Myeloablative chemo-/radiotherapy supported by transplantation of autologous bone marrow or blood progenitor cells for acute leukemia, lymphoma, or myeloma continues to be associated with a high relapse rate because of the infusion of malignant stem cells and the lack of an in vivo graft-vs.-leukemia (GVL) effect. Although various methods of purging are established for marrow, purging procedures for blood progenitor cell preparations have not been widely used primarily because of the technical challenges to process a higher number of cells. As a broadly applicable method for immunological purging, we tested whether highly cytotoxic cells from a natural killer (NK) cell line characterized previously (NK-92) could be used for immunological purging of blood preparations. The NK-92 cell line, which was established from a patient with non-Hodgkin's lymphoma, can lyse in vitro a broad range of leukemia, lymphoma, and myeloma cell lines even at very low effector:target (E:T) ratios; this lysis is superior to cytotoxicity obtained from normal peripheral blood mononuclear cells (PBMCs) stimulated for 4 days with interleukin (IL)-2. In an attempt to quantitate the purging achievable with NK-92 cells, normal PBMCs were spiked with 10% K562 cells that had been transfected with the neo(r) marker gene (K562-neo(r). Various numbers of NK-92 cells were then added to the cell mixtures, which were incubated for 4 or 48 hours at 37 degrees C with or without IL-2 (500 U/mL). In order to prevent their proliferation, NK-92 cells were irradiated with 1000 cGy (cesium source). This radiation dose was determined to suppress proliferation of NK-92 cells, but at the same time maintain full cytotoxic activity. After co-culture, the cells were plated in methylcellulose containing 0.8 mg/mL G418. The number of surviving K562-neo(r) colonies was counted under the microscope 7 days later and the results were considered a quantitative readout for the purging efficacy of NK-92 cells. No neomycin-resistant K562 colonies could be detected up to a ratio of NK-92:K562-neo(r) cells of 5:1 (effective NK-92:PBMC ratio of 0.5:1). The presence or absence of IL-2 during the culture period did not affect the results. At this ratio of NK-92:PBMC, the growth of normal clonogenic hematopoietic progenitor cells was not compromised as determined by a standard methylcellulose assay. Considering that K562 is a rapidly proliferating cell line and that the input number of K562 cells (10%) tested here was high, the data suggest that the cytotoxic NK-92 clone (after irradiation to prevent proliferation) could be used effectively for immunological ex vivo purging without compromising hematopoietic cell function.