Epstein-Barr Virus Encoded dUTPase Containing Exosomes Modulate Innate and Adaptive Immune Responses in Human Dendritic Cells and Peripheral Blood Mononuclear Cells

We have recently demonstrated that Epstein-Barr virus (EBV)-encoded deoxyuridine triphosphate nucleotidohydrolase (dUTPase) modulates innate immunity in human primary monocyte-derived macrophages through toll-like receptor (TLR) 2 leading to NF-κB activation and the production of pro-inflammatory cytokines. Our previous depletion studies indicated that dendritic cells (DCs) may also be a target of the EBV-encoded dUTPase. However, the role of EBV-encoded dUTPase in DC activation/function and its potential contribution to the inflammatory cellular milieu characteristic of EBV-associated diseases remains poorly understood. In the present study, we demonstrate that EBV-encoded dUTPase significantly altered the expression of genes involved in oncogenesis, inflammation and viral defense mechanisms in human primary DCs by microarray analysis. Proteome array studies revealed that EBV-encoded dUTPase modulates DC immune responses by inducing the secretion of pro-inflammatory TH1/TH17 cytokines. More importantly, we demonstrate that EBV-encoded dUTPase is secreted in exosomes from chemically induced Raji cells at sufficient levels to induce NF-κB activation and cytokine secretion in primary DCs and peripheral blood mononuclear cells (PBMCs). Interestingly, the production of pro-inflammatory cytokines in DCs and PBMCs was TLR2-dependent. Together these findings suggest that the EBV-encoded dUTPase may act as an intercellular signaling molecule capable of modulating the cellular microenvironment and thus, it may be important in the pathophysiology of EBV related diseases.


Introduction
Epstein-Barr virus (EBV) is a gamma herpesvirus that is implicated in the pathogenesis of a variety of human malignancies including Burkitt's lymphoma (BL), nasopharyngeal carcinoma (NPC), Hodgkin's disease (HD), chronic lymphocytic leukemia (CLL), diffuse large B-cell lymphoma, NK/T-cell lymphoma, and gastric carcinoma [1]. EBV infects a significant percentage (.90%) of the worldwide population and establishes a life-long persistent infection in memory B-cells. However, the various strategies that EBV employs to prevent its clearance and that allow for the establishment and maintenance of a persistent infection in immunocompetent individuals are poorly understood.
Dendritic cells (DCs) are professional antigen presenting cells (APC) and play a pivotal role in regulating the balance between immunological tolerance and immune responses that initiate innate and adaptive immunity. Given the importance of DCs in initiating an immune response against pathogens and the ability of EBV to establish persistent infections in the host, it might be expected that the virus has developed mechanism(s) to regulate the function of DCs as part of the virus strategy to evade immune surveillance. However, the interactions between DCs and EBV remain unclear and conflicting results have been reported concerning the ability of EBV to induce productive infections in DCs. Li et al [2] and Wang et al [3] reported that infection of monocytes by EBV results in apoptosis, thus, preventing their differentiation into DCs. Conversely, Walling et al [4] demonstrated that EBV established a latent infection in blood-borne mononuclear cells, which are precursors of Langerhans cells (LC) and that upon migration and differentiation into LC in the epithelium, EBV is reactivated establishing a productive (lytic) infection. These conflicting results suggest that while EBV may be able to infect monocytes, the effect of EBV on these cells varies with some precursors undergoing apoptosis, while infection of other precursors leads to the establishment of either a latent or productive infection.
The EBV-encoded deoxyuridine triphosphate nucleotidohydrolase (dUTPase), which was identified by our laboratory [5], is encoded by the BLLF3 gene and expressed as an early protein during lytic replication [5,6]. The EBV-encoded dUTPase protein has been detected using immunohistochemical techniques in the upper epithelial layers of oral hairy leukoplakia (HL) lesions, in lymphoid cells from tonsils of patients with infectious mononucleosis (IM) and in NPC tissue [7,8]. Expression of BLLF3 has also been detected in EBV genome positive tumor cell lines established from patients with nasal NK/T-cell lymphoma using microarray technology [9]. Our recent studies have demonstrated that the EBV-encoded dUTPase possesses novel functions in innate/ adaptive immunity, independent of its enzymatic activity, due in part to the stimulation of toll-like receptor (TLR) 2 and subsequent activation of NF-kB leading to the induction/secretion of proinflammatory cytokines [10][11][12]. These studies indicated that the primary cellular targets of the EBV-encoded dUTPase were monocytes/macrophages and DCs. To gain insight into the biological effects of EBV-encoded dUTPase on human DCs (hDCs) function we evaluated gene expression changes by microarray analysis following treatment of hDCs with the EBVencoded dUTPase and compared it to that of untreated cells using U133 Plus 2.0 Human genome GeneChips. The results from this study revealed that EBV-encoded dUTPase significantly altered the expression of genes involved in oncogenesis, inflammation and viral defense mechanisms. Additional proteome array studies of hDC supernatants following treatment with the EBV-encoded dUTPase demonstrated that the EBV-encoded dUTPase modulated DC function by activating cytokine/chemokine-receptor signaling pathways leading to the production of T H 1 and T H 17 pro-inflammatory cytokines. Finally, we demonstrate that the EBV-encoded dUTPase is secreted in exosomes of chemically induced Raji cells and that these exosomes induced the secretion of cytokines and chemokines through TLR2 in human primary cells. These results suggest that the EBV-encoded dUTPase by being secreted from cells undergoing lytic and/or abortive-lytic replication can modulate immune functions of neighboring cells and thus, alter the microenvironment to contribute to the pathophysiology of EBV-associated diseases.

