REVIEWS Epstein-Barr girus: adaptation to a life within the immune system Maria G. Masucci and lngemar Ernberg A fter primary infection of epithelium, all human- pathogenic herpes- viruses establish lifelong latency in neural or mesen- chymal cells. Intermittent reactivation, with productive replication at endothelial or epithelial surfaces, ensures spread to new hosts. This dual strategy has proved very suc- cessful, resulting in the wide dissemination of herpesviruses in all human populations with prevalences ranging from 35% to more than 90% (Ref. 1). One of them, Epstein-Barr virus (EBV), is probably the most common virus in humans. Epstein-Barr virus has developed multiple strategies to ensure its long-term persistence in the infected B cells of immunocompetent hosts. These include the establishment of cell-phenotype- specific programs of viral gene expression and the transduction of cellular genes that modulate immune responses. Cytotoxic T cells may specifically influence the evolution of this genetically stable virus. B blasts are immunogenic and elicit strong humoral and cellu- lar responses, and antibodies specific for viral antigens associated with growth trans- formation persist through life. Moreover, LCLs are potent stimulators of proliferative and cytotoxic reactions in vitro. Why are the infected B cells not eliminated by the immune response of the host? Recent observations have shed some light on these paradoxes. EBV seems to have solved the dilemma by learning how to live within the immune system itself. M.G. Masucci and I. Ernberg are in the Microbiology and Tumor Biology Center, Karolinska Institute, 171 77 Stockholm, Sweden. In contrast to other herpesviruses, EBV is trans- mitted to the vast majority of individuals without apparent disease. Only when primary infection is de- layed until adolescence or adulthood may it cause a benign lymphoproliferative disease known as infectious mononucleosis2. Yet EBV is the only herpesvirus that is consistently associated with human malignancies. The immunoblastic lymphomas of the immuno- suppressed3, endemic Burkitt’s lymphoma4 and un- differentiated nasopharyngeal carcinoma5 have the strongest association with this virus. EBV has also been demonstrated recently in some 40% of Hodgkin’s lymphomas, anaplastic large-cell lymphomas of T cell origin and peripheral T cell lymphomas6,‘. The non- pathogenic relationship between this ubiquitous and potentially oncogenic virus and our species presents an intriguing case of multiple strategies favoring the survival of virus-infected cells in immunocompetent hosts. The paradox of EBV persistence Although it is a member of the normal human microbial flora, EBV is the most potent growth-transforming agent known. Between 10 and 50% of B cells from any donor can be immortalized as lymphoblastoid cell lines (LCLs) after exposure to laboratory or wild-type virus isolates*. Proliferating EBV-carrying immuno- blasts are easily detected in the blood and lymphoid tissues of infectious mononucleosis patients9. Further- more, EBV-carrying LCLs can be established, with- out the need for the presence of exogenous virus, by in vitro cultivation of purified B cells from previously infected individualslo. How can the healthy host cope with these virus-infected cells? The immortalized The puzzling problem of, persistence in B cells In all herpesviruses, the part of the genome that organizes virus replication is very similar. In contrast, the genes controlling latency diverge in the neurotropic a-herpesviruses, the P-herpesviruses and the lympho- tropic y-herpesviruses . I1 This reflects the development of genetic programs that are adapted to the type of host cells where latency, and hence reactivation, takes place. The site of EBV latency has been a matter of con- troversy. Recent observations support the notion that infected B cells serve this purpose. Although multiple EBV variants are generated during the replicative cycle, the same isolate is detected in the saliva of an infected individual over a period of several years12. Thus, endogenous virus must be fed to the replicative com- partment from a latent reservoir that is not subject to the hazards of extensive virus replication. Prolonged treatment with acyclovir, which efficiently blocks the productive cycle in the oropharyngeal epithelium, does not significantly affect the number of virus-carrying B cells in the blood13. Maintenance of the infected B cell pool is therefore independent of continuous reinfection. More convincingly, the depletion of the recipient’s hematopoietic compartment that precedes bone mar- row transplantation can result in elimination of the resident virus. Successful engrafting with an EBV- positive marrow is accompanied by transfer of the virus of the donor to the transplant recipient14. The choice of B cells as the site of latency seems rather witless given the rapid cellular turnover of the lymphoid compartment. Studies in mouse and rat models suggest that as much as 1% of the total pool of B cells carrying surface immunoglobulin is renewed