TGF-beta induces apoptosis in human B cells by transcriptional regulation of BIK and BCL-XL

TGF-β potently induces apoptosis in Burkitt’s Lymphoma (BL) cell lines and in explanted primary human B lymphocytes. The physiological relevance and mechanism of TGF-β-mediated apoptosis induction in these cells remains to be determined. Here we demonstrate the requirement for TGF-β-mediated regulation of BIK and BCL-X_L to activate an intrinsic apoptotic pathway in centroblastic BL cells. TGF-β directly induced transcription of BIK and a consensus Smad binding element identified in the BIK promoter recruits TGF-β-activated Smad transcription factor complexes in vivo. TGF-β also transcriptionally repressed expression of the apoptosis inhibitor BCL-X_L. Inhibition of BCL-X_L sensitised BL cells to TGF-β-induced apoptosis while overexpression of BCL-X_L or suppression of BIK by shRNA, diminished TGF-β-induced apoptosis. BIK and BCL-X_L were also identified as TGF-β target genes in purified normal human centroblast B cells and immunohistochemical analyses of tonsil tissue revealed widespread TGF-β receptor-regulated Smad activation and a focal pattern of BIK expression. Furthermore, using a selective inhibitor of the TGF-β receptor we provide evidence that autocrine TGF-β signaling through ALK5 contributes to the default apoptotic program in normal human centroblasts undergoing spontaneous apoptosis. Our data suggests that TGF-β may act as a physiological mediator of human germinal centre homeostasis via regulation of BIK and BCL-X_L.

Two Epstein-Barr virus (EBV) oncoproteins cooperate to repress expression of the proapoptotic tumour-suppressor Bim: clues to the pathogenesis of Burkitt's lymphoma

Epstein-Barr virus (EBV) contributes to the development of several human cancers including the endemic form of Burkitt's lymphoma (BL). In culture, EBV induces the continuous proliferation of primary B cells as lymphoblastoid cell lines (LCLs) and if EBV-negative BL-derived cells are infected with EBV, latency-associated viral factors confer resistance to various inducers of apoptosis. Nuclear proteins EBNA3A and EBNA3C (but not EBNA3B) are necessary to establish LCLs and their expression may be involved in the resistance of BL cells to cytotoxic agents. We have therefore created recombinant EBVs from which each of the EBNA3 genes has been independently deleted, and revertant viruses in which the genes have been re-introduced into the viral genome. Infection of EBV-negative BL cells with this panel of EBVs and challenge with various cytotoxic drugs showed that EBNA3A and EBNA3C cooperate as the main determinants of both drug resistance and the downregulation of the proapoptotic Bcl-2-family member Bcl-2-interacting mediator of cell death (Bim). The regulation of Bim is predominantly at the level of RNA, with little evidence of post-translational Bim stabilization by EBV. In the absence of Bim, EBNA3A and EBNA3C appear to provide no survival advantage. The level of Bim is a critical regulator of B cell survival and reduced expression is a major determinant of lymphoproliferative disease in mice and humans; moreover, Bim is uniquely important in the pathogenesis of BL. By targeting this tumour-suppressor for repression, EBV significantly increases the likelihood of B lymphomagenesis in general, and BL in particular. Our results may also explain the selection pressure that gives rise to a subset of BL that retain expression of the EBNA3 proteins.

Physical and functional interactions between the corepressor CtBP and the Epstein-Barr virus nuclear antigen EBNA3C

CtBP has been shown to be a highly conserved corepressor of transcription. E1A and all the various transcription factors to which CtBP binds contain a conserved PLDLS CtBP-interacting domain, and EBNA3C includes a PLDLS motif (amino acids [aa] 728 to 732). Here we show that EBNA3C binds to CtBP both in vitro and in vivo and that the interaction requires an intact PLDLS. The C terminus of EBNA3C (aa 580 to 992) has modest trans-repressor activity when it is fused to the DNA-binding domain of Gal4, and deletion or mutation of the PLDLS sequence ablates this and unmasks a transactivation function within the fragment. However, loss of the CtBP interaction motif had little effect on the ability of full-length EBNA3C to repress transcription. A striking correlation between CtBP binding and the capacity of EBNA3C to cooperate with (Ha-)Ras in the immortalization and transformation of primary rat embryo fibroblasts was also revealed.

Activators of the Epstein-Barr virus lytic program concomitantly induce apoptosis, but lytic gene expression protects from cell death

Expression of the lytic cycle genes of Epstain-Barr virus (EBV) is induced in type I Burkitt's lymphoma-derived cells by treatment with phorbol esters (e.g., phorbol myristate acetate [PMA]), anti-immunoglobulin, or the cytokine transforming growth factor beta (TGF-beta). Concomitantly, all these agents induce apoptosis as judged by a sub-G1 fluorescence-activated cell sorter (FACS) profile, proteolytic cleavage of poly(ADP-ribose) polymerase (PARP) and terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) staining. However, caspase activation is not required for induction of the lytic cycle since the latter is not blocked by the caspase inhibitor ZVAD. Furthermore, not all agents that induce apoptosis in these cultures (for example, cisplatin and ceramide) induce the EBV lytic programme. Although it is closely associated with the lytic cycle, apoptosis is neither necessary nor sufficient for its activation. Multiparameter FACS analysis of cultures treated with PMA, anti-Ig, or TGF-beta revealed BZLF1-expressing cells distributed in different phases of the cell cycle according to which inducer was used. However, BZLF1-positive cells did not appear to undergo apoptosis and accumulate with a sub-G1 DNA content, irrespective of the inducer used. This result, which suggests that lytic gene expression is protective, was confirmed and extended by immunofluorescence staining doubled with TUNEL analysis. BZLF1- and also gp350-expressing cells were almost always shown to be negative for TUNEL staining. Similar experiments using EBV-positive and -negative subclones of Akata BL cells carrying an episomal BZLF1 reporter plasmid confirmed that protection from apoptosis was associated with the presence of the EBV genome. Finally, treatment with phosphonoacetic acid or acyclovir prior to induction with PMA, anti-Ig, or TGF-beta blocked the protective effect in Mutu-I cells. These data suggest that a late gene product(s) may be particularly important for protection against caspase activity and cell death.

