Extracellular Vesicles in Epstein-Barr Virus Pathogenesis

Abstract

Purpose of review

Epstein-Barr virus (EBV) is a known determinant for numerous malignancies and may contribute to autoimmune diseases. The underlining mechanisms behind EBV pathologies is not completely understood. Recently, extracellular vesicles (EVs) released from infected cells have been found to produce profound effects on cellular microenvironments. Therefore, in this review we sought to critically evaluate the roles of EVs in EBV pathogenesis and assess their potential therapeutic and diagnostic utility.

Recent findings

EBV-altered EVs are capable of activating signaling cascades and phenotypic changes in recipient cells through the transfer of viral proteins and RNAs. Moreover, several EV-associated microRNAs have encouraging prognostic or diagnostic potential in EBV-associated cancers.

Summary

Current evidence suggests that EBV-modified EVs affect viral pathogenesis and cancer progression. However, further research is needed to investigate the direct role of both viral and host products on recipient cells and the mechanisms driving viral protein and RNA EV packaging and content modification.

Introduction:

Epstein-Barr virus (EBV) is a member of the human Herpesviridae family that has established a persistent infection in more than 90% of the world’s population (1). However, in certain susceptible individuals, EBV can lead to the development of a wide range of malignancies including nasopharyngeal carcinoma, gastric adenocarcinoma, Hodgkin lymphoma, Burkitt lymphoma, post-transplant lymphoproliferative disease, NK/T-cell lymphoma, aggressive NK-cell leukemia, lymphomatoid granulomatosis, and diffuse large B cell lymphoma (2, 3). Currently, EBV is estimated to contribute to the development of roughly 200,000 new cancers world-wide each year and account for nearly 2% of all human cancers (4). There is also some evidence that EBV may be an underestimated contributing co-factor to the progression of other malignancies like cervical and breast cancers (5–7). However, more studies are needed considering it is often difficult to establish a causative role of EBV in a particular cancer since only a small percentage of people infected will ever develop malignancies. In addition to cancer, EBV has been linked with autoimmune diseases such as multiple sclerosis and systemic lupus erythematosus (8, 9).

Oncogenic viruses contribute to the development and progression of malignancy by fostering a pro-tumoral microenvironment (10). Extracellular vesicles (EVs) have emerged as key factors driving intercellular communications in tumor microenvironments. Depending on the type of cancer, EVs have been shown to function in cell growth, invasion, metastasis, angiogenesis, drug resistance, and immune cell regulation through the transfer of bioactive molecules (11). EVs contain a diverse repertoire of proteins and RNAs enclosed in a lipid bilayer enriched in sphingolipids and cholesterol (12). Several EV populations are recognized based on their intracellular site of origin and include exosomes and microvesicles (12, 13). Exosomes are small (50–150nm) vesicles that form on the limiting membrane of multi-vesicular bodies, whereas microvesicles bud off from the plasma membrane and are generally larger in size (100–1000nm). The overlap in size and density has made it difficult to get truly pure vesicle populations using methods currently employed in the field for EV isolation. Adding to the complexity is the existence of exosome and microvesicle sub-populations that likely have distinct cargo and functions (**14).

Once released from the cell, EVs can spread locally or enter bodily fluids and be transported to other sites in the body where they can be taken-up by target cells and release their cargo (11, 12). This EV-mediated intercellular molecular exchange can influence recipient cells and contribute to physiological and pathological processes (15–18). The unique properties of EVs has generated tremendous interest for their use as diagnostics, drug delivery vehicles, and therapeutic agents. However, before the full clinical utility of EVs can be achieved, a better understanding of the mechanisms driving biogenesis of EV subpopulations, EV-target cell interactions, and cargo release must be obtained. It is clear from our work and others, that viral infection alters the components and functions of EVs released from infected cells (10, 12, 19). In the case of EBV, it is evident that EVs contribute to pathogenesis of EBV-associated cancers (10).

