A Systematic Approach to PKPD Model Development to describe Sleep Effects of Compounds with Different Mechanisms of Action Using Semi-Mechanistic Markov Chain Models
Arianna Madrid(1), Nieves Vélez de Mendizabal(1,2), Kimberley Jackson(3), Andrew McCarthy(3), Dale M. Edgar(3), Brian J. Eastwood(3), Wesley Seidel(3), Iñaki F. Trocóniz(1).
(1) Department of Pharmacy and Pharmaceutical Technology; School of Pharmacy; University of Navarra; Pamplona 31080; Spain. (2) Indiana University School of Medicine; Indianapolis, IN, USA. (3) Eli Lilly and Company, UK.
Objectives: To describe the sleep effects of the non-benzodiazepine hypnotic agent Zopiclone (ZOP), and the selective 5-HT2A antagonist MDL-100,907 (MDL) using a semi-mechanistic pharmacokinetic-pharmacodynamic (PKPD) Markov-chain model previously developed for Zolpidem in healthy rats[1].
Methods: Experimental. Electroencephalogram (EEG) data were obtained in rats. For each 10 second interval, EEG data were converted into awake, REM or NREM stages representing non-ordered categories. The data consisted of a 12 h baseline where EEGs were monitored in the absence of any type of perturbation and a 12 h period during which methylcellulose (MC), ZOP or MDL were administered.Data analysis. The time course of the 9 possible transition probabilities between the 3 sleep stages was described using a non-homogeneous Markov chain model based on piecewise multinomial logistic functions[2], as previously described[1]. Literature PK data was used to generate concentrations of ZOP and MDL over time[3-5]. Analyses were performed using NONMEM VII v2 using the LAPLACIAN estimation method. Model evaluation was based on visual predictive checks (VPCs).
Results: Baseline model. A model selected previously[1] was used to generate VPCs for the baseline data from the new studies. Results indicated that this model was adequate to describe and predict the new data. MC model. The effects of MC administered orally or IP were incorporated using a Bateman function to reflect an increase in the transition probability from NREM to awake as observed in the data. Drug effect model. Exploration of the time course of transition probabilities revealed that both ZOP and MDL elicited a temporal decrease in the transition probability from NREM to awake indicating that sleep was promoted. ZOP exhibited a rebound effect approx. 8-10h after dosing, whereas such rebound phenomena were not observed in the data with MDL. ZOP effects were described using a turnover feedback model.[1, 6] For MDL, the PKPD models that best described the data were the link[7] or indirect response[8] (IDR) models.
Conclusions: The baseline response model used to describe the underlying physiological system (a non-homogeneous Markov chain model based on piecewise multinomial logistic functions) has been shown to be conserved across several studies, thereby supporting its application for future studies. Drug level effects need to be considered separately, contingent on their mechanism of action and the observed responses.
References:
[1] Arianna M, Nieves V, Kimberley J, Andrew M, Iñaki T. Modelling the sleep effects of Zolpidem in rats using non-homogeneous Markov chain models. PAGE 20 (2011) Abstr 2034 [www.page-meeting.org/?abstract=2034].
[2] Bizzotto R, Zamuner S, De Nicolao G, Karlsson MO, Gomeni R. Multinomial logistic estimation of Markov-chain models for modeling sleep architecture in primary insomnia patients. J Pharmacokinet Pharmacodyn. 2010; 37: 137-55.
[3] Drover R. Comparative Pharmacokinetics and Pharmacodynamics of Short-Acting Hypnosedatives Zaleplon, Zolpidem and Zopiclone. Clin Pharmacokinet 2004:227-238.
[4] Visser, S. A. G., F. L. C. Wolters, J. M. Gubbens-Stibbe, E. Tukker, P. H. Van Der Graaf, L. A. Peletier, and M. Danhof. Mechanism-Based pharmacokinetic/pharmacodynamic Modeling of the Electroencephalogram Effects of GABAA Receptor Modulators: In Vitro-in Vivo Correlations. J Pharmacol Exp Ther 2003; 304: 88-101.
[5] Scott D, Heath T. Investigation of the CNS penetration of a potent 5-HT2a receptor antagonist (MDL 100,907) and an active metabolite (MDL 105,725) using in vivo microdialysis sampling in the rat. J Pharm Biomed Anal 1998; 17: 17-25.
[6] Wakelkamp M, Alván G, Gabrielsson J, Paintaud G. Pharmacodynamic modelling of furosemide tolerance after multiple intravenous administration. Clin Pharmacol Ther 1996; 60: 75-88.
[7] Sheiner LB, Stanski DR, Vozeh S, Miller RD, Ham J. Simultaneous modeling of pharmacokinetics and pharmacodynamics: application to d-tubocurarine. Clin Pharmacol Ther. 1979 25: 358-371.
[8] Dayneka NL, Garg V, Jusko WJ. Comparison of four basic models of indirect pharmacodynamic responses.J Pharmacokinet Biopharm. 1993 21: 457-478.
