of CTLs generated ex vivo would be expected to exert a clinical effect superior to that of cancer vaccines. Based on this concept, various types of T-cell adoptive transfer have been performed and various degrees of clinical responses have been obtained. Donor lymphocyte infusion (DLI) has been widely performed after allogeneic hematopoietic stem cell transplantation for patients with therapy-resistant malignancies, especially chronic myelogenous leukemia (CML). The major concern regarding DLI is that its antitumor effect is strongly dependent on graft-versus-host disease (GVHD) [3]. To control severe GVHD induced by DLI, clinical trials of infusion of donor lympho cytes transduced with a suicide gene, such as herpes simplex thymidine kinase, have been performed [4]. Furthermore, a more defined type of adoptive T-cell transfer, involving infusion of virus-specific T cells, has been investigated. Infusion of Epstein–Barr virus (EBV)-reactive T cells, generated by stimulating peripheral blood lymphocytes with autologous EBV-transformed B cells, has been applied for the treatment of EBV-associated post-transplantation lymphoproliferative disease (PTLD) and EBVassociated malignancies, including Hodgkin’s lymphoma and naso pharyngeal carcinoma [5,6]. Previous studies have revealed that adoptive transfer of EBV-specific T cells as monotherapy is effective in inducing remission of EBV-associated PTLD with low-risk features. For high-risk EBVassociated diseases, a combination of adoptive transfer of EBV-specific T cells with conventional chemotherapies has yielded superior outcomes [7]. In addition, infusion of CTLs specific for antigens, uniquely or preferentially expressed by cancer cells, has received marked attention for its antitumor effectiveness with minimal adverse events including GVHD. More than 10 years ago, Falkenburg et al. reported that infusion Adoptive T-cell immunotherapy It is well known that the immune surveillance system plays an important role in resistance to tumor cell growth. Therefore, if the antitumor immune response can be augmented in vivo, improved survival of cancer patients would be expected. The immune surveillance system is harmonized with interactions between various kinds of immune cells, and among them, cytotoxic T lymphocytes (CTLs) play the most important role in antitumor effect. On the basis of this concept, various types of cell-mediated immunotherapy aimed at induction of tumorspecific CTLs have been developed. These include vaccination with peptides derived from tumor-associated antigens, or with dendritic cells pulsed with peptide or tumor cell lysate, adoptive transfer of tumor-specific CTLs, and gene-engineered T-cell immunotherapy. Among these approaches, various trials of cancer peptide vaccines have been ongoing in many countries [1]. The results obtained from these clinical trials have revealed no severe adverse events and peptide-specific immune responses have been detected in the majority of the vaccinated patients. However, the clinical responses have failed to meet expectations. A variety of immunosuppressive mechanisms, particularly those mediated by regulatory T cells and myeloid-derived suppressor cells, operate in cancer patients. This suppressive influence may play a role in limiting the antitumor effectiveness of cancer vaccines [2].