The effect of cell-free ascitic fluid from patients with epithelial ovarian carcinoma on the generation of lymphokine-activated killer cells (LAK) was compared to the activity generated in control medium containing 10% fetal bovine serum, using Daudi target cells. Samples of ascitic fluid from nine different patients tested inhibited LAK generation. Suppressive activity was evident as early as 24 h of incubation in the presence of ascitic fluid and increasing suppression developed with prolonged exposure. Suppression was concentration-dependent, present at 10%-20% and increasing with concentrations up to 80%. The suppressive effect of ascitic fluid was only partially reversed on increasing the concentration of interleukin-2 (IL-2) from 10 units to 1000 units/ml. Activated LAK appeared to maintain the majority of their activity on further culture in ascitic fluid in the presence of IL-2 but further enhancement of lytic activity was prevented. Fractionation of a suppressive sample by HPLC, using 0.1 M KCl/acetic acid buffer pH 2.6, revealed that the dominant peak of suppressive activity eluted at 25 kDa; with pH 7.0 TRIS-buffered saline, most of the activity was lost on the column. Antibody neutralization studies of the 25-kDa suppressive peak as well as on whole ascitic fluid have revealed that transforming growth factor beta (TGF beta) is the major suppressive factor present in ascitic fluid. Factors that suppress LAK generation in vitro were present in all samples tested. The effect on the lytic activity of activated LAK cells was minimal. This suggests that, in the clinical setting, the greatest impact would be achieved by activating LAK cells ex vivo and subsequently transferring them to the peritoneal cavity in the presence of IL-2 rather than by attempting to generate them in situ by injecting IL-2 into the peritoneal cavity. However, reversal of TGF beta-mediated suppression in situ may be necessary to allow local proliferation of LAK cells to achieve an effective killer-to-target ratio.