the biogenesis of exosomes is upregulated in type III latently infected cells compared with EBV-negative and type I latently infected cells. We also observed that viral and several specific host microRNAs were predominantly incorporated in the exosomes released from the cells in type III latency. We confirmed that multiple viral microRNAs were transferred to the epithelial cells cocultured with EBV-infected B cells. Our findings indicate that EBV infection, in particular in type III latency, modulates the biogenesis of exosomes and the profile of exosomal microRNAs, potentially contributing to phenotypic changes in cells receiving these exosomes.


Introduction
Epstein-Barr virus (EBV), a human gamma herpesvirus, contributes causally to lymphomas and epithelial malignancies, such as Burkitt lymphoma (BL), Hodgkin lymphoma (HL), nasopharyngeal carcinoma (NPC), gastric carcinoma (GC), and post-transplant lymphoproliferative disorders (PTLD) [1]. The expression pattern of EBV-encoded latent genes defines three latency types specific to individual EBV-associated tumors [1]. Latency type I is associated with BL and GC and is restricted to the expression of EBV-encoded nuclear antigen 1 (EBNA1), the EBV-encoded small RNAs (EBERs), and BamHI A rightward transcripts (BARTs). Latency type II, which is associated with HD and and the expression profile of exosomal miRNAs, which may contribute to phenotypic changes in recipient cells.

Infection with EBV Is Sufficient to Accelerate the Biogenesis of Exosomes in Mutu III Cells
To perform NGS analysis for exosomal miRNAs, we purified exosomes released from EBV-negative (Mutu − ), EBV-infected of type I (Mutu I), or type III latency (Mutu III) [exosome (−), exosome (I), or exosome (III), respectively] by ultracentrifugation through a sucrose cushion. Because some BL cell lines support the viral lytic cycle [51], we tested for EBV virions in the isolated exosomal fractions by electron microscopy ( Figure 1A) and real-time PCR for EBV-encoded DNA polymerase BARF5 gene ( Figure 1B,C) and confirmed their absence. We also measured the presence of exosome markers, such as CD63 [52] and ALG-2-interacting protein X (Alix) [53] in Mutu cells and exosomal fractions by western blot analysis. We observed the highest expression of CD63 and Alix in Mutu III and exosome (III) (Figure 2A). In contrast, exosome fractions were negative for calnexin (an endoplasmic reticulum marker) and GM130 (a Golgi marker), which are commonly used as negative controls of extracellular vesicles [54].

More EBV miRNAs Are in Mutu III Cells and Exosomes (III) Than in EBV-Negative and Type I Latently Infected Cells and Their Exosomes
To determine the effect of EBV infection on miRNA expression profile, a small RNA-enriched fraction was isolated from cells and exosomes, respectively. The isolated RNAs were then subjected to NGS, which yielded 119,769, 26,388, and 1,559,462 reads for RNAs isolated from Mutu − , Mutu I and Mutu III, respectively. On the other hand, 66,018, 535,527 and 2,445,816 reads were obtained from the RNAs isolated from exosome (−), exosome (I), and exosome (III), respectively. The alignment of the sequences to the reference genome resulted in the annotation of 12.64-67.34% of cellular miRNA reads and 44.11-80.96% of exosomal miRNA reads (Table 1). Because exosomes are known to contain a substantial amount of tRNA and its fragments, the sequences were also aligned to the human-tRNA database. 2.1-4.65% of cellular RNAs and 2.33-3.97% of exosomal RNAs were characterized as human tRNAs ( Table 1). The sequences were also aligned to miRBase. 7.14% and 3.85% of the annotated cellular and exosomal miRNAs, respectively, originated from EBV in Mutu I cells. In Mutu III, 14.44% of annotated miRNAs in the cells originated from EBV, whereas the 8.07% of annotated exosomal miRNAs were derived from EBV (Table 1). These results indicate that more EBV miRNAs are in Mutu III cells and exosomes (III) than in EBV-negative and type I latently infected cells and their exosomes.