Purification of the EBV-encoded dUTPase
Detailed methods for the purification of EBV-encoded dUT-Pase have been previously reported by our group [10][11][12]. All EBV-encoded dUTPase preparations were tested as described previously [10] and were free of detectable levels of LPS, peptidoglycan (SLP-HS), DNA or RNA. Protein concentration was determined with a Coomassie Brilliant Blue dye-binding assay (Bio-Rad Laboratories, Hercules, CA) using bovine serum albumin as the standard. The purified EBV-encoded dUTPase used in these studies was stored at280uC at stock concentrations of 0.2 and 0.5 mg/ml.

Cell Culture
Human dendritic cells (hDC/LCs; myeloid, plasmacytoid and Langerhan cells) were obtained from MatTek Corporation (Ashland MD). These cells were generated from CD34+ progenitor cells derived from human umbilical cord blood (HUCB) cells and cultured using specially formulated medium, DC-100-MM, (MatTek) containing a cytokine cocktail designed to induce differentiation of the CD34 + into DCs. These DCs express surface markers CD1a, HLA-DR, co-stimulatory molecules, Birbeck granules and surface markers characteristic of both plasmacytoid and myeloid DC [13].

Gene Expression Analysis in Hdcs
The effect of the EBV-encoded dUTPase on gene expression in hDCs was determined by treating hDCs with purified EBVencoded dUTPase (10 mg/ml), as we have described [12], for 4 h and then comparing the gene expression in these cells to untreated controls. A total of 5 samples (2 untreated controls and 3 EBVencoded dUTPase treated) were employed in this study. After treatment, total RNA was isolated using the RNAqueous-4PCR kit (Ambion, Grand Island, NY), following the procedure supplied by the manufacturer. Trace amounts of DNA were removed by treatment with DNase 1 and the RNA quantity and integrity was analyzed using a 2100 Bioanalyzer (Agilent, Santa Clara, CA). For gene expression studies, sample preparation and processing procedures were performed as described in detail in the Affymetrix GeneChip Expression Analysis Manual (Santa Clara, CA). Briefly, cRNA was synthesized from 3.5 mg of total RNA using a one-cycle reaction followed by biotin labeling of antisense cRNA and then hybridized to human genome U133 Plus 2.0 GeneChips (Affymetrix Santa Clara, CA). The GeneChips were washed using the automated Affymetrix fluidics station and the bound biotinlabeled cRNA detected using a Streptavidin-Phycoerythrin conjugate. Subsequent signal amplification was performed using a biotinylated anti-streptavidin antibody followed by Genechip scanning. All microarray data have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession number GSE46519 (http://www.ncbi.nlm.nih.gov/ geo/query/acc.cgi?acc = GSE46519).
Affymetrix GeneChip hybridization data were normalized using two methods: Robust Multi-Array expression using sequence information (GCRMA) 2 [14] and Model-Based Expression Indexes (MBEI) [15]. GCRMA normalization was performed using the online analysis tool ArrayQuest 2 [16] to implement the Bioconductor GCRMA package (http://www.bioconductor.org/). MBEI normalization was performed using the dChip analysis tool. Differentially expressed genes between control and treated samples were then identified for each normalized data set using the criteria fold change .2 and p,0.05 (un-paired t-test on the comparison between the expression values of control and treated samples). Genes overlapping between the GCRMA and MBEI gene lists were used for further analysis. GO and KEGG annotations were retrieved for these genes using the Affymetrix NetAffx analysis tool and gene expression patterns were clustered using dChip unsupervised clustering algorithm with centroid linkage [17].