Apoptosis induced by TGF-beta 1 in Burkitt's lymphoma cells is caspase 8 dependent but is death receptor independent

TGF-beta is a potent inducer of apoptosis in many Burkitt's lymphoma (BL) cell lines. In this study, we characterize this apoptotic process in the EBV-negative BL41 cell line. Induction of apoptosis was detected as early as 8 h after TGF-beta treatment, as assayed by TUNEL and poly(ADP-ribose) polymerase cleavage. FACS analysis demonstrates that this proceeds predominately from the G1, but also from the G2/M phases of the cell cycle. We observed no early detectable changes in the steady-state levels of Bcl-2 and several of its family members after TGF-beta treatment. We detected cleavage of caspases 2, 3, 7, 8, and 9 into their active subunits. Consistent with the involvement of these enzymes in TGF-beta-mediated apoptosis, the broad spectrum caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp(Ome)-flouromethylketone (ZVAD-fmk) blocked TGF-beta-induced apoptosis and revealed a G1 arrest in treated cells. Use of specific caspase inhibitors revealed that the induction of apoptosis is caspase 8 dependent, but caspase 3 independent. Activation of caspase 8 has been shown to be a critical event in death receptor-mediated apoptosis. However, TGF-beta treatment of BL41 cells was found not to affect the cell surface expression of Fas, TNF-R1, DR3, DR4, or DR5, or the steady-state expression levels of Fas ligand, TNF-R1, DR3, DR4, and DR5. Furthermore, blocking experiments indicated that TGF-beta-mediated apoptosis is not dependent on Fas ligand, TNF-alpha, tumor necrosis-like apoptosis-inducing ligand, or TNF-like weak inducer of apoptosis signaling. Therefore, it appears that TGF-beta induces apoptosis in BL cell lines via caspase 8 in a death receptor-independent fashion.

High level expression of deltaN-p63: a mechanism for the inactivation of p53 in undifferentiated nasopharyngeal carcinoma (NPC)?

Undifferentiated nasopharyngeal carcinoma (NPC) is an epithelial malignancy that is consistently associated with Epstein-Barr virus (EBV) but which very rarely has p53 gene mutations in primary tumours. Since the tumour suppressor p53 is mutated in most human cancers or the wild type protein is inactivated in a significant number of the remainder, here we have investigated cellular factors that could compromise p53 function in primary NPC. Twenty-five primary tumours were judged to carry only wild type p53 by SSCP analysis of all exons and sequence determination of exons 4-9. Only one tumour was found to express significant levels of hMdm2 and in 24/25 there were no detectable mutations or deletions in exons 1beta and 2 of the p14(ARF) gene. However, immunohistochemistry consistently revealed that all the tumour cells express substantial amounts of the p53-related protein p63. Semi-quantitative RT-PCR analysis of mRNA from tumour biopsies showed that the dominant species expressed was invariably the truncated deltaN-isotype. Since this can block p53-mediated transactivation, it is potentially a dominant-negative isoform. In normal nasopharyngeal epithelium the distribution of p63 was restricted to the proliferating basal and suprabasal layers. We suggest that deltaN-p63 is a good candidate as a suppressor of wild type p53 function in these tumours and also that it may prove to be a valuable diagnostic marker for undifferentiated NPC.

Resistance to TGF-beta1 correlates with a reduction of TGF-beta type II receptor expression in Burkitt's lymphoma and Epstein-Barr virus-transformed B lymphoblastoid cell lines

The pleiotropic cytokine TGF-beta1 is a member of a large family of related factors involved in controlling cell proliferation, differentiation and apoptosis. TGF-beta ligands interact with a complex of type I and type II transmembrane serine/threonine kinases and they transmit their signals to the nucleus via a family of Smad proteins. A panel of over 20 Burkitt's lymphoma (BL) cell lines has been compiled including those that are Epstein-Barr virus (EBV) negative, those that carry EBV with a restricted pattern of EBV latent gene expression (group I) and those that express the full range of latent EBV genes (group III), together with selected EBV-transformed lymphoblastoid cell lines (LCLs). Most of the EBV-negative and group I BL cell lines underwent apoptosis or a G(1) arrest in response to TGF-beta1 treatment. In contrast, group III cell lines and LCLs were completely refractory to these effects of TGF-beta1. All of the cell lines expressed the TGF-beta pathway Smads and the TGF-beta type I receptor. Lack of responsiveness to TGF-beta1 appears to correlate with a down-regulation of TGF-beta type II receptor expression. Studies of EBV-converted and stably transfected BL cell lines demonstrated that the EBV gene LMP-1 is neither necessary nor sufficient to block the TGF-beta1 response.

A common polymorphism acts as an intragenic modifier of mutant p53 behaviour

The p73 protein, a homologue of the tumour-suppressor protein p53, can activate p53-responsive promoters and induce apoptosis in p53-deficient cells. Here we report that some tumour-derived p53 mutants can bind to and inactivate p73. The binding of such mutants is influenced by whether TP53 (encoding p53) codon 72, by virtue of a common polymorphism in the human population, encodes Arg or Pro. The ability of mutant p53 to bind p73, neutralize p73-induced apoptosis and transform cells in cooperation with EJ-Ras was enhanced when codon 72 encoded Arg. We found that the Arg-containing allele was preferentially mutated and retained in squamous cell tumours arising in Arg/Pro germline heterozygotes. Thus, inactivation of p53 family members may contribute to the biological properties of a subset of p53 mutants, and a polymorphic residue within p53 affects mutant behaviour.

