Monoclonal antibodies are essential to the success of molecularly targeted therapies. Recently, numerous therapeutic antibodies have been developed for various diseases, including cancer and autoimmune diseases. Experimental systems to effectively evaluate these candidate antibodies are urgently needed. One of the mechanisms used by antibodies to kill tumor cells is antibody-dependent cellular cytotoxicity (ADCC), in which natural killer cells (NK) are the main mediator. The capacity to induce ADCC has conventionally been assessed in the human-mouse xeno-graft model, in which human peripheral blood mononuclear cells (PBMC), containing NK cells along with antibodies, are administered to tumor-bearing immunodeficient mice. However, contamination from other cellular populations often affects tumor growth, making it difficult to evaluate the antibody's effect. In this study, we established a new NK-dependent ADCC assay model using a supra-immunodeficient strain of mice, NOD/SCID/gamma c(null) (NOG). Our model system simply consisted of three elements: isolated human NK cells, a Burkitt's lymphoma cell line (Daudi), and an anti-CD20 antibody (Rituximab). In this experimental setting, human NK cells from healthy donors retained their killing activity and suppressed the growth of Daudi cells in NOG mice when they were administered along with Rituximab. This system, therefore, is useful for evaluating the in vivo function of human NK cells. (C) 2010 Elsevier Inc. All rights reserved.