DDEL-03. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELING OF HETEROGENEOUS DRUG PENETRATION INTO THE HUMAN BRAIN AND GLIOBLASTOMA

Abstract

Abstract BACKGROUD Insufficient penetration of potentially effective therapeutic agents across the human blood-brain barrier (BBB) is a hurdle to the effective treatment of brain cancer. The BBB in glioblastoma is disrupted to varying extents. This study was to develop a PBPK model platform for prediction of heterogenous drug penetration into the human brain and glioblastoma. METHODS A 7-compartment CNS (7-CNS) PBPK model platform was developed to predict drug concentrations in the brain blood, brain parenchyma, cranial CSF, spinal CSF, and 3 tumor compartments representing infiltrating tumor (T1) with a slightly disrupted BBB, bulk tumor (T2) with a leaky BBB, and necrotic core (T3) with no blood vessels. The model accounted for brain/tumor general anatomy and regional pathophysiological differences. Drug transfer and fluid balance between compartments were described by differential equations. Simulations were performed using R programming. The model was verified by comparisons of predicted and observed tumor PK data of ribociclib and abemaciclib in glioblastoma patients. RESULTS For ribociclib, the model-predicted mean unbound drug brain-to-plasma ratio (Kp,uu) was 0.17, 2.2, 8.2, 24.9 in normal brain, T1, T2, and T3, respectively; the predicted Kp,uu in T1 and T2 was consistent with observed Kp,uu in non-enhancing (median, 2.0) and enhancing tumors (median, 10.1) from 12 patients. For abemaciclib, the predicted mean Kp,uu was 0.88, 1.4, 6.2, and 3.1 in normal brain, T1, T2, and T3, respectively; the predicted Kp,uu in T1 and T2 agreed with observed Kp,uu in non-enhancing (median, 2.1) and enhancing tumors (median, 7.2) from 38 patients. CONCLUSION The developed PBPK model platform well predicts heterogeneous drug penetration and distribution in normal human brain and different regions of glioblastoma, as verified by ribociclib and abemaciclib. It provides a mechanism-based computational modeling tool to assist the development and design of improved drugs and dosing regimens for more effective treatment of glioblastoma.