Epstein Barr Virus Unit, Queensland Institute of Medical Research, The Bancroft Centre, 300 Herston Road, Brisbane, Queensland 4029, Australia

Transcription

1 International Immunology, Vol. 9, No. 11, pp Oxford University Press Selection of a diverse TCR repertoire in response to an Epstein Barr virus-encoded transactivator protein BZLF1 by CD8 cytotoxic T lymphocytes during primary and persistent infection Sharon L. Silins, Simone M. Cross, Suzanne L. Elliott, Stephanie J. Pye, Jacqueline M. Burrows, Denis J. Moss and Ihor S. Misko Epstein Barr Virus Unit, Queensland Institute of Medical Research, The Bancroft Centre, 300 Herston Road, Brisbane, Queensland 4029, Australia Keywords: Epstein Barr virus, infectious mononucleosis, repertoire selection, TCR β chain, viral replicative antigen Abstract We investigated the CD8 cytotoxic T lymphocyte (CTL) repertoire to an HLA B8-restricted peptide, RAKFKQLLQ, located in the Epstein Barr virus (EBV) immediate-early protein, BZLF1. Repertoire selection was monitored by determining the TCR β chain sequences of RAKFKQLLQ-specific CTL established from primary infected and healthy virus carriers. PCR analysis of spontaneous EBVtransformed lymphoblastoid cell lines (LCL) from three individuals with primary infection showed that two were infected with type A and one with type B EBV. Polyclonal and clonal CTL that were generated by stimulating peripheral blood mononuclear cells with an HLA B8 homozygous LCL lysed T cell blasts pulsed with the peptide, RAKFKQLLQ; lysis of certain HLA B8 LCL targets was associated with the abundance of BZLF1 transcripts. TCR β analysis showed that while there was loop length restriction in the putative peptide contact site of all responding β chains, diverse and unique (non-recurrent) TCR β clonotypes were selected in individuals during primary infection and continued to emerge after long-term virus exposure. TCR-contact site heterogeneity was excluded as the selective force in diversity generation since the epitope-encoded sequences were found to be identical within endogenous virus isolates. In this first study of TCR repertoire selection for an EBV lytic antigen, a BZLF1-reactive component of diverse clonotypes was identified in primary type A or type B EBV infection which was sustained in the EBV-specific memory response throughout life-long infection. This diversity selection is likely to play a critical role in maintaining a balanced viral load throughout EBV persistence. Introduction Epstein Barr virus (EBV), a ubiquitous human gammaherpes virus consisting of two subtypes, type A and type B (1), is the etiologic agent of the self-limiting, lymphoproliferative disease, acute infectious mononucleosis (IM) (2). Acute IM is characterized by intense viral replication in oropharyngeal epithelial cells (3), tonsillar B cells (4) and by a high prevalence of virus-infected B cells in the peripheral blood (5). Notwithstanding, a primary infection with EBV can result in an asymptomatic or mild to severe symptomatic seroconversion. Class I-restricted cytotoxic T lymphocytes (CTL) play an important immunosurveillance role in controlling the latent infection of B lymphocytes that is established following primary infection (reviewed in 6). A multicomponent memory CTL response targeting epitopes in the latent EBV nuclear antigens (EBNA) and latent membrane proteins has been well documented in healthy virus carriers (7,8), and more recently has been detected in individuals undergoing acute IM (9,10). Many of these EBV-encoded CTL epitopes are strictly specific Correspondence to: S. Silins Transmitting editor: P. Doherty Received 28 May 1997, accepted 11 August 1997

2 1746 TCR diversity selection in Epstein Barr virus infection for type A viral antigens, although certain epitopes are crossreactive (7). In contrast, very little is known about CTL epitopes in type B viral antigens (11), and only a few epitopes have been defined in proteins associated with viral replication (12,13), the immunogenicity and prevalence of which in primary infection have still to be determined. CTL recognize virus-infected cells through interaction between their TCR and the MHC viral peptide complex on the target cell surface. The specificity of this interaction is determined by the TCR, which is a clonally distributed αβ heterodimer generated through somatic recombination of variable (TCRAV and TCRBV), diversity (TCRBD) and joining (TCRAJ and TCRBJ) gene elements (reviewed in 14). Both restricted (oligoclonal) and diverse (polyclonal) antigenspecific TCR repertoires have been identified. A major unsolved question, however, is what determines the degree of structural TCR diversity for a given peptide and HLA restriction. The target peptide itself (15), the overall avidity of the interaction, including the affinity of the TCR for its ligand, and the density of ligand presented are likely to be important. We have also shown that tolerance to self-antigens can broaden the range of responding TCR to a viral epitope in humans (16). Beyond the primary selection event, progressive changes and refinement in TCR usage may also develop as a result of recurrent or persistent antigen challenge. Indeed, while diversity selection is likely to be important in maintaining the breadth of the TCR response, preferential selection of dominant clonotypes particularly well suited for specific recognition of target epitopes may be effective in the long-term memory response. Evidence of repertoire focusing upon repeat antigenic stimulation has recently been reported, with the selective accumulation of distinct groups of TCR in the memory population of cytochrome c-specific CD4 helper cells in mice (17). In the case of persistent infection with EBV where periodic reactivation and lytic replication of the virus throughout life ensures repeated exposure of the CD8 T cell pool to viral antigens, an extreme degree of TCR repertoire restriction has been found. A highly focused memory response is observed in long-term HLA B8 virus carriers to the latent EBV nuclear antigen 3 (EBNA3)-encoded epitope, FLRGRAYGL, in which a single invariant or public TCR dominates (18). Additional repertoire focusing is further evident from (i) the high frequency of alternative public TCR that arises in response to this determinant when the dominant clonotype, which is crossreactive with HLA B*4402, is inactivated