Before a monoclonal antibody (or other biological ligand) can label or kill a tumor cell, it must first reach that cell. For portions of a tumor more than a few microns from the nearest blood vessel, access may be limited by the rate at which the molecule can """"""""percolate"""""""" through the extracellular space. We are investigating the spatial and temporal profiles of immunoglobulin (Ig) distribution generated by diffusion and convection through tumors, taking into account specific binding, nonspecific binding, and metabolism. For this purpose, we developed theoretical models that splice together the global pharmacology and the microscopic percolation process. Significant predictions thus far include the following: (1) Antibody molecules may be prevented from penetrating a tumor by the very fact of their successful binding to antigen (the """"""""binding site barrier""""""""). Thus, lower affinity may sometimes be preferable. (2) The flux of non-binding control Ig is much less likely to be limited by diffusion or convection. Nonspecific Ig's penetrate more deeply and more quickly into the tumor. (3) Even with saturable binding (but not metabolism), the """"""""C times T"""""""" exposure of tumor cells to antibody will be the same throughout the mass. (4) Metabolism will decrease the relative """"""""C times T"""""""" exposure of cells farther from the blood vessel. This may be a major barrier to effective treatment of solid tumors with ligand molecules. Predictions of the model are being tested using a subcutaneous tumor in guinea pigs and micrometastases of human tumors in nude mice. The distribution of antibody is being determined by fluorescence techniques and autoradiography. Concepts arising from this study have been applied to the design of clinical studies with monoclonal antibodies. In addition to immunoglobulin and other ligands (e.g.,soluble CD4) as administered agents, we are considering the physiology of endogenous molecular species including the lymphokines and growth factors.
