Kota Toshimoto, Yuichi Sugiyama
Sugiyama Laboratory, RIKEN Innovation Center, RIKEN, Yokohama, Japan
Objectives:
This study shows the potential of virtual clinical study (VCS) to predict the effect of inter-individual differences on drug or metabolite exposures in blood and target organs leading to pharmacological and toxicological effects [1]. VCS is computational simulation with physiologically-based pharmacokinetic (PBPK), pharmacodynamic, and toxicodynamic models to generate virtual patients with the variability of physiological and pharmacokinetic parameters based on genetic polymorphism, ethnic differences, and inter- and intra-individual variability [2, 3]. Irinotecan is intravenously administered for various cancer treatment. SN-38 is an active metabolite of irinotecan and causes serious side effects including neutropenia and diarrhea. UGT1A1 polymorphism increases risk of the side effects because of higher SN-38 exposure due to decreased UGT1A1 metabolic activity, and similar results have been reported in different clinical studies [4]. Except for UGT1A1, there has been no clear association between the side effects of irinotecan and the effect of genetic polymorphism of other transporters including organic anion transporting polypeptide (OATP) 1B1, multiple drug resistance (MDR) 1, breast cancer resistance protein (BCRP), and multidrug resistance-associated protein (MRP) 2.
Methods:
Because there were many unknown parameters in the PBPK model, it was difficult to determine appropriate initial values for unknown parameters. Thus, conventional nonlinear least squares methods such as the Gauss–Newton method and Levenberg–Marquardt method could not be applied. To overcome this problem, Cluster Newton method (CNM) was newly introduced to optimize unknown parameters of PBPK model [7]. In CNM, multiple sets of initial values of parameters were generated at random from a certain range for each unknown parameter. Then, linear approximations of projection from parameters to objective function were used to determine the next iterations from initial sets. To perform VCS, not only the inter-individual variability for each physiological and physicochemical parameters of the PBPK model but also the genetic polymorphisms of enzymes and transporters were considered. A virtual patient was generated by performing Monte Carlo simulation given a frequency distribution (average and coefficient of variation) for each parameter and the activity ratio and the allele frequency for each genetic polymorphism obtained from experiments in vitro and in vivo [8-12]. VCS was performed to find whether the PBPK model that reproduced a blood concentration–time profile, could reproduce the results of a previously reported clinical study. The patients with high unbound plasma AUC were assumed to have neutropenia and those with high unbound enterocyte AUC were assumed to have diarrhea.
Results:
Using CNM, thePBPK model with approximately 30 sets of parameters could give good reproduction of the pharmacokinetics of irinotecan and its metabolites. The computational time for CNM optimization with the initial 10,000 sets of parameters was about 15 min using a workstation (CPU: Xeon E5-2640 v3 ´2; OS: CentOS 6.7 64 bit; RAM: 32 GB). This result shows that CNM is a powerful algorithm by which to find multiple sets of parameters for PBPK models quickly. The VCS confirmed that the genetic polymorphisms of UGT1A1 affected the SN-38 plasma concentration, and was associated with neutropenia. The VCS also indicated that “biliary index (= AUC(irinotecan) ´ AUC(SN-38) / AUC(SN-38 glucuronide)), [13]” is a better biomarker of diarrhea than the UGT1A1 polymorphism.
Conclusions:
The multiple sets of PBPK model parameters could reproduce the effects of genetic polymorphisms of UGT1A1 on the plasma concentration of SN-38 and side effects such as neutropenia and diarrhea using a VCS approach. To optimize the numerous biochemical parameters of irinotecan and its metabolites in the PBPK model, a CNM parameter optimization algorithm was introduced. The current VCS confirmed the importance of the biliary index as a better biomarker of irinotecan-induced diarrhea compared with only UGT1A1 polymorphism. In this study, VCS was performed to evaluate whether reported clinical studies could be reproduced by the PBPK model,in a “retrospective” approach. To examine the potential of VCS to apply it to the “prospective” prediction, VCS for another anti-cancer drug is on-going prior to the results of clinical study published.
