Clinical immunotherapy trials have been performed recently where ex vivo interleukin-2 (IL-2)-activated peripheral blood mononuclear cells (i.e., the "LAK" cells) have been transfused in addition to IL-2 infusions. In such protocols, patients have received highly heterogeneous cell suspensions and the nature of the effector cells that may have contributed to tumor regression has remained unclear. In certain animal models, it has appeared that natural killer lymphocytes were the effector cell type responsible for tumor regression. To test whether NK cells could eventually be relevant for the treatment of human tumors, we have performed a feasibility trial where purified lymphokine-activated natural killer (LANAK) cells have been prepared and transfused to a limited series of renal cell carcinoma patients receiving IL-2 (continuous infusions at 3 x 10(6) U/m2/day). Natural killer lymphocytes (1-2 x 10(6] were purified from peripheral blood mononuclear cells and expanded during 4-5 weeks in the presence of IL-2 on microtiter plates containing feeder layers cells. In vitro, the resulting LANAK cell suspensions were 100 times (range of 2 to 10(3] more efficient against Daudi target cells than their autologous LAK counterparts. Twelve patients were included; 9 received the two planned courses of treatment with LANAK cells and IL-2. Overall toxicity was relatively moderate. Besides occasional chills, there were no apparent secondary effects due to cell infusions. The mean number of LANAK cells transfused per patients was 45.1 x 10(9), ranging from 7 to 125 x 10(9). The biodistribution of LANAK cells was similar to that reported previously for LAK cells with no preferential localization to tumor sites. We conclude from this study that using well-defined populations of effector lymphocytes is a feasible cellular therapy approach that may lead to improved understanding and efficacy of the novel immunotherapy methods.