Epstein-Barr virus EBNA3C can disrupt multiple cell cycle checkpoints and induce nuclear division divorced from cytokinesis

Expression of EBNA3C is essential for the immortalization of B cells by EBV in vitro and, in co-operation with activated ras, EBNA3C has oncogenic activity in primary rodent fibroblasts. This suggested that this viral oncoprotein might disrupt the cyclin/CDK-pRb-E2F pathway, which regulates cell cycle progression at the restriction point (R-point) in G1 of the proliferation cycle. An assay was established in which transfected EBNA3C-positive cells could be sorted and simultaneously analysed for their distribution in the cell cycle. This revealed that in NIH3T3 fibroblasts compelled to arrest by serum-withdrawal, EBNA3C induces nuclear division that is often divorced from cytokinesis and so produces bi- and multinucleated cells. This was confirmed using the ecdysone-inducible system for expression of EBNA3C in human U2OS cells and by microinjection of expression vectors into NIH3T3 and U2OS. Further analysis revealed that in the inducible system, EBNA3C expression inhibits the accumulation of p27(K1P1) but not the dephosphorylation of pRb. Experiments using the microtubule destabilizing drug nocodazole, showed that EBNA3C could abrogate the mitotic spindle checkpoint.

Epstein-Barr virus suppresses a G(2)/M checkpoint activated by genotoxins

Several Epstein-Barr virus (EBV)-negative Burkitt lymphoma-derived cell lines (for example, BL41 and Ramos) are extremely sensitive to genotoxic drugs despite being functionally null for the tumor suppressor p53. They rapidly undergo apoptosis, largely from G(2)/M of the cell cycle. 5-bromo-2'-deoxyuridine labeling experiments showed that although the treated cells can pass through S phase, they are unable to complete cell division, suggesting that a G(2)/M checkpoint is activated. Surprisingly, latent infection of these genotoxin-sensitive cells with EBV protects them from both apoptosis and cell cycle arrest, allowing them to complete the division cycle. However, a comparison with EBV-immortalized B-lymphoblastoid cell lines (which have functional p53) showed that EBV does not block apoptosis per se but rather abrogates the activation of, or signalling from, the checkpoint in G(2)/M. Furthermore, analyses of BL41 and Ramos cells latently infected with P3HR1 mutant virus, which expresses only a subset of the latent viral genes, showed that LMP-1, the main antiapoptotic latent protein encoded by EBV, is not involved in the protection afforded here by viral infection. This conclusion was confirmed by analysis of clones of BL41 stably expressing LMP-1 from a transfected plasmid, which respond like the parental cell line. Although steady-state levels of Bcl-2 and related proteins varied between BL41 lines and clones, they did not change significantly during apoptosis, nor was the level of any of these anti- or proapoptotic proteins predictive of the outcome of treatment. We have demonstrated that a subset of EBV latent gene products can inactivate a cell cycle checkpoint for monitoring the fidelity and timing of cell division and therefore genomic integrity. This is likely to be important in EBV-associated growth transformation of B cells and perhaps tumorigenesis. Furthermore, this study suggests that EBV will be a unique tool for investigating the intimate relationship between cell cycle regulation and apoptosis.

Epstein-Barr virus nuclear antigen 3C interacts with histone deacetylase to repress transcription

EBNA3C can specifically repress the expression of reporter plasmids containing EBV Cp latency-associated promoter elements. Cp is normally the main promoter for EBNA mRNA initiation, so it appears that EBNA3C contributes to a negative autoregulatory control loop. By mutational analysis it was previously established that this repression is consistent with EBNA3C being targeted to Cp by binding the cellular sequence-specific DNA-binding protein CBF1 (also known as recombination signal-binding protein [RBP]-Jkappa. Further analysis suggested that in vivo a corepressor interacts with EBNA3C in this DNA binding complex. Results presented here are all consistent with a component of such a corepressor exhibiting histone deacetylase activity. The drug trichostatin A, which specifically inhibits histone deacetylases, relieved two- to threefold the repression of Cp induced by EBNA3C in two different cell types. Moreover, repression of pTK-CAT-Cp4x by EBNA3C was specifically enhanced by cotransfection of an expression plasmid for human histone deacetylase-1 (HDAC1). Consistent with these functional assays, in vitro-translated HDAC1 bound to a glutathione S-transferase (GST) fusion protein including full-length EBNA3C, and in the reciprocal experiment EBNA3C bound to a GST fusion with the N terminus of HDAC1. Coimmunoprecipitations also revealed an EBNA3C-HDAC1 interaction in vivo, and GST-EBNA3C bound functional histone deacetylase enzyme activity from HeLa cell nuclear extracts. The region of EBNA3C involved in the interaction with HDAC1 appears to correspond to the region which is necessary for binding to CBF1/RBP-Jkappa. A direct physical interaction between EBNA3C and HDAC1 was demonstrated with recombinant proteins purified from bacterial cells, and we therefore conclude that HDAC1 and CBF1/RBP-Jkappa bind to the same or adjacent regions of EBNA3C. These data suggest that recruitment of histone deacetylase activity makes a significant contribution to the repression of transcription from Cp because EBNA3C bridges an interaction between CBF1/RBP-Jkappa and HDAC1.

