Imaging of human tumor xenografts in nude mice with radiolabeled monoclonal antibodies. Limitations of specificity due to nonspecific uptake of antibody.

To determine if in vivo binding specificities of monoclonal antibodies to tumor nodules would reflect in vitro antibody specificities as determined by radioimmunoassay, two monoclonal antibodies were selected for imaging of human tumor xenografts in nude mice. By in vitro radioimmunoassay, antibody 436 binds to the M20 melanoma cell line, but not to the P3 carcinoma cell line; antibody 44 binds to P3 but not to M20. Both antibodies are IgG1 isotypes. Nude mice bearing an M20 melanoma in one flank and a P3 carcinoma in the other were injected intravenously with 5 to 25 micrograms of each antibody labeled with either I-125 or I-131; in separate animals the labels were reversed. Animals were imaged daily with a scintillation camera equipped with a pinhole collimator. On day 6 the animals were sacrificed, and binding of the antibodies to the tumors and normal tissue were compared. Antibody 436 had up to a two-fold binding advantage in vivo for the melanoma, whereas antibody 44 had up to a two-fold binding advantage for the carcinoma, thus confirming the in vitro specificities of each. However, imaging on day 4 and computer analysis of percent radioactivity in the tumors showed that tumor images were related directly to tumor size and relatively uninfluenced by antibody specificity. Thus, even though antibody specificity can be demonstrated by differential tissue counting, imaging of tumor deposits appears to involve a nonspecific phenomenon that is largely dependent upon tumor weight.