Quantitative Real-time PCR
The expression of a select group of differentially expressed genes identified by microarray analysis was validated using pre-designed and functionally validated TaqMan assays on demand (Applied Biosystems, Grand Island NY). RNA from untreated or EBVencoded dUTPase treated hDC/LCs for 4 h was examined by qRT-PCR for the expression of CCL20, IL-6, CXCL11, IFIT1, IFIT2, IFIT3, OASL, OAS2, TNFAIP6 and IF44, FOXO1 using ABI specific TaqMan gene expression assays/ABI primer reference (Hs00171125_m1, Hs00174131_m1, Hs00171138_m1, Hs00533665_m1, Hs00155468_m1, Hs01911452_m1, Hs00984390_m1, Hs00942650_m1, Hs01113602_m1, Hs00951344_m1, and Hs01054576_m1). Samples were normalized to GAPDH (Hs99999905_m1) and expressed as the mRNA expression levels relative to untreated control. All reactions were performed in triplicate on an ABI 7900HT System.

Western Blot Analysis
The expression of a select group of differentially expressed genes identified by microarray analysis was also validated by western blot. Whole cell lysates (25 mg) from untreated or EBV-encoded dUTPase-treated hDC/LCs for 4 h were size-fractionated in a polyacrylamide gel under standard SDS-PAGE conditions. Proteins were then transferred onto PVDF membranes (Bio-Rad Laboratories, Hercules, CA) and analyzed by immunoblotting Detection of exosomal marker proteins was determined by western blotting using the exosome antibody kit, as described by the manufacturer (SBI System Biosciences Mountain View, CA). Briefly, exosome extracts (25 mg) from non-induced or TPA and sodium butyrate induced Raji cells for 72 h, were separated in a polyacrylamide gel under standard SDS-PAGE conditions. Proteins were then transferred onto PVDF membranes and analyzed by immunoblotting as described above using anti-HSP70 (1:1000 dilution) and HRP-conjugated secondary goat anti-rabbit antibody (1:20,000 dilution), as recommended by the manufacturer.

Luciferase Reporter Gene Assays
TLR2-HEK293 expressing cells (2.5610 5 ) were seeded into 12well plates and transiently transfected 24 h later using lipofectamine 2000 transfection reagent (Invitrogen; Carlsbad, CA), as we have previously described [12,18]. Briefly, cells were transfected with pNFkB-Luc (0.5 mg) and pRL-TK (8 ng) reporter vectors. At 24-36 h following transfection, cells were treated with intact exosomes or exosomes extracts from non-induced or TPA and sodium butyrate induced Raji cells for 72 h, culture supernatants (30 ml) from non-induced or chemically induced cells concentrated 10-fold using Centricon PL-10, or purified EBV-encoded dUTPase (10 mg/ml) for 8 h or left untreated. Following treatment, cell lysates were prepared and reporter gene activities were measured using the dual-luciferase reporter system (Promega, Madison, WI). Data was normalized for transfection efficiency and reporter activity expressed as the mean relative stimulation 6 SD.

Cytokine/Chemokine Proteome Array
Human primary dendritic cells (3610 5 ), were treated with EBVencoded dUTPase protein (10 mg/ml) for 24 h and the levels of twelve cytokines, chemokines in cell culture supernatants of treated and control samples were measured using SearchLightH proteome arrays (Aushon BioSystems, Billerica, MA), as we have described [18]. Briefly, samples were incubated for 1 h on the array plates that were pre-spotted with capture antibodies specific for each protein biomarker. The bound proteins were detected with a biotinylated detection antibody, followed by the addition of streptavidin-horseradish peroxidase (HRP) and lastly, a chemiluminescent substrate. The plates were immediately imaged using the SearchLightH imaging system, and data was analyzed using SearchLightH Array Analyst software. The amount of luminescent signal produced is proportional to the amount of each protein present in the original standard or sample. Concentrations are expressed as pg/ml and represent the average 6 SD of an n of 4.

Purification of Exosomes
Raji cells were induced with TPA and sodium butyrate for 72 h. Induced or non-induced Raji cell cultures were centrifuged at 30006g for 159 to remove cells and cell debris and the supernatants were transferred to a sterile tube. Culture supernatants (30 ml) were either concentrated 10-fold using Centricon PL-10 or used for exosome purification using the ExoQuick-TC TM precipitation method. This method was employed based on literature reports [19] demonstrating that ExoQuick precipitation produces exosomal RNA and protein with greater purity and quantity than chromatography, ultracentrifugation and Dyna-Beads. Briefly, ExoQuick-TC TM precipitation solution (2 ml; SBI System Biosciences Mountain View, CA) was added to the supernatant (10 ml) from induced and non-induced cells, incubated at 4uC for 48 h and the exosomes were collected by centrifugation as described by the manufacturer. Exosomes were resuspended in PBS and either lysed by freezing and thawing three times or left intact and subsequently used for cell stimulation studies. Additional exosomes preparations were lysedusing a general lysis buffer (50 mM Tris-HCl, pH 7.5 containing 150 mM NaCl and 1% Nonidet-40) and used for dUTPase enzyme assays.

