(most material taken with permission from Prof.Dr. Jaap M. Middeldorp et al, Crit Rev Oncol Hematol. (2003) ;45(1):1-36)

Table of contents :

---

Epidemiology Genomics Proteomics Transmission Pathogenesis Symptoms & signs Laboratory examinations Differential diagnosis Therapy Experimental animal models Prognosis Web resources

---

Epidemiology : prevalence > 90%, first cultured in 1964^ref
Genomics : like other herpesviruses, EBV has a toroid-shaped protein core wrapped with DNA, a nucleocapsid, a protein tegument and an outer envelope. The envelope carries Gp spikes of which the most abundant structural glycoprotein (Gp) is Gp350/220. The EBV genome is a linear, double stranded DNA molecule of 172 kb and has reiterated 0.5 kb terminal repeats and a reiterated 3 kb internal direct repeat. The unique sequences also contain several repeat elements, often encoding repeat domains in proteins. The BamHI restriction fragments of the B95.8 laboratory strain genome have been completely sequenced^ref; EBV genes are named after the BamHI restriction fragment containing the RNA start site and their leftward or rightward transcriptional orientation. BamHI-A being the largest fragment, BamHI-B the second largest, with BALF2 being the second leftward reading frame on the BamHI-A fragment and so on. Subsequent genomic sequencing of additional EBV isolates have revealed the existence of 2 predominant EBV-strains, named A-type and B-type (or type-1 and -2, respectively), with significant sequence polymorphism in the EBNA2 (BYRF1) and EBNA3A-C (BERF1-3) genes^{ref1, ref2} and minor ‘sub-strain’ variations in the EBNA1 (BKRF1), LMP1 (BNLF1), Zebra (BZLF1) and various other genes^{ref1, ref2, ref3}. In previous studies the A-type virus was found to prevail in most EBV-associated diseases and showed a more efficient transformation of B-cells in vitro^ref. The relative oncogenicity and biological behaviour of different EBV-strains and variants remains a matter of debate^{ref1, ref2}. Most viral genomes remain in cells as episomes : rarely integrated in cell chromosomes. 3 geographic subtypes of EBV have been identified to date differing by their nuclear antigen EBNA2 :

• EBNA2 AC strains predominate in Asia
• EBNA2 AD strains predominate in the USA
• EBNA2 B strains have all been identified in black Africa.

EBV genome contains 5 miRNAs : computer predictions of the mRNA targets of the 5 micro RNAs include the tumor suppressor p53, retinoblastoma binding protein 3, transcription factor E2F1, and genes involved in the immune response, such as a BCL8 homolog and the ICOS ligand precursor, a gene expressed in B cells that activates the immune system. The findings could also explain why EBV is linked to cancers such as Burkitt's lymphoma (BL) and Hodgkin's lymphoma (HL). Micro RNAs are expressed quite well in latent infected cells : most EBV genes are not expressed in latent cells—one of the reasons that the virus becomes latent—with the exception of a few latency-associated transcripts (LAT). One of the micro RNAs, miR-BART2, has the potential to target one of the EBV genes that encodes a DNA polymerase : degradation of that transcript could conceivably contribute to either the establishment or maintenance of latency^ref. People have found conserved miRNA sequences in animals and humans and plants—and now they're also in viruses.
Proteomics : the EBV genome contains over 100 open reading frames (ORFs). These are named after the BamHI restriction fragment on which they are located as indicated above. Only about 50–60% of the gene products have been characterized to date and much remains to be learned about their pathogenic role in vivo. A number of technical developments, such as (multi-primed) RT-PCR and NASBA, have allowed a sensitive and reliable documentation of EBV transcriptional activity in most disease syndromes^{ref1, ref2}. Transcription of various subsets of these genes has been well documented in EBV⁺ cell lines and can be detected in cellular samples from patients with a multitude of EBV-associated acute, chronic and malignant disease syndromes. In most EBV proliferative syndromes EBV gene expression is rather limited and does not involve lytic genome replication or production of new virions. In EBV-associated malignancies, the virus genome is present in the individual tumor cells in its latent episomal (closed circular) form and the viral genome is replicated during each cell division by host DNA polymerase together with host chromosomes. The detection of latent episomal EBV-DNA can be used to define tumor and virus clonality and is taken as proof for early involvement of EBV in tumor development^{ref1, ref2}. Consequently, most EBV disease syndromes are associated with so-called EBV latency. During such latent infections, the virus remains transcriptionally active—albeit in a restricted fashion—and the genes expressed then are referred to as ‘latent genes’
EBV genes and their (putative) functions

&z.cirf;, Positive; , negative; &z.cirf;, positive in some cases; ?, not determined.

• latent gene products : comprise about 10 different gene products that are expressed in different transcription profiles (latency types). During latency, EBV expresses a limited number of viral genes, which are involved in tasks such as stimulating cell proliferation, inhibiting apoptosis, blocking viral lytic replication, and assuring accurate and equal partitioning of the episomal viral genome to daughter cells. Because most EBV-linked syndromes are associated with viral latency, the functions of latent gene products have been studied most extensively
• EBV-encoded RNAs (EBER1 and EBER2) are small, non-polyadenylated RNAs with an estimated abundance of 10⁵–10⁷ copies per cell that have a stable secondary structure with extensive intramolecular base-pairing and some homology to tRNAs^{ref1, ref2, Arrand , JR, 145–149}. They are localized in the nucleus, where they are associated with the cellular La antigen^{Clemens, ML, 107–111} and the EBER-associated protein^ref. It has been proposed that they play a role in splicing of other viral transcripts including primary EBNA and Latent Membrane Protein (LMP) transcripts. Furthermore, it has been suggested that EBER1, 2 neutralize the transcription block normally induced by eIF-2^ref, mediate prevention of apoptosis^ref, suppress antiviral effects of IFN-a and -g and induce IL-10 production^ref. These effects may be mediated by the ability of EBER1, 2 to interact with and inhibit the function of the double stranded RNA-activated protein kinase PKR, a crucial mediator of interferon-induced antiviral effects. Recently an independent role in malignant transformation was suggested^ref.
• EBV nuclear antigen (EBNA)1 transcripts can be derived from 4 different promoters. During the initial stages of primary infection, (i.e. at the time of recircularization of the genome), transcription is initiated in BamHI-W (Wp)^ref; once host cell transformation is established, there is a switch towards the promoter in BamHI-C (Cp)^ref. Primary transcripts from these 2 promoters are very long (ca. 50–70 kb) and contain coding sequences not only for EBNA1, but also for the other EBNAs^ref. It is thought that these polycistronic transcripts are spliced, resulting in bi- or monocistronic transcripts, dependent on the needs of the virus at a certain time point. In latency types I and II where EBNA1 is the only EBNA expressed, Cp and Wp are silenced by methylation^ref and transcription is initiated from a promoter in BamHI-Q (Qp)^ref. Qp is considered to be the true latent promoter for EBNA1 transcription and is regulated in a cell cycle dependent manner involving EBNA1 autoregulation as well as by cytokine-mediated regulation^{ref1, ref2}. The fourth promoter is localized in BamHI-F (Fp) and is activated upon entry in the lytic cycle^ref. These individual EBNA1 promoters provide different levels of EBNA1 transcripts, depending on the host cell type and its activation state^ref. EBNA1 is a nuclear protein consisting of a basic amino terminus, a Gly–Ala repeat segment of variable length, another short basic domain including a nuclear localization sequence and a long hydrophobic C-terminal domain with DNA-binding and dimerization activity^{ref1, ref2}. Recently the crystal structure of the EBNA1 C-terminal dimer domain, bound to its cognate DNA sequence, was resolved revealing unexpected strong structural homology with the Bovine Papillomavirus E2 protein^ref. EBNA1 dimers bind to DNA by interacting with
• a partial palindrome sequence [TAGGATAGCATA-TGCTACCCAGATCCAG] that is present at 2 sites in the EBV genome^ref. There are 2 high-affinity sites: a set (family) of 20 tandem repeats of the cognate sequence (FR), and a combination of 2 sequences in dyad symmetry and 2 in tandem (DS). FR and DS elements are separated by a 1 kb intervening sequence. Together these form oriP, the origin of plasmid replication^{ref1, ref2}. Binding of EBNA1 to these sites results in looping-out of the 1 kb sequence between FR and DS, thus enabling the replication and persistence of the episome in cycling cells^ref; at the same time, EBNA1 is associated with host-cell chromosomes during mitosis which is important for segregation of episomes into progeny nuclei^{ref1, ref2, ref3}
• the third EBNA1 binding site is downstream of the latency I promoter in BamHI Q. EBNA1 binding to this site represses Qp activity, but this repression can be overcome by the presence of E2F and thus activity of Qp is cell-cycle dependent^{ref1, ref2}.
• EBNA1 furthermore enhances transcription from cellular^ref and at least 2 key latency-associated EBV promoters (Cp and LMP1p), which is correlated with its ability to link distant DNA sequences^{ref1, ref2} and binding to host-cell nucleosomal transcription and splicing factors, such as P32/Tap and EBp2^{ref1, ref2, ref3, ref4}.
In addition to its role(s) in maintenance of the viral episome and promoter activation, an oncogenic potential has been ascribed to this protein: mice carrying an EBNA1 transgene under control of the Ig heavy chain enhancer develop lymphomas^ref. Furthermore, the ability of EBNA1 to induce expression of the recombinase-activating genes RAG-1 and -2 is thought to lead to genetic instability^{ref1, ref2}. These observations may be relevant in the multistep oncogenesis as discussed above. EBNA1 is the only EBV-protein that is thought to be expressed in all latently EBV-infected cells. Although EBNA1 is a foreign protein to the host, EBV-infected cells expressing EBNA1 are not killed by CTLs. This is due to a inhibitory effect of the protein's Gly/Ala repeat (GAr) domain on proteasomal processing thereby preventing endogenous MHC class I-restricted presentation. This is likely to prevent peptide production from defective ribosomal products (DRiPs) of EBNA1 translation, a  key source of MHC class I presented peptides^ref. Although EBNA1 is barely recognized by CD8⁺ T-cells, CD4⁺ T-cells and antibodies reactive with EBNA1 are readily detected in most healthy virus carriers^{ref1, ref2}. The latter most likely arise from cross-priming by dendritic cells loaded with EBNA1 that is acquired via digestion of apoptosed EBV-positive cells^{ref1, ref2}.
• EBNA2 is encoded in the long primary transcripts initiated from Wp and Cp^ref and is the first EBV protein to be expressed after delivery of the viral genome to the nucleus. EBNA2 is a nuclear protein that in vitro specifically trans-activates cellular genes, such as the B-cell activation markers CD23 and CD21^ref and the c-fgr proto-oncogene^ref, and viral genes including LMP1^ref, LMP2^ref and the cis-acting element upstream of Cp^ref. The interaction of EBNA2 with its responsive elements is not a direct one but instead occurs via the RBP J-k protein^ref, resembling the signaling exerted by the Notch protein pathway^{ref1, ref2}. Together with EBNA5, EBNA2 is involved in G₀-G₁ transition^ref. Significant sequence diversity exists between the EBNA2 proteins of EBV types 1 and 2, which directly relates to functional differences such as in vitro transforming activity^ref. Although EBNA2 is essential for initial growth transformation of EBV-infected B-cells in vitro^ref, its expression may be less relevant for malignant growth in vivo because EBNA2 is not expressed in most EBV-associated tumors, except those in immunocompromized individuals. Even there EBNA2 expression seems to be limited and rather heterogeneous at the single cell level^ref.
• EBNA leader protein (EBNA-LP, EBNA5) is encoded in the leader of the Wp driven EBNA mRNAs^ref and is expressed simultaneously with EBNA2 in freshly infected B-cells^ref. The ORF is encoded by repeating exons in BamHI-Y and 2 short exons in BamHI-Y leading to the synthesis of multiple protein species ranging from 20 to 130 kDa^ref. Distinct functional domains on LP were recently mapped^ref. Whether the protein is actually synthesized, depends on the splicing of the mature mRNA. EBNA-LP function and phosphorylation appear to be cell cycle dependent^ref [110]. In vitro experiments have indicated that EBNA-LP is essential for transformation and directly or indirectly (via EBNA2) upregulates the expression of host factors required for B-cell growth^{ref1, ref2} [103 and 111]. Moreover, EBNA-LP was shown to bind p53 and pRb and cellular PKCs HA95 and HAX1, which regulate transcription by specific phosphorylation pathways resembling B-cell receptor mediated activation^{ref1, ref2, ref3, ref4} [103, 112, 113 and 114], leading to G₀–G₁ transition of resting B-cells. This is an important step in the early stages of EBV-mediated transformation.
• EBNA3A (EBNA3), EBNA3B (EBNA4) and EBNA3C (EBNA6), like EBNA2 are encoded in the long Wp and Cp driven primary transcripts^ref [64]. Mature mRNA encoding these 3 proteins are the least abundant of all EBNA mRNAs. EBNA3A, B and C are nuclear proteins of 140, 165 and 155 kDa, respectively, and localize into large nuclear clusters associated with the nuclear matrix, but excluding the nucleolus^{ref1, ref2} [115 and 116]. The EBV type 1 and 2 strains show a sequence identity of only 84, 80 and 72%, respectively, mainly due to variations in the C-terminal repeat regions, thus allowing strain typing. Reverse genetics have shown that EBNA3B is dispensible for B-cell growth transformation^ref [107]. On the other hand EBNA3B expression correlates with upregulation of CD40 and downregulation of CD77 in vitro^ref [117]. EBNA3A–C regulate the expression of certain cellular genes and bind to a variety of host proteins including different isoforms of the cellular transcription factor RBP J-k^{ref1, ref2} [118 and 119], thus modulating the EBNA2-driven upregulation of cellular and viral promoters^ref [120].
• EBNA3 / EBNA3A induces the nuclear accumulation of F538, a novel human uridine kinase/uracil phosphoribosyltransferase that is part of the ribonucleotide salvage pathway. Increased intranuclear levels of UK/UPRT may contribute to the metabolic build-up that is needed for blast transformation and rapid proliferation^ref
• EBNA3C upregulates CD21 expression in vitro^ref [121], augments the EBNA2-driven upregulation of LMP1 expression^ref [122] and downregulates Cp-promoter activity^ref [123]. EBNA3C resembles Adenovirus E1A and Papilloma virus E7 proteins by its ability to disturb cell cycle checkpoints and can override the G₂M checkpoint imposed by UV irradiation or genotoxic drugs such as etoposide and cisplatin, thus contributing to accumulation of DNA damage^{ref1, ref2, ref3} [124, 125 and 126]. Because the expression of EBNA3C in EBV-associated neoplasms in vivo has received little attention thusfar, its role in supporting tumor outgrowth remains to be determined.
• LMP1 mRNA originating from the BNLF1 ORF is a highly abundant viral transcript in most latently infected B-cell lines. The LMP1 protein is encoded by 3 exons and is an integral membrane protein with its hydrophilic (small) N- and (large) C-terminus positioned in the cytoplasm flanking 6 short hydrophobic membrane-spanning -helices linked by three short reverse turns that are predicted to loop-out on the extracellular side of the membrane^ref [127]. LMP1 forms patches in the cell membrane^ref [128] and associates with various intracellular membrane vesicles^ref [129] possibly mediated via its association with the intermediate filament protein vimentin^ref [130]. Part of the LMP1 may be secreted from EBV-infected cells in the form of MHC-II containing secretory vesicles, called exosomes^ref [131]. The domains responsible for patch formation comprise the highly conserved N-terminus and first 2 membrane spanning domains^ref [132]. Newly synthesized LMP1 remains detergent soluble for about 2 h, before it becomes phosphorylated and relocates to detergent insoluble cytoskeleton-associated intracellular structures^{ref1, ref2} [128 and 130]. > 50% of LMP1 is located intracellularly and even in clonal and synchronized EBV⁺ B-cell populations LMP1 expression is highly heterogeneous from cell to cell^ref [133]. In some viral strains induction of the viral lytic cycle leads to a boost of LMP1 expression, which, in part, may be derived from a second start site on the mRNA, resulting in a 45 kDa truncated form, largely lacking the transmembrane domains and devoid of most characteristic functions of LMP1^ref [134]. When expressed at low level LMP1 has clear growth enhancing and oncogenic effects^ref [135], but when expressed at high levels LMP1 causes general cytostasis^ref [136]. The cytostatic activity of LMP1 is linked to the first 2 transmembrane domains of LMP1^ref [137], which coincide with the T-cell inhibitory domain LALLFWL located in the first transmembrane helix at AA34–40^ref [131]. The balanced expression of LMP1 and its relative positioning on intracellular and plasma membranes in EBV transformed cells may prove to be relevant for the cell's growth potential. The level of LMP1 expression in different EBV-infected cell lines in vitro shows considerable variation similar to the LMP1 expression in EBV carrying tumor cells in vivo^{ref1, ref2, ref3, ref4} [106, 138, 139 and 140]. The expression of high levels of LMP1 in vivo has been linked to a worse prognosis^ref [141]. LMP1 is considered to be the major EBV-encoded transforming gene for several reasons^ref [39]: transfection of LMP1 has growth transforming effects in rodent fibroblasts^ref [142] and induces many of the changes that are usually associated with primary EBV-infection and transformation of B-cells^ref [143]. Moreover, LMP1 inhibits apoptosis in B-cells by inducing bcl-2^ref [144] and A20^ref [145]. In addition, LMP1 was shown to upregulate the cytokine IL-10 that stimulates B-cell proliferation and inhibits local immune response^ref [146]. In addition, studies with LMP1 transgenic mice have demonstrated direct oncogenic potential for LMP1. Mice carrying an LMP1-transgene under control of the Ig heavy chain promoter and enhancer develop lymphomas^ref  [147]. Upregulation of A20 and Bcl-2 expression in these transgenic mice was observed, in line with the previously found in vitro effects of LMP1. Mice carrying the LMP1 transgene under control of the polyoma virus enhancer and promoter (which directs specific expression in epithelia) show hyperplasia of the epidermis^ref  [148]. It was suggested that LMP1 expression—probably in association with aberrant expression of other (host) genes—may predispose to carcinogenesis. LMP1 was shown to interact with cellular proteins that are mediators of cytoplasmic signaling from the family of TNFR^ref [149]. This interaction is crucial for LMP1 to stimulate cell proliferation, both in B-cells and epithelial cells^{ref1, ref2} [150 and 151]. Thus, LMP-1 mimics the CD40–CD40L mediated NF-kB signal transduction pathway by interacting with specific members of the TRAF's (esp. TRAF1 and TRAF3) and TRADD^{ref1, ref2} [152 and 153]. LMP1 and CD40 co-localize in lipid rafts, but LMP1-associated TRAF-mediated signaling seems to be more efficient^ref [154]. The TRAF1, 3 interaction leads to NF-kB activation resulting in morphological changes and enhanced expression of B-cell activation markers including CD23, CD39, CD40, CD44, MHC class II and the cellular adhesion molecules LFA-1 and ICAM-1 [155]. LMP1 also induces enhanced signaling through the JAK–STAT pathway via binding to JAK3 kinase and ERK MAPK signaling, which involves TRAF2 as signal mediator^ref [156]. All these effects are mediated by at least three distinct signal activating domains (CTAR1-3) located in the cytoplasmic C-terminal half of LMP1^ref [153]. The hydrophilic N-terminus and first hydrophobic membrane-anchoring domain of LMP1 are essential for LMP1 function^ref [132] and mediate LMP1 association with distinct glycosphingolipid-rich raft domains in various cell membrane compartments^{ref1, ref2, ref3} [154, 157 and 158]. LMP1 has distinct effects on cytokine and cytokine receptor synthesis^ref [159], that may affect angiogenesis^ref [160] and inflammatory responses at the single tumor cell level in vivo^ref [161], thus contributing to tumor growth and immune escape^{ref1, ref2, ref3, ref4} [159, 161, 162 and 163].
• LMP2A and LMP2B are encoded by spliced mRNAs transcribed from the region spanning the terminal repeats of the EBV genome. Therefore, LMP2A, B mRNA is only transcribed once the circular viral episome is formed^ref  [164]. They have separate promoters resulting in unique first exons with LMP2B missing most of the N-terminal domain; the remaining exons are common for LMP2A and B. The LMP2 proteins both have 12 putative transmembrane domains and a hydrophilic C-terminal domain and colocalize with LMP1 in raft domains^ref [158]. It was shown that the LMP2A 119 aa N-terminal cytoplasmic domain, which is lacking in LMP2B, interacts via multiple phosphotyrosines arranged in ITAM- and SH2-protein binding motifs with the tyrosine kinases Lyn and Syk^ref [165]. This interaction diminishes Syk and Lyn from binding to the cytoplasmic B-cell receptor signalling domains, thus preventing antigen-receptor mediated activation of EBV-infected B-cells, that would otherwise result in induction of the lytic cycle. LMP2 is not required for B-cell transformation and has no growth altering effects when expressed in differentiating epithelia^{ref1, ref2} [39 and 166]. Studies with LMP2A transgenic mice have shown that LMP2A—although not oncogenic in these mice—can provide an survival signal to B-cells even when these do not express a B-cell receptor^{ref1, ref2} [167 and 168]. It was suggested that this signal, in combination with the B-cell activation-preventing activity of LMP2A, is important for persistence of EBV in the host^ref [165]. This is further supported by recent data showing a well regulated co-expression of LMP1 and LMP2, together with EBNA1, in circulating EBV carrying memory B-cells travelling through lymphoid follicles of the tonsil in vivo. This tightly regulated expression of LMP1 and LMP2 appears to be important for maintaining long-term persistence, by providing temporal growth and survival signals during passage of these cells through lymphoid organs^{ref1, ref2, ref3} [169, 170 and 171].
• rightward transcripts of the BamHI-A region of the viral genome (BARTs) can be detected in all types of EBV latency, including EBV-infected peripheral blood B-cells and are expressed at particularly high levels in nasopharyngeal carcinomas^{ref1, ref2, ref3, ref4, ref5}[172, 173, 174, 175 and 176]. Interestingly, these transcripts are co-terminal and harbour the BARF0 open reading frame (ORF) at their 3?-ends^{ref1, ref2, ref3} [177, 178 and 179]. This ORF encodes a putative protein of 174 amino acids. By means of alternative splicing, the BARF0 ORF may be extended with 105 additional amino acids. The alternative ORF thus generated is called RK-BARF0^ref [180]. Whether (RK-)BARF0 protein is expressed in vivo is at present uncertain. The presence of a polyadenylation signal 5' to the stopcodon renders actual translation unlikely^{ref1, ref2} [177 and 179], but anti-BARF0 antibody [180] and cytotoxic T lymphocyte (CTL)^ref  [181] responses detected in patients suggest that synthesis of this protein does take place. However, a publication from the same group indicated that alternative splicing may remove the actual BARF0 coding sequences from BART's and suggested that the original antibody used to detect the (RK-)BARF0 proteins also detects a cellular component of similar molecular weight^ref [182]. Work from our own group, using newly developed monoclonal antibodies reactive in vitro with various native and denatured recombinant forms of the (RK-)BARF0 protein, did not result in any evidence for (RK-)BARF0 protein expression in vivo in multiple EBV⁺ tumor and EBV-transformed cell line specimens, despite the clear detection of abundant spliced and non-spliced RNA transcripts encoding the putative (RK-)BARF0 protein species. In addition, we could not reproduce the published antibody detection results using purified (RK-)BARF0 protein or defined peptide domains thereof as antigen^ref [183]. Therefore, there remains considerable dispute on a possible role for (RK-)BARF0 protein EBV-infection in vivo. Recently, in vitro functional studies were initiated using heterologous expressed recombinant (RK-)BARF0 protein. One of these studies showed a nuclear localization of the protein^ref [184]. Transfected RK-BARF0 was shown to induce the expression of LMP1, a mechanism dependent on the apparent interaction between RK-BARF0 and Notch^ref [185]. It was hypothesized that this mechanism guarantees expression of LMP1 in EBNA2-lacking latency types. The RK-BARF0 protein, however, appears to turn over rapidly with Notch^ref [185], which could be the reason why the protein could not be detected in EBV-associated diseases^ref [183]. At least 2 other putative ORFs were demonstrated in BARTs, called RPMS1 and RPMS2 / A73^{ref1, ref2, ref3} [173, 186 and 187]. RPMS1 encodes a putative nuclear protein^{ref1, ref2} [173 and 186] that is partially homologous to EBNA2, in particular the WWP containing domain that interacts with RBP J-k. It is tempting to speculate that RPMS1 acts as a substitute for EBNA2, for example in latency types I and II. However, recent data point to a negative influence of the putative RPMS1 protein on EBNA2 and Notch signalling^ref [188]. Although transcription of RPMS1 in latently infected B-cells from healthy donors was found^ref [173], neither transcription of RPMS1 in patient samples nor expression of RPMS1 protein have been reported. Even less is known about potential expression and function of the A73 protein which has been suggested to modulate integrin signalling pathways^ref [186]. Finally, a role for BARTs in transcriptional control was suggested, by virtue of their anti-sense orientation relative to several important early–late leftward gene transcripts, such as BALF5 (DNA-polymerase), BALF4 (the nuclear transport membrane protein gp125) and BILF1, a putative cytokine receptor^{ref1, ref2, ref3, ref4, ref5} [39, 175, 176, 177 and 186]. Thus, BARTs could play a role in maintenance of latency. BARF1 is a rightward transcript from the BamHI-A region that has its own promoter. Originally recognized as an early lytic phase transcript^ref [189]. However, recent data indicate that BARF1 is preferentially expressed in EBV carrying epithelial tumors but not in lymphoid tumors^{ref1, ref2, ref3, ref4, ref5} [50, 54, 190, 191 and 192]. This classifies the BARF1 gene as an interesting and useful carcinoma-specific marker. The BARF1 protein has transforming properties both when expressed in B-cells^ref [193] and in epithelial cells^ref [194]. The transforming domain of BARF1 may be located in the N-terminus and mediates Bcl2 upregulation^ref [195]. The extracellular domain of BARF1 may be selectively cleaved off and secreted from the cell and is considered to be a functional, soluble homolog of the human colony stimulating factor 1 (CSF-1) receptor^ref [196]. As such, BARF1 is hypothesized to play a role in local immune evasion by absorbing CSF-1 that is necessary for an effective immune response. Thus, BARF1 constitutes a protein with pleiotropic function, involved both in oncogenesis and immune escape with a possible significant role in EBV-associated carcinomas, like NPC and GC.
Expression patterns of latent genes :
• latency I (first detected in Burkitt's lymphoma (BL)) : EBNA1, BARTs and EBERs (10⁶ copies per cells)
• latency II (first detected in nasopharyngeal carcinoma (NPC) but this also proved to be the prevailing expression pattern in most other EBV-positive tumors that occur in the immunocompetent host, such as Hodgkin's lymphoma (HL) and T- and B-cell NHLs) : EBERs, BARTs, EBNA1 + LMP1 + LMP2A + LMP2B
• latency III (lymphoblastoid cell lines (LCLs) in vitro) : EBNA1 + EBNA2 + EBNA3A + EBNA3C +  LP + LMP1 + LMP2. However, in vivo, expression of the complete set of latent genes is only found in absence of a fully functional immune response, i.e. in lymphomas of AIDS patients and of transplant recipients. Under these circumstances individual tumor cells may display a considerable degree of expression heterogeneity, overall resembling latency-III^{ref1, ref2} [106 and 201]
• several other gene expression patterns are recognized that cannot be grouped with the known latency patterns^ref [191]
Overview of EBV protein expression (for which suitable antibodies are available) in EBV-associated diseases