Viral EV cargo in the context of EBV cancers

EBV has the capability of altering vesicle content and secretion, which can have substantial effects on tissue and tumor microenvironments. EBV packages various viral components, including proteins and RNAs, into EVs (10, 20). The viral manipulation of host secretory pathways has been demonstrated to aid in the evasion of the immune system while promoting cancer progression by modifying the angiogenic, invasive and metastatic potential within the tumor microenvironment (21, 22). The principal oncogene of EBV is latent membrane protein 1 (LMP1) as it is expressed in most EBV-related cancers and is critical for B-cell immortalization and cellular transformation (23). In figure 1, the viral EV cargo secreted by cell type and latency program are depicted and the functions of these products are summarized in table 1.

Figure 1.

EBV products packaged into extracellular vesicles. Assorted viral components have been found to be incorporated into EVs released from EBV infected cells. There are some differences in components depending on cellular origin and latency type. These EBV-modified EVs perform various functions in recipient cells, such as inhibiting immune responses or promoting cell growth.

Table 1.

Viral EV cargo secreted by cell type

Viral Product	Cell Lines	Functions	References
LMP1	BL, GC, LCL, NPC	Enhance cancer growth, metastasis, invasion, and malignancy, immune regulation	Shair (24), Sheu (27), Meckes (15), Aga (29), Dawson (25), Wheelock (31), Wakisaka (33), Tsai (34)
LMP2	BL, LCL, NPC	Promote EMT, enhance cancer progression, immune modulation	Moody (41), Scholle (42), Fukuda (43), Morrison (44), Cen (45)
EBER	BL, LCL, NPC	Alter recipient cell function	Ahmed (47), Zeng (48), Lee (49)
BHRF1	BL, HL, LCL, PTLD	Immune regulation	Cai (58), Skinner (59), and Pegtel (16)
BART	BL, HL, NPC, GC, NKTL, LCL, DLBCL, PTLD	Immune regulation Cancer progression	Pegtel (16) Meckes (15), Cai (58), Ramayanti (60), Kang (61), Ambrosio (63), Hooykaas (64), Wong (65)
mRNA	LCL and NPC	Enhance infection or cancer progression	Canitano (69)

BL- Burkitt’s lymphoma

HL- Hodgkin’s lymphoma

LCL- Lymphoblastoid Cell Line

PTLD- Post-transplant lymphoproliferative disease

NPC- Nasopharyngeal Carcinoma

GC- Gastric Carcinoma

NKTL- Natural Killer/T-Cell Lymphoma

DLBCL- Diffuse Large B-Cell Lymphoma

Latent membrane protein 1 (LMP1):

LMP1 is a constitutively active CD40 receptor mimic that triggers many pro-oncogenic consequences in cells by recruiting tumor necrosis factor receptor-associated factor (TRAF) proteins and other effector molecules to the C-terminal activating regions (CTAR). This recruitment allows LMP1 to initiate various signaling pathways that inhibit apoptosis and promote cell growth and survival (24–27). LMP1 has been identified in EVs and is capable of inducing similar cellular responses when taken-up by recipient cells (15, *28, 29). Specifically, several studies have demonstrated the ability of LMP1-containing EVs to activate PI3K/AKT and MAPK/ERK pathways in receiver cells as well as inhibit the function of immune cells (21, 22, 29). In fact, LMP1 expression increases levels of EGFR secreted by cells, which likely contributes to the ERK and AKT activating properties of the EVs (15). Along with the cargo modifications, LMP1 has also been found to increase the release of small CD63-containing EVs, perceived as predominately exosomes, from various cell lines and thereby floods tissue environments with virally-modified communicative vehicles (*28).

