Eur J Clin Pharmacol (2015) 71:617–624 DOI 10.1007/s00228-015-1834-y
PHARMACOKINETICS AND DISPOSITION
Evaluating a physiologically based pharmacokinetic model for prediction of omeprazole clearance and assessing ethnic sensitivity in CYP2C19 metabolic pathway Sheng Feng 1 & Yumi Cleary 2 & Neil Parrott 2 & Pei Hu 3 & Cornelia Weber 2 & Yongqing Wang 4 & Ophelia Q. P. Yin 5 & Jun Shi 1
Received: 30 November 2014 / Accepted: 10 March 2015 / Published online: 24 March 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose The purpose of this study is to evaluate the ethnicityspecific population models in the SimCYP Simulator® for prediction of omeprazole clearance with attention to differences in the CYP2C19 metabolic pathway. Methods The SimCYP® models incorporating Caucasian, Chinese, and Japanese population-specific demographic, physiological, and enzyme data were applied to simulate omeprazole pharmacokinetics. Published pharmacokinetic data of omeprazole after intravenous or oral administration in Caucasian, Chinese, and Japanese were used for the evaluation. Results Following oral administration, the ratio of the predicted to observed geometric mean of omeprazole clearance in Caucasian extensive metabolizers (EMs) was 0.88. The ratios in Chinese EMs were 1.16 and 0.99 after intravenous and oral administration, respectively. The ratios in Japanese EMs were 0.88 and 0.71 after intravenous and oral administration, respectively. Electronic supplementary material The online version of this article (doi:10.1007/s00228-015-1834-y) contains supplementary material, which is available to authorized users. * Jun Shi [email protected] 1
Roche Innovation Center Shanghai, Building 6, Lane 917, Ha Lei Road, Pudong, Shanghai, China
2
Roche Innovation Center Basel, Basel, Switzerland
3
Peking Union Medical College Hospital, Beijing, China
4
The First Affiliated Hospital, Nanjing Medical University, Nanjing, China
5
Chinese University of Hong Kong, Hong Kong, People’s Republic of China
Significant differences (2-fold) in the observed oral clearance of omeprazole were identified between Caucasian and Asian (Chinese and Japanese) EMs while the observed oral and intravenous clearances of omeprazole were similar between Chinese and Japanese EMs. Physiologically based pharmacokinetics (PBPK) models within SimCYP accurately predicted the difference in the observed oral clearance between Caucasian and Chinese EMs but overpredicted the difference between Caucasians and Japanese EMs due to underprediction of oral clearance in Japanese EMs. Conclusions The PBPK model within SimCYP adequately predicted omeprazole clearance in Caucasian, Chinese, and Japanese EMs and the 2-fold differences in clearance of omeprazole between Caucasian and Asian EMs. This may lead to early identification of ethnic sensitivity in clearance and the need for different dosing regimens in a specific ethnic group for substrates of CYP2C19 which can support the rational design of bridging clinical trials.
