Burkitt's lymphoma: the Rosetta Stone deciphering Epstein-Barr virus biology

Abstract

Epstein-Barr virus was originally identified in the tumour cells of a Burkitt's lymphoma, and was the first virus to be associated with the pathogenesis of a human cancer. Studies on the relationship of EBV with Burkitt's lymphoma have revealed important general principles that are relevant to other virus-associated cancers. In addition, the impact of such studies on the knowledge of EBV biology has been enormous. Here, we review some of the key historical observations arising from studies on Burkitt's lymphoma that have informed our understanding of EBV, and we summarise the current hypotheses regarding the role of EBV in the pathogenesis of Burkitt's lymphoma.

Fig. 1

Patterns of EBV latent protein expression in different forms of viral latency. The top panel (A) is a schematic illustration of the EBV genome, showing the location of coding exons for the nuclear antigens EBNA1, EBNA2, EBNA3A, EBNA3B, EBNA3C and EBNA-LP; the viral bcl-2 homologue, BHRF1; and the three latent membrane proteins LMP1, LMP2A and LMP2B. Also shown are the latent origin of replication (oriP), the BamHI W internal repeat region and the fusion of the terminal repeats (TR). Shown below (B) are the structures of viral mRNA transcripts expressed in different forms of latency. Promoters are identified by arrowheads, coding exons are coloured and non-coding exons are unshaded. Latency I, seen in most endemic BL tumours, is characterised by the restricted expression of a single latent antigen EBNA1 transcribed from Qp. The Latency II form of infection seen in EBV-positive Hodgkin's lymphoma and NPC tumour cells is characterised by expression of the LMP proteins in addition to Qp-initiated EBNA1. The classical growth-transforming Latency III infection, displayed by in vitro-transformed LCLs, is characterised by expression of all ten latent cycle proteins; the different EBNA mRNAs are generated by alternative splicing of long primary transcripts initiated either from the tandemly repeated Wp promoter or the upstream Cp promoter, while separate promoters in the BamHI N region transcribe the LMPs. BHRF1, which has recently been described as a tenth latent antigen, is also expressed from Wp-initiated transcripts in Latency III. A fourth type of infection, termed Wp-restricted latency, is characterised by expression of EBNA1, EBNA3A, EBNA3B, EBNA3C, EBNA-LP and BHRF1 in the absence of EBNA2 and the LMPs, and is seen in a subset of endemic BL tumours which carry EBV genomes deleted for the EBNA2 gene.

Fig. 2

Expression of non-coding EBV RNAs. Schematic illustration of the EBV genome showing the location of the non-coding EBER RNAs, BHRF1 miRNAs and BART miRNAs. The non-coding nuclear EBER1 and EBER2 RNAs are transcribed by RNA polymerase III and are the most abundant viral transcripts in latently infected cells. The BamHI A region gives rise to a family of highly spliced BamHI A rightward transcripts (BARTs) which are also expressed in all forms of latency. While the protein-coding function of these BART RNAs remains controversial, at least 21 miRNAs are generated from BART RNA-derived introns. A second set of miRNAs, derived from the BamHI H region, are thought to be generated from Wp/Cp-initiated transcripts. Also shown are the location of the latent cycle proteins, latent origin of replication (oriP) and terminal repeats (TR).

Fig. 3

Selective colonization of the memory B cell compartment during primary EBV infection. How EBV achieves selective colonization of the isotype class-switched memory B cell pool in vivo remains unclear. One hypothesis proposes that incoming virions preferentially infect IgM⁺IgD⁺ naïve B cells which, as a result of a transient growth-transforming infection mimicking antigen stimulation, are driven to form a germinal centre (GC). Thereafter the physiologic processes of somatic hypermutation and class switch recombination come into play and deliver latently infected IgG⁺CD27⁺ or IgA⁺CD27⁺ GC progeny cells into the long-lived memory compartment. An alternative hypothesis, based on studies of individual cells micro-dissected from infectious mononucleosis lymphoid tissues , questions the involvement of GC transit and instead envisages the preferential infection and/or survival of memory cells post-infection, compared to their naïve counterparts. These models are not mutually exclusive, and both share one key feature; that downregulation of viral antigen expression is linked to the normal differentiation of B cells.