by tolerance in HLA B8 /B4402 individuals (16), and (ii) the observation during acute EBV infection that distinct epitope-specific CTL clonotypes are expanded and preserved long-term in persistent infection (9). Collectively, these findings provide strong support for a model in which persistent EBV infection is conducive to the oligoclonal expansion of distinct groups of virus-specific CD8 cells. To further investigate CTL selection and TCR repertoire diversity in EBV infection, we studied the TCR β response to the replicative epitope, RAKFKQLLQ (residues ), in primary and persistent infection. This antigenic determinant is encoded within the immediate-early transactivator of virus replication, BZLF1 (also termed Zta or ZEBRA), of both type A and type B EBV, and is recognized by class I-restricted CTL in HLA B8 healthy virus carriers (12) at high precursor frequencies (19). In the present work, we show, for the first time, that CTL generated in culture from individuals undergoing a primary infection with type A or type B EBV recognize the lytic epitope, RAKFKQLLQ, with a diverse TCR β repertoire that is maintained long term in healthy virus carriers. Methods Cell donors Blood samples were taken from HLA B8 donors HB, LC and BT that were IgM for EBV viral capsid antigen (VCA) and IgG for EBNA, a serological profile indicating a primary infection with EBV (20). Donors HB and LC had been selected for a phase I EBV vaccine trial, but were excluded from the trial when they seroconverted following a primary infection prior to vaccination; LC was diagnosed with mild clinical symptoms and HB seroconverted asymptomatically. The third donor, BT, was diagnosed with acute IM based on serology and clinical symptoms that included swollen posterior cervical lymph nodes, fever and pharyngitis. Donor BT was studied longitudinally during the acute IM stage [referred to as BT(I)] and at 85 [BT(II)] and 196 days [BT(III)] post-diagnosis when the patient showed recovery from the disease and IgG EBNA antibodies developed (VCA IgG, EBNA IgG ). Blood was also taken from two healthy (VCA IgG, EBNA IgG ), longterm HLA B8 virus carriers, ISM and MS. Donors ISM and MS were infected with type A and type B virus respectively, based on T cell regression (21) and A- and B-specific CTL epitope responses. Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood by centrifugation over Ficoll-Paque (Pharmacia Biotechnology, Melbourne, Australia). The HLA B8 status of each donor was initially determined by FACScan analysis of PBMC using an anti-hla B8 mab (clone 59HA-1; One Lambda, Los Angeles, CA) and confirmed by serological typing of the donor s PBMC. Establishment and maintenance of cell lines Autologous LCL were established from donors HB, LC and BT by transformation of B cells with exogenous type A EBV (QIMR-WIL isolate) as described (22). Allogeneic LCL from an HLA B8 donor, JS, were transformed with the type A IARC-BL74, B95.8 or IARC-BL36 isolates, or type B Ag876 or L4 isolates as described (22). Spontaneous LCL that were infected in vivo with the donor s endogenous virus were generated from donors HB, LC and BT in limiting dilution cultures (23). LCL were routinely maintained in growth medium consisting of RPMI 1640 medium, 2 mm glutamine, 100 IU/ ml penicillin, 100 µg/ml streptomycin and 10% (v/v) heatinactivated FBS. PBMC were stimulated with phytohaemagglutinin (PHA) as described (24). PHA blasts were maintained in continuous culture for 6 8 weeks in growth medium containing 20 IU/ml ril-2 (25,26). EBV strain typing Spontaneous LCL containing endogenous EBV isolates were generated from donors HB, BT and LC as described (23). The virus infecting spontaneous LCL was identified by PCR

3 TCR diversity selection in Epstein Barr virus infection 1747 analysis using type-specific primers as previously described (27). Generation of polyclonal EBV-specific CTL in primary infection Polyclonal EBV-specific CTL from donors HB, LC and BT were established in PBMC cultures stimulated with the γ-irradiated (80 Gy) LCL from the type A EBV, HLA-A1/B8 homozygous individual BM at a stimulator:responder cell ratio of 1:20. The growth medium was supplemented with 20 IU/ml ril-2 to avoid apoptosis of in vivo activated T cells (28). Polyclonal CTL were amplified in culture with biweekly re-stimulation with ril-2 and γ-irradiated BM LCL. Agar cloning of T cells T cell colonies were generated from primary infected donors (HB, LC and BT) and healthy virus carriers (ISM and MS) as described (29), except that PBMC from the primary infected donors were exposed to growth medium containing 20 U/ml ril-2 after their isolation over Ficoll-Paque to avoid apoptosis of T cells (28). Briefly, ex vivo PBMC were activated by stimulation with the γ-irradiated (80 Gy) LCL from the type A EBV, HLA-A1/B8 homozygous individual BM at a stimulator:responder cell ratio of 1:20. After 3 days, dispersed cells were seeded in 0.35% agarose (SeaPlaque; FMC BioProducts, Rockland, ME) containing RPMI 1640 medium, 10% FBS, 25% (v/v) supernatant from MLA-144 cultures (TIB-201; ATCC, Rockville, MD) and 30 U/ml ril-2. Colonies were harvested after a further 3 days and amplified in culture with biweekly re-stimulation with ril-2 and γ-irradiated BM LCL. Colonies were routinely immunophenotyped using directly labelled mab to CD4 (Leu-3a FITC) and CD8 (Leu-2a phycoerythrin) (both from Becton Dickinson, Sydney, Australia), and flow cytometry was performed on a FACScan (Becton Dickinson). CTL colonies were initially screened using the rapid visual T cell T cell killer assay (30) to identify RAKFKQLLQ-specific CD8 CTL colonies. Cytotoxicity assay Polyclonal CTL and epitope-specific CTL colonies were used as effectors against BM PHA T cell blasts, with and without adsorbed peptide (100 µg/ml, 0.5 ml), and HLA B8 LCL. The EBNA3-encoded, HLA B8-restricted peptides, FLRGRAYGL and QAKWRLQTL (31), were included in each assay as specificity controls. Peptides were purchased from Chiron Mimotopes (Melbourne, Australia), dissolved in methyl sulphoxide and diluted in serum-free RPMI 1640 medium for use in cytotoxicity assays. Target cells were incubated with 100 µci of 51 Cr at 37 C for 90 min, with and without peptide, washed twice by centrifugation, and used in standard 4 h 51 Cr-release assays. The mean spontaneous lysis for targets in culture medium was 20%, the mean maximum lysis in 0.5% SDS