LMP1 is adept in activating a multitude of pathways that contributes to the progression of cancer (30). One of the recognized characteristics of malignancy progression is the epithelial-mesenchymal transition (EMT), which involves epithelial cells shifting to a more mesenchymal phenotype (31, 32). This phenotypic change is characterized by a cadherin switch, which is when cellular levels of E-cadherin fall while N-cadherin levels increase (31). LMP1 has been found to induce migration and produce a more malignant phenotype in cells by increasing integrin alpha-5 and N-cadherin levels (32). However, whether EV-associated LMP1 produces similar effects in recipient cells has yet to be evaluated. Others have found that LMP1 further promotes malignancy by upregulating hypoxia-inducible factor-1 alpha (HIF1α) and secreting EVs capable of promoting EMT and angiogenesis (29, 33). It is therefore likely that LMP1-modified EVs contribute to EMT and metastatic properties of EBV infected cancer cells.

In addition to enhancing migratory phenotypes, LMP1 supports tumor progression by regulating host immune functions. In a recent study by Tsai et al., LMP1 suppressed host humoral immune responses by blocking differentiation of B cells into antibody secreting cells (ASCs) (34). This study also discovered that LMP1 upregulated the expression of indoleamine 2,3-dioxygenase 1 (IDO1) which is involved in tryptophan production. The tryptophan metabolites inhibit B cell function in the surrounding cells thus further suppressing immunological responses (33). Interestingly, LMP1 was found to regulate programmed cell death protein 1 ligand (PD-L1) in NPC, which is involved in immune suppression, (35) and recently, PD-L1 was detected in exosomes (**36). Therefore, EBV LMP1 may increase packaging of PD-L1 into EVs, conceivably for the modulation of the host immune checkpoints within the tumor microenvironment. This could ultimately contribute to EBV immune evasion strategies and pathogenesis by preventing adequate tumor recognition by immune cells. Immune checkpoint inhibitors are now becoming more widely used for the treatment of certain cancers. Based on current evidence, EBV cancers may therefore be suitable candidates for these therapies.

Since LMP1 can modify EVs, presumably to manipulate the infected microenvironment, gaining a greater understanding of how LMP1 traffics to EVs could foster the generation of novel targeted therapies. Kobayashi et al. determined that farnesylation of the C terminus by ubiquitin C-terminal hydrolase-L1 (UCH-L1) is critical for sorting of LMP1 into EVs (37). Our group later demonstrated that the N terminus and transmembrane region 1 of LMP1 are sufficient for sorting into EVs (*38). Furthermore, a mutant lacking the N terminus and transmembrane domains 1 through 4 failed to be packaged, whereas other mutations resulted in enhanced packaging. These data suggests that there are regulatory mechanisms for LMP1 EV cargo sorting (*38). Alternatively, LMP1 may exists in distinct EV populations with different mechanisms of targeting and biogenesis. For example, LMP1 was detected in fractions consistent with larger microvesicles, as well as the small light and dense EVs described by Kowal and colleagues (**14, *28). In the same study, LMP1 packaging into exosomes requires the tetraspanin protein CD63 (*28, **39). Despite recent advances, the mechanisms controlling LMP1 EV trafficking remain incomplete. Regardless, accumulating evidence supports a role of LMP1-modified EVs as having a prominent involvement in the pathogenesis of EBV-associated cancers.

Latent membrane protein 2 (LMP2):

LMP2 was found by Ikeda and Longnecker to be secreted from cells in exosomes (40). EV-associated LMP2, unlike LMP1, has not been as well studied. Even though LMP2 is critical for maintaining viral latency in EBV infected cells, it is not required for B-cell immortalization (30). Yet, LMP2 has the ability to activate many oncogenic signaling pathways including PI3K/Akt, MAPK/ERK, and NF-kB. LMP2 is required for outgrowth of EBV-infected epithelial cells in vitro (41) and can induce anchorage-independent growth, enhance cell motility and adhesion, and induce EMT (42–44). When cholesterol is depleted in cells, LMP2A cellular levels and EV secretion were both enhanced, but LMP2A was no longer able to be endocytosed (40). LMP2A in the EV fraction is ubiquitinated and not phosphorylated, suggesting that ubiquitination may be important for exosomal loading (40).