++, Positivity in >75% of neoplastic cells in all tested tumors; +, positivity in 10–50% of neoplastic cells in most of the cases; ±, positivity in <5% of neoplastic cells in part of the cases. Abbreviations: (see also Chapter 1) ePTCL, extranodal peripheral T-cell lymphoma; GC, gastric carcinoma; nd, not determined; ±, weak positivity in part of the cases; immunohistochemistry in these cases shows signals in small percentage of neoplastic and reactive cells; in isolated cases, weak transcription is found, although not confirmed by immunohistochemistry; therefore, these signals are ascribed to EBV-infected reactive cells.

• lytic phase : during the lytic replication phase of the EBV life cycle, many more viral genes are expressed which encode proteins involved in viral DNA replication and viral particle synthesis
• kinases expressed only during the lytic form of infection
• thymidine kinase
• BGLF4 phosphorylates the viral early antigen EA-D^ref
• immediate early (IE) gene products : genes which are expressed immediately upon induction of the lytic cycle, independently of new protein synthesis, and which encode transcription factors that are crucial for activation (switching-on) of lytic phase genes; essentially considered to form the switch between latent and lytic cycle. However, IE gene expression cannot be considered synonymous to full viral lytic replication and on the other hand some late gene products may be expressed without concomitant IE gene expression^{ref1, ref2} [39 and 54].
• BZLF1 / Zebra (Z) : the best studied IE gene product; a powerful trans-activator of early EBV gene expression^ref [39]. The Zebra protein consists of a trans-activating domain, a basic domain that has homology to a conserved region of the c-jun/c-fos family of transcription factors^ref [197], and a domain that enables Zebra to interact with p53^ref [198]. Overexpression of p53 and g-irradiation have been shown to induce Zebra expression in a NPC-derived EBV carrying epithelial cell line leading to the (partial) expression of lytic cycle genes^ref [199]. In most EBV-associated malignancies sporadic tumor cells show detectable Zebra protein expression, probably reflecting partial (abortive) activation of the lytic cycle, because true late gene transcription is mostly absent^{ref1, ref2, ref3, ref4} [54, 200, 201 and 202]. Expression of Zebra may interfere with IFN signalling, by decreasing IFN-receptor expression and preventing IFN-induced STAT1 tyrosine phosphorylation together, resulting in abrogation of IFN-induced MHC-II upregulation and thereby contributing to immune escape^ref [203]. BZLF1 converts the virus from the latent to the lytic form of infection even when the viral genome is highly methylated. Methylation of CpG motifs in Z-responsive elements of the viral BRLF1 immediate-early promoter enhances Z binding to, and activation of, this promoter. Demethylation of the viral genome impairs Z activation of lytic viral genes. Z is the first transcription factor that preferentially binds to, and activates, a methylated promoter^ref.
• BRLF1, BRRF1 (R) and the BI-LF4 transactivator genes^{ref1, ref2} [39 and 204], have been studied only little in relation to EBV+ tumor development and will not be discussed here.
BZLF1 and BRLF1 both encode transcriptional activators, and together these proteins induce transcription of the entire lytic viral gene program. In latently infected B cells, the BZLF1 (Zp) and BRLF1 (Rp) promoters are inactive. However, ligation of the B-cell receptor^ref, phorbol ester treatment^ref, calcium ionophores^ref, TGF-ß₁^ref, demethylating agents^ref, and agents which induce histone acetylation^{ref1, ref2, ref3} are known to activate expression of the BZLF1 and BRLF1 IE promoters in at least a portion of EBV⁺ B-cell lines. In the case of the BZLF1 IE promoter (Zp), the two promoter elements which appear to be essential for stimulation by most, if not all, of these inducing factors are termed the ZI and ZII motifs^ref. Several of the ZI motifs are bound by the MEF2D cellular transcription factor, as well as by Sp1/Sp3^{ref1, ref2}, and MEF2D has been shown to be an important regulator of EBV infection in host cells^{ref1, ref2}. The ZII motif is a CRE site which is bound by CREB, ATF-1, c-Jun, and ATF-2^{ref1, ref2}. The ability of both gemcitabine and doxorubicin to activate BZLF1 transcription in EBV-negative cells requires both the ZI and ZII binding motifs of the BZLF1 promoter. Gemcitabine and doxorubicin also activated the BRLF1 IE promoter in EBV-negative cells. This effect was shown to require the presence of 2 EGR-1 (Zif268) binding sites in the promoter which were previously shown to be important for phorbol ester-induced activation of the promoter^ref.
• early gene products : genes of which the expression is not affected by inhibition of viral DNA synthesis including series of enzymes influencing the host cell nucleotide metabolism and DNA synthesis; early lytic genes are barely expressed in EBV-associated malignancies and are not considered to contribute to the oncogenic process, with the possible exception of BHRF1. However, their expression can be induced by chemical treatment, irradiation or membrane receptor triggering of latently infected cells, which is mediated by prior expression of the EBV IE transactivator protein Zebra^ref [39]. Enzymes included among the early lytic genes are potential targets for antiviral drugs which may lead to applications for future tumor therapy^{ref1, ref2, ref3} [199, 205 and 206].
• BHRF1 mRNAs are abundantly expressed from their own promoter in BamHI-H (Hp) during early lytic infection^ref [39]; the protein can also be detected then and relates to the 17 kDa EA-R antigen^ref [207]. However, the protein is not detectable in most latently infected cell lines^ref [208] and transcripts of BHRF1 in these cells include leader sequences found in Cp/Wp EBNA transcripts^ref [209]. In vivo, BHRF1 transcripts are predominantly found in EBV-associated B-cell lymphomas^ref [210], although expression at the protein level apparently does not occur significantly^{ref1, ref2} [211 and 212]. Low level BHRF1 transcription occasionally can be detected in Hodgkin's lymphoma (HL) and in T-cell NHLs and even in nasopharyngeal carcinoma (NPC) and BHRF1 mRNA and protein expression is particularly abundant in oral hairy leukoplakia (OHL)^{ref1, ref2} [54 and 212]. Interestingly, BHRF1 shows structural and functional homology to the host-encoded apoptosis inhibitor Bcl-2^ref [213] and is highly conserved among gamma herpesviruses, suggesting an evolutionary conserved important role for BHRF1-protein in vivo^{ref1, ref2} [214 and 215]. It is thought that BHRF1 mediated apoptosis-inhibition contributes to prevent early lysis of productively infected cells by cytotoxic T-cells thus prolonging virus production time. A deregulated expression of BHRF1 in EBV tumor cells may similarly contribute to enhanced tumor cell survival [216].
• late gene products : genes of which the expression relies on new linear genomic templates and whose expression is blocked by inhibition of lytic (linear) viral DNA synthesis, including
• most of the virion structural proteins. Most late genes that can be identified directly or based on their homology with other herpesvirus genes encode structural proteins. Amongst these are :
• VCA p18 (BFRF3) the small capsid protein
• VCA-p40 (BdRF1) the scaffold protein^ref [217] 
• Gp125 (BALF4) the nuclear membrane protein
They are strongly immunogenic in humans and serve as targets for serodiagnosis, because virtually all EBV carriers develop antibodies to these proteins^{ref1, ref2} [92 and 218].
BLLF1 encodes the major virion envelope glycoprotein Gp350/220 that mediates virion binding to CD21 / CR2 / EBVR and is the major target of the neutralizing antibody response^ref [219].
• non-structural genes :
• viral IL-10 (BCRF1)^{ref1, ref2} is of particular interest because of its close homology with human IL-10 (hIL10). The structural homology consists of nearly 90% colinear amino acid sequence identity^ref [220] and BCRF1 is also functionally homologous to hIL10^{ref1, ref2} [221 and 222]. They share the ability to modulate local immune responses, to inhibit the function of macrophages and NK cells and the production of interferon-. Similar to the BHRF1 encoded viral bcl-2 homologue, viral IL-10 probably serves to enhance survival of virus producing cells in a hostile inflammatory environment. Both genes were probably acquired by EBV as a consequence of close co-evolution with its human host within cells of the immune system providing a selective survival advantage. In vivo, vIL10 (BCRF1) expression is mainly restricted to OHL in AIDS-patients^ref [54]. OHL is a lesion with a unique combination of latent and lytic EBV gene expression and forms the only proliferative disorder associated with lytic EBV replication^ref [151]. In contrast, hIL10 expression seems to predominate in most EBV-associated malignancies, probably directly linked to the expression of LMP1^ref [161]. Overall IL-10 expression is elevated in most EBV-associated diseases and can be detected both in affected tissues and in serum and is shown to correlate with a bad prognosis^{ref1, ref2, ref3} [223, 224 and 225].

Transmission : direct or indirect through oropharynx ("kissing disease"), blood or vehicles. Sexually active teens and adults may be exposed to a larger dose of EBV through particularly "deep" kissing, or possibly through genital fluids, which can carry the virus. Acquisition of EBV is enhanced by penetrative sexual intercourse, although transmission could occur through related sexual behaviors, such as "deep kissing." EBV type 1 infection is significantly more likely to result in IM. A large EBV type 1 load acquired during sexual intercourse can rapidly colonize the B cell population and induce the exaggerated T cell response that causes IM. Thus, IM could, perhaps, be prevented with a vaccine that reduces the viral load without necessarily inducing sterile immunity^ref. Under normal circumstances, EBV-infection is restricted to humans, although some types of monkeys can be infected experimentally^ref. EBV enters epithelial cells in the oropharyngeal mucosa through 3 CD21-independent pathways:

• by direct cell-to-cell contact of apical cell membranes with EBV-infected lymphocytes
• by entry of cell-free virions through basolateral membranes, mediated in part through an interaction between b₁or a₅b₁ integrins and the EBV BMRF-2 protein
• after initial infection, by virus spread directly across lateral membranes to adjacent epithelial cells.

Pathogenesis : the finding that individuals with the heritable disorder X-linked agammaglobulinemia harbour no EBV in their blood or throat washings and do not have EBV specific memory cytotoxic T lymphocyte (CTL) response, indicates that B lymphocytes, and not oropharyngeal epithelial cells, are required for primary EBV-infection^ref [226]. At present, it is thought that primary EBV-infection occurs in the oropharynx via exchange of cell-free virus or productively infected cells in saliva^ref [241]. The critical steps underlying the initial EBV B-cell transformation are depicted below :

Schematic presentation of the initial events leading to EBV-induced transformation of human B-cells. The first steps consist of virion CD21 binding, cell penetration and capsid transport, host cell activation and viral gene expression following release of the linear viral genome in to the cell nucleus. Subsequently, a well-orchestrated series of molecular events is initiated, starting with the expression of EBNA2 and EBNA-LP and leading to the expression of LMP1 the major viral oncogene and multiple host genes as well and to the host-driven circularisation of the viral genome. Finally, expression of EBNA1 is induced allowing the episomal viral genome to replicate with dividing cells, thus completing the growth transformation process. Additional expression of EBERs, BARTs and LMP2a will modify the transformed state but are not essential for this process. The resulting EBV transformed cell will continue to grow indefinitely, expressing the latency-III program. In vivo this program is not tolerated by the immune system and EBV-infected cells will (be forced to) switch-off expression of the major immunogenic proteins. In vivo EBV persists in transcriptionally silenced memory B-cells which only express non-coding EBER1, 2 and BARTs and occasional low levels of LMP2 and EBNA1.