Novel p53 mutants selected in BRCA-associated tumours which dissociate transformation suppression from other wild-type p53 functions

Inheritance of germ-line mutant alleles of BRCA1 and BRCA2 confers a markedly increased risk of breast cancer and we have previously reported a higher incidence of p53 mutations in these tumours than in grade matched sporadic tumours. We have now characterized these p53 mutants. The results of these studies identify a novel class of p53 mutants previously undescribed in human cancer yet with multiple occurrences in BRCA-associated tumours which retain a profile of p53-dependent activities in terms of transactivation, growth suppression and apoptosis induction which is close or equal to wild-type. However, these mutants fail to suppress transformation and exhibit gain of function transforming activity in rat embryo fibroblasts. These mutants therefore fall into a novel category of p53 mutants which dissociate transformation suppression from other wild-type functions. The rarity of these mutants in human cancer and their multiple occurrence in BRCA-associated breast tumours suggests that these novel p53 mutants are selected during malignant progression in the unique genetic background of BRCA1- and BRCA2-associated tumours.

p53 mutation with frequent novel condons but not a mutator phenotype in BRCA1- and BRCA2-associated breast tumours

The status of p53 was investigated in breast tumours arising in germ-line carriers of mutant alleles of BRCA1 and BRCA2 and in a control series of sporadic breast tumours. p53 expression was detected in 20/26 (77%) BRCA1-, 10/22 (45%) BRCA2-associated and 25/72 (35%) grade-matched sporadic tumours. Analysis of p53 sequence revealed that the gene was mutant in 33/50 (66%) BRCA-associated tumours, whereas 7/20 (35%) sporadic grade-matched tumours contained p53 mutation (P<0.05). A number of the mutations detected in the BRCA-associated tumours have not been previously described in human cancer databases, whilst others occur extremely rarely. Analysis of additional genes, p16INK4, Ki-ras and beta-globin revealed absence or very low incidence of mutations, suggesting that the higher frequency of p53 mutation in the BRCA-associated tumours does not reflect a generalized increase in susceptibility to the acquisition of somatic mutation. Furthermore, absence of frameshift mutations in the polypurine tracts present in the coding sequence of the TGF beta type II receptor (TGF beta IIR) and Bax implies that loss of function of BRCA1 or BRCA2 does not confer a mutator phenotype such as that found in tumours with microsatellite instability (MSI). p21Waf1 was expressed in BRCA-associated tumours regardless of p53 status and, furthermore, some tumours expressing wild-type p53 did not express detectable p21Waf1. These data do not support, therefore, the simple model based on studies of BRCA-/- embryos, in which mutation of p53 in BRCA-associated tumours results in loss of p21Waf1 expression and deregulated proliferation. Rather, they imply that proliferation of such tumours will be subject to multiple mechanisms of growth regulation.

A transforming p53 mutant, which binds DNA, transactivates and induces apoptosis reveals a nuclear:cytoplasmic shuttling defect

The DG75 Burkitt lymphoma-derived human B cell line is heterozygous for p53, carrying wild type (WT) and mutant (Arg283His) alleles. The cells constitutively express high levels of both p53 proteins and also Mdm2. Arg283His transactivates the p21Waf1, Mdm2, bax, cyclin G and IGF-BP3 promoters in transient transfection assays equally as well as, if not better than WT p53. It also suppresses the outgrowth of SAOS-2 cells and specifically binds DNA like wild type protein. However, in primary rodent fibroblasts Arg283His fails to suppress transformation by HPV16-E7 and (Ha-)ras and even has modest transforming activity when transfected alone with (Ha-)ras. When Arg283His is transiently transfected into SAOS-2 cells it efficiently induces apoptosis, so - unlike mutants such as Arg175Pro - its behaviour in transformation assays does not clearly correlate with loss of the apoptosis function. Immunofluorescence staining of both REF transformants and transiently transfected SAOS-2 revealed that this unusual mutant becomes excluded from the nucleus and produces striking cytoplasmic fluorescence. The best correlation with transformation, therefore, appears to be the lack of nuclear retention of Arg283His. Since this mutation does not map to any known nuclear localization signal and its presence seems to result in aberrant exclusion from the nucleus, then it may prove very useful in exploring mechanisms involved in the nuclear:cytoplasmic shuttling of p53.

Epstein-Barr virus EBNA3C represses Cp, the major promoter for EBNA expression, but has no effect on the promoter of the cell gene CD21

EBNA3C is a potent repressor of transcription when bound to DNA as a fusion with the DNA binding domain (DBD) of GALA. A survey of promoters has revealed that the wild-type, unfused EBNA3C can specifically repress expression from reporter plasmids containing the Epstein-Barr virus Cp latency-associated promoter. Repression of Cp activity required amino acids 207 to 368, which encompasses a region resembling a basic DBD adjacent to a leucine zipper DNA binding motif and a site which binds to the cellular factor CBF1/RBP-Jkappa. However, amino acids 207 to 368 are dispensable when the protein is bound to DNA as a fusion with the GAL4 DBD, thus implicating this region in DNA binding. Mutation of the CBF1/RBP-Jkappa binding site in EBNA3C abrogated repression, strongly suggesting that CBF1/RBP-Jkappa is necessary for targeting the viral protein to Cp. Consistent with this result, mutation of the EBNA2 response element (a CBF1/RBP-Jkappa binding site) in Cp also prevented significant repression. In addition, amino acids 346 to 543, which were previously defined as important for the repressor activity of the GAL4-EBNA3C fusion proteins, also appear to be necessary for the repression of Cp. Since repression by these fusions was not observed in all cell types, it seems likely that EBNA3C either depends on a corepressor which may interact with amino acids 346 to 543 or is modified in a cell-specific manner in order to repress. These data are consistent with EBNA3C contributing to the regulation of EBNA expression in latently infected B cells through CBF1/RBP-Jkappa and another factor, but this need not directly involve EBNA2. Finally, although it has been reported that EBNA3C can upregulate CD21 in some B cells, we were unable to demonstrate any effect of EBNA3C on reporter plasmids which contain the CD21 promoter.

