Cytotoxic T-lymphocyte (CTL) responses to Epstein-Barr virus (EBV) tend to focus on a few immunodominant viral epitopes; where these epitope sequences are polymorphic between EBV strains, host CTL specificities should reflect the identity of the resident strain. In studying responses in HLA-B27-positive virus carriers, we identified 2 of 15 individuals who had strong CTL memory to the pan-B27 epitope RRIYDLIEL (RRIY) from nuclear antigen EBNA3C but whose endogenous EBV strain, isolated in vitro, encoded a variant sequence RKIYDLIEL (RKIY) which did not form stable complexes with B27 molecules and which was poorly recognized by RRIY-specific CTLs. To check if such individuals were also carrying an epitope-positive strain (either related to or distinct from the in vitro isolate), we screened DNA from freshly isolated peripheral blood mononuclear cells for amplifiable virus sequences across the EBNA3C epitope, across a different region of EBNA3C with type 1-type 2 sequence divergence, and across a polymorphic region of EBNA1. This showed that one of the unexplained RRIY responders carried two distinct type 1 strains, one with an RKIY and one with an RRIY epitope sequence. The other responder carried an RKIY-positive type 1 strain and a type 2 virus whose epitope sequence of RRIFDLIEL was antigenically cross-reactive with RRIY. Of 15 EBV-seropositive donors analyzed by such assays, 12 appeared to be carrying a single virus strain, one was coinfected with distinct type 1 strains, and two were carrying both type 1 and type 2 viruses. This implies that a small but significant percentage of healthy virus carriers harbor multiple, perhaps sequentially acquired, EBV strains.

The Epstein-Barr virus (EBV) nuclear antigen EBNA1 is the only viral protein detectably expressed in virus genome-positive Burkitt’s lymphoma (BL); recent work has suggested that viral strains with particular EBNA1 sequence changes are preferentially associated with this tumor and that, within a patient, the tumor-associated variant may have arisen de novo as a rare mutant of the dominant preexisting EBV strain (K. Bhatia, A. Raj, M. J. Gutierrez, J. G. Judde, G. Spangler, H. Venkatesh, and I. T. Magrath, Oncogene 13:177–181, 1996). In the present work we first study 12 BL patients and show that the virus strain in the tumor is identical in EBNA1 sequence and that it is matched at several other polymorphic loci to the dominant strain rescued in vitro from the patient’s normal circulating B cells. We then analyze BL-associated virus strains from three different geographic areas (East Africa, Europe, and New Guinea) alongside virus isolates from geographically matched control donors by using sequence changes in two separate regions of the EBNA1 gene (N-terminal codons 1 to 60 and C-terminal codons 460 to 510) to identify the EBNA1 subtype of each virus. Different geographic areas displayed different spectra of EBNA1 subtypes, with only limited overlap between them; even type 2 virus strains, which tended to be more homogeneous than their type 1 counterparts, showed geographic differences at the EBNA1 locus. Most importantly, within any one area the EBNA1 subtypes associated with BL were also found to be prevalent in the general population. We therefore find no evidence that Burkitt lymphomagenesis involves a selection for EBV strains with particular EBNA1 sequence changes.

In apparent contrast to earlier work on Epstein-Barr virus (EBV) carriage in the general Caucasian population, in vitro virus isolations from human immunodeficiency virus (HIV)-positive male homosexual cohorts have shown frequent examples of multiple EBV infection and an overall prevalence of type 2 EBV strains exceeding 30%. Here we ask to what extent these findings might hold true in another T-cell-immunocompromised cohort, HIV-positive hemophilic patients. Resident EBV strains were rescued within lymphoblastoid cell lines derived from the blood and throat washings of 39 such individuals, using the same in vitro protocols of virus isolation as for the homosexual cohort. A mean of 19 independent cell lines was made per patient, and in each case the resident virus was characterized by PCR-based viral genomic analysis and by immunoblotting to reveal the viral “EBNAprint.” By these criteria a significant proportion (14 of 39) of the hemophilic cohort carried more than one EBV strain, suggesting that T-cell impairment does indeed sensitize virus carriers to reinfection with new strains of exogenously transmitted virus. However, the overall incidence of type 2 EBV infection was 10%, which is close to that observed in the earlier work with healthy carriers and substantially lower than that seen in HIV-positive homosexuals. We infer that type 2 EBV is relatively rare in the general Caucasian population but has become endemic in the homosexual community.