Schematic representation of EBV RNA expression profiles in different EBV-infected B-cell populations in tonsils and peripheral blood of healthy EBV-carriers and their proposed relationship. Naive B-cells are infected by EB-virions that enter the oropharyngeal lymph nodes by crossing epithelial barriers. EBV glycoprotein gp350 binds to the receptor CD21, present on all B-cells. Under influence of CD21-triggering and the subsequent EBNA2-driven transcription program, naive B-cells differentiate into B blasts that express the full set of latent EBV RNAs and which are probably controlled by anti-EBV cytotoxic T-cell responses. The B blast further differentiates through the germinal center via centroblast to centrocyte. Both cell types express the latent membrane proteins LMP1 and LMP2, which provides them with growth and survival signals in absence of antigen, and EBNA1 which is essential for maintenance of the viral genome in the host cell. Finally, differentiated EBV-infected cells enter the circulation as memory B-cells, that are generally silent for viral RNA expression but may occasionally express LMP2 RNA. Infection of B lymphocytes is mediated through interaction of the viral Gp350/220 with CD21 / CR2 / EBVR, the physiological receptor for the complement factor C3d^{ref1, ref2} [227 and 228]. After CD21 binding the viral envelope fuses with the host cell membrane and triggers host cell activation^{ref1, ref2} [227 and 229]. The nucleocapsid is then transported to the nuclear boundary and subsequently degraded, releasing the linear viral DNA into the nucleus, which then leads to initial EBNA2 and EBNA-LP transcription^ref [39]. EBNA2 and EBNA-LP are essentially required for initial B-cell growth transformation, but their activity is modulated by the subsequent expression of EBNA3A–C. A second essential event is the recircularization of the viral genome at the terminal repeats, for which the host cell DNA repair machinery is used^{ref1, ref2} [230 and 231]. Although episomal forms are by far prevailing in EBV-associated malignancies in vivo, a small number of studies have described cell lines with integrated EBV^{ref1, ref2, ref3} [232, 233 and 234]. In the first stages after recircularization of the genome, the full spectrum of latent proteins is expressed^ref [39]. This complex event which leads to efficient growth transformation of the host cell includes the well coordinated transcription of poly-cistronic mRNAs derived from promoters in BamHI-C and -W, encoding EBNA1 together with one or more of the other EBNAs (EBNA2, -3a, -3b, -3c and -LP)^ref [39] and transcription of the major EBV oncogene LMP1, encoded in BNLF1. This transcription pattern allows rapid polyclonal expansion of the infected B-cells as lymhoblasts and is, therefore, referred to as the ‘growth program’^{ref1, ref2, ref3} [171, 235 and 236]. In vivo, incoming EBV predominantly infects subepithelial B-cells to become transformed B-blasts^{ref1, ref2} [39 and 237], but these are rapidly eliminated by the host T-cell response, which is mainly directed against viral lytic genes during early primary infection and against EBNA3a, -3b and -3c during lifelong persistence^{ref1, ref2, ref3, ref4} [237, 238, 239, 240 and 241]. The fulminant T-cell response against these freshly EBV transformed cells in adolescents and adults is the basis of the mononucleosis syndrome^ref [241]. Throughout life, outgrowth of EBV transformed B-cells is kept under control by a persistent alert immune response to latency-associated products, especially involving EBNA3A–C and LMP2^ref [240]. Part of the EBV-infected lymphoblasts proliferate and differentiate through a germinal center-type reaction and subsequently enter the peripheral B-cell pool as resting memory cells^{ref1, ref2, ref3, ref4, ref5} [171, 235, 236, 242 and 243]. It is likely that the continuous but limited expression of latent genes associated with the growth program is responsible for maintaining the high level of T-cell memory to these potentially dangerous cells. The pathogenic consequences of failing T-cell surveillance is seen in immunosuppressed individuals who are at high risk of developing lymphoproliferative disease and malignant lymphoma, that are predominantly EBV driven^ref [241]. Like other herpesviruses, EBV persists lifelong in its host. Total body irradiation of EBV-seropositive individuals awaiting bone marrow transplantation leads to eradication of the virus^ref [244]. This indicates that the cellular compartment where the infected cells reside must be associated with the hemopoietic tissue. More recent studies have shown that resting memory B-cells are the reservoir for latently present EBV^{ref1, ref2, ref3, ref4, ref5, ref6} [170, 171, 245, 246, 247 and 248]. In fact, it was already known for many years that it is possible to grow EBV transformed B-cells directly from the blood of virtually all EBV-seropositive individuals, provided that T-cells are depleted or suppressed in vitro^ref [241]. In vivo, these infected cells can escape CTL-mediated killing because the expression of immunogenic EBV-proteins such as EBNA3a, -3b and -3c is silenced once latent infection has been established^{ref1, ref2, ref3, ref4, ref5} [39, 246, 247, 248 and 249]. This silencing is accomplished by methylation of early latency promoters Cp and Wp^{ref1, ref2, ref3} [66, 250 and 251]. Because EBNA1 is indispensible for maintenance of the viral genome in the dividing host cells^{ref1, ref2} [39 and 77], its transcription is continued, but now initiated from an autoregulated promoter in BamHI-Q (Qp)^ref [82]. Interestingly, Qp-driven EBNA1 transcription is found in all EBV-associated malignancies of non-immunocompromized patients. Although EBNA1 is a foreign protein to the host, EBV-infected cells expressing EBNA1 are not killed by CTLs due to the inhibitory effect of the protein's Gly–Ala repeat on proteasomal processing and subsequent MHC class I-restricted presentation^ref [90]. This is considered to be an important mechanism by which EBV⁺ tumor cells escape CTL-mediated killing^ref [249]. Given the fact that memory CTLs reactive against most EBNAs remain clearly detectable for life^{ref1, ref2, ref3, ref4} [238, 239, 240 and 241], it is generally thought that there must always be a subset of B-cells present that express the growth program. Recently, this subset was shown to consist of naive (IgD⁺) B-cells in the mantle zones of the tonsil^{ref1, ref2} [246 and 247]. Circulating EBV⁺ B-cells may express defined homing receptors and cytokine response receptors, that preferentially direct them to the epithelial surfaces in the body, where these cells may be periodically triggered into the lytic cycle in order to maintain shedding of infectious virus in the oropharynx^ref [241]. It is thought that this switching process is influenced by signals that normally control B-cell behaviour, such as antigen-driven activation^{ref1, ref2, ref3, ref4} [171, 241, 242 and 243]. Indeed careful analysis of the localization of EBV⁺ B-cells in tonsils and salivary gland epithelia during infectious mononucleosis (IM) and latent carriership have revealed that these cells preferentially locate in the interfollicular region rather than in the germinal center and also accumulate around the crypts and subepithelial layers of the tonsil^{ref1, ref2, ref3} [237, 252 and 253]. The differential expression of defined cytokine receptors, such as CCR6, CCR7, CCR10 and CXCR4 and CXCR5 may be responsible for this phenomenon^ref [254]. Local antigen triggering and CD40 activation may subsequently lead to virus production and transepithelial secretion^ref [241]. The persistent (low level) production of infectious virus progeny is reflected by the life-long presence of IgG antibodies to the viral capsid antigen/membrane antigen (VCA/MA) complex^{ref1, ref2} [92 and 255].
Table 2. EBV latent gene expression patterns in EBV associated disorders

Table 3. Overview of EBV transcriptiona in EBV associated diseases
Important factors during virus-induced lymphomagenesis and carcinogenesis are growth transformation in combination with genetic instability, inhibition of apoptosis, angiogenesis and inhibition of (or evasion from) the local immune response. Individual EBV gene products were shown to exert such effects in vitro, and in concerted action are considered to play an important role in the genesis of EBV⁺ malignancies in vivo. Different EBV gene expression patterns are recognized, which are generally referred to as the latency programs (types) I, II and III. It is conceivable that these expression patterns are subject to the nature of the cells from which the respective malignancies are derived, or that they are in fact subject to the host immune response as is the case in PTLD and ARL. In addition to these well-established gene expression patterns, differential expression patterns were found among other EBV genes that encode homologs to human proteins involved in proliferation, differentiation, apoptosis inhibition and suppression of the local immune response. Some viral latency genes may predominantly provide their function in the initial B-cell transformation, like EBNA2 and EBNA-LP, and are subsequently downregulated or even switched off. Other latency genes, like EBNA1 and LMP2a, may be more essential in maintaining the long-term survival of the EBV genome in a resting cell environment, whereas LMP1 may provide the temporary growth kick when such cells pass through lymph nodes^ref [171]. The pleiotropic effects of LMP1 expression, together with external growth or differentiation inducing stimuli may well provide the basis for pre-malignant growth. It is fascinating to realize that long-term evolutionary co-existence have thought the virus to evade immune elimination by preventing the recognition of its most essential gene products, especially EBNA1 (via Gly–Ala proteasome inhibition) and LMP1, which is non-immunogenic and has distinct direct and indirect immunosuppressive functions^{ref1, ref2, ref3} [131, 351 and 437]. As indicated before the well regulated and well balanced, low level expression of LMP1 may be most crucial for persistence of EBV transformed B-cells. When activated in subepithelial layers additional viral genes can be expressed as well providing growth or survival functions. Transcripts encoding BHRF1 (a functional homolog of Bcl-2) were mainly found in B-cell lymphomas (both of immunocompetent and immunocompromized patients)^{ref1, ref2} [210 and 211]. Weak BHRF1 transcription signals were found in some T-cell lymphomas and HDs, but these signals are ascribed to the presence of EBV-positive reactive cells^{ref1, ref2, ref3} [54, 210 and 212]. BHRF1 transcription and protein expression at the single cell level in most of the lymphoid disorders has yet to be determined. In NPC the data suggest that BHRF1 expression in NPC is only limited. Preliminary data indicate that BHRF1 protein can also be detected by immunohistochemistry in this disorder^ref [212]. Particularly strong expression of early BHRF1 transcripts is confined to OHL^{ref1, ref2} [54 and 431]. It is thought that the putative BHRF1 protein may act in addition to the LMP1 induced Bcl-2 protein acts in prevention of host cell apoptosis, enhancing survival of the host cells during production of viral progeny ^{ref1, ref2}[438 and 439]. Newly synthesized virions released into the environment or by cell-cell contact with surrounded cells EBV may be transmitted. When the virus then is capable of entering an epithelial cell additional genes, that are not part of the B-cell program may become expressed, one of which is the BARF1 gene. BARF1 was originally recognized as an early transcript^ref [189]. Using a reversed northern blotting technique, the transcript was found to be preferentially expressed in epithelial but not in lymphoid cells^ref [190]. More recently, using NASBA, we found BARF1 transcription exclusively in NPCs and in OHL but not in lymphoid disorders^{ref1, ref2} ( Fig. 5) [50 and 54]. In addition to these studies, BARF1 transcripts were also detected in EBV-associated gastric carcinomas^ref [191]. Recently, expression of BARF1 at the protein level was detected in NPC biopsies by immunoblotting^ref [192]. These findings indicate that BARF1 expression is specific for EBV-associated epithelial malignancies. BARF1 has clear effects on epithelial cells in vitro^ref [194], in more than one aspect mimicking the role of LMP1 in B-cells^ref [440]. Transcription of BCRF1 (a functional homolog of human IL-10) occurs almost exclusively in OHL [54], which represents a productive infection. In this and other productive infections, the BCRF1 protein may play a role in the inhibition of the local immune response. These observations together suggest that multiple products encoded within the EBV genome may exert functions that may be relevant for pathogenic events in vivo. A number of genes with putative interesting functions, such as the cytokine receptor encoded in BILF1, BDLF2, a protein with homology to cyclin-B^ref [54] and the BFRF1 putative virion protein, remain to be explored^ref [441]. Furthermore, the in vivo expression and functional role (if any) of the intriguing but yet illusive proteins encoded in ORFs RPMS1, A72 and (RK-)BARF0 located within the BARTs that are abundantly transcribed in all EBV-associated malignancies remain to be defined. In addition, the relative importance and interactive collaboration by the individual gene products and the relevance of subtle mutations found in various EBV isolates recovered from directly from tumor tissues or patients with distinct clinical syndromes associated with EBV-infection largely remain to be established. Further unraveling of these viral functions and their ‘well orchestrated (inter)action(s)’ may permit the development of antiviral agents capable of interfering with these functions with options for future therapeutic intervention. EBV and some of its latent gene products have been demonstrated to modulate the expression of various cellular (proto-onco-) genes in vitro. The most obvious way to study the effect of the presence of EBV on the expression of these genes in vivo, is the comparison of their expression in EBV-positive and -negative cases of a single clinical entitiy. In this context, HD is a suitable model, because both EBV-associated HDs and EBV-negative HDs are recognized. Moreover, several findings suggest that EBV-associated and EBV-negative HDs have a different pathogenesis. For example, H-RS cells in EBV-associated HDs express high amounts of MHC class I molecules, whereas MHC class I appears to be dowregulated in EBV-negative HDs^ref [442]. MHC class II and a number of co-stimulatory molecules as well as intracellular TAPs associated with peptide transport after proteasomal degradation seems to be functionally intact^ref [362]. This is also found for NPC and T/NK cell lymphomas^{ref1, ref2} [336 and 394]. Most EBV-associated HDs express particularly abundant levels of LMP1 and LMP2, proteins which normally can evoke a CTL response. However, H-RS cells apparently are not killed by CTLs^{ref1, ref2} [363 and 364]. Several underlying mechanisms have been proposed for this phenomenon. It is thought that the presence of the CTLs results in a selection of H-RS cells that are resistant to CTL-mediated (and probably also therapy-mediated) apoptosis. HDs in which this has occurred, are likely to express low amounts of p53 and high amounts of bcl-2 in their H-RS cells^ref [368]. However, the expression of p53 (as determined by immunohistochemistry) is not related to the presence of EBV, and the expression of bcl-2 even shows an inverse relation to the presence of EBV^ref [368]. This indicates that induction of apoptosis resistance by modulation of p53 and bcl-2 expression is probably not a mechanism used by EBV in HD. This may also hold for NPC where high levels of p53 and bcl2 coexist, together with highly expressed PCNA^{ref1, ref2} [391 and 392]. Alternatively, p53 mediated apoptosis may be inhibited by upregulation of A20 expression via LMP1. We tested A20 expression in EBV-positive and -negative HD cases using NASBA, but we could not find a relation between A20 transcription and the presence of EBV (Brink et al., unpublished data). This seems to be in agreement with the finding that EBV apparently does not interfere with the normal function of p53 in HD^ref [204]. It would be best to confirm these data morphologically, but at present no antibodies are available that recognize human A20 in clinical material. To prevent CTL-mediated killing of EBV-infected H-RS cells, inhibition of the local CTL response by EBV is also a possible mechanism. EBV-associated HDs express significantly higher amounts of human IL-10^{ref1, ref2, ref3} [161, 362 and 443], which may contribute to immune evasion. We have shown that HDs do not express the viral homolog of Interleukin-10 (BCRF1) but that the detected IL-10 expression is of host cell origin^ref [161]. Upregulation of human IL-10 upon EBV-infection in vitro has been demonstrated^ref [365], and, interestingly, elevated human IL-10 expression levels have been detected in the H-RS cells of EBV-positive HDs^{ref1, ref2, ref3} [161, 362 and 443].
Table 6. Overview of cellular genes and their functions of which expression and/or function is modulated by EBV

None: no differences found for EBV+ and EBV? cases.

Using RT-PCR, we have also analyzed the expression of the proto-oncogene c-fgr. C-fgr was shown by others to be upregulated upon EBV-infection, and EBNA2 is most likely the EBV gene to induce c-fgr expression^{ref1, ref2} [96 and 444]. Moreover, alternative splicing of c-fgr transcripts was reported in EBV-positive cell lines^ref [445]. Using RT-PCR, we have performed a qualitative analysis of c-fgr transcription in EBV-positive (n=4) and -negative HDs (n=4), in EBV-associated PTDLs (n=3) and ARLs (n=4), and one IM case. We found no clear relation between the presence of EBV and c-fgr transcription in HD (Brink et al., unpublished data). Moreover, c-fgr RT-PCR signals in HD were relatively weak compared with positive controls. In most PTLDs and ARLs we found transcription of c-fgr, including the EBV specific alternative transcript. These data suggest that upregulation of the c-fgr proto-oncogene is not a mechanism used by EBV in HD, but may play a role in PTLDs, ARLs and IM. Moreover, these data strengthen the finding that EBNA2 (which is expressed to some extent in PTLDs, ARLs and IM but not in HD) is the EBV gene responsible for c-fgr upregulation. Future studies should aim to investigate c-fgr protein expression at the single cell level. In conclusion, differences in EBV gene expression patterns exist between the various EBV-associated diseases. This is true for lymphomas of immunocompromized versus immunocompetent patients, for lymphomas of B-cell versus T-cell origin, and most strikingly for lymphomas on the one hand and epithelial disorders on the other. For some of the differentially expressed genes, such as BARF1, it remains to be investigated whether the observed expression pattern is subject to host cell regulation factors or contributes actively to differences in pathogenesis. Therefore, future studies should not only aim to determine EBV gene expression in EBV-associated diseases, but should also include more fundamental research on expression regulation and functional interactions. Moreover, it is important to notice that high mRNA expression does not always coincide with expression at the protein level; we have shown this for BARF0 but it may well be that the EBV genome gives rise to other non-translated transcripts. In the future, it would be worthwhile to develop additional antibody reagents against the various (putative) EBV-proteins and use these for in situ expression analysis, in combination with mRNA profiling^ref. EBV-induced gene 3 (EBI3) is expressed in DCs and is part of the cytokine IL-27^ref that controls T_h cell development. However, its regulated expression in DCs is poorly understood. EBI3 is expressed in splenic CD8^-, CD8⁺, and plasmacytoid DC subsets and is induced upon TLR signaling. Cloning and functional analysis of the EBI3 promoter using in vivo footprinting and mutagenesis showed that stimulation via TLR2, TLR4, and TLR9 transactivated the promoter in primary DCs via NF-kB and Ets binding sites at -90 and -73 bp upstream of the transcriptional start site, respectively. Furthermore, NF-kB p50/p65 and PU.1 were sufficient to transactivate the EBI3 promoter in EBI3-deficient 293 cells. Finally, induced EBI3 gene expression in DCs was reduced or abrogated in TLR-2/TLR4, TLR9, and MyD88 knockout mice, whereas both basal and inducible EBI3 mRNA levels in DCs were strongly suppressed in NF-kB p50-deficient mice. In summary, EBI3 expression in DCs is transcriptionally regulated by TLR signaling via MyD88 and NF-kB. Thus, EBI3 gene transcription in DCs is induced rapidly by TLR signaling during innate immune responses preceding cytokine driven T_h cell development^ref. Increased numbers of EBV-infected cells in areas of active inflammatory bowel disease are secondary to influx or local proliferation of inflammatory cells & do not contribute significantly to local production of EBI3^ref
Release of progeny virions from polarized cells occurs from both their apical and basolateral membranes. No CPE, no productive replication in vivo, immortalization of B lymphocytes in vitro. Polyclonal activation of B cells => heterophilic IgMs.
Expression of IL-7Ra was lost from all CD8⁺ T cells, including EBV epitope-specific populations, during acute infectious mononucleosis (IM). Thereafter expression recovered quickly on total CD8⁺ cells but slowly and incompletely on EBV-specific memory cells. Expression of IL-15Ra was also lost in acute IM and remained undetectable thereafter not just on EBV-specific CD8⁺ populations but on the whole peripheral T and NK cell pool. This deficit, correlating with defective IL-15 responsiveness in vitro, was consistently observed in patients up to 14 years post-IM but not in patients after CMV-associated mononucleosis, nor in healthy EBV carriers with no history of IM, nor in EBV-naive individuals. By permanently scarring the immune system, symptomatic primary EBV infection provides a unique cohort of patients through which to study the effects of impaired IL-15 signalling on human lymphocyte functions in vitro and in vivo^ref.