Keywords PBPK . CYP2C19 . Omeprazole . Ethnic sensitivity
Introduction In recent years, there has been a dramatic increase in the number of global clinical trials [1, 2]. This trend offers opportunities for cost saving and recruitment acceleration as well as minimizing duplication of clinical data and shortening the drug approval gap among regions. To ensure successful global drug development, ethnicity-related factors on PK/PD, medical practice, and disease biology must be taken into consideration [3]. Physiologically based pharmacokinetics (PBPK) modeling is one of the key components of model-based drug development and is increasingly embraced within the industry and regulatory
618
authorities [4–6]. PBPK modeling can be a useful tool to simulate PK in various ethnic groups and may be used to assess ethnic sensitivity. Currently, commercially available PBPK modeling software tools use parameters gathered mainly in Caucasians while the population data in Chinese and Japanese is less complete [7, 8]. For instance, some physiological data, such as milligrams of microsomal protein per gram of liver (MPPGL), in Chinese and Japanese are not available, and so, the Caucasian data has to be used for simulations in these populations. Even the existing physiological data of Chinese and Japanese are often from limited sources, for example, the hepatic CYP2C19 abundance in the liver was only from 20 Chinese livers and 31 Japanese livers, respectively [8, 9]. Therefore, to facilitate prospective application of PBPK models for drug development and bridging studies, more model system data are needed and the models must be evaluated for the Chinese and Japanese populations. Omeprazole is a proton pump inhibitor (PPI) used in the treatment of gastroesophageal reflux and other diseases, which are caused by the excess secretion of stomach acid. Due to its safety and sensitivity, omeprazole is recommended as a probe drug of CYP2C19 by the FDA and others [10–12]. CYP2C19 is well known to manifest polymorphism, and while the differences in frequency of extensive metabolizers (EMs) and poor metabolizers (PMs) in the Caucasian and Asian population are well documented, the differences in activity still need to be investigated [13, 14]. CYP2C19*2 and CYP2C19*3 are the most common variants. Both are null alleles and have no activity in vivo, thus no activity in PMs [14, 15]. Therefore, we used data from EMs to investigate the CYP2C19 activity differences in the three ethnicity groups. The objective of the study was to compare observed and SimCYP predicted clearance of omeprazole in Caucasian, Chinese, and Japanese EMs and examine if ethnic differences in the CYP2C19 metabolic pathway were well captured.
Methods
Eur J Clin Pharmacol (2015) 71:617–624
inactivated by gastric acid, enteric-coated or delayed release formations are often used. To decrease the effect of formulation differences, the default omeprazole model for an entericcoated tablet was selected for model evaluation and was compared to clinical studies with this formulation. For distribution, a minimal PBPK model with one compartment was applied. The default value of volume of distribution at steady state (Vss) for omeprazole model in SimCYP was 0.15 L/kg. However, the Vss was reset to 0.23, 0.249, and 0.219 L/kg for Caucasians, Chinese, and Japanese, respectively [18–20]. The Vss for Caucasians and Japanese was the values reported in the literature while the Vss for Chinese was the mean of parameter estimates based on the individual concentration data. The SimCYP omeprazole model default values of intrinsic clearance for hepatic metabolic clearance due to CYP2C19 and CYP3A4 were used, and all other drug-specific parameter models were identical across the simulated populations.
Virtual population The main differences in the inputs for the three virtual populations are summarized in Table 1 (see also Online Resource 2). The demographic databases of Caucasians, Chinese, and Japanese were as provided in SimCYP. The intrinsic catalytic activity of CYPs 2C19 and 3A4 per unit amount of enzyme variant and tissue composition were assumed to be the same in Caucasians, Chinese, and Japanese. The default mean abundances of CYP2C19 in the liver in Caucasian and Chinese and Japanese EMs were 14, 8, and 1 pmol/mg in SimCYP, respectively. The abundance in Japanese EMs was adjusted from 1 to 4.77 pmol/mg based on the reported 2.9-fold difference in CYP2C19 abundance between Caucasian and Japanese EMs [21]. Table 1 Summary of main differences in the inputs for the three virtual populations for the omeprazole clearance prediction (default values in SimCYP) Parameters
General Simulations were performed using the SimCYP Populationbased simulator®, version 14 (SimCYP Ltd, Sheffield, UK). The omeprazole model file and virtual populations for Caucasians, Chinese, and Japanese as provided in the software were used. Omeprazole model The input parameters for the omeprazole model [16] are summarized in the Online Resource 1. The omeprazole model used the advanced dissolution, absorption, and metabolism (ADAM) model for absorption [17]. As omeprazole is partly
Caucasians
Chinese
CYP2C19 abundance in gastrointestinal tract (pmol/mg) Extensive metabolizers 1.5 0.85 Poor metabolizers 0 0 CYP3A4 abundance in 66.2 58 gastrointestinal tract (pmol/mg) CYP2C19 abundance in liver (pmol/mg) Extensive metabolizers 14 8 Poor metabolizers 0 0 CYP3A4 abundance in liver 137 120 (pmol/mg) Average liver weight (L) 1.65056 1.402976 a
Japanese
0.511a 0 59
4.77a 0 122 1.65056
It was adjusted from the default value based on the study reported by Inoue et al. [21]
Eur J Clin Pharmacol (2015) 71:617–624
619
Simulation Simulations were conducted using virtual populations generated by matching distributions of age, sex, dosage, and phenotype (EM or PM) to the actual study data. The simulations followed the reported study designs and conditions as closely as possible including numbers of subjects and dosing regimens, and ten trial simulations were performed for each study.
