Pharmacokinetic analysis of immunotoxin uptake in solid tumors: role of plasma kinetics, capillary permeability, and binding.

The delivery of cell-specific protein toxins to the interstitium of solid tumors was examined in athymic mice bearing s.c. human rhabdomyosarcoma (TE671) tumors. The toxins are diphtheria toxin (DT), Mr = 60,000, and an immunotoxin, Mr = 210,000. The immunotoxin is a chemical conjugate of a mutant DT defective in binding and a monoclonal antibody specific for the human transferrin receptor. The plasma, tumor, and muscle concentrations of DT, immunotoxin, and closely related nonbinding controls were measured 2, 6, and 24 h after i.v. injection into tumor-bearing mice. Both DT and immunotoxin are specific for the human xenograft in the mouse because DT is very toxic to human cells but not to murine cells and immunotoxin is directed against a human cell receptor. A compartmental pharmacokinetic model was developed for the analysis of the in vivo data to provide plasma-to-tissue transport constants (capillary permeability-area products), binding parameters (products of the association constant and the initial binding site concentration), and the interstitial fluid flow rate. The model also provides a simple mathematical framework for understanding the effect of these variables on the localization of macromolecules in tumors. The plasma-to-tissue transport constant of immunotoxin in TE671 tumor was 0.13 microliters/min/g, compared to 0.29 microliters/min/g for DT. However, despite the lower capillary permeability of the larger molecular weight toxin, the cumulative tumor exposure to immunotoxin was 80% higher than that to DT after 24 h. A longer plasma half-life and higher apparent in vivo binding parameter of immunotoxin compared to DT contributed to the higher tumor exposure. Plasma-to-tissue transport constants for tumor were 60 to 100% higher than those for muscle. This finding is consistent with observations by others that tumor vasculature is more permeable than are normal muscle capillaries. Also, the interstitial fluid flow of the tumor, 0.80 microliters/min/g, was higher than that of muscle, 0.58 microliters/min/g. The product of the binding affinity and binding site concentration for immunotoxin in vivo was 530 times lower than that predicted based on in vitro measurements. Lower expression of antigen binding sites, inaccessibility of binding sites in vivo, and degradation of the toxin are several possible factors that may account for the in vitro-in vivo differences in binding. This study illustrates the interrelationship of plasma kinetics, capillary permeability, and binding and their effects on toxin concentrations that are achieved in the tissue interstitium.