Symptoms & signs :

• largely subclinical in early childhood : latency II and III expression patterns in tonsils of healthy asymptomatic healthy carriers
• lymphoid diseases :
• infectious mononucleosis (IM) (18%; once termed Drusenfieber, i.e. Pfeiffer's glandular fever) after 30-60 days incubation (10-15 days in babies) : colonization of Waldeyer's ring => fever (66-76%, only in first 10 days), cough (14%), pharyngotonsillitis (1.6%; "kissing tonsils" : if severe bacterial overinfection and oedema => cortisone therapy); anterior and posterior laterocervical, inguinal and axillary lymphadenomegaly (5-94%, for 2-3 weeks), mild hepatomegaly (12%; increase in SGPT up to 100÷200 U/L at 37°C), eyelid edema (5.3%), infantile hepatitis syndrome (0.5%), jaundice (9%), splenomegaly (52%) due to red pulp hyperplasia (sometimes spleen upper pole may enlarge from the 7^th÷8^th intercostal space down to the left iliac cavity : rare splenic infarction and splenic rupture; avoid splenic injuries during recovery !), maculopapular exanthema (8-10%), mild sinusal tachycardia (1.6%), hyperhemic tonsils with no exudate, gross hematuria (0.5%), pericarditis and interstitial pneumonia. Genital ulcerations can be the initial manifestation of EBV in adolescents who are neither sexually active nor immunocompromised practicing cunnilingus^ref. IM can be considered the clinically manifest form of a primary EBV-infection. Its diagnosis relies in the detection of atypical lymphoid cells in the peripheral blood, the occurrence of so-called heterophile antibodies and EBV-seroconversion^ref [276]. IM is a benign disorder with expansion of the paracortex of lymphoid tissues. The proliferating cell populations are EBV-infected polyclonal B blasts^ref [277], accompanied by the growth of activated T-cells^ref [278]. Morphologically, the EBV-infected cells range from large immunoblasts, including H-RS like cells, to small lymphoid cells. In addition to latently infected cells, a small number of EBV-positive cells expressing lytic cycle antigens such as BZLF1 can be detected in tissues from IM patients. These are often found adjacent to crypt epithelium and are frequently positive for plasma cell markers^{ref1, ref2, ref3} [252, 253 and 279]. It is thought that EBV can replicate in these plasmacytic cells, thus generating a cellular source of infectious virus in the saliva of IM patients^{ref1, ref2} [241 and 253]. Further EBV gene expression analysis has shown a type III EBV latency pattern in IM [235], although the expression at the single cell level is more heterogeneous. Most cells are EBER positive but do not express EBNA2 or LMP1 (type I latency), whereas some large immunoblasts express LMP1 in the absence of EBNA2 (type II latency) and many small lymphocytes express EBNA2 but not LMP1^{ref1, ref2} [237 and 279].

Laboratory examinations : leukocytosis (> 4,500÷10,000 / mL) with relative reactive lymphocytosis due to proliferation of Downey cells / atypical lymphocytes (i.e. CD8⁺ T_s cells) :
• type I cells : kidney-shaped or lobulated nucleus with vacuolated, basophilic foamy cytoplasm
• type II cells contain plasmacytoid nuclei with less vacuolated and basophilic cytoplasm
• type III cells have a finer chromatin pattern and 1-2 nucleoli.
..., decrease in HCT, HGB and MCV (differential diagnosis with massive hemolysis, internal or external hemorrhages, hemoglobinopathies)
Prognosis : self-limiting within 2-4 weeks.
Complications :
• acute lymphadenitis (1.0%)
• subacute necrotizing lymphadenitis (0.5%)
• facial neuritis (1.0%)
• acute aplastic anemia (0.5%)
• viral myocarditis (2.6%)
• EBV-associated hemophagocytic syndrome (EBV-AHS) (1.6%)
• post-infective meningoencephalitis (PIE : < 1%). EBV CNS infections are divided into 2 groups :
• CNS syndromes associated with primary EBV or reactivated infection : diverse CNS syndromes can occur
• acute viral encephalitis (2.6%) : convulsion (1.6%)
• acute cerebellar ataxia
• acute disseminated encephalomyelitis (ADEM) (sometimes positive PCR from CSF)
• myelitis
• meningitis (sometimes positive PCR from CSF)
• CNS syndromes associated with chronic EBV infection
• encephalitis
• relapsing ADEM
• X-linked lymphoproliferative syndrome (X-LPS) / fatal mononucleosis : caused by hereditary mutations in the gene encoding the signalling lymphocyte activation molecule (SLAM)-associated protein (SAP) on position q25 of the X-chromosome^{ref1, ref2} [280 and 281]. The self-ligand SLAM protein is present on the surface of both B and T-cells. When interactions occur between SLAM molecules on the interface between T- and B-cells, signal transduction pathways are initiated. The T-cell protein SAP binds to SLAM and as such acts as a negative regulator^{ref1, ref2} [282 and 283]. Individuals that have inherited this trait are usually asymptomatic, but upon primary EBV-infection their immune response becomes overreactive because of the non-functional SAP protein. This usually results in a fulminant IM accompanied by a EBV-associated hemophagocytic syndrome (EBV-AHS) by which liver and bone-marrow are destroyed^ref [284]. The only curative treatment for this syndrome is allogeneic bone-marrow transplantation^ref [285]. However, patients who do survive the infection are prone to developing lymphomas later in life^ref [286].
• EBV-mild acquired immune deficiency syndrome (EBV-MAIDS) in postsurgical sinusitis

Therapy : periodic intramuscular IVIg
• autoimmunity may cause :
• autoimmune hemolytic anemia
• autoimmune thrombocytopenic purpura (ATP) (5.8%)
• autoimmune granulocytopenia
• Guillain-Barré syndrome
• type IA insulin-dependent diabetes mellitus
• polydermatomyositis
• rheumatoid arthritis (1.0% : RF in < 5%)
• chronic fatigue syndrome (CFS) : inability to demonstrate a role for reactivation of EBV in CFS, even in selected patients with high titers of antibody to VCA and EA of EBV
• systemic lupus erythematosus (SLE) (0.5%)
• Kawasaki disease (6.3%)
• immunoproliferative syndrome / lymphoproliferative disease (LPD) (latency III expression pattern) in stem cell and organ transplant recipients (Cohen, 2000) : the present view that lymphomas have to be considered neoplastic counterparts of reactions normally occurring in lymphoid tissues after antigenic stimulation, is helpful in understanding their pathogenesis. Lymphomagenesis is considered to be a multistep process^ref, during which an accumulation of genetic changes takes place. Lately, numerous studies have identified cytogenetic abnormalities that are characteristic of specific non-Hodgkin lymphomas (NHLs). These are frequently translocations involving the antigen receptor loci. Since transcriptional activity of the antigen receptor loci is inherent to the function of lymphoid cells, translocations involving these loci usually result in overexpression of an oncogene under the control of the antigen receptor expression regulation. A characteristic example is the t(14; 18) in follicular lymphomas, which results in overexpression of the apoptosis inhibitor bcl-2 under the control of the Ig heavy chain promoter^{ref1, ref2}. The finding that cells containing t(14; 18) can be detected in the peripheral blood of healthy individuals^ref indicates that this translocation may be an early event in the genesis of follicular lymphoma. Another translocation involving the Ig heavy chain locus is t(8; 14) which occurs in many of the Burkitt's lymphomas (BL). This translocation results in overexpression of the transcription factor c-myc under the control of the Ig heavy chain promoter^{ref1, ref2}. The difference in growth rate between follicular lymphomas and BL is clearly the consequence of the different genes involved in their respective genetic alterations. The t(14;18) involving bcl-2 results in an extended lifespan of the affected cells, allowing the accumulation of more genetic hits. By contrast, genetic changes resulting in overexpression of a transcription factor usually have a direct influence on proliferation and, therefore, are usually associated with rapidly growing lymphomas, as is the case with c-myc in BL. Besides cytogenetic abnormalities, chronic antigenic stimulation is thought to play an important role in lymphomagenesis. The presence of antigens not only stimulates proliferation (resulting in ‘fixation’ of genetic aberrations throughout the next cell generations) but also induces the process of Ig rearrangement, and thus increases the possibility of more aberrant rearrangements to take place^ref. This can also be illustrated from the c-myc translocations in BL. Recent findings show that all cases of BL harbor somatically mutated V region genes and are as such derived from B-cells that took part in a germinal center reaction^ref. It was proposed that most, if not all, c-myc/Ig translocations in endemic BL happen in a mutating germinal center B-cell and result from the process of hypermutation as such. In sporadic BL, c-myc translocations are usually targeted to the class-switch region of the Ig gene, and thus probably relate to class switch recombination that also take place in the germinal center^{ref1Lenoir GM, 173-206, ref2}. Some lymphoma subtypes are clearly related to the presence of a certain antigen. For example, gastric lymphomas are related to infection with Helicobacter pylori (H. pylori)^ref, and certain cutaneous B-cell lymphomas are associated with Borrelia burgdorferi infection^ref. In AIDS-related non-Hodgkin Lymphoma (ARNHL) similar chronic antigen exposure may be responsible for tumor outgrowth, but the triggering antigen remains elusive^ref. For the endemic form of BL, the antigens providing chronic stimulation are probably Epstein–Barr virus (EBV) and malaria^ref. The latter provides a strong and acute B-cell activation stimulus mediated by repetitive epitopes on the parasite surface, whereas the action of EBV in BL may be 2-fold: it functions as antigen and as a transforming agent. The presence of somatic hypermutations in the variable region of the Ig heavy chain in BL cells strengthens the hypothesis that exposure to antigens may be relevant for the pathogenesis of BL^{ref1Lenoir GM, 173-206, ref2}. Furthermore, the presence of Ig-gene hypermutations in the Reed–Sternberg cells of Hodgkin's disease^ref and the expression of functional differentiation markers such as granzyme-B and TIA-1 in all EBV-associated T-/NK-cell NHL's^ref also suggest that these tumors are initially driven by antigen stimulation. Similar to the situation in endemic BL, the presence of EBV genome may provide immortalizing functions thus contributing to the multistep oncogenic process^{ref1, ref2}.
• autoimmune lymphoproliferative syndrome (ALPS)
• Kikuchi-Fujimoto disease (KFD) / subacute or histiocytic necrotizing lymphadenitis (HNL) ?
• EBV-associated T/NK-cell LPD : target cell specificity, defects in host immune responses, and strain differences of EBV may account for ectopic EBV infections and for the unique clinical presentations characteristic of each illness^ref.
• chronic infectious mononucleosis (CIM) / chronic Epstein-Barr virus infection (CEBV) / chronic active EBV infection (CAEBV) in apparently immunocompetent hosts (extraordinary event after acute disease)

Symptoms & signs : chronic or recurrent infectious mononucleosis-like symptoms (intermittent fever, weight loss and liver abnormalities, rarely multiple nodular coagulation necrosis in the liver) persisting over a long time and by an unusual pattern of anti-EBV antibodies^{Rickinson AB, 13-14}. Patients with this disease have no evidence of any prior immunologic abnormalities or of any other recent infection that might explain their condition^{Rickinson AB, 13-14, ref2}. CAEBV is a disease with a high mortality and high morbidity with life-threatening complications, such as virus-associated hemophagocytic syndrome, interstitial pneumonia, lymphoma, coronary artery aneurysms, and central nervous system involvement^{ref1, ref2, ref3, ref4, ref5}. The 3 main criteria of CAEBV infection are^ref :
• (1) severe illness lasting > 6 months that began as a primary EBV infection and that was associated with grossly abnormal EBV antibody titers, antiviral capsid antigens (VCA) IgG > 5120, anti-early antigens (EA) IgG > 640, or anti-EB nuclear antigens (EBNA) < 2;
• (2) histologic evidence of major organ involvement such as interstitial pneumonia, hypoplasia of some bone marrow elements, uveitis, lymphadenitis, persistent hepatitisor splenomegaly; and
• chronic hepatitis might be a manifestation of chronic EBV infection in the lack of detectable immune deficiency; the expansion of CD28^-CD27^- and increase of functional EBV-specific CD8⁺ T cells being the only surrogate markers of viral activity^ref
• (3) increased quantities of EBV in affected tissues.
Subsequently, Okano and his colleagues^ref proposed similar criteria to diagnose severe CAEBV infection.
Importantly, many cases have been reported that do not satisfy the criteria described above. Some patients lack abnormal patterns of EBV-related antibodies, whereas other patients lack major organ involvement and have only skin symptoms, such as hypersensitivity to mosquito bites (HMB) or hydroa vacciniforme-like eruptions^{ref1, ref2}. On the other hand, patients with CAEBV infection had extremely high viral loads, as assessed by qPCR^{ref1, ref2}. Clonal expansion of EBV-infected T or natural killer (NK) cells could be associated with CAEBV infection^{ref1, ref2, ref3, ref4, ref5, ref, ref7}. Detailed clinical features of some patients have been described^{ref1, Morita M, 1485-1488, ref3}. An accumulating body of evidence suggests that clonal expansion of EBV-infected T or NK cells is associated with CAEBV infection^{ref1, ref2, ref3, ref4, ref5, ref6, ref7}. In healthy carriers, EBV exists latently in resting memory B cells^ref. It is unclear whether the invasion of blood cells other than B cells causes CAEBV infection or the invasion is an ordinary event, but the host's immunologic abnormalities allow the expansion of these cells. Recently, a defect in the SAP/SH2D1A gene was implicated in XLP, a fatal lymphoproliferative disorder^{ref1, ref2, ref3}. Patients with XLP are exclusively boys in whom primary infection with EBV causes lymphoproliferation and severe hepatitis, mimicking CAEBV infection^{ref1, ref2}. Therefore, we examined the SAP/SH2D1A gene in all the male subjects enrolled in our study, but did not find any abnormalities. However, it is possible that a defect in another gene essential for regulating lymphocyte activation and proliferation may be a cause of CAEBV infection. Such a defect or single nucleotide polymorphism might influence the function of virus-specific or nonspecific lymphocytes and thereby allow the expansion of EBV-infected T or NK cells. It has been reported that EBV-infected T and NK cells express a limited range of viral antigens^ref. EBV-infected B cells express at least 9 antigens, some of which are antigenic enough to be presented to CTLs. This is called latent infection type 1. EBV-infected T and NK cells express only EBNA1 and LMP1; this is called latent infection type 2^ref. The expression of viral antigens was examined in some CAEBV patients, and they were type 2^{ref1, ref2}. Because these 2 proteins are less antigenic, infected cells can evade from immune surveillance by CTLs. Therefore, they may proliferate and cause chronic infection. Chromosomal abnormality occurred in the lymph nodes of patients with CAEBV^ref : 50% of the patients examined had chromosomal aberrations in their peripheral blood cells and 79% of patients showed monoclonality of EBV. Because there were no specific patterns of genetic aberration, and some patients displayed several different aberrations, the chromosomal abnormalities seen in patients with CAEBV infection might only reflect chromosomal fragility. However, clonality of the EBV genome and chromosomal aberrations generally indicate clonal expansion of EBV-infected cells. These results indicate that clonal expansion is a common feature of CAEBV infection, and this disease might therefore be considered lymphoproliferative rather than infectious^ref. In fact, patients with CAEBV infection frequently develop neoplasms such as malignant lymphoma. It would therefore be better to call the cases presented here "chronic EBV-associated lymphoproliferative disorders." Approximately 10 out of 30 patients with CAEBV infection did not meet the previously accepted definition of CAEBV infection^{ref1, ref2}. These patients had symptoms typical for CAEBV infection and had extremely high EBV loads in their peripheral blood. High titers of EBV-related antibody are not always necessary for the diagnosis of CAEBV infection. On the other hand, all the patients had > 102.5 copies/µg EBV DNA in PBMC. Furthermore, it is of note that the copy number of EBV DNA in PBMC decreased below 102.5 copies/µg DNA in all 7 patients who underwent hematopoietic stem cell transplantation. These results indicate that viral loads greater than102.5 copies/µg DNA can be used not only as diagnostic factors for CAEBV infection, but also as predictors of therapeutic efficacy. Previous reports show that patients with CAEBV infection had cell-free EBV DNA in plasma, although the origin of the viral DNA is unclear.37,38 Usually, the plasma of healthy individuals does not contain EBV DNA^{ref1, ref2}. Therefore, the presence of EBV DNA in plasma may have significance for the diagnosis of CAEBV infection. However, it should be borne in mind that the plasma of patients with CAEBV infection did not always test positive for EBV DNA. Indeed, in this study, EBV DNA was not detected in plasma from 6 patients, whereas they had large amounts of EBV DNA in their PBMC. PBMC, and not plasma, should be used for diagnosing and monitoring patients with CAEBV infection. It is unclear why some patients did not have cell-free EBV DNA. Patients with negative plasma results sometimes turned positive at a later visit. Cell-free viral DNA may fluctuate from day to day. The other possibility is that inhibitors in plasma might influence PCR reactions and cause false-negative results. In conclusion, patients with CAEBV infection were divisible into 2 subgroups: T-cell and NK-cell CAEBV infection. Each group had different clinical features and prognosis. A high titer of EBV-related antibodies is not always a prerequisite for the diagnosis of CAEBV infection. Viral load, detected by quantitative PCR in PBMC, was useful for disease diagnosis and as an indicator of therapeutic efficacy. A viral load > 102.5 copies/µg DNA be used as a diagnostic criterion for CAEBV infection^ref.
• NK cell-type CAEBV : HMB and high IgE