was 90% of total uptake and the variation about the mean specific lysis was 5%. RNA isolation and cdna synthesis Total RNA was extracted from CTL or LCL using Total RNA Isolation Reagent (Advanced Biotechnologies, London, UK). For BZLF1 and TCR transcript analysis first strand cdna was synthesized using an oligo(dt) primer and an antisense TCRBC primer (C b1 ) (18) respectively. Firststrand cdna was synthesized from 2 5 µg of total RNA, and this was followed by RNA hydrolysis and the removal of excess primer as described previously (18). Semiquantitative PCR analysis of BZLF1 gene expression Semiquantitative PCR analysis of the relative amount of BZLF1 transcript was performed during exponential increase of PCR product [32 cycles for BZLF1; 22 cycles for β 2 -microglobulin (β 2 m)] using intron-spanning primers (exon 1 2/3) (32) specific for BZLF1 and β 2 m (33). Amplifications were performed in 25 µl reaction volumes using 2 µl cdna, 10 pmol of each 5 and 3 primer, 200 mm dntps, 20 mm MgCl 2, 1.25 U of Taq polymerase (Ampli-Taq), and a GeneAmp PCR 9600 system (Perkin-Elmer Cetus, Norfolk, CT) with cycles of 94 C for 20 s, 60 C for 40 s and 72 C for 40 s. Amplified products were separated on 2% agarose gels containing ethidium bromide and the fragments were quantified densitrometrically using ImageQuant software version 3.3 (Molecular Dynamics, Sunnyvale, CA). The single BZLF1 PCR product was directly sequenced to confirm its identity according to the conditions described below. Amplification and sequencing of rearranged TCR β sequences TCR β rearranged sequences were amplified with five sets of 5 TCRBV family-specific oligonucleotides (V β 1 5.1, V β 5.2 9, V β 10 14, V β and V β 20 24) and a 3 TCRBC (C β ) constant primer. The oligonucleotides V β 1 20/C β and V β were synthesized according to Panzara et al. (34) and Kalams et al. (35) respectively. Amplifications were performed in 25 µl reaction volumes using 0.5 µl cdna, 10 pmol of each V β and C β primer, 200 mm dntps, 20 mm MgCl 2,1.25UofTaq polymerase (Ampli-Taq), and a GeneAmp PCR 9600 system (Perkin-Elmer Cetus). The PCR conditions consisted of denaturation at 95 C for 15 s, annealing at 60 C for 40 s and extension at 72 C for 40 s for 35 cycles, followed by a 5 min final extension at 72 C. β-actin was also amplified as a control for cdna integrity (36). Positive reactions for a particular primer set were tested further by PCR using individual 5 V β primers and the 3 C β primer. PCR products were excised from 2.5 % (0.5 Tris-buffered EDTA) NuSieve GTG agarose gels (FMC BioProducts) and purified using a QIAEX gel extraction kit (Qiagen, Chatsworth, CA). Recovered PCR products were sequenced in both directions with a PRISM Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit and a 373A DNA sequencer (Applied Biosystems, Foster City, CA). Nucleotide sequence analysis of PCR products that revealed more than one specific sequence were subsequently ligated into the pgem-t vector system (Promega, Madison, WI). The nucleotide sequence of six clones was determined for each ligation. RAKFKQLLQ epitope sequencing of endogenous EBV isolates The EBV sequences within the RAKFKQLLQ epitope region in spontaneous LCL containing endogenous EBV isolates (23) were determined by RT-PCR amplification of BZLF1 transcripts using intron-spanning primers (32) followed by direct sequencing according to the conditions described above.

4 1748 TCR diversity selection in Epstein Barr virus infection Fig. 1. Type A (EBNA-2A) and type B (EBNA-2B) EBV PCR analysis of spontaneous LCL from primary infected donors BT, HB and LC. The lanes represent: kb (1 kb ladder); H 2 O (cdna PCR); A (type A LCL transformed with B95.8 isolate); B (type B LCL transformed with Ag876 isolate); K (K562, EBV cell line); BT, HB and LC (spontaneous LCL derived from donors BT, HB and LC). Results Identification of RAKFKQLLQ-specific CD8 CTL in the primary immune response to infection with type A or type B EBV In this study, the polyclonal and clonal CTL response to the HLA B8 -restricted peptide, RAKFKQLLQ, was investigated in primary infected donors. To determine the EBV type involved in the primary infection of donors HB, LC and BT, endogenous virus infecting their spontaneous LCL was identified by PCR analysis using type-specific primers. The data in Fig. 1 show unequivocally that the spontaneous LCL generated from donors BT and LC were infected with type A and the HB spontaneous LCL with type B EBV. The type B status of donor HB was also confirmed by a T cell regression assay showing that HB T cells prevented the outgrowth in culture of B cells transformed with type B, but not with type A EBV, and by their lack of a CTL response to type A-specific epitopes, FLRGRAYGL and QAKWRLQTL (data not shown). By contrast, donors BT and LC responded strongly to these type A- specific epitopes (data not shown). Thus, donors BT and LC were undergoing a primary infection with type A, and HB with type B EBV, but dual infections cannot be completely excluded. To demonstrate the presence of RAKFKQLLQ-specific CTL precursors in the primary immune response to EBV, polyclonal CTL were used as effectors against HLA B8 BM PHA T cell blasts pulsed with the RAKFKQLLQ peptide. The cytotoxicity data in Fig. 2(A) clearly show that the polyclonal CTL from donors BT, HB and LC strongly lysed peptide-coated BM blasts, but not uncoated control blasts. These findings were substantiated by analysing CTL colonies generated from the three donors. Initially, to identify RAKFKQLLQ-specific CTL colonies, CD8 ( 98%) colonies were screened using the rapid visual T cell T cell killer assay. The EBNA 3-encoded, HLA B8-restricted peptides, FLRGRAYGL and QAKWRLQTL, were included in each assay as specificity controls. RAKFKQLLQ-specific CTL colonies did not cross-react with Fig. 2. Specific lysis by (A) polyclonal CTL lines and (B) CTL colonies from primary infected donors BT, HB and LC of HLA B8 BM PHA blasts, with and without adsorbed peptide, RAKFKQLLQ. Effectors were tested in standard 4 h 51 Cr-release assays. Effector:target ratio: 5:1 (A) and 2:1 (B). these peptides (data not shown) and were used as effectors in 51 Cr-release assays against BM PHA T cell blasts, with and without adsorbed RAKFKQLLQ peptide, and HLA B8 LCL. In Fig. 2(B), the cytotoxicity profiles of representative CTL colonies from donors BT, HB and LC show that each colony lysed peptide-coated, but not uncoated, BM blasts. Moreover, three representative colonies [BT(I)142, HB39 and LC5] shown in Fig. 3(A) lysed type A BM, HB and LC targets, but the type A BT LCL target was only poorly lysed even though each LCL target was exogenously infected with the QIMR- WIL strain of EBV. A representative colony [BT(I)119] shown in Fig. 3(B) lysed type A and type B, HLA B8 LCL established from an allogeneic donor JS by exogenous infection with different strains of EBV confirming that the RAKFKQLLQ peptide was common to both viral types. Semiquantitative PCR analysis using BZLF1-specific primers was used to determine the relationship between abundance of BZLF1 transcript in target LCL and their sensitivity to lysis by BZLF1-specific CTL. The data in Fig. 4 show