Though the effects of LMP2 EVs on recipient cells have not been well studied, LMP1 research suggests that LMP2-modified EVs may both activate pathways and modulate the tumor microenvironment through the promotion of EMT in recipient cells. LMP2 also acts as a B-cell receptor mimic, but the distinct isoforms perform different functions in cells. LMP2a inhibits B-cell activity, while LMP2b activates B-cells (45). This contrasting activity may help maintain EBV latency. Further research is needed to determine if the unique isoforms also effect recipient cells differently. LMP2 promotes malignant progression of NPC therefore it is conceivable that EV-associated LMP2 may exhibit biological effects within the tumor microenvironment (46).

Epstein–Barr virus-encoded RNA (EBER):

In addition to proteins, EBV will package viral RNAs into EVs. Epstein-Barr virus-encoded RNAs (EBERs) are non-coding RNAs that are detected in all EBV infected cells and can be packaged into exosomes with the EBER binding protein La (47). Recently, EBER expression was found capable of distinguishing non-cancerous patients from NPC patients and higher EBER1 levels correlated with increased viability (48). Lee et al. discovered that EBER2 interacts with cellular transcription factor paired box protein 5 (PAX5), which is involved in B cell activation (49). These data suggest that EBV infected cells may use secreted EBERs to alter recipient cell function and could be used as a biomarker for cancer.

Epstein–Barr virus microRNAs:

MicroRNAs (miRNAs) are small noncoding RNAs, typically 20–25 base-pair long, which can act as key regulators of gene expression. The miRNAs can be packaged into EVs and secreted into the extracellular space for cell-to-cell transmission (20, 50). In the case of EBV, miRNAs have been found to play a large role in the pathogenesis of EBV-associated cancers. There are 44 mature EBV miRNAs. Four are mapped to the Bam HI fragment H rightward open reading frame 1 (BHRF1), and forty are mapped to BamHI-A rightward transcript (BART). Different latency stages have unique expression levels of these miRNAs. Nanbo et al. found that cells in type III latency produce the most EVs, when compared with type I latency or EBV-negative (51). Additionally, the cells in type III latency were found to incorporate many viral miRNAs, along with specific host miRNAs (51). EBV encoded miRNAs can regulate viral and host pathways to inhibit apoptosis, promote cell growth, and maintain a persistent infection (52, 53).

EBV miRNAs were found to suppress the release of proinflammatory cytokines, such as IL-12, and repress differentiation of CD4+ T cells. This reduces activation of cytotoxic EBV-specific CD4+ effect T cells (54). EBV infected cells also evade CD8+ T cells by releasing miRNAs that directly target the transported TAP2 and reduce levels of TAP1 and MHC class 1 molecules. The miRNAs are able to decrease levels of EBNA1, which is a target of CD8+ T cells (54). These pathways allow EBV to use EVs to evade the immune system.

Bam HI fragment H rightward open reading frame 1 (BHRF1):

BHRF miRNAs are not typically expressed in nasopharyngeal carcinoma (NPC), unless the cells undergo lytic reactivation, or exhibit a type III latency expression pattern (55). In diffuse large B-cell lymphoma (DLBCL), natural killer/T-cell lymphoma (NKTL), and gastric carcinoma (GC), expression of the BHRF1 miRNA cluster has not been detected (56, 57). However, Cai et al. found that BHRF miRNAs are expressed in cells exhibiting an EBV type III latency program (58). Interleukin 1 (IL-1) receptor 1, which is important for immune regulation, is suppressed by BHRF1–2-5p (59). A study by Pegtel et al. discovered that EVs secreted by EBV-infected B cells containing BHRF1–3 can reduce expression of certain genes, such as the immunostimulatory gene CXCL11, in uninfected recipient cells providing the first evidence of functional miRNA delivery by EVs (16).