Symptoms & signs : liver dysfunction
Therapy : vidarabine
• T-cell type CAEBV : fever and high titers of EBV-related antibodies. Although not statistically significant, patients with T-cell type of CAEBV infection had higher IgG levels than those with NK-cell type.. It is possible that EBV-infected T cells become activated and release inflammatory cytokines such as IFN-g, IL-6, or TNF-a, resulting in severe inflammation and fever. In EBV-related hemophagocytic syndrome, it has been shown that viral-infected T cells release TNF-a and activate macrophages, resulting in massive hemophagocytosis^{ref1, ref2}. These activated T cells might induce polyclonal B-cell activation through cytokine release and thereby induce high levels of IgG and EBV-related antibodies. On the other hand, it is unclear why HMB and high IgE were observed in patients with NK-cell type of CAEBV infection. We should emphasize that determining the cell type is important in predicting disease prognosis, because the survival rates were different between the 2 groups. Differences in survival rates are particularly important in assessing treatment choices. T
It remains unclear whether these 2 manifestations of disease represent different entities or simply appear different because of the nature of the infected cells
• severe chronic active EBV-infection (SCAEBV) (late 1970s)

Symptoms & signs : similar to those of chronic fatigue syndrome (CFS) but more severe in degree (pancytopenia, high fever, massive splenomegaly) => poor prognosis
Therapy : effect of in vitro-generated autologous EBV-specific CTLs or LAK cells  is limited or deteriorative
Therapy : to date, a treatment for CAEBV infection has not yet been established. Antiviral or immunomodulating agents, such as acyclovir, gancyclovir, vidarabine, IFN-g, and IL-2, have been tried^{ref1, Wakiguchi H, 2022-2025, ref3, ref4}. Adoptive transfer of virus-specific CTLs to a patient with CAEBV infection has been reported^ref. However, most of these reports were anecdotal and there are few confirmatory reports. Immunochemotherapy consisting of etoposide, steroids, and cyclosporin A was proposed for use in patients with advanced CAEBV infection^ref, but no evidence of its efficacy is available. Recently, successful treatment of CAEBV infection by allogeneic bone marrow transplantation has been reported^ref. Viral loads decreased in all patients receiving transplants, some of whom appeared to be cured. However, HSCT constitutes a considerable risk, because 4 of 7 patients died after transplantation. Prospective studies analyzing a larger number of patients with CAEBV infection are necessary to confirm the efficacy of transplantation and to establish safer conditioning regimens.
• EBV-associated hemophagocytic syndrome (HPS) (EBV-AHS)
• peripheral T-cell lymphoma
• aggressive NK-cell leukemia
• EBV-associated extranodal T-cell and natural killer (NK) cell lymphomas : the association of EBV with extranodal T-cell non-Hodgkin's lymphoma (T-NHL) is surprising^ref [326], because T-cells normally express the EBV receptor CD21 at a rate 10-fold lower than B-cells do^ref [327]. The association of EBV with T-NHL depends largely on site^{ref1, ref2} [326 and 328]. The expression of LMP1 in T/NK cell lymphomas is correlated with bad prognosis^ref [141] and a role for EBV induced IL-10-mediated local immunosuppression in the pathogenesis of nasal T/NK-lymphoma was recently suggested^ref [336].
• nasal type NK/T-cell lymphomas have a 100% association with EBV^{ref1, ref2} [329 and 330]. In NK/T-cell lymphomas in the nose, cytotoxic proteins like TIA-1 and granzyme B are even detected in up to 100% of the cases^ref [334]. Nasal (type) NK/T-cell lymphomas usually express CD56, showing their NK cell origin^ref [172]. These findings have led to the hypothesis that EBV is able to infect CTLs and NK cells during cell-cell contact with (productively) infected target cells by means of exchange of viral (virion) genetic material and subsequent transformation of the effector cell^ref [326]. EBV is actually present within the neoplastic cytotoxic cells, demonstrating that EBV is able to persist in these cells^ref [334]. One could speculate that predominantly mucosa-associated CTLs or NK cells become infected by EBV during the killing of EBV-infected target cells, because most productively infected B and epithelial cells are found in the oro/nasopharynx. In fact, EBV-infection by means of close cell–cell contact was described by others^ref [335]. The preferential homing of the infected cytotoxic cells would explain the presentation of EBV-associated T-NHL in the nasopharynx and the air and food passages
• intestinal T-NHL have a considerably lower association with EBV^ref [331] : the neoplastic cells frequently express cytotoxic proteins like TIA-1 and granzyme B^{ref1, ref2, ref3} [332, 333 and 334]
• chronic granular lymphocyte proliferative disorders (GLPD)
• NK cell leukemia/lymphoma (NKL)

Double staining for EBER RNA and surface markers in T-NHLs. Double staining for EBER RNA (red) and the B-cell marker CD20 (brown) (a) but not for EBER (dark blue) and the T-cell marker CD3 (brown) (b) can be observed in a nodal T-NHL. Double staining for EBER RNA (dark blue) and TIA-1 (brown) (c) can be observed in an intestinal T-NHL.
• nodal T-cell lymphomas :
• nodal angioimmunoblastic lymphadenopathy with dysproteinemia (AILD)-like lymphomas : using EBER RISH, EBV can be detected in up to 30% of cases^{ref1, ref2, ref3, ref4} [337, 338, 339 and 340]. These AILD-like lymphomas were shown to contain a clonal T-cell population^ref [337]. However, EBV in these lymphomas is not present in the neoplastic T-cells, but predominantly in B-cells which are usually polyclonal or oligoclonal^ref [339] and rarely become clonal in a later stage^ref [340]. This indicates that the proliferation of EBV-infected B-cells may be a secondary event and that EBV is not causally related to the pathogenesis of these lymphomas. The presence of EBV in B-cells in these lymphomas suggests that the expanded meshworks of follicular dendritic cells (FDCs), which are a hallmark of these lymphomas, stimulate proliferation of (EBV-infected) B blasts.
• anaplastic large cell lymphomas (ALCLs) : using EBER RISH, EBV can be detected in the majority of CD30⁺ neoplastic cells of a small percentage of cases^{ref1, ref2, ref3, ref4} [341, 342, 343 and 344]. These can, therefore, be considered as being EBV-associated. Moreover, EBV genomes in these ALCLs are clonal^ref [342], suggesting that EBV-infection is an early event and/or results in growth advantage. Interestingly, the majority of ALCLs appear to be derived from cytotoxic T-cells^{ref1, ref2} [343 and 344], suggesting a pathogenic mechanism resembling that of extranodal T-NHL.
• mosquito allergy with granular lymphocyte proliferative disorder (GLPD) / severe hypersensitivity to mosquito bites (HMB) (SHMB)

Epidemiology : mainly in Japanese patients (at least 58 patients) in the first 2 decades of life^ref
Pathogenesis : clonal expansion of EBV-infected NK cells (NK-GLPD) aberrantly expressing CD25 (augmented in vitro proliferative response to IL-2), Fas-L and Bcl-2, and resistance to in vitro Fas-induced apoptotic cell death (Fas-ACD). One of the unanswered questions is whether EBV-infected NK cell lymphoproliferative disease (LPD) precedes the onset of mosquito allergy among these patients. At 6 months old, a Japanese baby developed pancytopenia resulting in B-precursor acute lymphoblastic leukemia (ALL) without typical chromosomal abnormalities. Although she entered complete remission following our ALL protocol, continuous chemotherapy became difficult because of severe myelosuppression and repeated infections. Because she had no HLA-matched related or unrelated donors, we decided to perform CD34⁺ cell transplantation from her HLA-haploidentical father when she was aged 1 year and 11 months^{refYoshimoto T, 9-14}. At the time of transplantation she was negative for EBV serology and her father was positive. Preconditioning regimen consisted of total body irradiation (TBI) (12 Gy) and melphalan (140/m²), and 4.54 × 10⁶/kg of CD34⁺ cells were infused on February 6, 1996. Tacrolimus was used for GVHD prophylaxis. Since her posttransplantation clinical course was uneventful except for delayed platelet recovery, tacrolimus was stopped on day 64. Two months later, due to development of chronic GVHD of the skin, cyclosporine was started. Although tests for sex chromatin by fluorescence in situ hybridization (FISH) showed complete donor engraftment, poor recovery of platelet count still continued. Therefore additional 9.64 × 10⁶/kg of CD34⁺ cells from her father were reinfused on day 148, resulting in complete recovery of platelets. Because EBV serology was still negative after complete engraftment, we considered that there was no contamination of transmissible EBV-infected B cells in the purified CD34⁺ cells obtained from her father. 7 months after transplantation, the patient, aged 2 years and 6 months, developed primary EBV infection with atypical manifestations such as fever, eruptions, hypergammaglobulinemia (IgG, 3030 mg/dL; IgM, 300 mg/dL), liver dysfunction, abdominal lymph nodes swelling, and pancytopenia. Seroconversion against EBV (VCA IgG; 320 × , EA IgG; 10 × , EBNA; < 10 × ) was confirmed, and clinical symptoms as well as laboratory findings were resolved by stopping cyclosporin. 2 months later she developed fever and otitis media. Laboratory findings on this occasion revealed that  T cells were 16% in the PBMC (healthy control < 5%). She had been healthy until she developed mosquito allergy at age 5 years and 5 months. Upon admission at this time she had multiple skin lesions especially on the extremities and laboratory findings revealed liver dysfunction and a reactivation pattern of antibodies against EBV. gd T cells gradually increased with time :