Epstein-Barr virus nuclear antigen (EBNA)3C is an immortalizing oncoprotein with similar properties to adenovirus E1A and papillomavirus E7

Epstein-Barr virus (EBV) requires six genes to efficiently immortalize human B cells. We have shown that one of these, EBNA3C, can cooperate with activated (Ha-)ras in co-transfection assays to immortalize and transform rat embryo fibroblasts (REFs). EBNA3C also augmented transformation by (Ha-)ras and a mutant p53 to a similar extent as human papilloma virus E7. As with E7 this effect was not inhibited by cotransfection with the cyclin-dependent kinase inhibitor (CDKI), a p16INK4A, which can normally activate the retinoblastoma protein (pRb) and induce growth arrest. Also like E7/ras and E1A/ras transformed cells the EBNA3C/ras transformants are very susceptible to apoptotic cell death. In vitro EBNA3C binds to pRb in a manner which is dependent on the integrity of the pocket domain; this suggests that EBNA3C, even though it lacks the LXCXE pRb binding motif found in E7 and E1A, may interact with pRb in vivo. We conclude that EBNA3C functions as an oncoprotein which directs cell cycle progression through the G1 phase restriction point when conditions might signal arrest. For the first time this demonstrates that EBV encodes a protein, functionally but not necessarily mechanistically, similar to the pRb-neutralizing nuclear antigens encoded by the 'small' DNA tumor viruses.

Epstein-Barr virus nuclear antigen 3C is a powerful repressor of transcription when tethered to DNA

The expression of Epstein-Barr virus (EBV) nuclear antigen 3C (EBNA3C) is essential for the activation and immortalization of human B lymphocytes by EBV. EBNA3C consists of 992 amino acids and includes a potential bZIP motif and regions rich in acidic, proline, and glutamine residues. Thus, EBNA3C resembles several trans regulators of gene expression. It has recently been shown that a fragment of EBNA3C can activate reporter gene expression when fused to the DNA-binding domain of GAL4 (D. Marshall and C. Sample, J. Virol. 69:3624-3630,1995). Although EBNA3C binds DNA, a specific site for EBNA3C binding has not been identified; to test the ability of full-length EBNA3C to regulate transcription, EBNA3C (amino acids 11 to 992) was fused to the DNA-binding domain of GAL4. We show that this fusion protein does not transactivate but rather is a potent repressor of reporter gene expression. Repression is dependent on the dose of GAL4-EBNA3C and on the presence of GAL4-binding sites within reporter plasmids. Repression is not restricted to B cells nor is it species or promoter specific. Repression is independent of the location of the GAL4-binding sites relative to the transcription start site. A fragment of EBNA3C (amino acids 280 to 525) which represses expression in a manner which is nearly identical to that of the full-length protein has been identified; this fragment is rich in acidic and proline residues. A second, less potent repressor region located C terminal to amino acids 280 to 525 has also been identified; this domain is rich in proline and glutamine residues. We also show binding of EBNA3C, in vitro, to the TATA-binding protein component of TFIID, and this suggests a mechanism by which EBNA3C may communicate with the basal transcription complex.

DNA damage in human B cells can induce apoptosis, proceeding from G1/S when p53 is transactivation competent and G2/M when it is transactivation defective

Cisplatin treatment of Epstein-Barr virus-immortalized human B lymphoblastoid cell lines (LCLs) results in p53-mediated apoptosis which occurs largely in a population of cells at the G1/S boundary of the cell cycle. Cell cycle progression appears to be required for this apoptosis because arresting cells earlier in G1 inhibited apoptosis despite the accumulation of p53. Overexpression of wild-type p53 also induces apoptosis in an LCL. Therefore six mutant genes derived from Burkitt's lymphoma (BL) cells were assayed for their ability to induce apoptosis when similarly overexpressed. The same genes were analysed in transient transfection assays for their ability to transactivate appropriate reporter plasmids. A correlation between the ability of p53 to transactivate and induce apoptosis was revealed. The only mutant capable of transactivation also induced apoptosis. Further analysis of the BL lines in which p53 had been characterized showed that whereas some lines were essentially resistant to cisplatin, three were rapidly induced to undergo apoptosis. All three have a single p53 allele encoding a mutant which is incapable of transactivation or (for two tested) mediating apoptosis when expressed in an LCL. Cell cycle analysis revealed that this apparently p53-independent apoptosis did not follow G1 arrest but in fact occurred largely in cells distributed in the G2/M phase of the cell cycle. These data suggest the existence of a second checkpoint in the G2 or M phase which, in the absence of a functional p53, is the primary point of entry into the apoptosis programme following DNA damage.

Epstein-Barr virus efficiently immortalizes human B cells without neutralizing the function of p53

Epstein-Barr virus (EBV) efficiently converts resting human B cells into actively cycling, immortal, lymphoblastoid cell lines (LCLs). Here we show that LCLs expressing the full complement of latent viral genes are very sensitive to DNA-damaging agents such as cisplatin. The response includes a rapid accumulation of the tumour suppressor protein p53 and induction of the cellular genes mdm2 and WAF1/p21. Although the levels of Bcl2 protein and Bax mRNA appear unaltered by the activation of p53, within 24 h the majority of cells undergo apoptosis. Over-expression of wild-type p53 in an LCL also resulted in apoptosis; this was preceded by the dephosphorylation of the retinoblastoma gene product, pRb. Primary resting B cells showed no response to cisplatin and even after drug treatment, p53 remained undetectable. However, after infection with EBV, p53 gene expression was induced to a similar level to that found in mitogen-activated B cells. When the physiologically activated primary B cells were exposed to cisplatin, although p53 accumulated as in LCLs, the outcome was growth-arrest rather than gross cell death. We conclude that, in contrast to the transformation of fibroblasts by adenovirus, SV40 or HPV, when B cells become activated and immortalized by EBV they are sensitized to the p53-mediated damage response. When the resulting LCLs are treated with genotoxic agents such as cisplatin, they are unable to arrest like normal cells because they are driven to proliferate by EBV and consequently undergo apoptosis.