BamHI-A rightward transcript (BART):

Ramayanti et al. performed RNA sequencing on plasma RNA from patients with NPC and found that there were higher levels of EBV encoded miRNAs than endogenous miRNA (60). They also discovered that the miRNA profiles differed between patients and that BART13–3p was present in 97% of the samples, suggesting that this BART may be used as a diagnostic marker (60). Additionally, BART miRNAs have been found to be upregulated in GC with BART1 and BART4 being overexpressed (61). High levels of BART miRNAs were detected in DLBCL and NKTL, with the highest expression being BART7, BART22, and BART10 (62). BART miRNAs were found to be expressed at high levels in latently infected epithelial cells and low levels in B cells (58). IL-6 receptor B is targeted by BART6–3p (63), and CREB-binding protein, which is a coactivator in type 1 interferon signaling, is a direct target of BART16 (64). Taken together, these data imply that EBV-modified EVs are important for controlling innate and adaptive antiviral immune responses through miRNA transfer. Since the BART miRNAs can alter various pathways, such as PI3K/AKT and Wnt signaling, or bind to a multitude of tumor suppressor genes (65), it is likely that BART miRNAs in EVs will also target similar pathways in recipient cells.

lncRNA:

Cellular long non-coding RNAs (lncRNA), specifically H19 and H19 antisense, have been found to be packaged into exosomes and released from EBV-positive LCL cells (66). However, little is known about the role these lncRNA play in EBV pathogenesis. H19 was found to have increased levels in retinoblastoma cells and patients with high H19 expression were found to have decreased survival times. Also, knockdown of H19 suppresses cell proliferation and invasion (67). A recent study by Chen et al. additionally found that macrophages in breast cancer, specifically tumor associated (TAMs), secrete EVs containing HIF-1α-stabilizing lncRNA to alter glycolysis in cells within the tumor environment (68). It would be interesting to evaluate whether this occurs in EBV cancers or even what other influences cellular and viral lncRNAs have on the tumor microenvironment.

mRNA:

In addition to non-coding RNAs, EBV also secretes EV-associated mRNA including LMP1, LMP2, Epstein-Barr nuclear antigen 1 (EBNA1) and EBNA2 into the extracellular space (69). More research is required to determine if these mRNA are translated in recipient cells and their downstream effects. It is likely that these mRNAs are be secreted in EVs by infected cells to prime other cells for infection and may enhance cancer progression.

EV diagnostic markers

One of the foremost contributing factors towards improving prognosis and better health outcomes is having the tools and methods allowing for the early identification of diseases. Early diagnosis has been shown to vastly improve patient outcomes in many cancers, as exemplified by those patients diagnosed in stage I/II versus stage III/IV having significantly greater one-year survival rates (70, 71). This is especially true for nasopharyngeal cancers since N3 NPC has been shown to have a high risk of metastasis and low five-year survivability (72). Thus, the development of novel sensitive and specific biomarkers continues to remain a primary focus of various research endeavors. Within the biomarker field of study, EVs are now becoming more universally accepted as a superior source of biomarkers for many diseases, including cancer, and research into EVs as prospective biomarkers has dramatically increased in the past decade.

The benefits of utilizing EVs as biomarkers is that EVs are quite stable, typically minimally invasive to obtain, and have been shown to contain specific cargo relating to the disease and pathology (60, 73). Several recent studies have found the presence or enrichment of certain miRNAs in EVs correlated with particular diseases and even severity or disease stage.

In the case of EBV NPC, plasma EBV DNA and serological IgG assays have been the primary methods for the laboratory detection (74). However, a study by Ramayanti et al. just reported that EV BART13–3P was more sensitive and specific in differentiating EBV NPC from other head and neck cancers (60). As mentioned earlier, EBV BART 1 and 4 were found to be overexpressed in EBV gastric carcinoma. Another study by Tsai et al. found elevated levels of BART 4 as well as BARTs 11, 2, 6, 9, 18 by in situ hybridization of EBV GC surgical specimens (75). Interestingly, BART 9 sequence is homologous to miRNA-200a and miRNA-141, which have been associated with EBV GC EMT phenotype (75). It would be advantageous to examine the levels of these EBV miRNAs in EVs in comparison with the cellular levels since overexpression suggests enhanced secretion. Varying levels of EBV BART miRNAs could potentially not only indicate GC but may also provide additional information on GC phenotypes.