TcRg gene rearrangements were observed and EBV DNA was detected in the sorted  T cells. A skin biopsy specimen revealed infiltrating atypical lymphocytes around the small vessels (resembling angiocentric lymphoma) EBER⁺, CD3⁺, CD4, CD8, and CD56, in the skin lesion, which may play an important role in the development of mosquito allergy. The patient underwent chemotherapy and the skin lesions have improved and the number of  T cells has been reduced.
Symptoms & signs : following mosquito bites the skin lesion at bite sites is typically a erythema or bulla that develops into ulcer or scar, accompanied by general symptoms such as high fever and general malaise and subsequently may experience lymphadenopathy and hepatosplenomegaly
Prevention : vidarabine + foscarnet
Therapy : refractory to several conventional chemotherapies
Prognosis : 50% of the patients reported died of EBV-associated hemophagocytic syndrome / malignant histiocytosis^{ref1, ref2}, granular lymphocyte proliferative disorder, or lymphomas. On the other hand, nearly 33% of patients with chronic active Epstein-Barr virus (CAEBV) infection may also have mosquito allergy and NK-cell proliferation^{Kawa-Ha K, 52-55, ref2, Kawa K, 139-147, ref4}. More recently, a close relationship has been showed between mosquito allergy and EBV-infected NK-cell lymphoproliferation^ref
• hydroa vacciniforme accompanying with EBV-infected NK-cell proliferation^{ref1, ref2}
Therapy : reduction and elimination of EBV-infected T/NK cells seems to be essential for the treatment of EBV-associated T/NK-cell LPD and it is important to distinguish it from ordinary EBV-associated B-cell LPD^{ref1, ref2, ref3}.
• EBV-associated B-cell LPDs :
• lymphomas in immunocompromized patients :
• post-transplant lymphoproliferative disorder (PTLD) due to excessive depletion of donor-derived EBV-specific CTLs during GvHD prophylaxis. PTLDs are a feared complication in transplant recipients who are being treated with immunosuppressive drugs^ref [287]. The majority of these disorders is of B-cell origin and is associated with EBV^ref [288], although exceptions exist^{ref1, ref2} [289 and 290]. Generally, the EBV present in PTLDs of solid-organ transplant (SOT) recipients is of host origin. By contrast, in PTLDs of bone-marrow transplant (BMT) patients the EBV generally is of donor origin^{ref1, ref2} [241 and 291]. Most PTLDs occur in patients who are EBV-seronegative prior to the transplantation and who, therefore, are unable to timely mount an effective immune response to EBV transformed B-blasts^ref [288]. Thus, PTLDs are frequently associated with primary infection and, therefore, assessment of the EBV serostatus in donor and recipient is important for risk stratification and will allow early identification of patients at high risk for PTLD development^{ref1, ref2, ref3, ref4} [259, 260, 261 and 288]. Early PTLDs are mostly polyclonal expansions of EBV-infected B-cells that may regress spontaneously upon reduction of immunosuppresive therapy or T-cell infusion, but which may progress to polymorphic PTLDs and diffuse large B-cell lymphomas, both showing clonal Ig rearrangements and EBV genomes^ref [292]. Immune suppression allows outgrowth of EBV-infected B-cells, associated with the expression of immunodominant EBV-proteins. Indeed, (using non-morphological methods such as RT-PCR and western blotting) a latency type III-like EBV gene expression pattern can be found in early PTLDs^{ref1, ref2, ref3} [292, 293 and 294]. However, using immunohistochemical double stainings, we have shown the presence of a continuous spectrum of cells expressing different amounts of EBNA2 and LMP1 in relation to distinct morphological features: small neoplastic cells express only EBNA1 and sometimes also EBNA2; cells of intermediate size express EBNA1, EBNA2 and LMP1, and large blasts, sometimes resembling H-RS cells express EBNA1 and LMP1 Fig. 3^ref [106]. Whether identical biological events underly PTLD development in SOT and BMT remains to be defined^ref [295]. In BMT, lymphoid tissues are severely depleted from resident regulatory lymphoid cells that may influence the activation and outgrowth of EBV-positive B-cells, whereas in SOT, local regulatory (T-) cell function(s) is merely decreased by immunosuppressive medication leaving the overall architecture of lymphoid organs intact. The therapeutic role for EBV-reactive cytotoxic T-cell infusions in the prevention and treatment of BMT-associated PTLD is well established^{ref1, ref2, ref3} [295, 296 and 297], but this is more complicated and controversial in SOT^{ref1, ref2, ref3} [298, 299 and 300]. A recent therapeutic approach involves the use of humanized monoclonal antibodies, such as Rituximab, directed against the pan-B-cell marker CD20, which can effectively eliminate B-cells from the host, but is not without complication^{ref1, ref2} [301 and 302].
• AIDS-related lymphomas (ARLs) are mostly of B-cell origin and contain the patient's intrinsic EBV. Moreover, unlike PTLDs, most ARLs contain monoclonal EBV genomes [292 and 303] and both type 1 and type 2 EBV-strains are detectable, frequently co-infecting the host^{ref1, ref2} [304 and 305]. Recent data indicate a more equal distribution of EBV types 1 and 2 in AIDS patients and no prevalent association between EBV type 2 and ARL was found^ref [306]. 2 different types of ARL are recognized that differ amongst others in their EBV association and EBV gene expression patterns. Most of the cases of diffuse large B-cell lymphoma, especially those arising in the central nervous system, are associated with EBV and have an EBV gene expression pattern comparable to that of PTLD^{ref1, ref2} [292 and 303]. By contrast, only 30–40% of AIDS-related BL is associated with EBV [307, 308, 309 and 310]. These lymphomas never show EBNA2 expression and only rarely LMP1, suggesting a latency type I or II expression pattern. In addition, AIDS-related BL, like endemic and sporadic BL, carry the t(8; 14) translocation involving the c-myc gene^ref [292]. Overall, AIDS-NHL represents a spectrum of lymphomas with a somewhat different pathogenesis^ref [311]. The differences between diffuse large B-cell lymphomas and Burkitt-type lymphomas in AIDS-patients suggest that the pathogenesis of the former resembles that of PTLDs, with an EBV driven proliferation eventually leading to malignant lymphoma, while the pathogenesis of AIDS-related BL resembles that of sporadic BL. In addition to chronic antigenic stimulation, it is thought that HIV-induced B-cell proliferation plays a role in the pathogenesis of AIDS-related BL. Interestingly, the majority of ARL contains mutations in the Bcl-6 gene which is regarded as a marker of B-cell transition through the germinal center^ref [312]. Thus, it is conceivable that not only AIDS-related BL but also other AIDS-related NHL are derived from germinal center B-cells.
• lymphomas in immunocompetent patients
• non-Hodgkin's lymphoma (0.5%) :
• Burkitt's lymphoma (BL) is a B-cell lymphoma that was originally described in equatorial Africa where it accounts for approximately 50% of all childhood cancers^ref [313]. Histologically, the tumor is composed of monomorphic medium-sized cells, often with macrophages scattered among them (‘starry sky macrophages)^ref [314]. The first indication for a viral etiology was the climate-dependent distribution of the disease: BL had a high incidence in regions where malaria prevailed. EBV is present in approximately 95% of these endemic BLs. However, in Europe and the USA, the association is only 10–20%. This difference has been ascribed to the early age of EBV-infection in Africa compared with industrial nations^ref [315]. In the endemic tumors, the EBV latent gene expression pattern is very restricted: LMPs are not detected, and due to Qp promoter usage for EBNA1 transcription^ref [316], only EBNA1 but not the other EBNAs are expressed (latency type I^ref, [317]). Rarely, isolated tumor cells may express other EBV genes like EBNA2, LMP and BZLF1^ref [318]. Another important feature of both endemic and non-endemic BL is the presence of the t(8; 14) translocation, which juxtaposes the c-myc gene on chromosome 8 to the Ig heavy chain locus on chromosome 14, resulting in overexpression of c-myc^ref [319]. It was hypothesized that EBNA1, which can upregulate expression of RAG-1 and -2 may play a causal role for this translocation^{ref1, ref2} [88 and 89]. A recent study showed that RAG mRNA's could not be detected in EBV-associated BL, nor in other EBV-associated lymphomas^ref [320]. It cannot be excluded, however, and indeed it is likely, that RAG induction is mainly important during the initial stages of BL development.
• B-cell chronic lymphocytic leukemia (B-CLL)
• Evans syndrome (0.5%)
• the rate of detection of EBV in other B-NHL is relatively low (between 2 and 10%^{ref1, ref2} [321 and 322]). Only in pyothorax-associated pleural high-grade B-cell NHLs, clonal EBV genomes can often be demonstrated in the tumor cells^{ref1, ref2} [323 and 324]. However, these tumors relatively often show the presence of a presumably less transforming EBV-strain (EBV type 2). This, and the apparent latency type III expression pattern (EBNA2 and LMP1 positive) favor an underlying immune defect in these cases^ref [325]. The EBV gene expression pattern in B-NHL of immunocompetent patients is usually latency type II-like, with transcription of BHRF1 but no EBNA2^{ref1, ref2} [210 and 322]. At present, it is not clear whether BHRF1 is expressed at the protein level, nor whether these transcripts are expressed in the neoplastic B-cells or in EBV-positive reactive B-cells. This, among others, illustrates the difficulty of inferring latency patterns from RT-PCR analyses.
• Hodgkin's lymphoma (HL) : histologically, HD is characterized by mononuclear Hodgkin cells and their multinucleated variant, the Reed–Sternberg cells, together abbreviated as H-RS cells. These are embedded in a background of reactive cells, including lymphocytes, plasma cells, histiocytes and eosinophils^ref [345]. Most recent studies indicate that the H-RS cells in many (but not all) cases of HD are derived from B-cells^ref [346]. The most characteristic epidemiological feature of HD as seen in most western populations is the bimodal age-incidence curve. In these populations, very few cases occur among children; the incidence then increases, peaking at about age 25; the incidence decreases to a plateau level through middle age and increases again with age to a second peak. Males are more often affected than females, which is most clearly observed in the older age group. There is clear evidence that the risk for HD occurring from early childhood through middle age is associated with factors in the childhood environment that influence the age at which infection with EBV takes place, (i.e. social class, population density, geographic region)^ref (reviewed in [347]). Moreover, several independent cohort studies have shown that people who have had IM run a3-fold higher risk for getting HD. Interestingly, the majority of HDs in patients who have had IM is not associated with EBV (summarized in^{ref1, ref2} [348 and 349]). It has been suggested that EBV is an important cofactor in the pathogenesis of HD, but that in patients with relatively intact immune systems, EBV⁺ neoplastic cells are eradicated and only cells capable of virus-independent growth survive (hit-and-run mechanism^ref [350]). A role for anti-LMP1 antibody responses was recently suggested in this process, because in a large study of juvenile HD patients, those with highest anti-LMP1 responses had EBV^- tumors^ref [351]. Detection of monoclonal EBV within the H-RS cells of HD was first reported in 1989^ref [352]. One year later, the presence of EBER RNAs in H-RS cells of was reported^ref [353]. The outcome of numerous studies shows that in most industrialized countries EBV is associated with on average 40–60% of HD cases, depending on their histology^ref (reviewed in [347]). Depending on race and socio-economical status of the population studied, this percentage may approach 100% as found in some developing countries^{ref1, ref2} [349 and 354]. It is thought that different pathogenic mechanisms play a role in EBV-positive and EBV-negative HDs^ref [355]. Although patients with EBV-positive HD were shown to respond better to chemotherapy^ref [356], EBV genome status of HD is by itself not an independent prognosticator^ref [357]. The EBV gene expression pattern in H-RS cells resembles that in NPC: expression of LMP1 without detectable EBNA2, characteristic of latency type II^{ref1, ref2} [358 and 359]. Most intriguing is the finding that high levels of tumor infiltrating activated CTLs are associated with a bad prognosis of HD^ref [360]. This is surprising because one would expect that CTLs present in such large number would eradicate the neoplastic cells. However, clonotypic analysis of tumor infiltrating T-cells revealed an oligoclonal response, indicating that these cells were attracted to the tumor site in an antigen independent manner^ref [361]. It is possible that the CTLs present in HD may not be specific for the tumor antigens presented, but are rather attracted by the abundance of cytokines produced at the site of the tumor^ref [163]. An impairment of antigen presentation by the H-RS cells can probably be excluded because HD cell lines transfected with immunodominant EBV-proteins were able to process and present these endogenously synthesized proteins, and were subsequently killed by specific CTL clones^ref [362]. Alternatively, a local immune response inhibition may be present in vivo in the tumor, as was first shown by Frisan et al.^ref [363] and subsequently confirmed by Chapman et al^ref [364]. An active role for LMP1-mediated enhanced local production of IL-10 in this process is suggested^{ref1, ref2} [365 and 366], but quantitative immunohistochemical analysis failed to reveal a detectable influence of IL-10 production on local T-cell populations in EBV^- cases of HD^ref [161]. Another challenging possibility is the direct immunosuppression by LMP1, secreted from H-RS cells in the form of exosomes^ref [131]. On the other hand, a (long-term) process of in vivo selection for apoptosis resistance of the H-RS cells may determine the malignant phenotype of HD and determine the long-term prognosis, either by providing resistance to CTL-(granzyme-B) mediated killing or by providing resistance to therapy induced cell damage. Apoptosis resistance may be conferred by genetic defects in the major enzymatic pathways mediating apoptosis, by upregulation of multidrug-resistance gene-expression or by the expression in the tumor cells of PI-9, a direct inhibitor of granzyme B function^ref [367]. Given the in vitro properties of LMP1, its expression may result in apoptosis resistance via upregulation of host-cell encoded anti-apoptotic factors bcl-2^ref [144] and A20^ref [145]. However, several studies have shown that bcl-2 expression in the Reed–Sternberg cells of EBV⁺ HD cases is not significantly higher than that in EBV^- HD cases^{ref1, ref2} [368 and 369]. Finally, LMP1 expressed by EBV⁺ H-RS cells may have additional tumor growth enhancing effects by inducing MMPs that can affect the function of infiltrating T-cells^{ref1, ref2} [370 and 371] and by inducing the production and secretion of IL6, IL-8 or COX-2 stimulating autocrine growth and neo-vascularization^{ref1, ref2, ref3} [372, 373 and 374]. Delayed exposure to infection may increase risk of EBV-positive HL in young adults, but risk patterns differ in younger and older women for both EBV⁺ and EBV^- HL. Late EBV infection does not appear relevant to risk, suggesting that other pathogens impact HL etiology in affluent female populations^ref.
Therapy : EBV proteins expressed during latency may serve as targets for novel chemotherapeutic agents. EBV-transformed B cells grow in tight clumps in vitro. EBV gene expression differs among malignancies associated with the virus. EBV⁺ Burkitt lymphoma tissue shows expression of EBNA-1, but not the other latency-associated proteins (Rowe et al, 1987). Hodgkin's disease tissues show expression of EBNA-1, LMP-1, and LMP-2 (Deacon et al, 1993) while nasopharyngeal carcinoma (Fahraeus et al, 1988) and T-cell lymphomas (Anagnostopoulos et al, 1992) express EBNA-1, LMP-2 and have variable expression of LMP-1. EBV lymphomas in immunocompromised persons generally show expression of each of the 9 latency associated proteins. While the treatment for some EBV-associated malignancies has improved in recent years, newer approaches to therapy are needed. LMP-1 is the latency-associated protein that has been most directly linked to oncogenesis by EBV. Expression of LMP-1 in B cells of transgenic animals results in B-cell lymphomas (Kulwichit et al, 1998). LMP-1 upregulates the expression of a large number of proteins on the surface of virus-infected B cells, including ICAM-1, LFA-1, and LFA-3 (Figure 1A) (Peng and Lundgren, 1992). Expression of LMP-1 in lymphoma cells induces clumping of the cells (Wang et al, 1990). LMP-1 acts as a functional homologue of a constitutively active form of CD40. LMP-1 oligomerises on the surface of virus-infected cells and binds to TRAFs 1,2,3 and 5, TRADD, RIP, and JAK3 (Devergne et al, 1996). The interaction of LMP-1 with these proteins results in activation of NF-kB, c-jun N-terminal kinase, STATs, the p38 MAP kinase pathway, and stress-activated kinases. This results in constitutive B-cell proliferation and inhibition of apoptosis. LMP-1 also interacts with p85 to activate the PI3K/Akt pathway and increase cell survival (Dawson et al, 2003). LMP-1 upregulates several other antiapoptotic proteins including A20, Mcl-1, bcl-2, and bfl-1. EBV-associated B-cell lymphomas in humans show activation of NF-kB, and LMP-1 colocalises with TRAF-1 and TRAF-3 (Liebowitz, 1998). In addition, TRAFs 1, 2, and 3 and NF-kB are expressed in post-transplantation lymphoproliferative disorders (Ramalingam et al, 2003). Furthermore, these lesions show expression of adhesion molecules upregulated by LMP-1 including LFA-1 (Hamilton-Dutoit et al, 1993). Thus, LMP-1 is critical for B-cell proliferation and development of lymphomas in vivo
• treatment of EBV-transformed B cells with NF-kB inhibitors (e.g. IB mutant, Bay11-7082) has been shown to induce apoptosis of the cells (Cahir-McFarland et al, 2000, 2004)
• treatment of EBV-transformed cells with simvastatin also inhibits NF-kB and induces apoptosis of EBV-transformed B cells (Katano et al, 2004). Statins inhibit 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase. 6 of these compounds, atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin, are approved by the FDA for use in humans. Each of these compounds is used to treat elevated serum cholesterol. HMG-CoA reductase catalyses the conversion of HMG-CoA to mevalonate, which ultimately leads to synthesis of cholesterol. Therefore, statins reduce the level of mevalonate with a subsequent reduction in cholesterol. Statins have a number of other activities related to their inhibition of HMG-CoA reductase (reviewed in Raggatt and Partridge, 2002). Mevalonate is a precursor for isoprenoids, including geranyl pyrophosphate and farnesyl pyrophosphate. Statins reduce the levels of these compounds. Post-translational modification of proteins by farnesylation or geranylgeranylation results in their association with cell membranes and activation (reviewed in Bellosta et al, 2000). These modified proteins include members of the nuclear laminin family, ras, inositol triphosphate 5-phosphatase, which are farnesylated, and Rho, Rac, cdc42, Rab, Rap, and G-proteins, which are geranylgeranylated. These changes have pleotrophic effects including inhibition of smooth muscle proliferation (Corsini et al, 1993), inhibition of MHC class II complexes on antigen-presenting cells (Kwak et al, 2000), increase in bone morphogenetic protein (Mundy et al, 1999), suppression of T- and B-cell responses (Kurakata et al, 1996), reduced NK cell activity (Hillyard et al, 2004), reduced synthesis of chemokines (Waehre et al, 2003), and growth arrest of certain transformed cells (Graaf et al, 2004). These effects can be reversed with the addition of mevalonate. Certain statins have activities that are unrelated to their inhibition of HMG-CoA reductase. Kallen et al (1999) showed that lovastatin binds to the I-domain of LFA-1 and blocks its interaction with ICAM-1. LFA-1 (L2 integrin) is an adhesion molecule that promotes diapedesis of leucocytes across the endothelium and is a costimulatory molecule on activated T cells. The I-domain of LFA-1 is separate from the site of its binding to its ligand ICAM-1. Weitz-Schmidt et al (2001) subsequently showed that simvastatin, mevastatin as well as lovastatin (but not pravastatin) can bind to LFA-1 and block its binding to ICAM-1. This interaction is independent of HMG-CoA and is not reversed by mevalonate. Simvastatin and lovastatin concentrations of ~10 M are required to block these interactions; this is in contrast to the nanomolar concentrations required to inhibit HMG-CoA. A synthetic statin (LFA703) which lacks HMGCoA reductase inhibitory activity, but which has increased LFA-1 binding activity, blocks LFA-1-induced costimulation of T cells and suppresses the inflammatory response to thioglycollate in a mouse model of peritonitis. Statins have been shown to affect replication of HIV. HIV virions have ICAM-1 on their surface, which can bind to LFA-1 on target cells and enhance virus attachment (Fortin et al, 1997). Lovastatin inhibits replication of HIV by inhibiting the interaction of ICAM-1 on virions with LFA-1 on the surface of target cells (Giguere and Tremblay, 2004). Lovastatin inhibits HIV-induced Rho GTPase activation, which is important for HIV infection of cells (del Real et al, 2004). Entry into and exit from HIV-infected cells is blocked by lovastatin and this effect can be reversed by treatment with mevalonate. Treatment of HIV-infected severe combined immunodeficiency (SCID)-hu mice or humans with lovastatin results in a reduction in HIV RNA loads and increased CD4⁺ T-cell counts. Treatment of EBV-transformed B cells with 2 M of simvastatin, atorvastatin, or lovastatin resulted in dissociation of cell clumps and death beginning 5 days after treatment with the drug (Katano et al, 2004). Cell death induced by simvastatin was due to apoptosis as demonstrated by detection of fragmented DNA. While simvastatin induced cell death in EBV-positive Burkitt lymphoma cells (such as Akata, Mutu-1, Mutu-3, and P3HR-1 cells), EBV-negative B cells, and EBV-negative T cells, the concentration of simvastatin required to kill these cells (4 M) was higher than for cells transformed with EBV in vitro. Simvastatin and lovastatin block the interaction of LFA-1 with ICAM-1, while pravastatin does not. In contrast to apoptosis of EBV-transformed B cells induced by simvastatin, treatment of these cells with up to 16 M of pravastatin did not dissociate cell clumps or induce cell death (Katano et al, 2004). In addition, antibody to LFA-1, which can activate lymphocytes (Perez et al, 2003), prevented cell death induced by simvastatin. LMP-1 is present in lipid rafts on the cell membrane (Higuchi et al, 2001). Lipid rafts are microdomains in the membrane that are rich in cholesterol and sphingolipids and are resistant to detergent extraction. They are important for signal transduction in B cells by CD40 or by immunoglobulin on the surface of the cells. The carboxy terminal domain of LMP-1 is activated when targeted to lipid rafts where it induces signalling and activation of NF-B-mediated transcription (Kaykas et al, 2001). A mutation in one of the transmembrane domains of EBV LMP-1 results in lack of LMP-1 localisation to rafts, failure to bind TRAF 3, and loss of activation of NF-kB (Yasui et al, 2004). Treatment of EBV-transformed B cells with 2 M simvastatin for 3 days reduced the amount of LMP-1 present in lipid rafts by 85% (Katano et al, 2004). While this effect might have been attributable to the depletion of cholesterol by the statin, it occurred at a relatively low level of simvastatin and treatment of cells with 8 M pravastatin did not result in a reduced amount of LMP-1 in rafts. Treatment of EBV-transformed B cells with 2 M simvastatin for 3 days inhibited NF-B activation. Cells treated with 8 M pravastatin did show reduced activation of NF-kB. Since LMP-1 also activates the phosphatidylinositol 3-kinase/Akt pathway (Dawson et al, 2003), displacement of LMP-1 from lipid rafts may reduce activation of this pathway and inhibit survival of EBV-transformed B cells. While LMP-2 has also been shown to be present in lipid rafts (Dykstra et al, 2001) and simvastatin might have an effect on LMP-2, the observation that LMP-2 is dispensable for B-cell transformation by EBV (Speck et al, 1999), suggests that the effect of simvastatin in killing transformed B cells is unlikely to be mediated through LMP2. The effects of simvastatin on EBV-transformed B cells cells could be due the ability of the drug to block adhesion molecule interactions on the surface of B cells, or to displace LMP-1 from rafts and inhibit NF-kB (Figures 1B and 2B). Examination of multiple EBV-positive and EBV-negative cell lines showed that induction of cell death by simvastatin correlated best with expression of LMP-1 and activated NF-kB in the cells prior to treatment with drug (Katano et al, 2004). Cells expressing the highest levels of LMP-1 and NF-kB (EBV-transformed B cells) were most susceptible to simvastatin-induced cell death; one cell line (Mutu-3 Burkitt lymphoma cells) expressing lower, but detectable levels of LMP-1 and NF-kB showed an intermediate sensitivity to death by simvastatin. However, other EBV-positive cells (Akata, P3HR-1, and Mutu-1 Burkitt lymphoma cells) that expressed low levels of NF-kB and no detectable LMP-1 were much less sensitive to simvastatin. One of these latter cell lines (P3HR-1) expressed levels of LFA-1 and ICAM-1 that were similar to those in EBV-transformed B cells. Taken together, these finding suggest that inhibition of NF-B by simvastatin may be more important than its ability to block the interaction of LFA-1 with ICAM-1. Intraperitoneal injection of EBV-transformed B cells into SCID mice results in development of EBV-positive lymphomas that resemble the tumours seen in immunosuppressed persons (Rowe et al, 1991). These tumours show a pattern of gene expression similar to that in patients with EBV lymphoproliferative disease with EBNAs, LMP-1, and adhesion molecules detected in the tumours. Oral treatment of SCID mice with simvastatin beginning 3 days prior to injection with EBV-transformed B cells resulted in a statistically significant improvement in survival rate compared to animals not given the drug (Figure 3) (Katano et al, 2004). While there was a trend for longer survival for mice treated with simvastatin beginning 7 days after injection with EBV-transformed B cells, the difference with untreated mice was not statistically significant. Some tumours from mice that were treated with simvastatin showed downregulation of LFA-1 on their surface, compared with the EBV-transformed B cells that had been used to inject the animals. The dose of simvastatin used to treat these mice (250 mg/kg/day) is estimated to result in serum levels that would be 4 to 8 times that of humans receiving the maximum dose of simvastatin (80 mg/day) used to lower serum cholesterol. However, similar high serum levels have been achieved in humans treated with large doses of statins in cancer therapy trials). Statins have been shown to induce apoptosis in several proliferating tumour cell lines, including certain leukaemia, lymphoma, astrocytoma, pancreatic carcinoma, and neuroblastoma cell lines (reviewed in Wong et al, 2002). This effect is due to inhibition of HMG-CoA reductase since it can be inhibited by mevalonate. Lovastatin reduced viability of EBV-positive or EBV-negative Burkitt lymphoma cells by >75% due to inhibition of geranylgeranylation (van de Donk et al, 2003). Statins have been used in animal models and have reduced the tumour burden associated with melanoma, hepatoma, neuroblastoma, pancreatic cancer, and lung cancer (reviewed in Wong et al, 2002). In other animal models, statins have been used as adjunctive therapy in combination with cytotoxic agents in mouse models including carmustine for melanoma and with doxorubicin for lung cancer. A number of studies have tested the role of oral statin therapy in patients with tumours (reviewed in Wong et al, 2002). Lovastatin given in doses ranging from 2 to 45 mg/kg/day for 1 out of every 4 weeks was used to treat a variety of tumours including astrocytoma, glioblastoma, and prostate, breast, ovarian, and lung cancer (Thibault et al, 1996). Doses of 25 mg/kg/day for 7 days were well tolerated. Serum levels of the drug ranged from 0.1 to 3.9 M. One patient with an astrocytoma had a minor response to therapy. Since mice metabolise statins more rapidly than humans, a dose of 25 mg/kg/day-1 is comparable to ~250 mg/kg/day in mice. While large doses of statins have been used in patients with cancer, might standard doses of statins prevent development of haematologic malignancies? Large trials of simvastatin for the prevention of coronary events have evaluated the incidence of cancer in study recipients. No significant decrease in the overall incidence of cancer or of haematologic malignancies was noted in the MRC/BHF Heart Protection Study involving 20 536 patients randomised to 40 mg of simvastatin per day or placebo for a median of 5 years (Heart Protection Study Group, 2002). Similarly, there was no significant decrease in the incidence or mortality for cancer, or for lymphatic or haematopoeitic malignancies in the Scandinavian Simvastatin Survival Study in which 4444 patients were randomised to simvastatin (20-40 mg per day) vs placebo for a median of 5.4 years (Strandberg et al, 2004). Thus, simvastatin at the standard doses to prevent coronary events did not have a significant effect on the risk of haematologic cancers. EBV lymphomas occur in immunocompromised patients such as organ or stem cell transplant recipients, patients with HIV, or patients with congenital immunodeficiencies. These tumours generally express each of the EBV latency proteins and the tumours are driven by these viral proteins. Central nervous system lymphomas account for 20% of lymphomas in AIDS patients and express EBV LMP-1 and other latency proteins (MacMahon et al, 1991). Immunoblastic lymphomas account for about 60% of cases of lymphoma in patients with AIDS and usually express EBV LMP-1 and other latency proteins (Hamilton-Dutoit et al, 1993). These tumours generally occur in AIDS patients late in the course of disease when CD4⁺ T cell numbers are low. Tumours in immunocompromised patients can occur in lymph nodes, but frequently present at extranodal sites such as the gastrointestinal tract, central nervous system, liver, lung, bone marrow, or transplanted organ (Cohen, 2000). The tumours can be polyclonal or monoclonal and usually lack chromosomal translocations. Patients who lack immunity to EBV at the time of transplant and develop primary EBV infection after transplant are more likely to develop EBV-driven lymphomas. Transplant recipients who receive HLA-mismatched or T-cell-depleted bone marrow or infusions of antilymphocyte antibodies for rejection are also at increased risk for development of lymphomas. EBV viral DNA is often elevated in peripheral blood mononuclear cells of these patients prior to, and at the onset of lymphoma, indicative of the EBV-driven B-cell proliferation. Patients with these tumours have impaired T-cell immunity to EBV resulting in failure to regulate EBV-driven B-cell proliferation. Treatment for these lymphomas includes reduction in immunosuppression when possible. Resection of localised lesions, especially in the gastrointestinal tract has been effective in some patients. Monoclonal anti-CD20 antibody (rituximab) results in remissions in about 50% of patients. Interferon- has been effective in some patients, but may increase the risk of rejection of the transplanted organ. Lymphomas in stem cell transplant recipients are usually of donor origin; infusions of donor T cells (which are HLA-matched) have been effective in many cases of EBV lymphoma in these patients. Lymphomas in organ transplant recipients are usually recipient in origin; infusions of autologous or HLA-matched T cells have been effective. Radiation therapy, especially for central nervous system lesions, and cytotoxic chemotherapy are used for refractory cases. The latter two therapies are frequently used for lymphomas in AIDS patients whose immune systems are less responsive to immunologic-based therapies. Statins may have a role as adjunctive therapy in some patients with EBV-driven lymphomas. These might include stem cell transplant recipients in whom donor T cells are not available, organ transplant recipients whose lymphomas are not responsive to reduction of immunosuppressive therapy and in whom HLA-matched T cells are not available. In these settings, statins might be used in combination with other therapies (e.g. rituximab or interferon-) which by themselves result in remissions in about 50% of patients. Statins might be tried in AIDS patients with EBV-driven lymphomas, especially those with very low CD4T cell counts who tend to respond less well to chemotherapy. Statins might also be considered for patients with EBV-positive Hodgkin's disease who are refractory to chemotherapy and radiation therapy. Tumours from these patients usually express LMP-1 as well as adhesion molecules (Sandvej et al, 1993), and statins might reduce the viability of the tumour cells. While nasopharyngeal carcinoma and T-cell lymphomas frequently express LMP-1, the tumours show more variable expression of the protein and therefore might be less susceptible to statins. Finally, Burkitt lymphomas do not express LMP-1 and therefore would be unlikely to respond to statins at doses that are effective for cells transformed with EBV in vitro. Identification of signalling pathways for EBV-mediated transformation has helped to identify new targets and potential treatments for these tumours. Certain statins have been shown to inhibit the interactions of adhesion molecules and block NF-B activation in EBV-transformed cells, resulting in apoptosis. Simvastatin delays the development of EBV-lymphomas in SCID mice inoculated with EBV-transformed B cells. The dose of statin needed to induce apoptosis is much higher than that required for lowering serum cholesterol, but such doses have been tolerated by patients in clinical trials. Statins may have a role in the treatment of EBV-driven lymphomas, most likely as part of combination therapy for these lymphomas.
• intentional induction of the lytic form of EBV infection combined with ganciclovir treatment^{ref1, ref2, ref3, ref4, ref5}. GCV and AZT are not generally effective for treating EBV-positive tumors because most tumor cells are infected with the latent form of EBV and therefore do not express the kinases which activate these drugs. Virally encoded kinases (thymidine kinase and BGLF4) which are expressed only during the lytic form of infection convert GCV into its active, cytotoxic form. The ability of various chemotherapy drugs to induce lytic EBV induction is clearly cell type dependent :
• in EBV-positive epithelial cell tumors
• previous studies showed that certain chemotherapy agents (fluorouracil [5-FU] and cis-platinum) efficiently induce a lytic infection in vivo and in vitro and that the addition of GCV greatly enhances the ability of both 5-FU and cis-platinum to inhibit NPC tumor growth in nude mice^ref
• in EBV-positive B-cell tumors, however, EBV infection is primarily latent (in contrast to epithelial cells, in which the virus infection is often lytic) and consequently it has been more difficult to identify agents that can efficiently induce a lytic EBV infection in B-cell tumors.
• although g-irradiation induces lytic EBV infection and expression of the viral TK in EBV-positive lymphoblastoid cell lines (LCLs) both in vitro and in vivo^{ref1, ref2} and might be useful for treating localized EBV-positive lymphomas combined with GCV, this method cannot be used to treat widely disseminated EBV-positive B-cell lymphomas
• gemcitabine and doxorubicin (but not 5-azacytidine, cis-platinum, or fluorouracil [5-FU]) induce lytic EBV infection in EBV-transformed B cells in vitro and in vivo. Gemcitabine and doxorubicin both activated transcription from the promoters of the 2 viral immediate-early genes, BZLF1 and BRLF1, in EBV-negative B cells. This effect required the EGR-1 motif in the BRLF1 promoter and the CRE (ZII) and MEF-2D (ZI) binding sites in the BZLF1 promoter. GCV enhanced cell killing by gemcitabine or doxorubicin in lymphoblastoid cells transformed with wild-type EBV, but not in lymphoblastoid cells transformed by a mutant virus (with a deletion in the BZLF1 immediate-early gene) that is unable to enter the lytic form of infection. Most importantly, the combination of gemcitabine or doxorubicin and GCV was significantly more effective for the inhibition of EBV-driven lymphoproliferative disease in SCID mice than chemotherapy alone. In contrast, the combination of zidovudine and gemcitabine was no more effective than gemcitabine alone^ref
• another approach for inducing lytic EBV infection in tumors would be to introduce the BZLF1 or BRLF1 genes under the control of a strong heterologous promoter by gene delivery methods. It was recently shown that direct delivery of the 2 EBV IE genes into nasopharyngeal carcinomas by using adenovirus vectors induces lytic EBV gene expression in the tumors and inhibits tumor growth^ref. However, adenovirus delivery of the 2 EBV IE genes to B-cell lymphomas would be limited by the fact that B cells express very little (if any) of the major receptor (CAR) for adenovirus, although this hurdle could potentially be overcome by the use of bispecific antibody techniques^ref.
• previous reports have shown that sodium butyrate (which induces histone acetylation) effectively induces the lytic form of EBV infection in at least some Burkitt's lymphoma cells either alone^ref, or associated with radiation^ref, and the combination of arginine butyrate and GCV to treat EBV lymphomas in patients is currently being studied^ref
Nevertheless, drug-based strategies for inducing lytic EBV infection in widely disseminated EBV-induced lymphoproliferative disease are more likely to be successful clinically than gene delivery strategies
• epithelial diseases :
• respiratory tract infection (40%)
• Gianotti-Crosti syndrome (GCS) / papular acrodermatitis of childhood (PAC)
• oral hairy leukoplakia (OHL) is a white epithelial lesion of the tongue which can be found in immunocompromized individuals. EBV-DNA was first detected in this lesion in 1985, with localization restricted to the superficial epithelial cells^ref [414]. Expression of viral lytic cycle antigens has been shown in OHL and abundant production of virus particles has been demonstrated using TEM, indicating that epithelial cells can support viral replication^{ref1, ref2} [273 and 275]. The detection of linear EBV genomes in the absence of episomal DNA^{ref1, ref2} [415 and 416], the exclusive localization of the virus in upper epithelial cells, and the sensitivity of the lesion to acyclovir treatment^ref [417], support this notion. Expression of latent proteins in OHL has also been reported but EBER RNAs are expressed at very low levels^ref [416]. Recent data indicate that LMP1 may contribute to the proliferation and apoptosis resistance of OHL epithelial cells^ref [151].
• epithelial malignancies : as discussed above, the pathogenesis of lymphomas is a multistep process. This also holds true for the genesis of carcinomas^ref; in fact, similar pathogenic mechanisms play a role in the genesis of various types of carcinomas. These are among others inhibition of apoptosis^ref, inactivation of tumor suppressor genes^ref and genetic instability^ref. Environmental and dietary influences are considered to play an important role in the accumulation of genetic damages leading to disregulated control of epithelial cell differentiation and proliferation, ultimately leading to carcinoma formation. In addition, the presence of certain micro-organisms is thought to contribute to carcinogenesis. Interestingly, several types of carcinomas are specifically associated with certain micro-organisms as is the case for lymphomas. For example, the causal relation between cancer of the uterine cervix and high-risk human papillomavirus infection is well established. Furthermore a (in)direct role for of hepatitis viruses (HCV, HBV) in the genesis of hepatocellular carcinoma and H. pylori in the outgrowth of gastric cancer is clearly indicated^{Parkin, DM 5-33}. EBV is most clearly associated with nasopharyngeal carcinomas^ref and a substantial percentage of gastric adenocarcinomas^{Parkin, DM 5-33, ref2}, but is also found in rare salivary gland and thymic carcinomas^ref. Finally, EBV-specific DNA, RNA and proteins have been detected in hepatocellular carcinoma^ref and invasive breast cancers^{ref1, ref2}, but the relevance of these findings remains to be determined^{ref1, ref2}. How EBV acquires access to epithelial cells remains a matter of debate, but this may involve low level CD21 (the EBV receptor) expression on IFN-activated epithelial cells in combination with BZLF2–gp42 MHC-II mediated membrane fusion or IgA-mediated transepithelial transport^{ref1, ref2, ref3}.
• anaplastic or undifferentiated nasopharyngeal carcinoma (NPC) / lymphoepithelioma : latency II expression pattern. NPC usually arises from the fossa of Rosenmuller and Eustachian cushions, or in the roof of the nasopharynx and very rarely in its anterior and lateral walls^ref [375]. Unfortunately, NPC is often detected only after dissemination to cervical lymph nodes which is associated with a significantly worse prognosis^ref [376]. The WHO histopathological classification of NPC recognizes keratinizing squamous-cell carcinoma (well, moderately or poorly differentiated), differentiated non-keratinizing carcinoma and undifferentiated carcinoma (including lymphoepitheliomas). Especially undifferentiated NPCs are EBV-associated (up to 100%)^{ref1, ref2} [376 and 377]. NPC shows a characteristic racial and geographical distribution with the highest incidence among males in Southern China [378] and natives of the Arctic region^ref [379]; intermediate rates of NPC are seen among Southeast Asian peoples of both Chinese and Malayan backgrounds, and among Arabs and Negroid populations in Kuwait and Northern/Central Africa^ref [378]. Subsequent generations of individuals from these high-risk groups who migrate to a low-risk area have a decreased risk for NPC^ref [380], and vice versa^ref [381]. As a result of these distribution and migration studies, numerous factors have been proposed to play a role in the genesis of NPC. These include consumption of Cantonese-style salted fish and other preserved foods, exposure to (tobacco) smoke, heavy alcohol use, use of Chinese herbal drugs and Chinese nasal oil. Moreover, certain HLA types are associated with a higher risk of NPC, and other hereditary factors may also be of importance^{ref1, ref2} (summarized in [348 and 382]). EBV-DNA in NPC biopsies was first detected by zur Hausen in 1970^ref [25]. The epidemiological and molecular linkage between EBV and NPC is confirmed in many studies^{ref1, ref2} [241 and 383]. Clonal EBV-infection was demonstrated in biopsy samples from NPC in China and the USA and in premalignant lesions of the nasopharynx, suggesting involvement of EBV prior to the carcinomatous state^ref [384]. The EBV latent gene expression pattern in NPC includes BamHI-A rightward transcripts (BARTs), EBER, EBNA1 and LMP2 and sometimes LMP1 (latency type II)^{ref1, ref2, ref3, ref4, ref5} [385, 386, 387, 388 and 389]. The heterogeneous expression of LMP1 is intriguing, but remains unexplained thusfar. In addition, also the pathogenic differences between LMP1 positive and negative NPC tumors remain to be explored. The role of EBV variant strains, that are frequently detected in NPC also remains unclear, because similar variants are encountered in healthy regional controls. Like in HD, bcl2 expression is enhanced in NPC, thus providing apoptosis resistance to the tumor cells. This may allow for high levels of p53 and concomitant high PCNA expression, contributing to the malignant phenotype [390, 391 and 392]. On the other hand, unlike HD, NPC cells rarely show upregulated interleukin expression^ref [374]. The detection of tumor-associated IL-10 has been linked to bad prognosis, but a relation between LMP1 expression and IL-10 detection at the tumor cell level was not clearly demonstrated^ref [223]. In contrasts to EBV-associated lymphoid malignancies, NPC is characterized by the abundant transcription of the BARF1 ORF ( Fig. 5)^{ref1, ref2, ref3} [50, 192 and 386]. An oncogenic role for BARF1 in epithelial cells is indicated by a number of experimental results in vitro and in vivo^{ref1, ref2, ref3} [190, 192 and 393]. It has been suggested that to a certain extent BARF1 expression may compensate for the reduced expression of LMP1 in NPC, for example by inducing apoptosis resistance via upregulation of bcl-2^ref [195]. As described for HD, NPC is characterised by an abundant influx of CTL, but the functional relevance of these CTL remains unknown^ref [394]. NPC patients are characterised by high IgG and IgA responses to proteins of the early and late lytic phase and to EBNA1^{ref1, ref2} [218 and 241]. However, except for EBNA1 these proteins are not expressed in the tumor itself. Therefore, these responses possibly reflect active viral replication at distinct (sub-epithelial) sites of the body or, alternatively, are directed against sporadic differentiating epithelial cells within the tumor, that consequently switch-on the EBV lytic cycle^ref [265].


