Mechanistic Multilayer Quantitative Model for Nonlinear Pharmacokinetics, Target Occupancy and Pharmacodynamics (PK/TO/PD) Relationship of D-Amino Acid Oxidase Inhibitor, TAK-831 in Mice

Abstract

Purpose: TAK-831 is a highly selective and potent inhibitor of D-amino acid oxidase (DAAO) currently under clinical development for schizophrenia. In this study, a mechanistic multilayer quantitative model that parsimoniously connects pharmacokinetics (PK), target occupancy (TO) and D-serine concentrations as a pharmacodynamic (PD) readout was established in mice.

Methods: PK, TO and PD time-profiles were obtained in mice and analyzed by mechanistic binding kinetics model connected with an indirect response model in a step wise fashion. Brain distribution was investigated to elucidate a possible mechanism driving the hysteresis between PK and TO.

Results: The observed nonlinear PK/TO/PD relationship was well captured by mechanistic modeling framework within a wide dose range of TAK-831 in mice. Remarkably different brain distribution was observed between target and reference regions, suggesting that the target-mediated slow binding kinetics rather than slow penetration through the blood brain barrier caused the observed distinct kinetics between PK and TO.

Conclusion: A quantitative mechanistic model for concentration- and time-dependent nonlinear PK/TO/PD relationship was established for TAK-831 in mice with accounting for possible rate-determining process. The established mechanistic modeling framework will provide a quantitative means for multilayer biomarker-assisted clinical development in multiple central nervous system indications.

Keywords: D-amino acid oxidase inhibitor; PK/PD, multilayer quantitative model; TAK-831; target occupancy.

Fig. 1

Chemical structures of (a) TAK-831 (asterisk donates ¹⁴C-labeled position) and (b) PGM019260.

Fig. 2

Schematic description of developed mechanistic multilayer model structure of PK/TO/PD of TAK-831 in mice. Plasma PK was described by 1-compartment model with linear elimination and connected with cerebellum TO with mechanistic binding kinetics model to target enzyme, DAAO. TO of TAK-831 was measured as the magnitude of tracer replacement. The PD modulation was described by indirect response model with inhibition of PD elimination dependent on TO. Where Cp, Vp and CLp represent concentration, oral clearance and apparent distribution volume of TAK-831 in plasma PK compartment, respectively. Kon, Koff and BRmax represent second-order association rate constant, first-order dissociation rate constant and maximum binding ratio of TAK-831 in cerebellum TO compartment, respectively. BLtr represents baseline concentration of tracer in cerebellum. Kin, Kout, BL represent zero-order generation rate constant, first-order elimination rate constant and baseline concentration of D-serine in cerebellum PD compartment, respectively. Imax represents maximum inhibitory effect of TO on Kout.

Fig. 3

The empirical observations and mechanistic multilayer model analysis on PK/TO/PD relationship after a single oral administration of TAK-831 in mice. Observed and model simulated (a) mean plasma concentration-time profiles of TAK-831 and (b) mean concentration-time profiles of the specific binding of PGM019260 in cerebellum after a single oral administration of TAK-831 at 0.3, 1, 3 and 10 mg/kg followed by a single intravenous administration of PGM019260 at 20 min before blood and brain sampling in mice (n = 4). (c) Observed and model simulated mean concentration-time profiles of D-serine in cerebellum after a single oral administration of TAK-831 at 0.3, 1, 3 and 10 mg/kg in mice (n = 7). In each panel, symbols, error bars and lines represent the mean observed data, standard deviations and the model simulations, respectively. Obs, observed; sim, simulated.

Fig. 4

Concentration- and time-dependent brain distribution of TAK-831 in mice. Mean plasma concentration-time profiles of radioactivity in plasma, cerebellum (target region) and front part of cerebrum (reference region) at (a) 0.3 mg/kg and (b) 3 mg/kg, and (c) their correlation between cerebellum or front part of cerebrum and plasma after a single oral administration of [¹⁴C]TAK-831 at 0.3 and 3 mg/kg in mice (n = 2). In each panel, symbols represent the mean observed data (n = 2).

Fig. 5

The established mechanistic multilayer PK/TO/PD model-based simulation of (a) plasma concentration of TAK-831, (b) target occupancy of TAK-831 in cerebellum and (c) D-serine concentration in cerebellum after multiple oral administrations of TAK-831 at 0.3, 1, 3 and 10 mg/kg once daily in mice.