Exosome-Mediated Transfer of EBV miRNAs to Target Epithelial Cells
Finally, we examined whether EBV miRNAs expressed in Mutu cells are transferred to recipient cells via exosomes. We cocultured Mutu cells with EBV-negative human GC AGS cell lines separately through a trans-well for 3 days. Total RNA was isolated from cocultured AGS cells and subjected to stem-loop real-time RT-PCR to quantify the copy numbers of miR-BHRF1-1, miR-BART7, miR-BART10, and miR-BART15. As expected, no EBV miRNAs were detected in uncocultured AGS cells and AGS cells cocultured with Mutu − cells ( Figure 5). miR-BHRF1-1 and miR-BART10 were predominantly detected in AGS cells cocultured with Mutu III (Figure 5A,C). miR-BART7 was detected more efficiently in AGS cells cocultured with Mutu I cells relative to those cocultured with Mutu III ( Figure 5B). Little miR-BART15 was observed in the cocultured recipient cells ( Figure 5D). The frequency of the expression of miRNAs in the recipient cells was consistent with their expression levels in the exosomes, as determined by NGS (Figure 3). To confirm the importance of the exosome secretion from Mutu cells, the cocultured cells were treated with GW4869, an inhibitor of sphingomyelinase that markedly reduces exosome secretion [60]. GW4869 inhibited transfer of BART10 to AGS cells cocultured with Mutu III (Figure 5E), indicating that transfer of EBV miRNAs to recipient cells was mediated by exosomes. Taken together, these data indicate that type III latency of EBV infection modulates the expression profile of host and viral miRNAs in cells and exosomes, which may contribute to induction of EBV-associated malignancies or other pathologies in the recipient cells.
predominantly detected in AGS cells cocultured with Mutu III (Figure 5A,C). miR-BART7 was detected more efficiently in AGS cells cocultured with Mutu I cells relative to those cocultured with Mutu III (Figure 5B). Little miR-BART15 was observed in the cocultured recipient cells ( Figure 5D). The frequency of the expression of miRNAs in the recipient cells was consistent with their expression levels in the exosomes, as determined by NGS (Figure 3). To confirm the importance of the exosome secretion from Mutu cells, the cocultured cells were treated with GW4869, an inhibitor of sphingomyelinase that markedly reduces exosome secretion [60]. GW4869 inhibited transfer of BART10 to AGS cells cocultured with Mutu III (Figure 5E), indicating that transfer of EBV miRNAs to recipient cells was mediated by exosomes. Taken together, these data indicate that type III latency of EBV infection modulates the expression profile of host and viral miRNAs in cells and exosomes, which may contribute to induction of EBV-associated malignancies or other pathologies in the recipient cells.

