Bispecific murine monoclonal antibodies that target tumor and Fc gamma RIII (CD16) can promote relevant tumor lysis by large granular lymphocytes. For these antibodies to be clinically useful, their properties should be maintained in vivo, where competing human immunoglobulin, shed target antigen, and shed CD16 may be encountered. At a minimum, bispecific antibody antitumor effects should be preserved in whole blood. Furthermore, potentiation of tumor lysis should be reflected by demonstrating the ability of bispecific antibody-retargeted effector cells to infiltrate and mediate lysis of organized tumor. If these characteristics are demonstrated, and there is evidence of in vivo efficacy of bispecific antibody-based therapy in a relevant animal model, further clinical development of such antibodies would be warranted. In this report the ability of CL158 bispecific antibody supernatants to mediate lysis of SW948 tumor growing in monolayer is shown to be preserved in the presence of interleukin 2-activated whole blood. When SW948 cells were grown in vitro as multicellular human tumor spheroids, incubation with interleukin 2-activated lymphocytes (LAK cells) and CL158 led to structural and widespread necrosis. This was dependent on CL158 and resistant to competition by pooled human immunoglobulin or interleukin 2-exposed whole blood. These effects were not promoted by the monospecific antibodies produced by the parent clones of CL158 and were not observed when the IgG2a variant of CA19-9 antibody, which mediates conventional antibody-dependent cellular cytotoxicity, was used instead of its bispecific derivative. To examine the efficacy of bispecific antibody-based treatments on in vivo tumor, scid mice bearing early s.c. SW948 xenografts were treated with interleukin 2 for 5 consecutive days, supplemented by three i.v. injections of 10(7) human LAK cells and various antibodies. Treatment of mice bearing SW948 tumors with LAK cells did not retard tumor growth, but when CL158 was added, significant delays in tumor growth were observed. Tumor growth delay required treatment with both LAK cells and the bispecific antibody. Treatment with the IgG2a variant of CA19-9 antibody, alone or with LAK cells, had no effects on tumor growth. Although the mechanisms of these antitumor effects require further study, it is clear that human LAK cell treatment of animals bearing early, established s.c. tumors is enhanced by the addition of bispecific antibodies with relevant binding characteristics. When compared with the IgG2a isotype variant of CA19-9 monoclonal antibody, this bispecific antibody offers the advantages of preservation of activity in physiological conditions, infiltration and disruption of organized tumor in vitro, and antitumor effects in a relevant xenograft model.(ABSTRACT TRUNCATED AT 400 WORDS)