Fig. 5. BARF1 NASBA analysis on NPC and HD samples. BARF1 signals are only found (after blotting and radioactive hybridization of NASBA products) in the NPC material (in 6 of 7 cases). BJAB, EBV-negative BL cell line; Louckes BARF1, EBV-negative Louckes BL cell line transfected with BARF1 expression construct.
• gastric adenocarcinomas : using PCR and DNA- and EBER-ISH, EBV was first detected in gastric adenocarcinomas in the USA^ref [395], in 16% of the cases. Studies in other areas including Japan, where the incidence of gastric carcinoma is very high, showed a generally lower proportion of EBV-associated gastric adenocarcinomas (between 2 and 8%)^{ref1, ref2, ref3} [396, 397 and 398]. Worldwide, the overall incidence of EBV⁺ tumors in regular gastric adenocarcinomas is 10% (6–16%). Gastric stump carcinomas show a higher frequency of EBV-association, up to 30%^ref [399]. Viral genomes in EBV-associated gastric carcinomas are monoclonal^ref [400]. The EBV gene expression pattern in gastric carcinomas is exceptional: Qp driven EBNA1 transcripts are detected, and EBER, LMP2 but not LMP1 or EBNA2 are transcribed. Moreover, transcription of BARF1 is detected and like in NPC it is thought that BARF1 may act as the alternative viral transforming factor [27].
• lymphocyte-rich carcinomas / lymphoepitheliomas often harbour EBV in their neoplastic cells^ref [401]. This was shown among others for . Furthermore,
• lymphoepithelioma-like carcinoma (LELC) of the esophagus
• lymphoepithelioma-like carcinoma (LELC) of the stomach
• lymphoepithelioma-like carcinoma (LELC) of the palatine tonsil
• lymphoepithelioma-like carcinoma (LELC) of the oral cavity
• lymphoepithelioma-like carcinoma (LELC) of the salivary glands^ref [403]
• lymphoepithelioma-like carcinoma (LELC) of the lung^ref [404]
• lymphoepitheliomas of the middle ear^ref [402]
• lymphoepitheliomas of the thymus^ref [405]
• oral squamous cell carcinomas ^{: ref} [408], and cervical cancer^{ref1, ref2} [410 and 411]. However, using EBER RISH in oral squamous cell carcinomas^ref [409] and cervical cancer^{ref1, ref2} [412 and 413], any signals present were confined to EBV-infected reactive cells and were not found in neoplastic cells. In addition, the absence of other EBV transcripts in these tumors indicates that EBV does not play a role in their pathogenesis.
• breast carcinoma in French women^ref, but it may have been present in infected lymphocytes in the breast (quantitative real-time PCR analysis of EBV⁺ breast carcinoma tissues suggested that < 0.1% of the cells contained viral genomes : sporadic lytic EBV infection may contribute to PCR-based detection of EBV in traditionally nonvirally associated epithelial malignancies), or it is possible that EBV may enhance the action of HHMMTV, because it is known that some retroviruses remain dormant unless activity is promoted by other herpesviruses^{ref1, ref2, ref3}. Evidence against^ref. Recent studies reported the presence of EBV in breast cancer biopsies using either PCR and in situ hybridization^ref [31], or PCR, EBNA1 immunohistochemistry and Southern blotting analysis^ref [30]. We and others could not detect EBERs or any other EBV transcript in neoplastic cells in breast cancer samples, although we did find EBV DNA in a small percentage of cases using PCR^{ref1, ref2} [32 and 33]. An EBER-less gene expression pattern, however, cannot be excluded, since it was also found in EBV-RT-PCR positive hepatocellular carcinomas^ref [29].
• hepatocellular carcinoma (HCC)^ref
• smooth-muscle tumors that arise with an increased frequency in transplant recipients^ref [406] and AIDS patients^ref [407] are often associated with EBV.

Laboratory examinations :

• Paul-Bunnel hemagglutination modified by Davidsohn looks for heterophilic IgM able to agglutinate goat and horse RBCs and to lyse bull RBCs after pre-absorption with guinea pig kidney extract and bull RBCs (nowadays it has been replaced by Mono-spot, in which anti-capsule Abs are looked for). Heterophile antibody tests have similar specificity and are cheaper, but are less sensitive in children or in adults during the early days of the illness (SOR: C, based on validating cohort study)
• serology is most sensitive, highly specific, and also the most expensive for diagnosing IM (strength of recommendation [SOR]: C, based on validating cohort study) :
• viral capsid Ag (VCA)
• early Ag - diffuse (EA-D)
• early Ag - restricted (EA-R)
• Epstein-Barr nuclear Ag (EBNA)
specific IgG avidity index (AI)^ref
• EBV-DNA
• in the past usually detected by means of Southern blotting. The disadvantages of this technique are its relative insensitivity (especially when the % of EBV-infected cells in a certain sample is low) and the need for relatively large amounts of DNA. However, Southern blotting is still the appropriate means to determine the clonality of a population of EBV-infected cells^{ref1, ref2} [52 and 256]. In addition, Southern blotting can be used to determine whether EBV in a given sample is episomal or integrated.
• PCR^ref [257] has greatly enhanced the sensitivity with which EBV-DNA and its transcripts are detected in cell lines and clinical samples. Thus, PCR-based techniques and not Southern blotting are recommended for mere detection of the presence of EBV in (clinical) samples. In addition, quantitative PCR techniques for the determination of EBV load have been described^{ref1, ref2} [258 and 259] that have proven very useful for monitoring EBV-DNA load in body fluids and lymphocytes and are currently being incorporated in routine clinical diagnostic monitoring of high risk patients^{ref1, ref2, ref3} [260, 261 and 262]. PCR is more sensitive than the heterophile antibody test in children, is highly specific, but is also expensive (SOR: C, based on validating cohort study).
• % of atypical lymphocytes and total lymphocytes on a complete blood count provide another specific and moderately sensitive, yet inexpensive, test (SOR: C, based on validating cohort study).
• morphological detection of EBV : the actual identification of the EBV-infected cells is only possible using morphological techniques. First, DNA in situ hybridization was used. However, this technique is not very sensitive because it only detects the viral genome, which is present in limited numbers in most EBV-linked diseases except in OHL. Alternatively, the high abundance of the EBER1, 2 RNAs in most if not all EBV⁺ tumor cells renders them a much better target for RNA in situ hybridization (RISH) assays^ref [263]. At present, EBER RISH is often performed non-radioactively^ref [264] and it is considered to be the most reliable technique to detect EBV in clinical materials with preservation of morphology^ref [52]. In recent year a variety of mAbs have become available for the detection of functional EBV gene expression in tissues and in circulating B-cells as well^{ref1, ref2, ref3, ref4, ref5, ref6, ref7, ref8} [106, 140, 200, 201, 265, 266, 267, 268 and 269]. Combination of EBER RISH and mAb staining and double staining for different EBV antigens has revealed that EBV-latency antigen expression at the single cell level may be more heterogeneous than previously appreciated^{ref1, ref2} [106 and 201]. These findings are in concordance with current ideas on the defined and well-regulated gene expression in circulating EBV₊ B-cells^{ref1, ref2} [170 and 171] and may have important consequences for understanding the genesis of lymphoproliferative diseases.


