To investigate GVL effects, Balb/c mice (H-2d) received 5 x 10(5) A20 (B cell leukemia) or 1 x 10(6) WEHI-3 (myelomonocytic leukemia) cells. These cell lines lead to death after a median of 19 (WEHI-3) or 30 days (A20). A lethal dose of total body irradiation followed by syngeneic BMT resulted in significantly prolonged survival of leukemia-bearing animals. Transplantation of (C57 x Balb/c)F1 (H-2bxd) allogeneic, but GVH-non-reactive marrow grafts differentially influenced the relapse rates in the two leukemia models. Whereas allogeneic BMT reduced the relapse rates in A20-bearing mice, the leukemia-free survival was not improved in mice bearing the leukemia WEHI-3 compared with syngeneic BMT. Pre-treatment of allogeneic (C57 x Balb/c)F1 or syngeneic Balb/c marrow cells with 200 U/ml IL-2 for 24 h did not reduce relapse rates in animals inoculated with A20 leukemia cells compared with unmanipulated bone marrow. In contrast, IL-2 treatment of syngeneic or allogeneic GVH non-reactive donor marrow significantly decreased the relapse rate in mice inoculated with WEHI-3 leukemia cells. The NK cell-mediated lysis of cultured leukemia cells was determined in vitro using a conventional 56Cr-release assay. Our data revealed a strong correlation between the level of natural killer activity determined in vitro and GVL activity in vivo.