dUTPase Activity in Exosomes
The dUTPase activity present in exosome lysate preparations and/or in culture supernatants from induced and non-induced Raji cells, obtained as described above, was determined using the standard assay [10] in the presence and absence of human serum from a patient with EBV genome positive diffuse large B-cell lymphoma, which we previously demonstrated contained specific neutralizing antibodies against the EBV-encoded dUTPase.

Cytokine Profile Induced By Exosomes
For cytokines experiments with exosomes, dendritic cells and PBMCs were seeded at a density of 3610 5 in 24-well plates and cultured in AIM-V serum-free medium supplemented with Lglutamine (2 mM), streptomycin (50 mg/ml) and gentamycin (10 mg/ml). The next day, cells were treated with either intact exosomes or exosomal extracts from non-induced or chemically induced Raji cells, EBV-encoded dUTPase (0.1 mg/ml or 10 mg/ ml), Pam3csk4 (0.1 mg/ml) or left untreated for 24 h. Following treatment, cell supernatants were collected and the levels of 7 cytokines in treated and control samples were measured by ELISA (MSD Multi-array and Multi-spot human cytokine kit) using a Sector Imager 2400 instrument and the MSD Data Analysis Toolbox software. Concentrations are expressed as pg/ml and represent the average 6 SD of an n of 4.

hTLR2 Blocking Experiments
For blocking experiments, hDCs and PBMCs were seeded at a density of 3610 5 in 24-well plates and cultured in AIM-V serumfree medium supplemented with L-glutamine (2 mM), streptomycin (50 mg/ml) and gentamycin (10 mg/ml). The next day, cells were pretreated with (10 mg/ml) anti-human TLR2 monoclonal antibody (anti-TLR2 MAb; IgG2a, clone TL2.1) or IgG2a MAb isotype control for 1 h at 37uC and subsequently exposed to either intact exosomes or exosomal extracts from non-induced or chemically induced Raji cells, Pam3csk4 (0.1 mg/ml) or left untreated for 24 h. Following treatment, cell supernatants were collected and the levels of TNF-a in treated and control samples were measured using the MSD human cytokine assays kit, a Sector Imager 2400 instrument and the MSD Data Analysis Toolbox software. Concentrations are expressed as pg/ml and represent the average 6 SD of an n of 4.

Statistical Analysis
The statistical comparisons between different study groups were carried out using Student's t test and p,0.05 reported when significant. Values represent the average of at least three independent experiments.

EBV-encoded dUTPase Induced gene Changes in hDCs
Gene expression patterns provide insight into complex biological networks in which viruses and host cells interact. Dendritic cells are the most potent antigen presenting cells of the immune system and are crucial for the initiation of T-cell responses to viral pathogens. Previous studies performed in our laboratories demonstrated that EBV-encoded dUTPase, an early protein produced during lytic replication of EBV, triggers a signaling cascade through TLR2 that results in the activation of NF-kB and increased secretion of pro-inflammatory cytokines [12]. In addition, the results from our depletion studies [10] indicated that hDCs might be a likely target of the EBV-encoded dUTPase. To better understand the biological effects of the EBV-encoded dUTPase on gene modulation in cells that play a central role in innate and adaptive immune responses, microarray gene expression profiling studies were performed on untreated or EBVencoded dUTPase treated hDCs using the human genome U133 Plus 2.0 GeneChip, as described in Materials and Methods. The results of this study identified 894 differentially expressed genes between control and EBV-encoded dUTPase treated samples for each normalized data set using the criteria fold change .2 and p,0.05. As shown in Table 1, the main pathways affected by EBV-encoded dUTPase in hDCs include cytokine/chemokine and receptor interaction, Toll-like receptor signaling, cell cycle/ apoptosis/proliferation pathways, exosome formation and oncogenesis.
A representative group of EBV-encoded dUTPase induced genes from the microarray profiling study was validated by qRT-PCR and western blot analysis. As shown in Figure 1A, EBVencoded dUTPase induced the expression of CCL20, IL-6, CXCL11, IFIT1, IFIT2, IFIT3, OASL, OAS2, TNFAIP6, IFI44 and FOXO1 in hDCs following a 4 h treatment relative to untreated cells as determined by qRT-PCR using specific TaqMan gene expression assays (ABI) ( Figure 1A). Samples were normalized to GAPDH and expressed as the mRNA expression levels relative to untreated control. Furthermore, western blot analysis of CCL20, CXCL11, IFIT3 ( Figure 1B) using whole cell lysates from untreated or EBV-encoded dUTPase-treated and protein specific Abs further confirmed the results obtained by microarray gene profiling studies. All together, these data provide a modular view of unique hDC responses to EBV-encoded dUTPase and suggests mechanisms by which EBV-encoded dUTPase may modulate immune responses.

