Epstein–Barr virus replicating in epithelial cells

Epstein–Barr virus (EBV), a ubiquitous and potentially oncogenic human herpesvirus (1), was discovered 50 y ago this year in a lymphoid cell line grown from a Burkitt lymphoma (2). Since that time, virus latency and replication have been intensively studied in B cells, which are easily infected and maintained in vitro. Nine years after its discovery in B cells, it was shown that EBV is also harbored in the epithelial cells of nasopharyngeal carcinoma (3). Study of the interaction of the virus with this cell type has, however, been significantly hampered by difficulties in establishing reproducible and robust infection in vitro. In PNAS, Temple et al. (4) report efficient replication of EBV in stratified epithelium. The authors generated organotypic or “raft” cultures from primary gingival or tonsil epithelial cells and successfully infected them from the apical surface either by overlaying cultures with EBV-producing B-cell lines or by addition of cell-free virus. The replicating virus spread throughout the suprabasal epithelium, expressing latency proteins along with those of the lytic cycle. No cells were found expressing latency proteins alone, and no virus was seen in the basal epithelium.

Models of EBV Latency in B Cells and Replication in Epithelial Cells

Current models of EBV infection at the organismal level propose that virus is transmitted orally in saliva, replicates in or is transcytosed across epithelial cells (5), and accesses lymphocytes in Waldeyer’s ring, the lymphoid tissue that surrounds the oropharynx (6). Infection of B cells drives proliferation as the result of expression of at least nine virus latency genes and additional noncoding RNAs, but infected cells ultimately enter or return to the resting long-lived memory compartment where no virus proteins are made (7). Terminal differentiation of an infected memory cell triggers reactivation of virus into lytic replication (8). Virus is then amplified in epithelial cells for shedding in saliva or for infection of more B cells to replenish the latent reservoir, establishing a cycle of persistence that endures for the life of a healthy carrier.

Many of the details of B-cell infection; latency, which in vitro leads to B-cell transformation; and reactivation have been amenable to study in freshly isolated B cells or EBV-infected B-cell tumors. In vivo, B cells infected with EBV have been studied in the tonsil and in the blood. Epithelial cell infec

Temple et al. report efficient replication of EBV in stratified epithelium.

tion has, however, been much more elusive. The only disease associated with primary infection by EBV is infectious mononucleosis, and it has been estimated that the incubation period is somewhere between 30 and 50 d (9). It has therefore proven impossible to determine if replication in epithelial cells truly precedes infection of B lymphocytes. Amplification of virus in epithelial cells before shedding in saliva has at least some evidential support as virus shed almost daily in the saliva of carriers has the glycoprotein composition of virus made in an epithelial cell rather than a B cell (10). However, again it has been difficult to visualize in vivo. Seeking foci of replication in what amounts to a vast area of oral epithelium presents many challenges. Visible lesions driven by EBV replication are readily apparent only in oral hairy leukoplakia (OHL), a disorder seen almost exclusively in the late stages of AIDS (11).

The difficulty of studying virus replication in epithelial cells in vivo has been compounded by the difficulties of recapitulating it in vitro. Epithelial cells that express complement receptor type 1, CD21, which the virus can use as an attachment receptor, can be infected quite readily (12), but many epithelial cells in vivo lack its expression (13). The majority of infections achieved in vitro have, in addition, resulted in latency and not productive replication. There has been evidence of virus spread within monolayers of polarized epithelium (14), but no reports of production of high levels of infectious virions such as that now provided by Temple et al.

Demonstration of Direct Entry of EBV into Lytic Replication in an Epithelial Cell

It has long been suspected that part of the reason for the failure to develop models of EBV replication in epithelial cells relates to a failure to establish cultures of normal cells in an appropriate state of differentiation. Work done when OHL lesions were more readily available indicated that replication in these lesions was restricted to the upper, more differentiated layers (15, 16). More recently, virus expression has been shown to be limited to cells expressing B-lymphocyte-induced maturation protein 1 (Blimp1), which is up-regulated during epithelial differentiation, and the promoter of one of the immediate early genes of EBV is responsive to Blimp1 (17). The work of Temple et al. now provides further support for a model in which replication is linked to differentiation. It also raises anew the question of whether EBV is ever latent in a normal epithelial cell or only establishes latency in epithelial cells that have undergone malignant or premalignant changes. The authors found no cells expressing latency proteins alone and no infected basal cells, although they raise the possibility that this may be an artifact of the system. A second possibility is that infection of basal cells occurs more readily from reactivating B cells delivering virus to the basal surface than from epithelial cells replicating virus in the suprabasal layers. This might then support a model of epithelial persistence in which EBV behaves more like a papillomavirus, maintaining latency until the environment in the cell changes (18). The argument against this model has been the failure to detect latent infection in the basal layers of OHL (15), but OHL, which typically occurs on the sides of the tongue, may represent another situation where virus accesses the epithelium from the apical rather than the basolateral surface. A reexamination of the issue is also probably overdue in light of the current availability of more sensitive techniques.

It has never before been possible to study the direct entry of EBV into a lytic cycle of replication. All that we understand about lytic replication of the virus comes from study of cells reactivating from latency. The current achievements of Temple et al. thus open up an entirely new window of opportunity to learn more about a virus that infects almost all of us and, in the unfortunate few, causes serious disease. More than 200,000 new cases of EBV-associated tumors are reported annually, and a vaccine is needed (19). Those seeking to determine what virus proteins should be targeted to prevent initial replication in a human now have a new set of tools.