Human effective dose prediction of analgesic drug using translational PK-PD modeling
Seongmee Jeong(1), Hyungjun Kim(1), Jung-woo Chae(1),(2)＊ , Hwi-yeol Yun(1),(2)*
(1) College of Pharmacy, Chungnam National University, Daejeon, South Korea, (2) Bio-AI Convergence Research Center, Chungnam National University, South Korea *Those of authors contributed equally a s correspondence.
Objective: Translational PK–PD modeling is aimed to predict exposure-response relationships in human based on preclinical data[1]. As such translational research and the application of PK–PD models on the interface between drug discovery and early clinical development provides a sound basis for support of dose selection and optimization of dose regimens for early studies conducted in patient[2,3]. There are several disease models to evaluate the effectiveness of analgesics on Osteoarthritis. These models are used to verify drug efficacy by dose and can be used to predict drug efficacy in human. The purpose of this study is to help in the drug development stage by establishing a translation PK-PD model that can predict the effect of analgesics in human.
Methods: Osteoarthritis' disease model used MIA or IA induced model for mouse and rat. The data used for model establishment were collected from the literature. The study drugs are celecoxib, naproxen, and tramadol. The PK/PD analyses were performed using NONMEM with the first order conditional estimation method with interaction (FOCE INTER) or FOCE. A population PK model of analgesic drug was established based on data from literature. For describing PK-PD relationship, the PK and PD results in MIA or IA induced rodent was used to determine the PK and PD of analgesic agents and population PK/PD model in rodent was developed. To predict PD response in human, developed PK-PD model in animal was conducted by human PK modeling based on assumption what there was no change on the PK-PD relationship among inter-species. The PD evaluation in human was compared and evaluated by simulating the PD according to the scenario evaluated in the literature.
Results: The PK model of the drugs was established including active metabolite known to have pharmacological effects. In the PK study for rodent and human, the analgesic agent’s PK properties were described by a one compartmental model with first order absorption and elimination. The human population pharmacokinetic parameter estimated values for celecoxib (mean) were: ka = 0.129 h-1, Vc(Volume of distribution) = 0.466 L, CL(Clearance) = 0.0605 L/hr. The human population pharmacokinetic parameter estimated values for naproxen (mean) were: F(bioavailability) = 0.745, ka= 1.04 h-1, Vc,p (Volume of distribution for parent drug) = 257 L , Vc,m(Volume of distribution for metabolite) = 143, CL,p (Clearance for parent drug) = 29.4 L/hr, CL,m (Clearance for metabolite) = 414, Rpm (parent drug to metabolite ratio) = 0.259. The human population pharmacokinetic parameter estimated values for naproxen (mean) were: ka = 0.138 h-1, Vc = 0.27 L, CL = 0.0373 L/hr.
PD properties, tested by weight bearing, was characterized using the sigmoidal Emax model. EC50 was fixed with unboun fraction in plasma (Fup) and the half-maximal effective concentration, which had an analagesic effect in in vitro experiments. To establish a translational PK-PD model, population human PK model was integrated with rodent PD model. Weight bearing PD response in rodent was converted by human Visual Analogue Scales (VAS) for pain, assumed linear extrapolated relationships between them. As a result of evaluating PD prediction in humans, a model that explains observation well was established.
Conclusions:A translational PK-PD model for analgesics has been developed to predict the efficacy in human. Traditional PK-PD modeling showed its value by successfully bridging preclinical to clinical The limitation of this research is that drugs are limited, and PK/PD data are insufficient for each drug, so the predictive power of various doses may be reduced. As a further study, to overcome these limitations, more drug information is collected to improve the predictive power of the model.
References:
[1] [1] Danhof, Meindert et al. “Mechanism-based pharmacokinetic-pharmacodynamic (PK-PD) modeling in translational drug research.” Trends in pharmacological sciences vol. 29,4 (2008): 186-91. doi:10.1016/j.tips.2008.01.007
[2] Gabrielsson, J, and A R Green. “Quantitative pharmacology or pharmacokinetic pharmacodynamic integration should be a vital component in integrative pharmacology.” The Journal of pharmacology and experimental therapeutics vol. 331,3 (2009): 767-74. doi:10.1124/jpet.109.157172
[3] Gabrielsson, Johan et al. “Early integration of pharmacokinetic and dynamic reasoning is essential for optimal development of lead compounds: strategic considerations.” Drug discovery today vol. 14,7-8 (2009): 358-72. doi:10.1016/j.drudis.2008.12.011