Discussion
Here, we have characterized the expression profile of cellular and exosomal miRNAs derived from cell lines originating from the same African BL patient with different states of EBV-infection by next-generation sequencing.
Both the formation of MVBs and the biogenesis of exosomes were upregulated in Mutu III cells, which have only a low number of EBV genomes (Figure 2). Hurwitz and colleagues demonstrated that CD63 plays a critical role in LMP1-mediated enhancement of exosome production [58]. The same group recently observed that CD63 coordinates the autophagic and endosomal pathways to regulate LMP1-mediated signals and secretion of exosomes [61]. Previously we demonstrated that EBV-infected cells require a certain threshold number of EBV genomes for their optimal growth under selection [62], suggesting that maintenance of limited copy numbers of EBV is sufficient to accelerate LMP1-mediated exosome production.
Exosome (III) contain more viral miRNAs than exosome (−) and exosome (I). Moreover, multiple specific cellular miRNAs were predominantly incorporated into exosomes (III) ( Table 3). Although EXOmotifs were frequently identified in the highly concentrated miRNAs in exosome (III) ( Table 4), the numbers of EXOmotifs varied among these miRNAs and no significant correlation was found between sorting efficiency of miRNAs to the exosomes and their number of EXOmotifs, suggesting that EXOmotifs-independent mechanism(s) for sorting of miRNA to exosomes are likely involved. For instance, Kosaka et al. demonstrated that the neural sphingomyelinase 2 (nSMase2) upregulates the efficiency of sorting of miRNAs to the exosomes [63]. Other studies suggest a possible mechanism involving miRNA sorting in a miRNA 3 end nucleotide or miRNA induced silencing complex (miRISC)-dependent manner [37].
Multiple specific cellular miRNAs, such as miR-143, miR-877, miR-4516-5p, miR-6087-5p, and miR-7704-5p were incorporated into exosome (III) ( Table 3). miR-143 has been characterized as a tumor-suppressive factor by targeting several oncogenes, including Kirsten rat sarcoma viral oncogene homolog (KRAS) and extracellular signal-regulated kinases 5 (ERK5) [64]. Two independent reports demonstrate a role for miR-877 as a tumor suppressor in renal cell carcinoma by targeting eukaryotic elongation factor-2 kinase (eEF2K), and in myofibroblast differentiation and bleomycin-induced lung fibrosis by targeting targets Smad7 [65,66]. miR-4516 has been shown to down regulate the STAT3-signaling pathway, which results from the induction of UV-induced apoptosis in keratinocytes [67]. miR-6087 is incorporated in human adipose mesenchymal stem cell-derived exosomes, which exhibited an anti-proliferative effect for ovarian cancer cell lines [68]. miR-7704 is upregulated in macrophages treated with a combination of interleukin-27 (IL-27) and macrophage colony stimulating factor. The same study identified putative targets genes for miR-7704, such as GF0D1, the membrane-associated RING-CH3 (MARCH3), and the Src homology 2 domain containing transforming protein C3 (SHC3). miR-7704 has been shown to target the open reading frames of a number of viruses, including Herpes simplex virus (HSV)-1, HSV-2, and human herpesvirus (HHV)-8, suggesting that miR-7704 may partially contribute to the anti-viral properties of IL-27 against these viruses [69]. Although the functional properties of these miRNAs in EBV's lifecycle remain unclear, EBV may exploit the mechanism of active sorting of these miRNAs, which possess tumor suppressive, anti-viral, or anti-apoptotic functions, into exosomes to induce EBV-associated tumors. Moreover, these unique miRNAs could also lead to potential biomarkers, enabling earlier and more accurate diagnoses and accelerating the development of therapeutics for EBV-associated tumors. It is also essential to perform NGS for other sets of BL cell lines to confirm the generality of our findings.
Previously we demonstrated that exosomes derived from all three Mutu cells were internalized into the target cells via caveolae-dependent endocytosis in a similar fashion. Moreover, we observed that exosomes (III) up-regulated proliferation and expression of intercellular adhesion molecule 1 (ICAM-1) in the recipient cells more significantly than exosome (−) and exosome (I). We also identified LMP1 as a gene responsible for induction of ICAM-1 expression [56].
Taken together, our findings indicate that EBV infection, in particular in type III latency, modulates the biogenesis of exosomes and expression profile of exosomal miRNAs, which along with LMP1 may contribute to the induction of EBV-associated tumors by modulating cell and virus functions.

Cell Culture
Mutu I and Mutu III cells, which are type I and type III latency EBV-infected B cell lines, respectively, were established from the identical BL tumor [70]. EBV-negative subclone (Mutu − ) was isolated from Mutu I by the limiting dilution methods [71]. BJAB cell is an EBV-negative BL cell line [72]. Namalwa cell is a BL cell line possessing two copies of EBV DNA integrated into host genome [73,74]. Mutu − , Mutu I, Mutu III, BJAB, and Namalwa cells were maintained in RPMI-1640 medium containing 10% fetal bovine serum (FBS) and antibiotics. EBV-negative human GC epithelial AGS [75,76] were grown in high-glucose Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS and antibiotics.

Electron Microscopy
A 6-µL aliquot of the fractions containing exosomes was absorbed onto glow-discharged 300-mesh heavy-duty carboncoated Cu grids (Veco grids, Nisshin EM, Tokyo, Japan) for 2 min and the excess was blotted on filter paper (Whatman, GE Healthcare, Piscataway, NJ, USA). The grids were then washed twice with MilliQ water and negatively stained with 2% uranyl acetate. Data were collected using an HT7700 transmission electron microscope (Hitachi, Tokyo, Japan) operating at 80 kV and 10,000× magnification.