Double staining for EBNA2 (black) and LMP1 (brown) in a post-transplantation lymphoproliferative disorder. This staining shows a continuous spectrum of smaller tumor cells staining strongly for EBNA2 but relatively weak for LMP1, via intermediate phenotypes to larger, more blastic tumor cells even resembling (mononuclear variants of) Reed–Sternberg cells, staining positively for LMP1, but negatively for EBNA2.
• EBV gene expression analysis : EBV transcripts were detected in the past using northern blotting analysis; however, the RT-PCR developed more recently allows much more sensitive detection of even low-abundant mRNAs, and is less material-consuming. A multi-primed adaptation of the current RT-PCR protocol allows for even more efficient use of small lymphoma biopsies, and in addition, allows for better quality controls because control (‘housekeeping gene’) transcripts and the transcripts of interest are reverse transcribed simultaneously^ref [49]. However, we need additional, less sensitive, RNA quality controls. Routine gel electrophoresis of the RNA preparations is recommended to check for the presence of 18S/28S rRNA, which can be used as markers for the integrity of the RNA in clinical samples^{ref1, ref2} [49 and 50]. In addition, nucleic acid sequence based amplification (NASBA^ref [270]) proved to be especially suitable for the specific amplification of non-spliced EBV RNAs, such as BCRF1 (vIL10) and BARF1, even in a background of genomic DNA^{ref1, ref2, ref3} [50, 54 and 191]. To study the expression of mRNA at the single cell level, RNA in situ hybridization can also be used, but its routine use is rather limited to relatively abundant transcripts. Moreover, in order to definitively demonstrate RNA-mediated protein expression in the tumor cells, monoclonal antibodies specific for a variety of EBV proteins such as EBNA1^{ref1, ref2} [73 and 267], EBNA2^ref [271], LMP1^{ref1, ref2, ref3} [138, 140 and 269], LMP2^ref [272], Zebra^ref [273], EA^{ref1, ref2, ref3} [151, 212 and 274] and VCA^{ref1, ref2} [265 and 275] are available. These antibodies allow the detection of the related proteins both by western blot analysis and by in situ immunohistochemistry.

Differential diagnosis with other atypical lymphocytoses. A lymphocyte–white blood cell count (L/WCC) ratio > 0.35 had a specificity of 100% and a sensitivity of 90% for the detection of mononucleosis and against acute purulent tonsillitis^ref.
Therapy :

• inhibitors of viral DNA polymerase : most of them act as competitive alternative substrates, competing with GTP on the substrate-binding site, and as DNA chain terminators, by incorporating into the growing DNA chain and blocking its elongation due to their acyclic structure^ref
• acyclic nucleoside analogues (aciclovir, ganciclovir, penciclovir, as well as their prodrugs valaciclovir (Purifoy DJ, Beauchamp LM, de Miranda P et al. Review of research leading to new anti-herpesvirus agents in clinical development: valaciclovir hydrochloride (256 U, the L-valyl ester of acyclovir) and 882C, a specific agent for varicella zoster virus. J Medical Virol 1993; Suppl. 1: 139–45), valganciclovir^ref and famciclovir^ref, respectively)
• acyclovir^{ref1, ref2, ref3, ref4, ref5} is a potent inhibitor of EBV replication in cell culture, but clinical trials of aciclovir for the treatment of patients with IM have failed to detect benefit^{ref1, ref2}. There are several possible explanations for these outcomes
• the symptoms of IM are insidious in onset, which, coupled with a long incubation period (4–6 weeks), make for late diagnosis—in contrast, for example, to herpes labialis or chickenpox
• EBV is shed in the saliva. The levels of aciclovir achieved in the oropharynx, particularly after oral administration of the drug, may be inadequate as judged in part by failure to suppress virus titres in the saliva^{ref1, ref2}, in contrast to the reduction in titres that are produced by aciclovir given intravenously^ref. Presumably, the orally administered aciclovir prodrug, valaciclovir (Purifoy DJ, Beauchamp LM, de Miranda P et al. Review of research leading to new anti-herpesvirus agents in clinical development: valaciclovir hydrochloride (256 U, the L-valyl ester of acyclovir) and 882C, a specific agent for varicella zoster virus. J Medical Virol 1993; Suppl. 1: 139–45), would do almost as well as intravenously administered aciclovir, since the prodrug produces high blood levels of aciclovir
• most of the symptoms and signs of IM are due not directly to viral cytopathology in infected tissues, but to immunopathic responses to EBV-infected cells, particularly EBV-infected B-lymphocytes that circulate in the blood and infiltrate tissues of affected organs. The virus is usually restricted to infection of epithelium in the oropharynx and the cervix and of B-lymphocytes, initially in the tonsillar region. In immunocompetent hosts, the virus does not usually infect liver cells, neural cells or haematological cells other than lymphocytes, although liver, neural and bone marrow tissues and cells may be affected indirectly by immunopathic responses in complications of IM. Indeed, the atypical lymphocytosis (up to 40% or more of total circulating white blood cells) characteristic of IM consists of T cells, not EBV-infected B cells, and signifies the massive cell-mediated immune response mounted in the course of infection to the proliferating infected B cells (Pagano JS. Epstein-Barr Virus and the Infectious Mononucleosis Syndrome. In: Humes HD, Dupont HL, eds. Kelley's Textbook of Internal Medicine, 4th edn. Philadelphia: Lippincott-Raven Publishers, 2000; 2181–5).
• Ideally then it seems apparent that antiviral therapy coupled with an immunomodulatory drug might be effective. Corticosteroids are used empirically by physicians in treating IM, especially if there is severe pharyngitis, tonsillar swelling or airway obstruction^ref, and in fact a controlled trial of prednisolone given with aciclovir for treatment of IM was carried out, but without apparent benefit^ref. However, the design of this trial may not have been optimal, as indicated by the fact that prednisolone alone did not produce the expected results^ref. Also, use of aciclovir may not have produced adequate levels of the drug, although there was suppression of shedding of virus in the oropharynx
• ganciclovir.
• acyclic nucleotide analogues (cidofovir^ref and adefovir^ref)
• pyrophosphate analogues [phosphonoformic (foscarnet)^ref and phosphonoacetic^ref acids] interact directly with the pyrophosphate-binding site of the enzyme
• 4-oxo-dihydroquinolines (4-oxo-DHQs) (as represented by PNU-182171 and PNU-183792)^{ref1, ref2} showed a broad spectrum of anti-herpesvirus activity. Although non-nucleosides, the compounds seem to affect viral DNA polymerase and inhibit HSV-1 and -2, VZV, HCMV, kaposi's sarcoma herpesvirus (KSHV)^ref and EBV^ref in cell culture. The mechanism of action of these drugs has not been reported, but it is clearly different from that of aciclovir and related drugs. Resistance to 4-oxo-DHQs has been mapped to V823 of the viral DNA polymerase, a residue that is conserved in the DNA polymerases of six (HSV-1 and -2, VZV, HCMV, EBV and KSHV) of the 8 human herpesviruses, which implies that the compounds specifically inhibit these enzymes^ref.
Acyclic nucleoside and nucleotide analogues require phosphorylation to their triphosphorylated forms in order to become active. However, in the case of the nucleoside analogues, the antiviral specificity is based in part on the fact that viral kinases are more efficient in catalysing the first step in the drug intracellular metabolism (monophosphorylation), and so these compounds are more specific. EBV encodes two kinases: a thymidine kinase (BXLF1 gene product) and a protein kinase (BGLF4 gene product). However, in contrast to all the other herpesviruses, it is unclear which of these enzymes is responsible for the monophosphorylation and activation of nucleoside analogues^{ref1, ref2}. Specificity is also imparted by the ultimate target of these drugs, viral DNA polymerase, which explains why the phosphonated nucleoside analogues like cidofovir are still specific. It is unclear if herpesvirus DNA polymerases differ in sensitivity to these drugs since rigorous comparative determinations of K_m's and K_i's of the different herpesvirus DNA polymerases for any given drug have not been attempted, in part because of issues of enzyme purification (Pagano JS. Epstein-Barr virus: Therapy of active and latent infection. In: Jeffries DJ, De Clercq E, eds. Antiviral Chemotherapy. John Wiley & Sons Ltd. 1995; pp. 155–95). While usable, all these compounds suffer from a number of drawbacks (i.e. toxic side effects, poor oral bioavailability and risk for emergence of drug-resistant virus strains).
• compounds of mixed nature :
• benzimidazole D-ribonucleosides inhibit HCMV^ref but not EBV replication^ref. These compounds prevent the processing and maturation of concatameric viral DNA to monomeric genomes^{ref1, ref2}.
• 2,5,6-trichloro-1-ß-D-ribofuranosylbenzimidazole (TCRB)
• 2-bromo-5,6-dichloro-1-ß-D-ribofuranosylbenzimidazole (BDCRB)
• maribavir^ref is an L-ribonucleoside. Surprisingly, this novel compound has a mechanism of action that is different from that of BDCRB and in fact not yet fully understood. Maribavir is active against both HCMV^{ref1, ref2, ref3, ref4, ref5} and EBV^{ref1, ref2, ref3, ref4}, and its effects on these viruses seem to involve direct or indirect inhibition of viral protein kinases^{ref1, ref2, ref3, ref4, ref5} as well as possible interference with nuclear egress of virions during viral maturation^ref. Maribavir resistance in HCMV has been mapped to UL97^ref and UL27^{ref1, ref2} genes. In contrast, the EBV BGLF4 gene product, a protein kinase that is the homologue of HCMV UL97, was not inhibited directly by the drug^{ref1, ref2}, and EBV does not encode a homologue of UL27. However, phosphorylation of the EBV DNA-processivity factor, EA-D, by the EBV protein kinase is inhibited by maribavir in infected cells^ref. Maribavir is the only drug in this group for which Phase I clinical trials have been completed, viz., for HCMV infection^{ref1, ref2}, and Phase II trials were initiated in July 2004^ref.
• ß-L-5-iododioxolane uracil^ref
• indolocarbazoles have emerged recently as potential inhibitors of HCMV^ref and EBV^ref replication. The mechanism of action of these compounds seems to involve the inhibition of HCMV UL97 protein kinase^{ref1, ref2}; however, the same compounds (except for one, K252a) failed to directly inhibit EBV BGLF4 protein kinase^ref. These observations suggest that although some compounds may affect different groups of viruses, the mechanism of their action might be different for each of these groups.
• NIGC-I^{ref1, ref2, ref3, ref4}
Mechanisms of action of these drugs are still under study, but they do not involve inhibition of the viral DNA polymerase.
• EBV-specific CTLs specific for  have been successfully used in immunotherapy of PTLD, Hodgkin's disease, and in HIV/AIDS patients who also have high risk of developing EBV-associated lymphoma^ref

Patients given ampicillin or amoxicillin develop a rash (a.k.a. Dave Rouse disease)
Except for OHL, which is benign and represents a purely lytic infection with large amounts of EBV replicating freely in the lesions, all the other EBV diseases are malignancies characterized by latent infection that relies on cellular enzymes for EBV episomal DNA synthesis. Therefore, it is not expected that antiviral drugs directed at viral synthetic processes would affect EBV in the latent phase. In latent infection, the linear double-stranded genomes characteristic of productive lytic infection and encapsidated in virus particles are not made. Rather infection persists via controlled replication of viral episomes that are found only in the nucleus in a nucleosomal form. These circular supercoiled genomes are replicated once during cell division and perpetuated in progeny cells indefinitely. Replication is mediated by host DNA polymerase (and other enzymes of the cellular replicative machinery). Episomal copy number is tightly regulated and remains constant while expression of viral genes is greatly limited to several latency genes. None of these processes is affected by, or sensitive to, antiviral drugs to any degree, as shown by their lack of effect on latently EBV-infected cell lines or tumours (Pagano JS. Epstein-Barr virus: Therapy of active and latent infection. In: Jeffries DJ, De Clercq E, eds. Antiviral Chemotherapy. John Wiley & Sons Ltd. 1995; pp. 155–95). A possible exception might be the initially polyclonal EBV lymphoproliferative disorders, in which occasional cells do exhibit lytic rather than latent infection; in such cases, theoretically antiviral therapy might prevent secondary infection of a new population of B cells. There is only anecdotal evidence for such an effect clinically, and it is scant. However, Chodosh et al. have reported experiments in which treatment of latently EBV-infected cells with hydroxyurea led to loss of episomes^ref, and in a subsequent case report detailed apparently successful use of hydroxyurea in a patient with an EBV-related CNS lymphoma^ref. It should be noted though, that a number of groups are intensively studying means to induce a switch from latent to cytolytic infection as a therapeutic approach for the treatment of EBV⁺ neoplasms, the rationale being that viral replication causes cytolysis of the infected cells and could do so in tumours^ref :

• chemotherapeutic agents gemcitabine and doxorubicin^{ref1, ref2}
• arginine butyrate^{ref1, ref2, ref3, ref4}
• radiation therapy^ref

... all of which alone or in combination induce viral reactivation. Combinations of such inducers with ganciclovir, which is activated by EBV PK (or TK) and thus toxic for infected and some neighbouring cells, has led to some progress in treatment of EBV⁺ malignancies in animal models^{ref1, ref2, ref3} and in Phase I/II clinical trials^ref. In quite another experimental approach to treatment of EBV-associated malignancies, phosphonated nucleoside analogues have been used to cause apoptosis of human EBV⁺ anaplastic or undifferentiated nasopharyngeal carcinoma (NPC) / lymphoepithelioma grown in athymic mice. Since NPCs are latently infected, again the antitumour effect must be independent of the antiviral effects of these drugs^{ref1, ref2} (Wakisaka N, Yoshizaki T, Murono S et al. Ribonucleotide reductase inhibitors enhance cidofovir-induced apoptosis in EBV⁺ nasopharyngeal carcinoma xenografts. Int J Cancer 2005; 116: in press). Interestingly, cidofovir has been reported to produce complete responses in vivo in another virus-associated neoplasm, laryngeal papillomatosis, which is caused by human papillomavirus^ref (Snoeck R, Andrei G & De Clercq E. Cidofovir in the treatment of HPV-associated lesions. Verh K Acad Geneeskd Belg 2001; 63: 93–120, discussion 120–2). Such effects of cidofovir might be linked to decreased expression of EBV LMP1 and consequent aberrations in apoptotic mechanisms^ref, but there is evidence that weighs against these conclusions^ref, and so how these nucleotide analogues affect neoplastic growths is far from clear and remains an important topic for continued study^ref
Experimental animal models : at present it is not known to what extent the mechanisms described in vitro also play a role in vivo. To gain more insight into the in vivo situation, several animal models have been developed. The pathogenic aspects of separate EBV gene products have been studied by means of transgenic mice. These studies clearly indicate putative pathogenic roles for EBNA1^ref [87], LMP1 (both in lymphoid^ref [47] and epithelial^ref [48] tissue), and LMP2A^ref [168] in driving oncogenic cell transformation and inhibition of cellular differentiation. However, the implications of these findings for the pathogenesis of EBV-infection and malignant transformation in humans, where EBV persists in the face of a highly responsive immune system, remain to be clarified by further studies. In addition, animal models have been developed to study the biology of EBV-infection. Especially cottontop tamarins are susceptible for EBV-infection: administration of a high intravenous dose of the EBV B85-8 strain rapidly produces gross, multiple, malignant lymphomas in 100% of these animals^ref [418]. However, this pathogenesis is clearly not representative for the situation in humans, where several types of lymphomas develop only after a long period of latent EBV-infection. A more closely related animal model was developed by Wedderburn et al. using the common marmoset which developes a persistent infection more closely resembling EBV-infection of humans. However, only limited experience exists in this model^ref [419]. From lymphoma tissue arising in various monkey species EBV-like viruses have been recovered that have subsequently been used to develop an animal model for lymphomagenesis^{ref1, ref2, ref3, ref4} [420, 421, 422 and 423]. An excellent alternative model was developed by Moghaddam et al.^ref [424], who did not use EBV but a naturally endemic rhesus lymphocryptovirus. This animal model represented several aspects of human EBV-infection, including oral transmission, atypical lymphocytes, lymphadenopathy, latent infection in peripheral blood and serologic responses to latent and lytic viral antigens. Interestingly, tamarins were shown to be effectively protected from lymphomas after vaccination with a gp340 vaccine containing the native fully glycosylated Gp^ref [425], although in protected animals virus-carrying cells could still be detected which showed a restricted gene expression pattern^ref [426]. Similar studies have recently been undertaken in the common marmoset, showing reduced shedding by EBV challenge in gp340 immunized animals^ref [427]. In humans, during natural infection gp340 also elicits virus-neutralizing antibodies which remain present in healthy virus carriers throughout life^ref [428]. Indeed a small human trial with recombinant vaccinia virus expressing gp340 showed promising results in delaying primary infection in virus-naïve infants^ref [429]. In addition to studies in monkeys, EBV-mediated lymphomagenesis has also been studied in SCID mice and in nude mice^{ref1, ref2, ref3} [430, 431 and 432]. This model also proved a valuable tool for anti-tumor therapeutic studies of EBV lymphomas^{ref1, ref2} [433 and 434] and carcinomas^{ref1, ref2} [435 and 436]. It is, however, difficult to extrapolate these data to EBV-mediated lymphomagenesis or carcinogenesis in humans, because these nude mice and SCID mice lack (part of) the immune responses which play a role in the establishment and control of EBV-infection in humans.
Prognosis : establishes a life-long persistent infection in > 90% of the world's population
Web resources : International Association for Research on EBV and Associated Diseases

---