Epstein-Barr virus nuclear antigen EBNA3C/6 expression maintains the level of latent membrane protein 1 in G1-arrested cells

The Epstein-Barr virus in the Burkitt lymphoma-derived cell line Raji has a deletion in the EBNA3C gene. When Raji cells are allowed to grow to high density and most of the cells become growth arrested in the G1 phase of the cell cycle, the level of detectable latent membrane protein 1 (LMP1) is substantially reduced. After dilution of the cells with fresh growth medium, within 8 h, there is a large increase in LMP1 mRNA, and by 12 h, LMP1 is expressed to a high level (H. Boos, M. Stoehr, M. Sauter, and N. Mueller-Lantzch, J. Gen. Virol. 71:1811-1815, 1990). Here we show that in Raji cells which constitutively express a transfected EBNA3C gene, the down-regulation of LMP1 in growth-arrested cells does not take place. Furthermore, we show that in wild-type Raji cells, low-level LMP1 expression occurs when most of the cells are arrested at a point(s) early in G1 (or G0) when the product of the retinoblastoma gene, pRb, is hypophosphorylated. The dramatic synthesis of LMP1 coincides with the progression of these cells to late G1 when pRb becomes hyperphosphorylated. Thus, in Raji cells, the LMP1 gene is apparently regulated in a cell cycle- or proliferation-dependent manner, but when EBNA3C is present, sustained LMP1 expression occurs as it does in a lymphoblastoid cell line. EBNA3C appears to either relieve the apparent repression of LMP1 in cells progressing through early G1 or possibly alter the stage at which the cells growth arrest to one where they are permissive for LMP1 expression.

Epstein-Barr virus (EBV) nuclear antigen 6 induces expression of the EBV latent membrane protein and an activated phenotype in Raji cells

Epstein-Barr virus (EBV) nuclear antigen (EBNA) 6 (also known as 3c) is a latent nuclear protein with an M(r) of about 160K which is invariably expressed in EBV-immortalized B cells. It includes a putative basic leucine zipper domain; as such it is a good candidate for a regulator of viral gene expression. More than 75% of the EBNA 6 coding sequence is deleted from viral genomes carried in the Burkitt's lymphoma (BL) tumour-derived cell line, Raji. Thus although Raji cells express normal levels of the remaining five EBNAs and low levels of latent membrane protein (LMP), EBNA 6 protein is completely absent. In this study we have established Raji clones stably expressing EBNA 6 after cotransfection of an EBNA 6 gene under the control of the simian virus 40 early promoter with a selectable marker. Analysis of these clones has revealed that EBNA 6 induces a significant increase in the expression of LMP. In addition the cells have undergone a number of morphological and phenotypic changes consistent with blast-activation of normal B lymphocytes. The Raji cells expressing EBNA 6 show ruffling of the cell membrane and the development of a polarity defined by multiple villous ('spiky') projections at one end of the cell. This morphological change is associated with a dramatic increase in the expression of the cytoskeletal protein, vimentin. The EBV-associated B cell activation marker CD23 (blast 2) is induced to high levels although other activation markers such as CD30 and CD39 are unaffected. All these changes appear to be independent of the precise levels of EBNA 6 protein expressed. EBNA 2 has been shown previously to trans-activate the LMP gene and in the control Raji cells, EBNA 6-positive Raji cells and in B lymphoblastoid cells similar levels of EBNA 2 are expressed. Our findings are therefore most consistent with a model in which EBNA 6 either augments or complements the action of EBNA 2 in the induction of LMP and the cascade of gene expression which leads to B cell activation and immortalization by EBV.

CpG methylation of viral DNA in EBV-associated tumours

In EBV-immortalized lymphoblastoid cell lines (LCLs) a small number of "latent" proteins are expressed. These are the EBV nuclear antigens, EBNAs 1-6, and a latent membrane protein, LMP. We have investigated the expression of these proteins in a variety of EBV-associated tumours and cell lines. Whereas transplant and B-cell lymphomas from cotton-top tamarins appear to express the full range of antigens found in LCLs, we and others have found that in Burkitt's lymphomas (BL) and a nasopharyngeal carcinoma (NPC) isolate, EBNA expression is restricted to EBNA-I. (In NPC, but not in BL, LMP may also be expressed). In order to ask what restricts the expression of EBNA 2-6 in NPC and BL cells it seemed reasonable to consider the possibility that the DNA sequences normally regulating expression of these antigens could be chemically modified. In this analysis, a tight inverse correlation between methylation of CpG dinucleotides in the 5' flanking region of the EBNA-2 gene and the expression of EBNAs 2-6 has been revealed. In the NPC tumour, CpG methylation within the gene is also observed, as is specific methylation over the EBNA-I region I and II binding sites (in oriP). The significance of these observations is considered.

Immunohistology of Epstein-Barr virus-associated antigens in B cell disorders from immunocompromised individuals

Proliferating B cell lesions developing in a series of immunosuppressed organ transplant recipients and patients with X-linked lymphoproliferative syndrome were examined for Epstein-Barr virus and cellular gene expression using immunocytochemistry and immunoblotting techniques. Results indicate that all the lesions examined from the patients in this series expressed Epstein-Barr virus gene products that were consistent with a latent, nonproductive type of infection. No lytic cycle antigens associated with productive viral infection were detected. This pattern is similar to the viral gene expression in normal B cells immortalized by Epstein-Barr virus in vitro. The demonstration in this study of Epstein-Barr virus viral gene expression in posttransplant and X-linked proliferative syndrome B cell disorders provides important new evidence for the primary role of Epstein-Barr virus in the development of these lesions. This is in contrast to the subsidiary role that the Epstein-Barr virus has in the etiology of Burkitt's lymphoma.