Quantitative PCR
A 10 µL aliquot of exosomes was treated with 5 U DNase at 37 • C for 15 min. DNase was inactivated by incubation at 70 • C for 5 min in the presence of 2.5 mM EDTA. DNA was extracted by use of the DNeasy Blood & Tissue Kit (Qiagen Hilden, Germany). As positive control, progeny EBV was obtained by treatment of EBV-infected Akata cells (Akata + ) with 1% goat anti-human IgG (αhIgG) (DAKO, Glostrup, Denmark) for 48 h [51,77,78]. After treatment with DNase I, EBV DNA in the supernatant was isolated as described above. For detection of EBV DNA, real-time PCR was carried out as previously described [29,79] with slight modifications by means of sense (CGGAAGCCCTCTGGACTTC) and antisense (CCCTGTTTATCCGATGGAATG) oligonucleotides and probe (TGTACACGCACGAGAAATGCGCC) specific for the EBV-encoded BARF5 sequence.
As an internal control, cellular Rhodopsin was analyzed using sense (ATCAGGAACCATTGCCACGTCCTA) and antisense (AGGCCAAAGATGGACACACAGAGT) oligonucleotides and probe (AGCCTCTAGTTTCCAGAAGCTGCACA). The copy number of EBV genomes were quantified with a standard curve of dilution of the plasmid encoding EBV BARF5 gene followed by the normalization by the number of cellular genomes quantified against a standard curve of dilution of the plasmid encoding Rhodopsin.

Fluorescent In Situ Hybridization (FISH)
FISH analysis was performed as described previously [62,80]. Briefly Mutu cells were treated with 0.075 M KCl for 20 min at 37 • C, fixed in methanol:acetic acid (3:1) for 30 min at room temperature and spread on the slides. Slides were treated with 4 × SSC (1 × SSC; 0.15 M NaCl, 0.015 M sodium citrate) containing 0.5% (v/v) nonidet P-40 for 30 min at 37 • C, dehydrated in a cold ethanol series (70, 80, 90%) for 2 min each, air dried and denatured in 70% formamide-2 × SSC for 2 min at 72 • C. Slides were dehydrated in a cold ethanol series and air dried. Hybridization probes for detection of EBV genomes were generated by nick translation with the plasmid containing EBV BamHI-WWYH fragment using biotin-11-dUTP (Roche, Basel, Switzerland). 20 µg of a probe was precipitated by ethanol in the presence of 6 µg salmon sperm DNA (Eppendorf, Hamburg, Germany) and 4 µg human Cot-1 DNA (Thermo Fisher Scientific, Waltham, MA, USA), resuspended in hybridization buffer (2 × SSC, 50% formamide, 10% dextran sulfate) and incubated for 10 min at 70 • C, for 5 min at 4 • C and for 1 h at 37 • C. A hybridization mix containing 5 ng probe was placed on each sample and incubated overnight at 37 • C in a moist chamber. Slides were washed in 2 × SSC containing 50% formamide for 30 min at 50 • C and in 2 × SSC for 30 min at 50 • C. After blocking in 4 × SSC containing 1% BSA, the hybridized probe was revealed by incubation with streptavidin conjugated to FITC (Sigma-Aldrich, St. Louis, MO, USA) for 30 min at 37 • C. Slides were washed twice in 4 × SSC containing 0.05% Triton X-100 for 5 min at room temperature. The nucleus was counterstained with Hoechst 33342 (Cell Signaling Technology). The copy numbers of EBV genome were analyzed by a confocal laser scanning microscope. (Fluoview FV10i, Olympus, Tokyo, Japan) and acquired by using FV10-ASW software Ver. 4.2 (Olympus, Tokyo, Japan).

