OKT3, a mouse anti-human CD3 mAb, is a potent immunosuppressive agent used in clinical transplantation to prevent or treat allograft rejection. Associated with this therapy is the systemic release of several cytokines that result in a series of adverse side effects. This release of cytokines is dependent on the cross-linking mediated by OKT3 between T cells and the Fc gamma R-bearing cells. To generate an anti-human CD3 mAb with reduced activating properties as compared with OKT3, we have transferred the complementary determining regions of OKT3 onto human IgG frameworks and then performed point mutations that reduce the affinity of the "humanized" anti-CD3 mAbs for Fc gamma Rs. Initial, in vitro, studies showed that whereas OKT3 and the parental humanized anti-CD3 mAbs activated T cells similarly, a humanized Fc variant failed to do so. Both the Fc variant and the activating anti-CD3 mAbs induced comparable modulation of the TCR and suppression of cytolytic T cell activity, in vitro. In the current study, we exploited an experimental model in which human splenocytes from cadaveric organ donors were inoculated into severe combined immunodeficient mice (hu-SPL-SCID mice) to test the activating and immunosuppressive properties of these anti-human CD3 mAbs in vivo. Unlike injection of OKT3 or of the parental humanized mAb, administration of the Fc variant did not result in T cell activation in vivo, as evidenced by the lack of induction of surface markers of activation, and of systemic human cytokines, including IL-2. Importantly, similar prolongation of human allograft survival was achieved with all anti-CD3 mAbs, indicating that the nonactivating anti-CD3 mAbs retained significant immunosuppressive properties in vivo. Thus, the use of an Fc variant in clinical transplantation should result in fewer side effects than observed with OKT3, while maintaining its clinical efficacy.