Although it is accepted that regulatory T cells (T regs) contribute to cancer progression, most studies in the field consider nonantigenspecific suppression. Here, we show the presence of tumor antigenspecific CD4 ؉ T regs in the blood of patients with metastatic melanoma. These CD4 ؉ T regs recognize a broad range of tumor antigens, including gp100 and TRP1 (melanoma tissue differentiation antigens), NY-ESO-1 (cancer/testis antigen) and survivin (inhibitor of apoptosis protein (IAP) family antigen). These tumor antigen-specific T regs proliferate in peripheral blood mononuclear cells (PBMC) cultures in response to specific 15-mer peptides, produce preferentially IL-10 and express high levels of FoxP3. They suppress autologous CD4 ؉ CD25 ؊ T cell responses in a cell contactdependent manner and thus share properties of both naturally occurring regulatory T cells and type 1 regulatory T cells. Such tumor antigen-specific T regs were not detected in healthy individuals. These tumor antigen-specific T regs might thus represent another target for immunotherapy of metastatic melanoma.M olecular identification of human cancer antigens in the last decade, for example MART-1/MelanA or NY-ESO-1, has ushered in a new era of antigen-specific cancer immunotherapy targeting these antigens. Despite promising data in animal models, the use of therapeutic cancer vaccines in humans has not yet lived up to its promises. The improved understandings of antigen presentation, and most specifically in dendritic cell (DC) biology, represent a new avenue for research in the field. Ex vivo-generated and antigen-loaded DCs have been used as vaccines to improve immunity in patients with cancer (1). We have vaccinated patients with metastatic melanoma with ex vivo-generated dendritic cells loaded with either HLA-A*0201-restricted peptides derived from melanoma-associated antigens (2) or killed allogeneic melanoma cells (3). Despite some remarkable clinical responses, the overall objective tumor regression rate remains low (1). Because insufficient antitumor immunity might be due to immunosuppression driven by regulatory T cells (T regs), we engaged in identifying such cells in our patient population.T regs are specialized in the control of responsiveness to self. They are composed of subsets with distinct ontogeny and functions. Naturally occurring CD4 ϩ CD25 high T regs are produced in the thymus (4), and express FoxP3, a transcriptional factor critical for their development and function (5, 6). Their depletion results in the development of autoimmune diseases in murine models (7,8). Conversely, their adoptive transfer inhibits the occurrence of autoimmune diseases in disease-prone mice (9). CD4 ϩ CD25 high T regs suppress the immune responses in a cell contact-dependent manner through mechanisms that remain to be identified (10). T regs are also generated in the periphery from nonregulatory T cells (11)(12)(13)(14)(15)(16). These include regulatory type 1 (Tr1) (12) and Th3 (11) cells, both of which preferentially secrete regulatory cytoki...