EBV-encoded dUTPase Induces the Secretion of T H 1 and T H 17 Cytokines in Human DCs
We have previously demonstrated that the EBV-encoded dUTPase induced the production of pro-inflammatory cytokines in PBMCs and human monocyte derived macrophages [10,11]. To further examine the biological properties of EBV-encoded dUTPase protein on the production of pro-inflammatory T H 1 and T H 17 cytokines, proteome array studies were performed in hDCs treated with EBV-encoded dUTPase, control human dUTPase proteins or left untreated for 24 h, as described in Materials and Methods. As shown in Figure 2, treatment of primary hDCs with EBV-encoded dUTPase recombinant protein resulted in a statistically significant increase in the production of cytokines TNF-a, IL-23, IL12p40, IL-1b, IL-6, IL-10 and TGF-a in comparison to untreated hDCs or hDCs treated with the human nuclear dUTPase control protein (data not shown). While IL-17 levels were slightly elevated in the supernatant of EBV-encoded dUTPase treated hDCs, they were not significant (data not shown). In addition, EBV-encoded dUTPase strongly induced the production of CCL20 and RANTES (p,0.05). These results suggest that the EBV-dUTPase has the potential to modulate T H 17 cell responses.

EBV-encoded dUTPase is Released from Chemically Induced Raji Cells in Exosomes
To determine whether EBV-encoded dUTPase could be secreted from EBV-infected cells in exosomes, Raji cells, an EBV-genome positive Burkitt's lymphoma nonproducing cell line [20], were employed. Raji cells were induced with TPA-sodium butyrate for 72 h and the culture supernatants were either concentrated or used for exosome purification, as described in Materials and Methods. Exosomal extracts from non-induced and induced Raji cells were routinely analyzed for the presence of the exosomal marker HSP-70 by western blot ( Figure 3A) prior to being used in the experiments. Culture supernatants and exosomes obtained from non-induced Raji cells served as controls. As shown in Table 2, there was a 2.1-fold (p,0.05) and 4.8-fold (p,0.01)  increase in dUTPase activity in culture supernatants concentrates and exosomes obtained from induced Raji cells, respectively, when compared to non-induced controls. Furthermore, these data demonstrate that there is greater than a 700-fold enrichment of dUTPase activity in exosomes obtained from induced Raji cells when compared to culture supernatants from these cells (total dUTPase activity in concentrated supernatant 0.114 unit/ml vs 8.19 unit/ml in exosomes). Since chemical induction of latent EBV in Raji cells causes abortive-lytic replication, there is no cell lysis, this further supports the data presented in this study and the premise that the EBV-encoded dUTPase is secreted in exosomes. Furthermore, since a previous study had reported that the nuclear isoform of the human dUTPase is secreted in exosomes of B-cells [21], neutralization experiments were performed to determine whether the dUTPase activity detected in the culture supernatant concentrate and exosomes from chemically induced Raji cells was specifically the EBV-encoded dUTPase. The results of this study showed that the dUTPase activity in the culture supernatant concentrate and exosomes from chemically induced Raji cells was inhibited 71% and 63%, respectively, by human serum from a patient with EBV genome positive diffuse large B-cell lymphoma, which we have demonstrated contained specific neutralizing antibodies against the EBV-encoded dUTPase (unpublished data). Conversely, there was less than 11% and 2% inhibition of the dUTPase activity in concentrated culture supernatants and exosomes derived from non-induced Raji cells, respectively ( Table 2). We next investigated whether the levels of EBV-encoded dUTPase protein present in the exosomes of chemically induced Raji cells were sufficient to induce TLR2-mediated activation of NF-kB. As shown in Figure 3B, exposure of TLR2-HEK293 cells to exosomal extracts from induced Raji cells resulted in a statistically significant (p,0.05) low level but dose-dependent activation of the NF-kB reporter gene when compared to non-treated controls.  Conversely, no activation of the NF-kB reporter gene was observed in HEK293 lacking TLR2 in response to exosomal extracts derived from induced Raji cells (data not shown), which suggests that NF-kB activation is TLR2 dependent.

