Development of a pharmacokinetic-pharmacodynamic model to describe blood hepcidin levels in patients with myelofibrosis and assess target engagement of zilurgisertib, an ALK2 inhibitor
Ikhlas Guermi (1), Beatriz Guglieri-Lopez (1), Phil Wang (2), Betty Lamothe (2), Amanda McBride (2), Jennifer Sheng (2), Xuejun Chen (2), Yan-ou Yang (2)
(1) Pumas-AI, USA, (2) Incyte Research Institute, USA
Introduction: Anaemia is a predominant feature of myelofibrosis (MF), a type of myeloproliferative neoplasm, and is associated with poor prognosis [1] and inferior quality of life [2]. Elevated serum levels of the hormone hepcidin in patients with MF is associated with reduced haemoglobin levels, increased requirement for red blood cell transfusions, and reduced overall survival [3]. Zilurgisertib is a novel and selective ALK2 inhibitor under clinical development for the treatment of anaemia due to MF.
Objectives: The objective of this analysis was to develop a pharmacokinetic-pharmacodynamic (PK-PD) model to describe plasma zilurgisertib and blood hepcidin concentration levels to assess zilurgisertib target engagement.
Methods: Zilurgisertib concentration-time data from two phase 1 studies (INCB 00928-101, INCB 00928-102) in healthy volunteers, and one phase 1/2 study (INCB 00928-104) in patients with MF were analysed. Plasma hepcidin concentration data were available from all 3 studies. Data related to zilurgisertib at doses ranging from 10 mg to 600 mg administered once daily were studied. A population PK-PD model was developed sequentially to describe the relationship between zilurgisertib plasma concentrations and blood hepcidin concentrations using non-linear mixed-effect modelling implemented in Pumas software (version 2.4.1; Pumas-AI, Inc., Dover, DE, USA) [4]. Estimation was performed by first-order conditional estimation with interaction (FOCEI). Model selection was based on the objective function value (OFV), precision and plausibility of the parameter estimates, and goodness-of-fit diagnostic plots.
Results: A 3-compartment model with first-order absorption and linear elimination was used to describe the PK of zilurgisertib. A transit absorption compartment and dose-dependent bioavailability were included in the modelling to capture the lack of dose proportionality following oral administration of increasing doses of zilurgisertib. The PD response, represented by hepcidin blood concentration levels over time, was described using an empirical model that incorporated hepcidin’s circadian rhythm. Zilurgisertib concentrations were linked to hepcidin concentrations by means of a power function, with an effect compartment added to capture the delay observed between zilurgisertib administration and the subsequent decrease in hepcidin concentrations. Inherent circadian variation of hepcidin blood concentration as a function of time was described by a cosine function. Inter-individual variability was estimated for the absorption rate constant, clearance, and volume of distribution of the central and first peripheral compartments (Vc and Vp1) of zilurgisertib. It was also estimated for the slope (ie, the proportionality constant linking zilurgisertib concentrations in the effect compartment and hepcidin blood concentrations), baseline hepcidin levels (mean hepcidin concentration in the absence of the zilurgisertib), as well as the amplitude and acrophase of the cosine function describing the circadian fluctuations of hepcidin. Disease status (healthy individuals or patients with MF) was found to be a significant covariate on the slope, the power parameter, and baseline hepcidin concentrations. Plasma ferritin level was also found to be a significant covariate on baseline hepcidin concentrations. Despite considerable inter-individual variability, the PK-PD model adequately described the observed data with no systematic bias, as demonstrated by the goodness-of-fit plots and the prediction-corrected visual predictive check.
Conclusions: The developed PK-PD model adequately described the relationship between plasma zilurgisertib and blood hepcidin concentration levels in both healthy individuals and patients with MF. The addition of the circadian rhythm component helped capture the natural fluctuations of hepcidin blood levels. The PK-PD model was used to inform decision-making for zilurgisertib dose escalation during clinical development in patients with MF.
References:
[1] Cervantes F, Pereira A. Curr Hematol Malig Rep. 2012;7:43-49.
[2] Tefferi A, et al. Clin Ther. 2014;36:560-566.
[3] Pardanani A, et al. Am J Hematol. 2013;88:312-316.
[4] Rackauckas C, et al. bioRxiv. 2022. Preprint posted online March 20, 2022.