References:
[1]. Toshimoto K, Tomaru A, Hosokawa M, Sugiyama Y. Virtual Clinical Studies to Examine the Probability Distribution of the AUC at Target Tissues Using Physiologically-Based Pharmacokinetic Modeling: Application to Analyses of the Effect of Genetic Polymorphism of Enzymes and Transporters on Irinotecan Induced Side Effects. Pharm Res. 2017:34(8):1584-1600.
[2]. Yamada A, Maeda K, Kiyotani K, Mushiroda T, Nakamura Y, Sugiyama Y. Kinetic Interpretation of the Importance of OATP1B3 and MRP2 in Docetaxel-Induced Hematopoietic Toxicity. CPT Pharmacometrics Syst Pharmacol. 2014;3:e126.
[3]. Ito M, Kusuhara H, Ose A, Kondo T, Tanabe K, Nakayama H, Horita S, Fujita T, Sugiyama Y. Pharmacokinetic Modeling and Monte Carlo Simulation to Predict Interindividual Variability in Human Exposure to Oseltamivir and Its Active Metabolite, Ro 64-0802. AAPS J. 2017;19(1):286-297.
[4]. Fijiwara Y, Minami H. An overview of the recent progress in irinotecan pharmacogenetics. Pharmacogenomics. 2010;11(3):391-406.
[5]. Rhodes KE, Zhang W, Yang D, Press OA, Gordon M, Vallböhmer D, Schultheis AM, Lurje G, Ladner RD, Fazzone W, Iqbal S, Lenz HJ. ABCB1, SLCO1B1 and UGT1A1 gene polymorphisms are associated with toxicity in metastatic colorectal cancer patients treated with first-line irinotecan. Drug Metab Lett. 2007;1(1):23-30.
[6]. Lara PN Jr, Natale R, Crowley J, Lenz HJ, Redman MW, Carleton JE, Jett J, Langer CJ, Kuebler JP, Dakhil SR, Chansky K, Gandara DR. Phase III trial of irinotecan/cisplatin compared with etoposide/cisplatin in extensive-stage small-cell lung cancer: clinical and pharmacogenomic results from SWOG S0124. J Clin Oncol. 2009;27(15):2530-5.
[7]. Aoki Y, Hayami K, Sterck HD, Konagaya A. Cluster Newton Method for Sampling Multiple Solutions of Underdetermined Inverse Problems: Application to a Parameter Identification Problem in Pharmacokinetics. SIAM J. Sci. Comput. 2014;36:B14-B44.
[8]. Wynne HA, Cope LH, Mutch E, Rawlins MD, Woodhouse KW, James OF. The effect of age upon liver volume and apparent liver blood flow in healthy man. Hepatology. 1989;9(2):297-301.
[9]. Berg UB. Differences in decline in GFR with age between males and females. Reference data on clearances of inulin and PAH in potential kidney donors. Nephrol Dial Transplant. 2006;21(9):2577-82.
[10]. Lin YH, Shen TY, Chang CA. Reduction of the interferences of biochemicals and hematocrit ratio on the determination of whole blood glucose using multiple screen-printed carbon electrode test strips. Anal Bioanal Chem. 2007;389(5):1623-31.
[11]. Kato M, Chiba K, Ito T, Koue T, Sugiyama Y. Prediction of interindividual variability in pharmacokinetics for CYP3A4 substrates in humans. Drug Metab Pharmacokinet. 2010;25(4):367-78.
[12]. Sai K, Kaniwa N, Itoda M, Saito Y, Hasegawa R, Komamura K, Ueno K, Kamakura S, Kitakaze M, Shirao K, Minami H, Ohtsu A, Yoshida T, Saijo N, Kitamura Y, Kamatani N, Ozawa S, Sawada J. Haplotype analysis of ABCB1/MDR1 blocks in a Japanese population reveals genotype-dependent renal clearance of irinotecan. Pharmacogenetics. 2003;13(12):741-57.
[13]. Gupta E, Lestingi TM, Mick R, Ramirez J, Vokes EE, Ratain MJ. Metabolic fate of irinotecan in humans: correlation of glucuronidation with diarrhea. Cancer Res. 1994;54(14):3723-5.
Reference: PAGE 27 (2018) Abstr 8575 [www.page-meeting.org/?abstract=8575]
Poster: Drug/Disease Modelling - Absorption & PBPK