Combined immunohistochemical and autoradiographic analyses of antigen/antibody interactions in tumor xenograft models.

A study was conducted to establish optimal conditions which would allow for the simultaneous localization of a carcinoma antigen and its complementary radiolabeled antibody. Immunoperoxidase staining was used to identify the tumor distribution of antigen, while tissue localization of the radiolabeled antibody was identified by autoradiography. The tumor associated glycoprotein-72 (TAG-72) antigen and the high affinity murine monoclonal antibody, CC49 IgG were used as the model antigen/antibody pair. Athymic female mice bearing either CX-1 or LS-174T human colorectal adenocarcinoma xenografts were used as animal/tumor test systems. Experimental mice each received a bolus intravenous injection of the CC49 antibody which was labeled with 125I (specific activity, 0.17 to 0.26 microCi/microgram). Control mice were given a bolus injection of MOPC-21 IgG monoclonal antibody (tumor irrelevant antibody) which was also radiolabeled with 125I (specific activity, 0.24 to 0.35 microCi/microgram). At 24 hours postinjection, all tumors removed, counted for radioactivity, and fixed in formalin. The avidin/biotin immunoperoxidase complex technique was used to identify TAG-72 antigenic sites on slide-mounted tissue sections. Nonradiolabeled CC49 IgG (0.5 micrograms/ml) was used as the specific antigen binding primary antibody in the immunostaining procedures. Nonradiolabeled MOPC-21 IgG (0.5 micrograms/ml) served as the negative control. Immunohistochemically stained tissue sections were coated with photographic emulsion and processed for autoradiographic localization of 125I-CC49 or 125I-MOPC-21. After an optimal exposure time of 6 days, slides were processed and examined under a light microscope. Results of the biolocalization experiment revealed that the % of injected dose/gram of 125I-CC49 in both LS-174T and CX-1 tumors (30.4 +/- 5.2% and 20.6 +/- 5.4%, respectively) were significantly greater (p greater than 0.01) than those for 125I-MOPC-21 (4.9 +/- 0.5% and 5.1 +/- 0.7%, respectively). In both tumor lines from mice injected with 125I-CC49, dense clusters of silver grains were found over those regions which were positive for TAG-72 immunoreactivity. These dual-labeled structures were also found in contact with, or in close proximity to the microvasculature. Tumors from mice which were injected with the control radioconjugate showed a random distribution of silver grains within stromal tissue but no specific localization to TAG-72 positive regions. We conclude that intravenously administered 125I-CC49 IgG localizes specifically to antigen-containing sites in the LS-174T and CX-1 tumor models. The methods described herein should serve as useful tools for the direct study of antigen-antibody interactions in tumor biology.