The effect of remodeling of a glycoantigen such as the α-Gal epitope, Galα1, 3Ga1β1, 4GlcNAc-R, by the introduction of glycosyltransferase genes on natural killer (NK) cell-mediated direct cytotoxicity was investigated using human peripheral blood mononuclear cells (PBMC) or an NK-like cell line, YT cells, as an effector, and swine endothelial cells (SEC) as a target. Several SEC transfectants were established by transfection with the genes for β1, 4-N-acetylglucosaminyltransferase III, α2, 3-sialyltransferase and α1, 2-fucosyl-transferase. These transfections led to dramatic reductions in both direct and indirect NK cell-mediated cytotoxicity, by 72-94% in the case of PBMC and 27-72% in that of YT cells, in addition to an effective reduction in xenoantigenicity, which is substantially caused by the α-Gal epitope, to human natural antibodies. The NK cell-mediated direct cytotoxicity was remarkably blocked by an anti-α-Gal epitope monoclonal antibody or GSI lectin which preferentially binds to the epitope. Furthermore, treatment of the parental cells with α-galactosidase resulted in a significant reduction in cytotoxicity. These results suggest that the α-Gal epitope is involved not only in hyperacute rejection and acute vascular rejection, but also in NK cell-mediated direct cytotoxicity. Thus, the genetic remodeling of the α-Gal epitope and probably other glycoantigens as well can be expected to represent a new approach for overcoming not only indirect but also direct immunity to xenografts.