In addition to EBV miRNAs, analyzing host EV miRNA and other non-coding RNAs, may prove worthwhile since cellular expression levels in the context of cancer have been studied extensively. For instance, elevated levels of miRNA-155 as well as the long non-coding RNA PVT1 have both been associated with poor outcomes and survivability in NPC (76, 77). There are ample other host RNA associations in various diseases and analyzing all of the differential expression for each disease type can quickly become complex. Yet, examining EV RNAs could perhaps help narrow down these in order to establish EV profiling assays for each disease. In fact, Taylor et al. established the first multiplexed miRNA EV profiling for ovarian cancer that could not only aid in patient diagnosis but also could be used in screening assays of asymptomatic women (78). Additionally, these multiplex assays potentially increase the prognostic capabilities of the assay which can help better direct patient therapies.

EV therapeutic applications

A key limitation to many types of therapeutics has been effective drug delivery. Recently, there has been a growing interest in nanoparticles as therapy delivery vehicles due to their competent targeting as well as permeability and retention capabilities (79). Nanoparticle delivery systems, including liposomes and mesoporous silica nanoparticles, typically range from 10 to 1000 nm in diameter and can be loaded with chemical therapeutics and miRNAs (79, 80). However, EVs have lately been explored as a possible superior vehicle. Unlike fabricated nanoparticles, EVs have greater stability, are more likely to evade compliment immune responses, and less prone to triggering negative off-target effects (81, 82). Several studies have reported that coupling antigens with EVs increases the immunogenicity and efficacy of potential cancer vaccines (83, 84). This perhaps could be a method for targeting LMP1-induced pathogenic effects in both EBV cancer and autoimmune diseases.

In addition to EV loading methods, the therapeutic application of EVs from specific cell types is being explored for clinical purposes. MSCs, or mesenchymal stem cells, have been researched considerably for potential therapies but recently the EVs from these cells appear more promising because vesicle treatments induce the beneficial effects in target locations without the risk of uninhibited cell proliferation and differentiation (85). A recent study by Yuan et al. found that MSC EV express TNF-related apoptosis-inducing ligand (TRAIL) and could be selectively target and induce apoptosis in vitro (86). TRAIL was previously shown to induce apoptosis in numerous transformed cell types but its utilization has been problematic due to low bioavailability and difficulties with therapeutic delivery (86, 87).

EVs from immature dendritic cells (DC) may also prove to have therapeutic applications. DC-EVs have been found to induce beneficial anti-inflammatory effects that could be utilize in treating autoimmune diseases (88). A study by Kim et al. discovered that DC-EVs exhibited anti-inflammatory effects in vivo and reduced collagen-induced arthritis occurrence as well as severity in mice (89). In addition, DC-EVs are being researched as a possible antitumor therapies and several have made it into clinical trials (90, 91). In a Phase I trial for advanced non-small cell lung cancer, EVs containing MAGE tumor antigens were harvested from patients DCs and readministered (91). The therapy was well tolerated and some patients had positive immune responses and long-term disease stability (91). It may be beneficial to explore both MSC and DC

EVs as therapeutics for EBV associated diseases

Despite promising EV therapeutic uses, there are still obstacles that need be overcome before the widespread clinical application of EV therapies. EV isolation can be arduous and this is further complicated by the inability of current isolation methods to purely separate vesicle types (**14, 85, 92). EVs biogenesis pathways also have yet to be fully appreciated, consequently isolations may contain vesicles from different origins with highly varying cargos (**14, 93). The in vivo consequences of these EVs could pose safety risks so adequate research should be performed prior to the application of these treatments in the clinical setting.