5 TCR diversity selection in Epstein Barr virus infection 1749 Fig. 4. Semiquantitative PCR detection of BZLF1 in type A LCL. The lanes represent: kb (1 kb ladder); BM, BT, HB and LC (LCL transformed with QIMR-WIL EBV isolate and established from donors BM, BT, HB, and LC). Fig. 3. Specific lysis by RAKFKQLLQ-specific CTL colonies from primary infected donors BT, HB and LC. The target panel consisted of (A) exogenously infected autologous and HLA B8 allogeneic type A LCL, and (B) HLA B8 BM PHA blasts, with and without adsorbed peptide RAKFKQLLQ, and HLA B8, allogeneic type A and type B LCL. The LCL targets were exogenously transformed with (A) type A QIMR-WIL isolate and (B) type A IARC-BL74 (designated PUY), B95.8, or IARC-BL36 (designated LOU) isolates, or type B Ag876 or L4 isolates as described (22). Effectors BT(I)142, HB39, LC5 (A) and BT(I)119 (B) were tested in standard 4 h 51 Cr-release assays. Effector:target ratio: 2:1 (A) and 4:1 (B). that the BM, HB and LC LCL that were prone to lysis, in three separate experiments, consistently expressed 16- to 20-fold more abundant BZLF1 transcripts than the poorly lysed BT LCL. These findings were substantiated by using an anti- BZLF1 mab (37) in an immunofluorescent assay to show differences in BZLF1 protein expression between the different LCL (data not shown). Selection of TCR β diversity in the RAKFKQLLQ-specific response during acute infection TCR repertoire selection was analysed by TCR β chain determination of RAKFKQLLQ-specific CTL from primary infected and long-term healthy HLA B8 virus carriers. The TCR β (V D J C) rearrangements expressed by these colonies were identified using TCR V β family-specific PCR followed by direct sequencing. Only the β chain of the heterodimer was determined since, unlike the α chain, the control of β expression by allelic exclusion usually results in only a single in-frame rearrangement in each T cell clone (38). Analysis of the TCR β sequences from the three donors, HB, LC and BT, undergoing primary EBV infection showed that most colonies expressed a single productively rearranged TCR β transcript confirming their clonality (Fig. 5). Considerable heterogeneity was observed in the T cells responding to RAKFKQLLQ both within and among the three different donors. Although for mixed colonies that expressed more than one productively rearranged TCR β chain, the specific antigen receptor could not be assigned, heterogeneity selection was also evident from the non-recurrence of the transcripts. Of the RAKFKQLLQ-specific TCR, no preferential usage of specific TCRV or TCRJ genes was found; five of seven used unique TCRV gene elements and three of seven used unique TCRJ gene elements. Of the few clones that shared either TCRV [LC50 and BT(I)173] or TCRJ [LC5 and BT(I)142; LC34 and BT(I)173] gene usage, none were found to share both TCRV or TCRJ usage. The complementarity determining region 3 (CDR3), that directly contacts the peptide epitope (39 41), was also unique to each T cell clone having no structural motifs in common with any other β chain. However, a CDR3 length restriction of 9 13 amino acids was observed in all of the clones, indicating structural constraints on the molecular interactions between the TCR and the peptide MHC target [reported CDR3 lengths range of 5 17 amino acids (42)]. Another notable feature of the RAKFKQLLQ memory response in these donors was the lack of detection of recurrent clonotypes, reinforcing the heterogeneity of the responding repertoire. Maintenance of TCR β diversity in the RAKFKQLLQ-specific CTL response following acute infection and in healthy virus carriers Having established TCR diversity selection during primary infection for the RAKFKQLLQ response we next monitored the effects of persistent virus infection on β chain development. RAKFKQLLQ-specific colonies were generated at two addi-

6 1750 TCR diversity selection in Epstein Barr virus infection Fig. 5. TCR β junctional region sequences of RAKFKQLLQ-specific CTL clones and colonies isolated from EBV donors HB, LC and BT undergoing primary infection with EBV. A translated amino acid sequence is shown above each corresponding nucleotide sequence which extended at least 70 nucleotides further 5 of the sequence shown. Designations for TCRBV and TCRBJ gene elements follow that of Arden et al. (68) and Toyonaga et al. (69) respectively. For each clone, the deduced amino acid sequence of the CDR3 loop, defined according to Chothia et al. (70), is shown putatively supported by two framework branches (FW) and the CDR3 amino acid length is reported. TCRBJ germline sequences are underlined and in normal print. Putative TCRBD germline sequences are italicized and underlined. A non-productively rearranged TCR β chain was also detected in clones HB40 and LC50. Colonies HB40, HB39 and BT(I)66 were found to be mixed with the expression of more than one productively rearranged transcript. tional time points after primary infection in donor BT; clones BT(II) were generated 85 days post clinical diagnosis and BT(III) after 196 days. The RAKFKQLLQ specificity of these colonies was confirmed by the rapid T cell T cell assay and 51 Cr-release assays (data not shown). In all colonies except one [BT(III)36] a single in-frame TCR β rearrangement was detected (Fig. 6A). Each time point analysis revealed a unique set of TCR responding to RAKFKQLLQ, with no patterns of TCRV, TCRJ or CDR3 amino acid conservation. The only structurally similar clones were BT(I)74 (Fig. 5) and BT(II)114 (Fig. 6A) that shared BV6 gene usage and the CDR3 amino acid sequence, SLILXG. Interestingly, the