The EBV-encoded dUTPase Present in Exosomes Induces the Secretion of Cytokines from hDCs and PBMCs
We next investigated whether the EBV-encoded dUTPase levels present in exosomes were sufficient to induce the secretion of cytokines in hDCs and PBMCs. To address this possibility, cytokine analyses were performed by ELISA using culture supernatants of hDCs treated with either intact exosomes or exosomal extracts from non-induced or chemically induced Raji cells or left untreated for 24 h, as described in Materials and Methods. EBV-encoded dUTPase (0.1 mg/ml or 10 mg/ml) and Pam3csk4 (0.1 mg/ml; a ligand for TLR2) were used as controls. As shown in Table 3, treatment of primary hDCs with either intact or lysed exosomes obtained from chemically induced Raji cells resulted in a statistically significant (p,0.01) increase in the production of cytokines IL-10, IL-12p70, IL-1ß, IL-6 and TNF-a in hDCs. Conversely, with the exception of IL-10, there was no significant increase in the production of IL-12p70, IL-1ß or IL-6 by either intact or exosomal extracts from non-induced Raji cells. There was, however, a statistically significant increase (p,0.05) in IL-10 induction by the exosomal extract obtained from non-induced Raji cells. More importantly, with the exception of IL-8 and TNF-a the level of cytokine IL-6, IL-1b, IL-12p70 and IL-10 induced by intact exosomes and the exosomal extracts were similar to that observed in hDCs treated with recombinant EBVencoded dUTPase (0.1 mg/ml; 10 mg/ml).
Similar results were obtained in a parallel study with PBMCs treated with either intact exosomes or exosomal extracts (Table 4). There was no significant production of any of the cytokines examined by exosomes or exosomal extracts from non-induced Raji cells. Conversely, there was a significant induction (p,0.05) of IL-10, IL-12p70, IL-1ß, IL-8 and TNF-a, by intact exosomes and the exosomal extracts from chemically induced Raji cells. These results demonstrate that the EBV-dUTPase secreted in exosomes has the potential to modulate hDCs and PBMCs immune responses.

TNF-a Secretion in hDCs and PBMCs by EBV-encoded dUTPase in Exosomes is Mediated by TLR2
As shown above, EBV-encoded dUTPase secreted in exosomes from chemically induced Raji cells stimulates the secretion of proinflammatory cytokines in hDCs and PBMCs. To determine whether this process is TLR2 dependent, blocking experiments were performed. hDCS and PBMCs were incubated with anti-TLR2 or isotype control (IgG2a, IgG1) antibodies for 1 h, followed by treatment with intact exosomes, exosomal extracts from non-induced or chemically induced Raji cells or left untreated for 24 h, as described in Materials and Methods. Pam3csk4 (0.1 mg/ml; a ligand for TLR2) was used as a control. After 24 h, culture supernatants from control and treated samples were collected and analyzed for TNF-a levels by ELISA. We chose to measure TNF-a as a representative marker of EBV-encoded dUTPase mediated induction of this group of cytokines. As can be seen by the data presented in Table 5, treatment of hDCs and PBMCS with anti-TLR2 antibody resulted in a statistically significant (p,0.01) decrease in the production of TNF-a from hDCs and PBMCs treated with either the exosomal extract or intact exosomes obtained from induced Raji cells. However, preincubation of cells with the isotype control antibody did not inhibit exosomes (intact or extract)-mediated stimulation of TNF-a production in hDCs or PBMCs.
The results from these experiments demonstrate that the EBVencoded dUTPase is released in exosomes from EBV-infected cells undergoing lytic or abortive/lytic replication of EBV and can induce the secretion of pro-inflammatory cytokines in neighboring immune cells through TLR2.