Autoimmune diseases

EBV continues to be implicated in multiple autoimmune diseases, but directly linking infection with disease development can be problematic. Although EBV appears to be a prerequisite for some syndromes, not all who are infected will develop the disease. Regardless, EBV has been connected with several autoimmune diseases but the evidence is strongest for multiple sclerosis (MS), rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) (94–97). Previous research revealed a near 100% EBV-seropositivity in adult and pediatric patients who developed MS (96,95). Recently, over 200 gene loci have been associated with MS risk using genome wide association studies and EBV infection was found to regulate many of these genes, suggesting that EBV infection could be a contributing factor for MS development or pathogenesis (98). RA patients have been found to have higher circulating anti-EBNA1 antibodies (99), elevated blood EBV DNA load (100) and other viral products (i.e. EBERs, LMP1) in synovial fluid (101, 102). Also, SLE patients have elevated circulating EBV DNA, antibodies to EBV EBNAs and viral capsid antigen (VCA) however previously contracting infectious mononucleosis does not lead to an increase chance of developing the disease (103, 104).

One of the accepted means by which viruses are thought to promote the development of autoimmune diseases is through antigenic mimicry. Some viruses have amino acid sequences in their proteins that closely resemble host proteins. This sequence mimicry can lead to immune cells developing antibodies or T cell receptors that can cross-react with host proteins following the viral infection. In the case of EBV, antigenic mimicry has been documented in MS, RA, and SLE (105). Pathogenesis in MS occurs mainly as a result of T cell induced CNS demyelination and chronic inflammation. The MHC class II allele DRB1*1501 has previously been reported as a risk factor for MS but Lang et al. demonstrated the close similarity of the crystal structures of DRB1*1501-MBP (myelin basic protein) peptide and DRB5*1010-EBV peptide presented for T cell recognition (105, 106). For RA and SLE, there are reports of mimicry of EBNA1 peptides with synovial protein and RO/Sm autoantigens respectively (107–109).

It is evident that EVs have many roles in both normal and pathological states. One function of EVs is to serve as a source of antigen presentation complexes that can be taken up by antigen presenting cells (APC) or even possible directly activate T cells (110, 111). EBV has established capabilities in utilizing host EV machinery for the production, processing and release of viral products (15). Interestingly, CD63 is known to have a role in autophagy and trafficking of MHC class molecules for antigen presentation as well as being required for the vesicle secretion of certain viral products (*28, **39, 112). Therefore, it is plausible that EVs displaying the viral mimicry peptides are secreted and could serve as a source for autoimmune clonal expansion of B and T cells. Since EBV LMP1 protein has been shown to increase vesicle secretion, increasing levels of autoimmunogenic EVs may be released with LMP1 expression especially during latency switching (15, *28). As highlighted earlier, LMP1 has been detected in synovial fluid of afflicted joints of RA patients which may also contribute to the release of these mimicry-containing EVs that promote inflammation and pathogenesis (101).

Additionally EVs are capable of stimulating angiogenesis, which can induce damage to cartilage and bone in the joints of RA patients and vasculitis in SLE patients (113). LMP1 is known to increase VEGF expression and LMP1 containing EVs have been reported to promote angiogenesis (15, 114). Increased EBV infected cells (as inferred by elevated DNA levels) in autoimmune patients, may also contribute to increase secretion of pathogenic EVs that are capable of promoting inflammation and immune dysfunction. Altogether the literature supports a pathogenic role for EBV modified EVs in autoimmune diseases.

Conclusions

Viruses have evolved with us and consequently many of the tactics that our bodies have employed to combat infection over the course of time have been met with effective counter adaptive strategies. It can be argued that EBV and other herpesviruses in particular, have been some of the most successful viruses in countering host defenses and exploiting cellular pathways. EVs are essential for the normal function of cells but unfortunately are also exploitable by viruses for the maintenance and spread of infection. Yet, EVs appear to be an opportunity for truly innovative therapeutic interventions that not only can target the viral infection but also the associated pathologies. EVs clearly have a role in the severity of many EBV induce pathologies but EV research may be the means by which we can improve early diagnosis and improve patient outcomes.