Jae Eun Ahn (1), Richann Liu (2), Eva Hajos-Korcsok (1), Terence Fullerton (1), Ruolun Qiu (1)
(1) Pfizer Inc. USA, (2) Pfizer Inc. USA during the study conduct
Objectives: Gamma secretase modulators (GSMs) shift the cleavage of amyloid-β (Aβ) peptides in favor of shorter forms (such as Aβ37 and Aβ38) without inhibiting overall amyloid precursor protein processing, thus preferentially lower the production of longer, amyloidogenic forms (Aβ42 and Aβ40). This study aimed to develop a population pharmacokinetic/pharmacodynamic (PK/PD) model to characterize the differential effects of PF-06648671 on four Aβ species simultaneously. The developed model was intended to inform the dose selection for future studies.
Methods: Clinical PK/PD data were obtained from 133 healthy subjects in three Phase 1 studies. In the first-in-human single ascending dose study, PK data from 2, 4, 12, 40, 120, 240, 360 mg doses and placebo were obtained. From the single dose cerebrospinal fluid (CSF) biomarker study, PK as well as serial CSF samples were collected up to 36 hr post-dose (150, 300 mg doses, and placebo) by lumbar catheterization. In the multiple dose study, PK and PD samples were collected (40, 100, 200, 360 mg doses and placebo for 14 days) but only one post-baseline CSF sample was collected at trough, in addition to the baseline. A previously developed population PK model was used (2-compartment model with first-order absorption and elimination), fixing population PK parameters with PK data still included (PPP&D approach [1]). Drifts in baseline (a rising tendency) were observed in placebo subjects from the single dose biomarker study, which was empirically described in the PK/PD model. An indirect response model was implemented allowing PF-06648671 plasma concentrations to change the production rate of CSF Aβ peptides (decreased Aβ42 and Aβ40; increased Aβ37 and Aβ38). Simple inhibitory Emax (Imax) and sigmoidal Emax models were applied for longer and shorter forms of Aβ, respectively. Baseline correlation among Aβ species was also accounted for in the model. NONMEM 7.3 [2] was used for population PK/PD modeling.
Results: The PK/PD model described the observed Aβ responses from both single and multiple dose studies reasonably well. The inhibitory effect on CSF Aβ42 was estimated to be greater than on Aβ40 (Imax and IC50: 0.744 and 421 ng/mL for Aβ42 vs. 0.725 and 1228 ng/mL for Aβ40) whereas the stimulatory effect on Aβ37 was greater than on Aβ38 (Emax and EC50: 5.63 and 1285 ng/mL for Aβ37 vs. 0.869 and 1765 ng/mL for Aβ38). Precision of the parameter estimates was reasonable (relative standard errors, RSEs, were less than 40 %) except for Aβ38, for which a step-like response was observed. The PK/PD model suggested CSF Aβ42 average reductions from baseline at steady state of approximately 50%, 59, and 65%, after 75, 150, and 300 mg q.d. administrations respectively. The reductions in ratios of Aβ42 to shorter peptides (Aβ42: Aβ40, Aβ42: Aβ38, and Aβ42: Aβ37) were predicted to reach a plateau after a daily dose of approximately 150 mg.
Conclusions: An indirect response model was developed to characterize differential and exposure-dependent effects of PF-06648671 on CSF Aβ peptides in healthy subjects. The PK/PD model was used to predict the drug effect on individual Aβ peptides as well as the ratio of Aβ peptides in CSF across a range of PF-06648671 doses. The model-predicted changes in CSF concentrations of different Aβ species and ratios of long to short fragments were in keeping with the expected pharmacological actions of a GSM. The results of this effort have utility in informing dose selection for potential future clinical trials of PF-06648671.
References:
[1] Zhang L, Beal SL, and Sheiner LB, 2003, Simultaneous vs. sequential analysis for population PK/PD data i: best-case performance. Journal of pharmacokinetics and pharmacodynamics vol. 30:387-404.
[2] Beal S, Sheiner LB, Boekmann A, and Bauer Rj, NONMEM User’s Guides. 1989-2015. ICON Development Solutions, Ellicott City, MD, USA.
Reference: PAGE 27 (2018) Abstr 8419 [www.page-meeting.org/?abstract=8419]
Poster: Drug/Disease Modelling - CNS