Discussion
Primary infection of EBV, which occurs in the oropharynx, primarily in the tonsils, results in the establishment of a latent infection in memory B-cells in a significant proportion of the adult population [22]. Periodic reactivation results in productive replication of EBV in plasma cells as well as epithelial cells in the tonsils, which is important for transmission of the virus [22,23]. hDCs play an important role in pathogen recognition, antigen presentation and modulation of both innate and adaptive immune responses and various subtypes of hDCs are reported to be found in tonsils [24]. In the case of EBV, plasmacytoid DCs [25] and a novel subset of langerin positive immature hDCs [26] have been reported to have potential roles in controlling EBV infections.  with NF-kB reporter gene and pRL-TK transfection control plasmids. After 24-36 h, cells were exposed for 8 h to equal amounts of exosome extracts (13-39 mg) from induced or non-induced Raji cells and luciferase reporter activity measured, as we have described [12]. Data was normalized for transfection efficiency and reporter activity expressed as the mean relative stimulation 6 SD of an n = 4. *p,0.05, **p,0.01. doi:10.1371/journal.pone.0069827.g003 However, there is essentially nothing known regarding the ability of various EBV-encoded proteins, which are expressed during lytic or abortive-lytic replication of EBV, to modulate DC function. We have previously reported that the EBV-encoded dUTPase activated NF-kB in human monocyte-derived macrophages through TLR2, which resulted in the increased secretion of various pro-inflammatory cytokines [12]. Because of the potential role that hDCs may have in controlling EBV infections, we next determined the biological properties of EBV-encoded dUTPase on gene expression in hDCs and the modulation of the innate and adaptive immune responses.
Recent studies have demonstrated that several EBV-encoded early proteins (BGLF5-alkaline deoxyribonuclease, BNLF2a and BILF1-viral G protein coupled receptor) act as immune evasion molecules by altering MHC-class 1 antigen presentation [27]. The alkaline deoxyribonuclease has also been reported to downregulate the expression of TLR9, which has been implicated in sensing EBV by primary monocytes, plasmacytoid DCs [28], and B-cells [29]. Somewhat surprising is that there have not been any studies, to the best of our knowledge, to determine whether any of the early EBV-encoded proteins may modulate DC function. Type 1 interferons play a major role in establishing an antiviral state by inducing the expression of many genes involved in innate immunity. Studies to elucidate the mechanisms by which viruses are recognized by the innate immune system have focused primarily on viral nucleic acids, since it was assumed that viral encoded proteins lacked conserved motifs that would activate pathogen recognition receptors (PRR) [30][31][32]. These studies have shown that viral nucleic acids induce Type 1 interferon production through TLR7 and TLR9, only in plasmacytoid DCs [28][29][30].
Recently, Barbalat et al [33] demonstrated that inactivated vaccinia virus and mouse cytomegalovirus (MCMV) stimulated  the production of Type 1 interferons through a TLR2-dependent pathway in Ly6C hi inflammatory monocytes. Since TLR2 has been reported to recognize several DNA viruses, including members of the herpesvirus family [34][35][36], it suggests that these viruses must contain a protein(s), which is part of the virion, that can target the innate immune system [33]. In the present study, we demonstrate, using microarray gene expression analyses (Table 1), that the EBV-encoded dUTPase up-regulates interferon ß (11.15fold) as well as five effector pathways of the IFN-mediated antiviral response: the MX1 GTPase pathway, specifically the IFN induced proteins MX1, GBP1, GBP4 and GBP5; the OAS pathway; the ISG20 pathway, specifically members of the tripartite-motifcontaining proteins PML, TRIM 19 as well as TRIM 2, 8, 21, 25 and 56; the ISG15 ubiquitin-like protein modifier pathway, which includes HERC5 and USB18 proteins and the IFIT pathway [37][38][39]. Furthermore, microarray analyses demonstrated that the EBV-encoded dUTPase increased the expression of the interferon-inducible T-cell attracting chemokine CXCL-11 by 62fold in hDCs. Interestingly, these are novel findings mediated by an early protein expressed during lytic or abortive-lytic replication of EBV and suggest that the EBV-encoded dUTPase could prime hDCs to mount an antiviral immune response. Proteome array analysis of supernatants from untreated or EBV-encoded dUTPase treated hDCs demonstrated that the EBV-encoded dUTPase strongly induced the secretion of proinflammatory T H 1/T H 17 cytokines including, IL-6, TNF-a, IL-23 and IL-12p40 as well as the chemokines CCL20, IL-8 and RANTES. These chemokines are important in the trafficking of lymphocytes and neutrophils. Interestingly, increased secretion of IL-8, RANTES and CCL20 are known to be up-regulated in EBV associated malignancies [40]. CCL20 is the major chemoattractant for immature DCs, effector/memory T-cells and B-cells through its receptor CCR6. The CCL20-CCR6 axis has been suggested to play a critical role in the initiation of immune responses, especially in the skin and mucosal surfaces [41]. The findings in this study and our previously published work [12] suggest that engagement/stimulation of TLR2 and subsequent activation of NF-kB may lead to increased CCL20 expression, which is consistent with previous studies demonstrating that the expression of CCL20 is mediated by NF-kB [42]. While the role that CCL20 has in the natural EBV-host relationship/interaction remains to be determined, one possibility is that CCL20 enhances the transmission of EBV from B-cells/plasma cells to epithelial cells by promoting the trafficking of B-cells/plasma cells infected with EBV to a microenvironment containing a cellular milieu that is favorable to allowing lytic replication of the virus. Furthermore, these results suggest that CCL20 either alone or in combination with IL-6 may contribute to the proliferation of EBV growth transformed B-cells.