CDR3 loop length restriction that was evident during primary infection continued to be selected for in persistent infection, with all the colonies having a CDR3 length of 8 11 amino acids. Altogether, this longitudinal analysis showed no dramatic RAKFKQLLQ clonotype selection, but rather the maintenance of TCR β diversity. TCR β chain sequence analysis was also extended to an HLA B8 type A and type B healthy virus carrier, ISM and MS respectively (Fig. 6B). The RAKFKQLLQ specificity of these colonies was confirmed by the rapid T cell T cell assay and 51 Cr-release assays (data not shown). As was the case in primary infection, non-recurrent rearrangements with a diverse TCR β repertoire usage and a restricted CDR3 length were identified in all of the clones from these donors except in the case of ISM52 and ISM28 which expressed identical β chain transcripts. Overall, comparison of all the RAKFKQLLQ-reactive clones and colonies from the five HLA B8 donors showed that each individual responded with a diverse and unique or private TCR β repertoire that showed no dramatic signs of focusing after long-term virus exposure. These findings are different to our previous TCR studies of EBV-specific CD8 memory clonotypes in which oligoclonal expansions have readily been identified from similar numbers of clones analysed in latent viral epitope responses from donors, some of which are included in this study (BT and ISM) (9,16,18). One possible explanation for the selection of this diverse TCR β response is the presence of sequence variation in potential TCR contact sites within the viral epitope. Variation in the RAKFKQLLQ epitope at non-anchor residues not affecting CTL recognition has been reported (12,43,44). In the present study epitope variation was investigated by sequencing the RAKFKQLLQ region of endogenous EBV isolates from acute donors HB and BT (Table 1). As shown in Table 1, no sequence variation was detected in the HLA B8-restricted BZLF1 epitope of virus isolates from donors HB and BT during the course of primary infection. Thus, it is unlikely that the diversity and lack of

7 TCR diversity selection in Epstein Barr virus infection 1751 Fig. 6. TCR β junctional region sequences of RAKFKQLLQ-specific CTL clones and colonies isolated from donor BT after primary infection (A), and from the long-term virus carriers ISM and MS (B). Clones BT(II) and BT(III) were generated 85 and 196 days after primary infection respectively. A non-productively rearranged TCR β chain was also detected in clones BT(II)61, BT(II)119, ISM52, ISM28 and MS27. TCRBV, TCRBJ and TCRBD gene segments, and CDR3 region loops are presented and assigned as outlined in the legend of Fig. 5. Colonies BT(III)36, ISM58 and MS7 were found to be mixed with the expression of more than one productively rearranged transcript. detection of public β chain rearrangements in the RAKFKQLLQ memory response of HLA B8 individuals is the result of stimulation by different geographical isolates in vivo. Discussion The immediate-early protein, BZLF1, is potentially a vital target antigen for EBV-specific CTL surveillance. Firstly, BZLF1 is a pivotal protein in the viral replication cascade (45), and is expressed early during the lytic cycle in epithelial cells of oral hairy leukoplakia and productively infected malignant or transformed B cells (46,47). Therefore, lytically infected cells could potentially be eliminated by BZLF1-specific CTL before the release of mature virions during the course of a natural infection. This immune strategy is essential to keep a balanced virus load during the life-long persistence of EBV in B cells (48,49). Indeed, the elevated levels of serum antibodies to BZLF1, or the increased frequency of spontaneous LCL outgrowth in culture, in individuals with EBV-associated malignancies or disease (50 52) could be attributable to a dysfunctional CTL control of EBV replication. Secondly, our finding that RAKFKQLLQ-specific CTL are present in individuals undergoing a primary infection with either type A or type B virus shows that the epitope segregates independently of intertypic variation. Instead, segregation is more likely to correlate with the observed geographic distribution of strain variation in BZLF1 (12,44). Since the incidence of type B virus in many populations is probably considerably higher than initially appreciated (1,53), any effective peptide-based antiviral therapy will need to include immunodominant peptide epitopes common to both viral types. The RAKFKQLLQ peptide satisfies these criteria as we have shown that the CTL precursor frequency to this peptide in healthy type A or type B virus carriers is immunodominant, indicating antigendriven selection in the memory T cell compartment (19). CTL targeted to epitopes in latent EBV determinants have recently been detected in the peripheral circulation of individuals affected by IM (9,10). It has been suspected that the active replication of EBV in the lymphoid and epithelial cell compartments in IM also provides a focus for CTL induction against antigens associated with viral replication. Recently, a class II antigen-restricted CTL epitope has been mapped to

8 1752 TCR diversity selection in Epstein Barr virus infection Table 1. Sequences of the RAKFKQLLQ epitope in EBV isolates from HLA B8 virus carriers Virus isolate a Epitope sequence b R A K F K Q L L Q B95.8 cgg gcc aag ttt aag caa ctg ctg cag BT(I) d BT(II) f BT(III) a BT(III) b BT(III) d HB a The nucleotide sequence of the prototype B95.8 strain is shown for comparison. Endogenous EBV isolates were amplified in spontaneous LCL generated from donor BT at different time points during [BT(I)] and following [BT(II) and BT(III)] primary infection and from donor HB. b Dashes indicate identity. the EBV-encoded BHRF1 protein, a 17 kda component of the restricted early antigen complex (13). The present study extends the repertoire of known EBV-specific CTL reactivities in the primary antiviral immune response, to a diverse population of BZLF1-specific CD8 clonotypes. It is possible, at least in vitro, that the CD8 RAKFKQLLQspecific CTL recognize both exogenously and endogenously processed peptides in the context of HLA B8 molecules. Since LCL cultures usually contain 1% of cells in the lytic cycle, the failure of RAKFKQLLQ-specific CTL clones to effectively lyse BT LCL may have occurred because of a limiting