DDEL-11. PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELING OF SPATIAL HETEROGENEITY OF DRUG PENETRATION AND EXPOSURE IN THE HUMAN BRAIN AND BRAIN TUMORS

Abstract

Abstract BACKGROUD The difficulty in early, reliable prediction of drug penetration and exposure in the human brain and brain tumors makes development of new drugs and use of existing drugs for treating brain cancer a challenging and often unsuccessful task. We developed a physiologically based pharmacokinetic (PBPK) model platform for prediction of spatial pharmacokinetics in the central nervous system (CNS) following systemic drug administration. METHODS A 9-compartment CNS (9-CNS) PBPK model was developed to predict drug concentrations in brain blood, 2 brain parenchyma compartments (brain tissue adjacent to the CSF tract and deep brain parenchyma), 3 CSF compartments (ventricular CSF, cranial and spinal subarachnoid CSF), and 3 tumor compartments (tumor rim, bulk tumor, and tumor core). The pharmacokinetics in the CNS compartments were driven by plasma concentrations and governed by drug properties and CNS pathophysiology. The 9-CNS model was developed and validated with 4 drugs (abemaciclib, ribociclib, ceritinib, and temuterkib). RESULTS When considering tumor heterogeneity and inter-individual plasma pharmacokinetic variability following clinical standard dosing regimens, the model-predicted steady-state unbound drug tumor concentrations varied from 7.1 to 356 nM for abemaciclib, 74 to 4577 nM for ribociclib, 0.66 to 97 nM for ceritinib, and 0.8 to 288 nM for temuterkib, which were consistent with the heterogeneous drug tumor concentrations observed in glioblastoma patients. Drug penetration into the deep brain parenchyma was significantly less than that into tumors and CSF-adjacent brain tissue, particularly for strong ABCB1/ABCG2 substrate drugs. The ventricular and cranial subarachnoid CSF shared similar pharmacokinetics, while the spinal CSF pharmacokinetics was different. CONCLUSION A novel 9-CNS PBPK model was developed and validated for mechanistic prediction of spatial heterogeneity of drug penetration and exposure in the human brain, brain tumors, and CSF. It provides a valuable computational modeling tool to assist development and design of improved drugs or dosing regimens for brain cancer treatment.