Exosomes are membrane nanovesicles of variable composition that are secreted from multiple cell types into the extracellular space where they have been implicated in several biological processes, including immune surveillance, and are emerging as a potent mechanism of intercellular communication [21,[43][44][45][46][47][48][49][50]. Furthermore, numerous studies suggest that various pathogens Table 5. TNF-a production by EBV-encoded dUTPasecontaining exosomes in hDCs and PBMCs is TLR2-mediated.  modulate the immune response through the increase production of exosomes, which activate various signaling cascades through TLRs, including TLR2 [51][52][53][54][55]. Several studies have demonstrated that macromolecules encoded by EBV are secreted in exosomes from EBV-transformed lymphoblastoid B cells (LBC) and nasopharyngeal carcinoma cells (NPC) [56][57][58][59]. Most importantly, these studies revealed that the EBV-encoded macromolecules, which include gp350 [59], latent membrane protein 1 (LMP1) [56,57] and miRNA [58], modulate cells in the microenvironment. While human and EBV-encoded dUTPases lack consensus secretory signal domains, several studies have reported that the nuclear isoform of the human dUTPase is released from stressed cells [60,61] and it was recently shown that the human dUTPase is secreted in exosomes from B-cells [21]. Interestingly, the human adenovirus type 9 E4-ORF1 protein, which encodes for an ancestral dUTPase [62], is also targeted to membrane vesicles [63]. A critical and novel finding in our study is the demonstration that the EBV-encoded dUTPase protein is being secreted from chemically induced Raji cells in exosomes. This is supported by at least a 700-fold enrichment of dUTPase activity in exosomal fractions when compared to concentrated culture supernatants from chemically induced Raji cells and a 4.8fold increase (p,0.01) in dUTPase activity in exosomes derived from chemically induced Raji cells relative to non-induced cells. Perhaps more importantly, these data not only demonstrate that the EBV-encoded dUTPase is being released from EBV-infected cells in exosomes, but that the EBV-encoded dUTPase containing exosomes interact with hDCs through TLR2 to enhance the production of various cytokines at levels that are similar to those observed following treatment of hDCs with purified EBV-encoded dUTPase protein.
While EBV is considered to be a B lymphotrophic virus, it infects a variety of cell types including monocytes/macrophages, DCs, neutrophils, natural killer cells and T-cells, but there is no evidence that productive infection occurs in these cells [2][3][4][64][65][66][67][68]. Likewise, it is well established that productive lytic replication occurs in epithelial and plasma cells in vivo. However, the majority of infected cells undergo abortive-lytic replication [2,3,64,65], which does not result in cell lysis but rather apoptosis [22,[69][70][71]. Thus, the primary mechanism for release of EBV encoded macromolecules in cells undergoing abortive-lytic replication may be through exosomes. The data presented in this study demonstrates that the EBV-encoded dUTPase is released in exosomes from B-cells chemically induced to enter an abortivelytic replicative cycle. The data also demonstrate that the EBVencoded dUTPase, either free or in exosomes, modulates hDC and PBMC functions through TLR2 by activating a diverse group of signaling pathways that may lead to inflammation, impaired adaptive immunity and cellular proliferation/cell survival ( Figure 4). Furthermore, since the EBV-encoded dUTPase can induce chemokines, it has the potential to modulate cell trafficking and attract cells that could become infected by the virus, resulting in its dissemination to other sites within the host. Thus, the EBVencoded dUTPase may play an important role not only in local but also systemic cell-cell communications by modifying the microenvironment to support/promote the establishment/maintenance of a persistent infection. Although additional studies are necessary to better understand the complex interplay between EBV, EBV-encoded proteins and the host immune system, the data presented in this study suggest that the EBV-encoded dUTPase may play an important role in contributing to the modulation of the EBV life cycle and to the pathology of EBV associated diseases by modulating immune cells' (hDCs, PBMCs) functions. Interestingly, EBV has been implicated as a possible trigger in several autoimmune diseases including multiple sclerosis [72][73][74], rheumatoid arthritis [75,76], and systemic lupus erythematosus [77][78][79][80], as well as in the immune-mediated disease chronic fatigue syndrome [81]. However, the mechanism(s) remains to be elucidated. Our studies open up the possibility of the EBV-encoded dUTPase released in exosomes as a potential mechanism by which this protein could contribute to the altered immune responses that occur in patients with these diseases. Finally, based upon this study and our previous work concerning the EBV-encoded dUTPase [10][11][12], as well as our recent findings with the human endogenous retrovirus encoded dUTPase [18], the data suggest that some virus-encoded dUTPases possess novel immunomodulatory properties that may contribute to the development of various diseases and therefore, they could be potential targets for the development of novel therapeutics.