amount of BZLF1 protein available for endogenous and exogenous processing. Thus, RAKFKQLLQ-specific CTL may recognize the BZLF1 peptide that has been processed (i) endogenously by LCL in the lytic cycle and (ii) exogenously by co-resident cells in the latent cycle following their uptake of BZLF1 protein from disintegrating, lytically infected cells in culture. Certainly it has been shown that CD8 CTL can recognize exogenously processed peptides presented by class I MHC molecules (54). Moreover, the concentration of processed peptide presented on the surface of LCL is obviously important in determining the lytic outcome (55). Our results demonstrate that TCR β chain usage in the RAKFKQLLQ primary and memory responses is diverse and privately expressed in different HLA B8 individuals. The general lack of detection of recurrent clonotypes in any one individual, even from the same in vitro stimulation, adds support to heterogeneity selection within the responding RAKFKQLLQ-specific repertoire. This is in contrast to the selective clonotype expansions that have previously been defined for two other HLA B8 -restricted latent EBV epitopes, FLRGRAYGL and QAKWRLQTL (9,16,18). One possible factor contributing to repertoire focusing in persistent EBV infection is lymphocyte competition and preferential selection of highavidity clonotypes over repeated antigen exposure. Oligoclonal T cell expansions have also been found in the epitopespecific memory responses of other persisting or reinfecting viruses such as HIV (35), HTLV (56) and influenza (57), strengthening the possibility that with successive antigen encounters the responding antiviral repertoire is selectively amplified. In the case of RAKFKQLLQ, antigen-driven selection was evident from the CDR3 loop length restriction that was maintained in TCR β chain usage throughout both primary and persistent infection. The clonotype diversity observed may be strongly dependent upon the RAKFKQLLQ epitope s structural conformation and on the stimulation of only low-avidity TCR rearrangements. An analogous situation is thought to occur in the private repertoire selection to hen egg lysozyme in mice; the apparently random emergence of private TCR is thought to be linked with their low affinity for antigen (58). Alternatively, the association of BZLF1 with the lytic viral cycle, its localization and distribution in vivo may be important. Compared to the restricted expression of EBNA3 in transformed B cells, BZLF1 is more widely expressed by productively infected B cells, oropharyngeal epithelial cells (59) and bone marrow cells (60). Moreover, a high frequency of BZLF1 transcription has been reported in PBMC of healthy virus carriers (60). Consequently, abundant antigen presentation within the different immunological milieus may allow for noncompetitive re-stimulation of a broad avidity CTL response at high precursor CTL frequencies. Under such conditions, persistent infection, epitope immunogenicity and immunodominance may not influence repertoire diversity selection. The concept of anatomical environment diversity in shaping the phenotype of viral immunity has been put forth by Doherty (61). In addition, the repertoire may be influenced by exposure to other antigens that stimulate cross-reactive T cell responses, in which case a broad low-affinity TCR response may be more susceptible to cross-reactions. Reactivation of memory CD8 CTL by heterologous virus infections and resulting disruption to precursor CTL frequencies has recently been demonstrated (62). Taken overall, the CTL control of EBV appears to involve selection of both restricted and diverse epitope-specific CD8 TCR repertoire responses. Despite the potential pressures of persistent virus challenge, this study shows that polyclonality in responding TCR can be maintained throughout life-long infection. The apparent plasticity of the TCR repertoire in maintaining highly restricted and diverse antiviral clonotypes is likely to be important in the balance of immune control between viral replication, viral persistence and elimination by CTL surveillance. In immunocompromised individuals, an increased level of virus shedding has been correlated with a higher incidence of EBV-infected B cells in the peripheral blood (63). Moreover, an increased virus load is indicated in

9 TCR diversity selection in Epstein Barr virus infection 1753 primary and secondary immunocompromised individuals by the elevated levels of serum antibodies to EBV replication antigens (64,65). Such individuals are at higher risk of developing EBV-associated lymphomas (66). The infusion of autologous EBV-specific CTL into bone marrow transplant recipients has resulted in the regression of EBV-associated immunoblastic lymphomas that usually express the full array of EBV latent proteins (67). Likewise, it may be opportune to adoptively transfer autologous BZLF1-specific CTL into individuals with an excessive EBV load to alleviate disease and reduce the risk of developing EBV-associated malignancies that can evade CTL surveillance. Acknowledgements We thank Scott Burrows and Dr Rajiv Khanna for their discussions and critical review of the manuscript. We thank Chiron Minotopes for the generous gift of ril-2. This work was supported by grants from the National Health and Medical Research Council of Australia. Abbreviations CDR CTL EBNA EBV HLA IM LCL PBMC PHA VCA References complementarity determining region cytotoxic T lymphocyte EBV nuclear antigen Epstein Barr virus human leukocyte antigen infectious mononucleosis lymphoblastoid cell line peripheral blood mononuclear cell phytohaemagglutinin viral capsid antigen 1 Sixbey, J. W., Shirley, P., Chesney, P. J., Buntin, D. M. and Resnick, L Detection of a second widespread strain of Epstein Barr virus. Lancet ii: Henle, G., Lennette, E. T., Alspaugh, M. A. and Henle, W Rheumatoid factor as a cause of positive reactions in tests for Epstein Barr virus-specific IgM antibodies. Clin. Exp. Immunol. 