Clinical data Clinical pharmacokinetic data for omeprazole following intravenous and oral administration in Caucasian, Chinese, and Japanese subjects were searched in the PubMed online database using omeprazole and pharmacokinetics as the search terms. Reports were excluded if they did not give phenotype information or reported clearance and area under the plasma concentration-time curve (AUC) that could not be used to calculate the geometric mean of clearance. Furthermore, only studies using enteric-coated or delayed release formulations were considered. Eight papers (three for Caucasians, two for Chinese, and three for Japanese) were considered eligible for the study [19, 20, 22–27]. Whereas both intravenous and oral studies were found for Chinese and Japanese, only studies with an oral administration of omeprazole were found for Caucasians. Lastly, the authors of the two Chinese papers [19, 25] kindly provided us with the individual data.
Population PK parameters are generally considered to be log normally distributed. Hence, values of geometric mean (GM) and 95 % confidence intervals (CIs) were used to compare the clearance. GMs and 95 % CIs were calculated from population mean of μ and standard deviation of σ, the natural logs of GM and geometric standard deviation, using Eqs. 1–4. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi σ ¼ Ln 1 þ CV2 ð1Þ μ ¼ LnðMeanÞ−0:5 σ2 ð2Þ
95% CIs ¼ exp
pffiffiffiffiffiσ μ 1:96 n−1
When data from more than one study or group were available for the same route of administration in the same ethnicity, the pooled geometric mean (PGM) was calculated using Eq. 7. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑n nj j¼1 PGM ¼ GM1n1 GM2n2 ⋯ GMj n j ð7Þ For those papers only reporting the mean and SD values of AUC0–inf [20, 23, 24] or AUC0–24 h [27], firstly, the GM of AUC was calculated using the Eqs. 1–3. Then, the GM of clearance was calculated by dose / AUC. The clearance was calculated from a non-compartment analysis using Phoenix software (WinNonlin models version 6.2, Pharsight Corp, Cary, NC, USA) if the individual data could be obtained, including observed data [19, 25] and SimCYP simulated data. The figures were drawn using R software (version 3.1.0). The mean concentration–time data was extracted with the Plot Digitizer software (version 2.6.3).
Results Subject characteristics The characteristics of the observed subjects including ethnicity, phenotype, the number of the observed subjects, and demographic data can be seen in Table 2.
Statistical analysis
GM ¼ expμ
When the mean and 95 % CI were reported, SD was calculated using the Eq. 6. pffiffiffi ðUpper Limit−Lower LimitÞ n ð6Þ SD ¼ 2 1:96
ð3Þ ð4Þ
where n=the number of the subjects in the study. Coefficient of variation (CV) was calculated from reported mean and standard deviation (SD) values in the published studies using Eq. 5. Standard deviation CV ¼ Mean
ð5Þ
Prediction of clearance in Caucasian EMs Plasma concentration–time profiles of omeprazole after 20 mg of oral administration in Caucasian EMs were available from two studies [23, 24]. Simulated and observed plasma concentration–time curves were compared for the ten virtual trials in Fig. 1a, b (n=9 and n=12, respectively). The plasma concentration–time profiles of omeprazole were generally well recovered for both studies. The ratios of predicted to the observed GM of CLPO in the entire virtual population of each study were 1.13, 0.79, and 0.75 (Table 2) and within 0.8–1.25 in 5, 4, and 5 out of the 10 trials for the studies reported by Baldwin et al. [22], Shirasaka et al. [23], and Andersson et al. [24], respectively (as shown in Online Resource 3). Prediction of clearance in Chinese EMs Simulated and observed plasma concentration–time curves after i.v. and oral administration of 40 mg omeprazole in Chinese EMs were compared for the ten virtual trials in
620 Table 2 Ethnicity
Eur J Clin Pharmacol (2015) 71:617–624 Details of the in vivo studies of observed and predicted geometric mean clearances in healthy Caucasian, Chinese, and Japanese subjects (EM) Phenotype n
Caucasian EM EM EM Chinese Japanese
Weight (kg)
Age (years)
11 69 (60–109) 33 (21–60) 9 Not given 12 68–86
Dose (mg) Route Observed CLa (L/h) Predicted CLa (L/h) Ratiob Reference 40
p.o.