Immunofluorescence Staining
Mutu cells were fixed with 4% paraformaldehyde in PBS for 10 min at room temperature, permeabilized with PBS containing 0.05% Triton X-100 for 10 min at room temperature and blocked in PBS containing 1% bovine serum albumin (BSA) for 20 min at room temperature. The cells were incubated with rabbit anti-CD63 polyclonal antibody (Abcam, Cambridge, UK, 1:200 dilution) and anti-Alix monoclonal antibody (BioLegend, 1:200 dilution) for 1 h at room temperature. After washing twice in PBS, the cells were incubated with Alexa Fluor 488-labeled anti-mouse IgG (Thermo Fisher Scientific, 1:1000 dilution) for 1 h at room temperature. After washing twice in PBS, the nuclei were counterstained with Hoechst 33342. Images were collected by a confocal laser scanning microscope. Same samples were quantitatively analyzed by means of flowcytometry.

RNA Extraction
Total RNA was extracted from cells or exosomes using the mirVana miRNA Isolation Kit (Thermo Fisher Scientific). The extracted RNA was analyzed using small RNA chips on a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Small RNA profiles contained small RNA consists on between 4 and 40 nucleotides in length, which is consistent with miRNA.

Next-Generation Sequencing (NGS)
Small RNA library for sequencing was prepared by the TruSeq Small RNA-Seq Sample Prep Kit (Illumina, San Diego, CA, USA). The quality and yield after sample preparation were analyzed by use of the Bioanalyzer with a High Sensitivity DNA kit (Agilent Technologies) and corresponded to the expected 150 bp. DNA sequencing was carried out using Miseq (Illumina) with MiSeq reagent kit v3. Sequence reads were analyzed with CLC Genomics Workbench (version 11.0; Qiagen).
After adaptor trimming, reads of less than 15 or more than 26 nucleotides in length were removed, and all reads of 15-25 nucleotides in length were analyzed against miRBase release 21 retrieved from the miRNA database (http://www.mirbase.org/). Homo_sapiens.GRCh37.57.ncrna was used as a comprehensive noncoding RNA database (http://www.ncrna.org/). Human tRNA was annotated using GRCh37/hg19 tRNA database (GtRNAdb, http://gtrnadb.ucsc.edu/). Reads matching pre-miRNA were counted as miRNA reads. The ratio of read numbers of mature miRNAs in total annotated miRNAs were analyzed between the samples. NSG analysis was duplicated for intracellular miRNAs derived from Mutu cells, and exosomal miRNAs from Mutu III, and triplicated for exosomal miRNAs derived from Mutu − and Mutu I.

Exosome Transfer Assay
Exosome transfer assay was performed as described previously [56]. Briefly, AGS cells (5 × 10 4 /well) were grown in the basolateral chamber of 24-well trans-well plate (Corning, Toledo, NY, USA). Mutu − , Mutu I, or Mutu III (1 × 10 5 , each) were added to the membrane inserts with pore size of 0.4 µm and incubated for 3 days. For the inhibitor treatment of exosome secretion, DMSO or 10 µM GW4869 (Sigma-Aldrich) was added to the coculture and incubated for 3 days Total RNA was isolated from AGS cells after cocultivation and subjected to Stem-loop real-time PCR.

Stem-Loop Real-Time PCR
For analysis of copy numbers of individual EBV miRNAs, total RNA was reversed transcribed using TaqMan MicroRNA Reverse Transcription Kit (Thermo Fisher Scientific) followed by real-time PCR reaction with TaqMan Universal Master Mix, no UNG AmpErase (Thermo Fisher Scientific) as described previously [29,81].

Accession Number
Sequence data of the small RNAs analyzed by NGS in this study were deposited in the DNA Data Bank of Japan (DDBJ; accession number: DRA006656, Bioproject PRJDB6853).

Conclusions
Here, we characterized the expression profile of exosomal miRNAs derived from EBV-negative, type I and type III latently infected cell lines, which were established from the same African BL tumor. Taken together, our findings indicate that EBV infection in type III latency modulates the biogenesis of exosomes and the profile of exosomal miRNAs, potentially contributing to induction of EBV-associated tumors by phenotypic modifications in adjacent recipient cells.