Comment on Straathof et al, page 1898

Nasopharyngeal carcinoma (NPC) is the most frequent Epstein-Barr virus (EBV)–associated malignancy worldwide. In this issue, Straathof and colleagues take aim at EBV-associated NPC showing that adoptive immunotherapy with EBV-specific T cells expanded in vitro can be successfully administered to NPC patients and is associated with potential therapeutic efficacy.

Epstein-Barr virus (EBV) immunotherapy has proved its mettle in stem cell transplantation where donor T cells are stimulated with autologous EBV-immortalized B cells in vitro and then transferred into the patient to prevent or treat uncontrolled proliferation of EBV-infected B cells, or so-called posttransplantation lymphoproliferative syndrome (PTLD). In these cases, EBV gene products expressed in the malignant B cell are ideal tumor-associated antigens marking tumor cells for killing by EBV-specific cytotoxic T cells (CTLs). It is not known which cell populations in the polyclonal T-cell mixture are responsible for therapeutic efficacy, but CD8+ CTLs specific for the latent infection viral antigens expressed in immortalized B cells are presumed to be important.

Successfully applying this therapeutic approach to EBV-associated NPC is not without potential issues. The most obvious problem is a significant mismatch between the repertoire of CD8+ CTLs and the pattern of EBV gene expression in NPC. EBV-immortalized B cells express 9 latent infection viral proteins—6 nuclearproteins(EBVnuclearantigens;EBNAs) and 3 latent membrane proteins (LMPs)—and CTLs stimulated in vitro by EBV-infected B cells most frequently recognize EBNA-3A, -3B, or -3C. However, NPC cells usually express only EBNA-1, LMP1, and LMP2, so these tumor cells fail to express the EBNA-3 targets recognized by most of the therapeutic CTLs. LMP1 is expressed in only 65% of the tumor cells and is not a very immunogenic protein, and LMP1-specific CTLs are uncommon. LMP2A protein is detected in only 50% of tumor cells, but LMP2A-specific T cells are a subdominant population following the EBNA-3s. EBNA-1 is important for maintenance of the viral episome, protein can be detected in every tumor cell, and it would be an ideal tumor marker. However, the EBNA-1 glycine-alanine repeat region has a potential immune evasion mechanism that prevents proteosome-dependent degradation and efficient presentation of EBNA-1 peptides through the HLA class I pathway, so that EBNA-1–specific CTLs fail to efficiently kill EBV-infected B cells in tissue culture.

Effective therapy also requires that the therapeutic T cells migrate to the tumor. The mucosal location and infiltrating T-cell milieu typically associated with NPC may provide additional obstacles for transferred effector T cells. NPC patients also present different logistic issues versus transplant patients. T cells from transplant donors can be prepared ahead of time, but this head start is not available for NPC patients.

The results reported by Straathof and colleagues show that these logistic problems can be overcome, and EBV-specific T-cell populations can be successfully generated for treatment of NPC patients. Cells were safely administered, and there was a trend for improved outcomes among the treated patients. These results are extremely encouraging since the protocol involved no significant alteration from the transplant patients. As expected, EBNA-3s were the dominant targets of the transferred T cells so that the bulk of the T cells may not have been expected to target the NPC tumor cells. The LMP2A-specific T cells represented less than 10% of the transferred T cells, so that specific manipulations to increase LMP2A-specific T cells may increase efficacy. The number of EBV-specific or LMP2A-specific T cells in the peripheral blood of treated patients did not significantly change, suggesting that a “full” lymphocyte pool may inhibit the in vivo amplification of the transferred EBV-specific T cells. The authors suggest that gentle immunodepletion prior to adoptive transfer may result in longer duration and efficacy of the EBV-specific T cells. An intriguing correlation was the best clinical response in a patient with a strong EBNA-1–specific response in the T-cell preparation. Recent studies indicate that EBNA-1–specific CTLs may have more effector activity against EBV targets in tissue culture than previously believed, so it may be premature to rule out EBNA-1 as a potential target for immunotherapy.

The demonstration that EBV-specific adoptive immunotherapy established for PTLD patients can be successfully applied to a significant number of NPC patients is a logical and important first step. The encouraging clinical results despite a number of theoretic obstacles indicate that it is time to “T” off on NPC. ▪