Modeling recombinant immunotoxin efficacies in solid tumors.

Effectiveness of cancer therapy is improved by the use of recombinant immunotoxins (RITs) that target membrane proteins unique to malignant tumor cells. Although RIT antitumor activity in vivo can always be improved with larger doses, clinical restriction on the dose toleration makes it critical to explore how RIT antitumor activity can be maximized without resorting to dose elevation. In this work, a mathematical model was developed to explore functional correlations between the properties of several recombinant immunotoxins and their antitumor efficacies in vivo. Simulations were compared with experimental data of human tumor xenografts grown on nude mice to assess parameters critical to optimal antitumor activity. We dissected out or held constant as many parameters of the model as possible to investigate the effect of the remaining parameters on the behavior of the system as a whole. Empirical correlations between immunotoxin binding affinity and the target binding site density were obtained for several recombinant immunotoxins targeting either human A431 carcinoma or CD46 Burkitt's lymphoma. Simulations reinforced the idea of binding site barrier for drug diffusion and suggested that optimal antitumor activity was achieved when the binding affinity is logarithmically dependent on the target binding site density.