Deregulated c-myc expression in Epstein-Barr-virus-immortalized B-cells induces altered growth properties and surface phenotype but not tumorigenicity

Endemic Burkitt's lymphoma (eBL) is characterized by the presence of Epstein-Barr virus (EBV) and a chromosomal translocation which results in deregulation and constitutive expression of the c-myc proto-oncogene. In order to examine the role played by activation of c-myc in determining the eBL phenotype, we have introduced into EBV-immortalized lymphoblastoid cells (LCL) plasmids which permit constitutive expression of c-myc. The resulting cells show a reduced serum dependence, reduced homotypic cell aggregation, and changes in surface characteristics. In particular, levels of the cell adhesion molecule, LFA-I, are greatly reduced. However, the cells continue to express all the EBV latent antigens associated with the LCL phenotype and they remain nontumorigenic. These results suggest that, whilst constitutive expression of c-myc may contribute to the malignant phenotype, it is insufficient to induce tumorigenicity.

Immortalization of Epstein-Barr virus-infected CD23-negative B lymphocytes by the addition of B cell growth factor

Epstein-Barr (EB) virus-immortalized B lymphocytes coexpress the EB viral latent gene products (EB viral nuclear antigens 1 to 6, the latent membrane protein and the terminal protein gene products) and the cellular activation antigen CD23. Immortalized B cells can be separated from those which are infected but not immortalized on the basis of CD23 expression as early as 2 days after in vitro infection. In the present report we have confirmed these data, but show that if left in culture for 7 days after infection before separation the CD23-negative cells show a donor-related ability to become CD23-positive and immortalize. CD23-negative cells separated 2 days after infection can be induced to immortalize by the addition of low Mr B cell growth factor but not by the addition of recombinant interleukin 1, 4 or soluble CD23. At 2 to 3 days after infection the EB viral nuclear antigens 1, 2 and the high Mr species 3, 4 and 6, as well as the latent membrane protein can be detected in the CD23-positive fraction. In contrast at this time only nuclear antigens 1 and 2 could be detected in the CD23-negative fraction. This difference in gene expression may account for the inability of the CD23-negative fraction to immortalize. In the light of these observations the mechanism of viral persistence in vivo is discussed.

Activation and immortalization of leukaemic B cells by Epstein-Barr virus

We have studied the responses of chronic leukaemic B cells to infection by Epstein-Barr virus (EBV). Our results define one population of B lymphocytes, represented by prolymphocytic leukaemic (PLL) cells, which are highly susceptible to immortalization by EBV. Another B-cell type, represented by chronic lymphocytic leukaemic (CLL) cells, can be readily infected by the virus but is resistant to immortalization. Comparative studies of viral and cellular related events early after infection in these 2 cell types reveal that both express the EB viral nuclear antigen (EBNA) complex, but the immortalization-resistant CLL cells fail to express the latent membrane protein (LMP), which can be detected in PLL cells 48 hr after infection. Circularization of the linear viral genome could be detected at 7 days post infection in the PLL cells, but only in 2 out of 4 CLL cells tested. Both CLL and PLL cells show increased surface expression of CD23 and HLA-DR molecules after infection but, whereas PLL cells show an increase in size, together with RNA and DNA synthesis indicative of cell cycle progression, CLL cells appear to be arrested in the G1/S phase of the cycle. The results suggest that the outcome of infection by EBV is determined by the nature of the target cell rather than by random virus-related events. The correlation between the block in immortalization of CLL cells and their failure to express LMP suggests that expression of this protein is essential for in vitro immortalization of B cells. The failure to detect circularization in some EBV-infected CLL cells suggests that this, as well as LMP expression, may be dependent on prior activation of the B cell by EBV, an event which may vary between the different CLL samples tested.

EBV gene expression in an NPC-related tumour

A nasopharyngeal carcinoma tumour (designated C15) propagated in nude mice has been used to generate a large cDNA library that we have analysed for Epstein-Barr virus (EBV) gene expression. No gross alterations exist in viral DNA from C15 relative to other human isolates and the large deletion present in the B95-8 'prototype' viral strain established in marmoset cells is not found; C15 contains no linear virion DNA. In the cDNA library, of the six EBV nuclear antigens (EBNAs) expressed in latently infected B-lymphocytes, only clones for EBNA-1 are found. These data are confirmed by immunoblotting. Sequence analysis shows the EBNA-1 mRNA splicing pattern in the carcinoma to differ from that observed in B-lymphocytes. Further, contrary to observations with B-cell lines, most viral transcription in the tumour is localized onto the 'rightmost' region of the conventional EBV physical map. Transcripts identified corresponding to known genes include those for the latent membrane protein (LMP), the alkaline DNA exonuclease and probably the terminal protein; major transcripts are also derived from the BamHI D fragment and the region deleted in B95-8 EBV DNA. Novel transcripts have also been identified that proceed in an anti-sense direction to genes encoding functions associated with replication, such as the viral DNA polymerase. They contain a large, hitherto unidentified, open reading frame in the viral genome that is complementary to the putative function known as BALF3 and a smaller open reading frame complementary to BALF5 (the DNA polymerase gene). From the present studies we can conclude that: (i) EBV transcription patterns in the epithelial cells vary markedly from those identified previously in B-cells, reflecting differential use of promoters or splicing patterns. (ii) Transcription is tightly regulated and restricted in the C15 tumour with many latent genes, notably EBNAs 2-6, being 'switched off.' (iii) A family of cytoplasmic RNAs are transcribed in an antisense direction to a number of existing open reading frames in the EBV genome. (iv) There are a number of mutations in C15 transcripts relative to the B95-8 genome, some of which could result in amino acid alterations in proteins.