36: Sixbey, J. W., Nedrud, J. G., Raab-Traub, N., Hanes, R. A. and Pagano, J. S Epstein Barr virus replication in oropharyngeal epithelial cells. N. Engl. J. Med. 310: Anagnostopoulos, I., Hummel, M., Kreschel, C. and Stein, H Morphology, immunophenotype, and distribution of latently and/or productively Epstein Barr virus-infected cells in acute infectious mononucleosis: implications for the interindividual infection route of Epstein Barr virus. Blood 85: Robinson, J. E., Brown, N., Andiman, W., Halliday, K., Francke, U., Robert, M. F., Andersson-Anvret, M., Horstmann, D. and Miller, G Diffuse polyclonal B-cell lymphoma during primary infection with Epstein Barr virus. N. Engl. J. Med. 302: Khanna, R., Burrows, S. R. and Moss, D. J Immune regulation in Epstein Barr virus-associated diseases. Microbiol. Rev. 59: Khanna, R., Burrows, S. R., Kurilla, M. G., Jacob, C. A., Misko, I. S., Sculley, T. B., Kieff, E. and Moss, D. J Localization of Epstein Barr virus cytotoxic T-cell epitopes using recombinant vaccinia: implications for vaccine development. J. Exp. Med. 176: Murray, R. J., Kurilla, M. G., Brooks, J. M., Thomas, W. A., Rowe, M., Kieff, E. and Rickinson, A. B Identification of target antigens for the human cytotoxic T-cell response to Epstein Barr virus (EBV): implications for the immune control of EBVpositive malignancies. J. Exp. Med. 176: Silins, S. L., Cross, S. M., Elliott, S. L., Pye, S. J., Burrows, S. R., Burrows, J. M., Moss, D. J., Argaet, V. P. and Misko, I. S Development of Epstein Barr virus-specific memory T cell receptor clonotypes in acute infectious mononucleosis. J. Exp. Med. 184: Steven, N. M., Leese, A. M., Annels, N. E., Lee, S. P. and Rickinson, A. B Epitope focusing in the primary cytotoxic T cell response to Epstein Barr virus and its relationship to T cell memory. J. Exp. Med. 184: Kerr, B. M., Kienzle, N., Burrows, J. M., Cross, S., Silins, S. L., Buck, M., Benson, E. M., Coupar, B., Moss, D. J. and Sculley, T. B Identification of type B-specific and cross-reactive cytotoxic T-lymphocyte responses to Epstein Barr virus. J. Virol. 70: Bogedain, C., Wolf, H., Modrow, S., Stuber, G. and Jilg, W Specific cytotoxic T lymphocytes recognize the immediate-early transactivator Zta of Epstein Barr virus. J. Virol. 69: White, C., Cross, S., Kerr, B., Kurilla, M., Schmidt, C., Misko, I., Khanna, R. and Moss, D Recruitment of CD3 CD4 CD8 T cells during infectious mononucleosis which recognize EBV early antigen BHRF1. Virology 219: Hedrick, S. M. and Eidelman, F. J T lymphocyte antigen receptor. In Paul, W. E., ed., Fundamental Immunology, p Raven Press, New York. 15 Bowness, P., Moss, P. A. H., Rowland-Jones, S., Bell, J. I. and McMichael, A. J Conservation of TCR usage by HLA B27- restricted influenza-specific cytotoxic T lymphocytes suggests a general pattern for antigen-specific MHC class I-restricted responses. Eur. J. Immunol. 23: Silins, S. L., Burrows, S. R., Moss, D. J., Khanna, R., Misko, I. S. and Argaet, V. P T cell receptor repertoire for a viral epitope in humans is diversified by tolerance to a background major histocompatibility complex antigen. J. Exp. Med. 182:1. 17 McHeyzer Williams, M. G. and Davis, M. M Antigenspecific development of primary and memory T cells in vivo. Science 268: Argaet, V. P., Schmidt, C. W., Burrows, S. R., Silins, S. L., Kurilla, M. G., Doolan, D. L., Suhrbier, A., Moss, D. J., Kieff, E., Sculley, T. B. and Misko, I. S Dominant selection of an invariant T- cell antigen receptor in response to persistent infection by Epstein Barr virus. J. Exp. Med. 180: Elliott, S. L., Pye, S. J., Schmidt, C. W., Cross, S. M., Silins, S. L. and Misko, I. S Dominant cytotoxic T lymphocyte response to the immediate-early transactivator protein, BZLF1, in persistent type A or B Epstein Barr virus infection. J. Infect. Dis., in press. 20 Henle, G., Henle, W. and Horwitz, C. A Antibodies to Epstein Barr virus-associated nuclear antigen in infectious mononucleosis. J. Infect. Dis. 130: Burrows, S. R., Misko, I. S., Sculley, T. B., Schmidt, C. and Moss, D. J An Epstein Barr virus-specific cytotoxic T-cell epitope present on A- and B-type transformants. J. Virol. 64: Moss, D. J., Misko, I. S., Burrows, S. R., Burman, K., McCarthy, R. and Sculley, T. B Cytotoxic T-cell clones discriminate between A- and B-type Epstein Barr virus transformants. Nature 331: Rickinson, A. B., Finerty, S. and Epstein, M. A Comparative studies on adult donor lymphocytes infected by EB virus in vivo or in vitro: origin of transformed cells arising in co-cultures with foetal lymphocytes. Int. J. Cancer 19: Burrows, S. R., Sculley, T. B., Misko, I. S., Schmidt, C. and Moss, D. J An Epstein Barr virus-specific cytotoxic T-cell epitope in EBV nuclear antigen 3 (EBNA 3). J. Exp. Med. 171: Wang, A., Lu, S. D. and Mark, D. F Site-specific mutagenesis of the human interleukin-2 gene: structure-function analysis of the cysteine residues. Science 224: Rosenberg, S. A., Grimm, E. A., McGrogan, M., Doyle, M., Kawasaki, E., Koths, K. and Mark, D. F Biological activity of recombinant human interleukin-2 produced in Escherichia coli. Science 223: Burrows, J. M., Khanna, R., Sculley, T. B., Alpers, M. P., Moss, D. J. and Burrows, S. R Identification of a naturally occurring recombinant Epstein Barr virus isolate from New Guinea that encodes both type 1 and type 2 nuclear antigen sequences. J. Virol. 70:4829.

10 1754 TCR diversity selection in Epstein Barr virus infection 28 Bishop, C. J., Moss, D. J., Ryan, J. M. and Burrows, S. R T lymphocytes in infectious mononucleosis. II. Response in vitro to interleukin-2 and establishment of T-cell lines. Clin. Exp. Immunol. 60: Misko, I. S., Pope, J. H., Hutter, R., Soszynski, T. D. and Kane, R. G HLA-DR-antigen-associated restriction of EBV-specific cytotoxic T-cell colonies. Int. J. Cancer 33: Burrows, S. R., Suhrbier, A., Khanna, R. and Moss, D. J Rapid visual assay of cytotoxic T-cell specificity utilizing synthetic peptide induced T-cell T-cell killing. Immunology 76: Burrows, S. R., Gardner, J., Khanna, R., Steward, T., Moss, D. J., Rodda, S. and Suhrbier, A Five new cytotoxic T cell epitopes identified within Epstein Barr virus nuclear antigen 3. J. Gen. Virol. 75: Tierney, R. J., Steven, N., Young, L. S. and Rickinson, A. B Epstein Barr virus latency in blood mononuclear cells: analysis of viral gene transcription during primary infection and in the carrier. J. Virol. 68: Mathes, T., Werner-Favre, C., Tang, H., Zhang, X., Kindler, V. and Zubler, R. H Cytokine mrna expression during an in vitro response of human B lymphocytes: kinetics of B cell tumor necrosis factor α, interleukin (IL)6, IL10, and transforming growth factor β1 mrnas. J. Exp. Med. 178: Panzara, M. A., Oksenberg, J. R. and Steinman, L The polymerase chain reaction for detection of T-cell antigen receptor expression. Curr. Opin. Immunol. 