Application of population PK/PD modeling and simulation to inform the design of a dose-finding study in patients with schizophrenia
Jan Berkhout (1), Teun M. Post (1), Lin Xu (2), Liming Zhang (2), Jens Wendland (2), Helene Faessel (2), Majid Vakilynejad (2)
(1) LAP&P Consultant, Leiden, The Netherlands; (2) Takeda Pharmaceuticals, Inc., Cambridge, MA, USA
Objectives:
TAK-831 is a highly selective and potent inhibitor of d-amino acid oxidase (DAAO), an enzyme that degrades d-serine and is highly expressed in glia and neurons within the mammalian brain[1,2]. Inhibition of DAAO increases levels of d-serine, a co-agonist of N-methyl-d-aspartate (NMDA) glutamate receptors, and may improve NMDA-dependent glutamatergic hypofunction[1], such as in cerebellar ataxia and schizophrenia. Clinical studies were performed to assess safety, pharmacokinetics (PK), pharmacodynamics (PD), and brain enzyme occupancy (EO) in healthy volunteers. TAK-831 is being developed for the treatment of Friedreich ataxia (FRDA) and as an adjunctive therapy for negative symptoms of schizophrenia.
The aim of this work was to describe the relationship of TAK-831 plasma PK with plasma d-serine, TAK-831 in cerebrospinal fluid (CSF), and CSF d-serine time courses. In addition, the TAK-831 plasma exposure versus brain EO was described. Simulations were performed to provide insights into the TAK-831 PK/PD relationships and inform dose selection for a dose-finding study in two proof-of-concept (POC) studies in patients with schizophrenia or FRDA.
Methods:
PK and PD d-serine data from 4 phase 1 studies of TAK-831 given as a single daily oral dose (10 to 1200 mg) or as multiple daily oral doses (30 to 400 mg) were pooled for analysis. A population PK model was developed using a 3-compartment model with 6 transit compartments to account for the observed differences in the absorption profile of TAK-831 given as an oral suspension or as a tablet and with or without food. The population PK/PD analyses were performed by means of nonlinear mixed effects models using NONMEM (v.7.3) and PsN (v4.6). An indirect response model was used to describe the inhibitory effect of TAK-831 on d-serine production in both plasma and CSF. A direct response model was used for linking TAK-831 plasma PK to brain EO.
Results:
A total of 149 healthy subjects with 2495 PK and 2270 PD observations were included for model building. The established models well described the PK/PD relationships among TAK-831 exposure, brain EO, and downstream d-serine increases in plasma and CSF. Brain EO and CSF d-serine levels reached a plateau, with reduced interindividual variability at the highest doses tested; the sustained CSF d-serine elevation during 24 h supported a once-daily dosing regimen. On the basis of the simulations, the highest dose associated with the maximum d-serine increase and 2 lower doses were recommended for a phase 2 POC study to fully describe the PK/PD relationships and subsequently understand how this relates to clinical benefit in the patient population.
Conclusions:
This integrated PK/PD and PK/EO modeling analyses provided the quantitative basis for model-informed dose selection in early clinical development. It was found that daily dosing of TAK-831 resulted in an exposure-dependent d-serine increases, with CSF steady-state levels remaining constant over 24 h.
References:
[1] Kakegawa W, Miyoshi Y, Hamase K, et al. d-Serine regulates cerebellar LTD and motor coordination through the delta2 glutamate receptor. Nat Neurosci. 2011;14(5):603-11.
[2] Sacchi S. D-Serine metabolism: new insights into the modulation of D-amino acid oxidase activity. Biochem Soc Trans. 2013;41(6):1551-6.