Epstein-Barr virus latent gene expression during the initiation of B cell immortalization

Epstein-Barr virus (EBV) has the capacity to immortalize a subpopulation of resting B lymphocytes. Lymphoblastoid cell lines (LCL) established in this way carry the latent EBV genome as multiple copies of an extrachromosomal episome. Viral gene expression in LCLs is highly restricted; products identified correspond to a membrane protein (latent membrane protein; LMP), a nuclear antigen complex (Epstein-Barr nuclear antigens; EBNAs 1 to 6), two small RNA species (EBERs 1 and 2) and RNA species thought to encode a second membrane-associated polypeptide designated terminal protein (TP). Here we have investigated the temporal sequence of expression of the characterized 'latent' proteins during the initiation of immortalization when resting B cells are stimulated to enter and traverse the cell cycle. The analysis has been carried out on prolymphocytic leukaemia cells infected in vitro with either the immortalizing B95-8 strain of virus or the non-immortalizing P3HR1 strain. The results reveal that following B95-8 infection, a sequence of EBV expression is initiated within approximately 8 h with the synthesis of detectable levels of EBNA 2 shortly followed by EBNAs 1, 3, 4, 5 and 6. There is then a delay of approximately 40 h until the expression of LMP completes the latent pattern of proteins found in LCLs. P3HR1 infection, however, produces only transient expression of some EBNA species in a small percentage of cells after approximately 48 h. These observations suggest the failure of P3HR1 virus to immortalize may not be due solely to the absence of EBNA 2 expression and that cellular and/or virus-mediated events occur after EBNA synthesis which then facilitate efficient LMP expression and immortalization.

EBNA-1: a virally induced nuclear antigen of primate lymphocytes and its expression in Drosophila cells

EBNA-1 is a nuclear antigen of lymphocytes infected by Epstein-Barr virus and whose size polymorphism correlates only with the strain of infecting virus and the length of the glycine-alanine copolymer encoded by the third internal repeat of the viral genome. The major antigenic determinant(s) also appear to reside in this region. We have been able to obtain efficient expression of this nuclear antigen in cultured Drosophila cells transfected with a cosmid carrying the EBNA-1 coding region, indicating that insect mechanisms recognise control sequences and transcripts of the herpes virus. The association of a vimertin-like protein of mol. wt. 46,000 with Drosophila cell nuclei has been found to vary with culture conditions and heat shock. We now find that the level and nuclear association of this protein also increase after transfection with either EBNA-1 or yolk protein DNA.

Prediction and demonstration of a novel Epstein-Barr virus nuclear antigen

The protein sequence predicted by the Epstein Barr virus (EBV) BERF4 open reading frame includes a tetrapeptide, Lys-Arg-Pro-Arg (KRPR), shown for other proteins to be a component of a signal for rapid nuclear localization. A subgenomic fragment of EBV DNA containing BERF4 has been incorporated into an expression vector, transfected onto primate cells and the nuclear distribution of the resulting protein established by immunofluorescence using EBV positive human sera. These sera contained high titres of antibodies to a fusion protein, produced in E. coli, consisting of beta-galactosidase and the C-terminal 167 amino acids of BERF4. Immunoaffinity purified antibodies reactive with the EBV component of the fusion show the molecular weight of this antigen in EBV immortalized B-cell lines to be about 160 kD. The demonstration that BERF4 contains an exon encoding a nuclear protein identifies a new EBNA gene (EBNA-6) and suggests that KRPR is a signal sequence common to a number of viral and cellular nuclear polypeptides which bind to nucleic acids and may therefore be of predictive value in identifying karyophilic proteins.

Role of epithelium in EBV persistence and pathogenesis of B-cell tumours

EBV has long been thought to be primarily a B-lymphotropic virus. A manifestation of this tropism is the association of the virus with a variety of B-cell lymphoproliferative disorders and tumours. However, there is now considerable evidence to suggest that the permissive cell type for EBV replication is epithelial and that infection of B cells may be a secondary, and, from the point of view of the virus, an unimportant consequence. A re-evaluation of the role that epithelium plays in viral persistence and of the importance of the immune response in the development of persistent infection indicates that T-cell-mediated immunity to the epithelial infection is critical in maintaining the normal delicate balance between virus and host. It also suggests that uncontrolled EBV replication in pharyngeal epithelium may be central to the evolution of some (or all) EBV-associated lymphomas.

Efficient expression of an Epstein-Barr nuclear antigen in Drosophila cells transfected with Epstein-Barr virus DNA

In a search for exogenous promoters which function in cultured Drosophila cells, we have co-transfected a D. melanogaster cell line with an Epstein-Barr virus (EBV) cosmid clone which encodes the Epstein-Barr nuclear antigen (EBNA-1). Here we report that Drosophila cells containing stably integrated copies of EBNA-1 encoding DNA synthesise a polypeptide of mol. wt. identical to that of authentic EBNA-1, which is detectable with EBNA-positive but not EBNA-negative human serum. As in EBV-transformed lymphoblastoid cells, this neo-antigen is associated with the nucleus of transfected cells suggesting that cellular localisation signals which operate in mammalian cells are also recognised in insect cells.

Epstein-Barr virus nuclear antigen (EBNA): size polymorphism of EBNA 1

The mol. wt. of the polymorphic Epstein-Barr virus (EBV) nuclear antigen (EBNA) molecule (EBNA 1) encoded by the BamHI K fragment of the EBV DNA has been determined in 14 EBV-carrying lymphoblastoid and Burkitt's lymphoma cell lines. There is no obvious correlation between the size of this polypeptide and any properties of the cells from which it is derived, other than those related to the strain of transforming virus. We confirm that the polymorphic region of this molecule is the glycine-alanine copolymer encoded by the third internal repeat of the EBV genome (IR3) and we consider the significance of this domain.