4: Kalams, S. A., Johnson, R. P., Trocha, A. K., Dynan, M. J., Ngo, H. S., D Aquila, R. T., Kurnick, J. T. and Walker, B. D Longitudinal analysis of T cell receptor (TCR) gene usage by human immunodeficiency virus 1 envelope-specific cytotoxic T lymphocyte clones reveals a limited TCR repertoire. J. Exp. Med. 179: Tan, S. S. and Weis, J. H Development of a sensitive reverse transcriptase PCR assay, RT-RPCR, utilizing rapid cycle times. PCR Methods Appl. 2: Young, L. S., Lau, R., Rowe, M., Niedobitek, G., Packham, G., Shanahan, F., Rowe, D. T., Greenspan, D., Greenspan, J. S., Rickinson, A. B. and Farrell, P. J Differentiation-associated expression of the Epstein Barr virus BZLF1 transactivator protein in oral hairy leukoplakia. J. Virol. 65: Malissen, M., Trucy, J., Letourneur, F., Rebai, N., Dunn, D. E., Fitch, F. W., Hood, L. and Malissen, B A T-cell clone expresses two T-cell receptor alpha genes but uses one alpha beta heterodimer for allorecognition and self MHC-restricted antigen recognition. Cell 55: Jorgensen, J. L., Esser, U., Fazekas de St Groth, B., Reay, P. A. and Davis, M. M Mapping TCR peptide contacts by variant peptide immunization of single-chain transgenics. Nature 355: Garcia, K. C., Degano, M., Stanfield, R. L., Brunmark, A., Jackson, M. R., Peterson, P. A., Teyton, L. and Wilson, I. A An alpha beta T cell receptor structure at 2. 5 angstrom and its orientation in the TCR MHC complex. Science 274: Garboczi, D. N., Ghosh, P., Utz, U., Fan, Q. R., Biddison, W. E. and Wiley, D. C Structure of the complex between human T-cell receptor, viral peptide and HLA-A2. Nature 384: Roldan, E. Q., Sottini, A., Bettinardi, A., Albertini, A., Imberti, L. and Primi, D Different TCRBV genes generate biased patterns of V D J diversity in human T-cells. Immunogenetics 41: Chang, Y., Chen, M. I., Wu, R., Chang, P., Shu, C. and Liu, S Sequence analysis of the BZLF 1 gene of Epstein Barr virus derived from a new variant isolated in Taiwan. Nucleic Acids Res. 20: Packham, G., Brimmell, M., Cook, D., Sinclair, A. J. and Farrell, P. J Strain variation in Epstein Barr virus immediate-early genes. Virology 192: Countryman, J. and Miller, G Activation of expression of latent Epstein Barr herpesvirus after gene transfer with a small cloned subfragment of heterogeneous viral DNA. Proc. Natl Acad. Sci. USA 82: Becker, J., Leser, U., Marschall, M., Langford, A., Jilg, W., Gelderblom, H., Reichart, P. and Wolf, H Expression of proteins encoded by Epstein Barr virus trans-activator genes depends on the differentiation of epithelial cells in oral hairy leukoplakia. Proc. Natl Acad. Sci. USA 88: Bogedain, C., Alliger, P., Schwarzmann, F., Marschall, M., Wolf, H. and Jilg, W Different activation of Epstein Barr virus immediate-early and early genes in Burkitt lymphoma cells and lymphoblastoid cell lines. J. Virol. 68: Gratama, J. W., Oosterveer, M. A., Zwaan, F. E., Lepoutre, J., Klein, G. and Ernberg, I Eradication of Epstein Barr virus by allogeneic bone marrow transplantation: implications for sites of viral latency. Proc. Natl Acad. Sci. USA 85: Yao, Q. Y., Rickinson, A. B. and Epstein, M. A A reexamination of the Epstein Barr virus carrier state in healthy seropositive individuals. Int. J. Cancer 35: Joab, I., Triki, H., De Saint Martin, J., Perricaudet, M. and Nicolas, J. C Detection of anti-epstein Barr virus trans-activator (ZEBRA) antibodies in sera from patients with human immunodeficiency virus. J. Infect. Dis. 163: Joab, I., Nicolas, J. C. Schwaab, G. de-thé, G. Clausse, B. Perricaudet, M. and Zeng, Y Detection of anti-epstein Barr-virus transactivator (ZEBRA) antibodies in sera from patients with nasopharyngeal carcinoma. Int. J. Cancer 48: Tedeschi, R., Foong, Y. T., Cheng, H. M., Depaoli, P., Elfborg, T. and Dillner, J The disease associations of the antibody response against the Epstein Barr virus transactivator protein ZEBRA can be separated into different epitopes. J. Gen. Virol. 76: Apolloni, A. and Sculley, T. B Detection of A-type and B- type Epstein Barr virus in throat washings and lymphocytes. Virology 202: Bachmann, M. F., Oxenius, A., Pircher, H., Hengartner, H., Ashton- Richardt, P. A., Tonegawa, S. and Zinkernagel, R. M TAP1- independent loading of class I molecules by exogenous viral proteins. Eur. J. Immunol. 25: Levitsky, V., Zhang, Q., Levitskaya, J. and Masucci, M. G The life span of major histocompatibility complex peptide complexes influences the efficiency of presentation and immunogenicity of two class I-restricted cytotoxic T lymphocyte epitopes in the Epstein Barr virus nuclear antigen 4. J. Exp. Med. 183: Utz, U., Banks, D., Jacobson, S. and Biddison, W. E Analysis of the T-cell receptor repertoire of human T-cell leukemia virus type 1 (HTLV-1) Tax-specific CD8 cytotoxic T lymphocytes from patients with HTLV-1-associated disease: evidence for oligoclonal expansion. J. Virol. 70: Lehner, P. J., Wang, E. C. Y., Moss, P. A. H., Williams, S., Platt, K., Friedman, S. M., Bell, J. I. and Borysiewicz, L. K Human HLA-A0201-restricted cytotoxic T lymphocyte recognition of influenza A is dominated by T cells bearing the V β 17 gene segment. J. Exp. Med. 181: Cibotti, R., Cabaniols, J. P., Pannetier, C., Delarbre, C., Vergnon, I., Kanellopoulos, J. M. and Kourilsky, P Public and private V beta T cell receptor repertoires against hen egg white lysozyme (HEL) in nontransgenic versus HEL transgenic mice. J. Exp. Med. 180: Rickinson, A. B. and Kieff, E Epstein Barr Virus. In Fields, B. N., Knipe, D. M. and Howley, P. M., eds, Fields Virology, p Lippincott-Raven, Philadelphia, PA. 60 Gonnella, R., Angeloni, A., Calogero, A., Farina, A., Santarelli, R., Gentile, G., Arcese, W., Martino, P., Mandelli, F., Frati, L., Faggioni, A. and Ragona, G Transcription of latent and replicative Epstein Barr virus genes in bone-marrow and peripheral-blood mononuclear cells of healthy donors. Int. J. Cancer 70: Doherty, P. C Anatomical environment as a determinant in viral immunity. J. Immunol. 155: Selin, L. K., Vergilis, K., Welsh, R. M. and Nahill, S. R Reduction of otherwise remarkably stable virus-specific cytotoxic T lymphocyte memory by heterologous viral infections. J. Exp. Med. 183: Yao, Q. Y., Rickinson, A. B., Gaston, J. S. and Epstein, M. A In vitro analysis of the Epstein Barr virus: host balance in long-term renal allograft recipients. Int. J. Cancer 35:43.