35.1 (21.5–57.3)
39.6 (33.6–46.6)
1.13
Baldwin et al. [22]
18–50
20
p.o.
53.9 (38.9–74.8)
42.6 (35.7–50.9)
0.79
Shirasaka et al. [23]
22–32
20
p.o.
58.5 (44.2–78.2)
43.8 (38.0–50.4)
0.75
Andersson et al. [24]
EM
16 61.9 (51–72) 26 (22–31)
40
i.v.
11.8 (9.69–14.3)
13.7 (13.0–14.3)
1.16
Wang et al. [19]
EM
19 61.7 (54–71) 21 (20–24)
40
p.o.
23.7 (18.5–30.5)
23.4 (22.1–24.9)
0.99
Yin et al. [25]
EM
17 57.7
20
i.v.
13.8
10.5 (9.64–11.3)
0.76
Ishizawa et al. [20]
25
EM
14 63.0 (42–90) 28.6 (21–44) 20
i.v.
10.7
11.2 (10.2–12.2)
1.05
Uno et al. [26]
EM EM
14 63.0 (42–90) 28.6 (21–44) 40 11 43–72 22–38 20
p.o. p.o.
19.3 30.0
17.4 (15.4–19.7) 15.7 (13.6–18.2)
0.90 0.52
Shirai et al. [27]
Data are shown as geometric mean values (95 % CI in brackets) for clearance and as arithmetic mean values (ranges in brackets, if reported) for weight and age EM extensive metabolizers, n number of the subjects, CL clearance (which is calculated as dose / AUC0–inf; in the study of Shirai, it is calculated as dose / AUC0–24 h), p.o. oral, i.v. intravenous a
CLPO for oral administration, CLPO: apparent clearance after oral administration
b
Ratios of predicted to observed CL
Fig. 1c, d, respectively. The simulated plasma concentration– time curves were in a good agreement with the observed data. The ratios of predicted to observed GM CLIV and CLPO were 1.16 and 0.99 (Table 2), and 7 and 10 out of 10 trials had the ratio within 0.8–1.25. These simulations demonstrated that the Chinese population model was able to recover CLIV and CLPO of omeprazole very well.
Prediction of clearance in Japanese EMs For Japanese, two papers were found for each intravenous and oral administration. Simulated and observed plasma concentration–time curves after 20-mg omeprazole intravenous administration were compared for the ten virtual trials in Fig. 1e, f (n=17 and n=14, respectively). The ratios of the simulated to the observed GM of CLIV in the entire virtual population were 0.76 and 1.05 (Table 2) and within 0.8–1.25 in 3 and 9 out of the 10 trials for the studies reported by Ishizawa et al. [20] and Uno et al. [26], respectively. Simulated and observed plasma concentration–time curves after oral administration of omeprazole in Japanese EMs were compared for the ten virtual trials in Fig. 1g, h (40 mg, n=14 and 20 mg, n=12, respectively). The ratios of the simulated to the observed GM of omeprazole CLPO in the entire virtual population were 0.90 and 0.52 (Table 2) and within 0.8–1.25 in 7 and 0 out of the 10 trials for the studies reported by Uno et al. [26] and Shirai et al. [27], respectively. The predicted values of omeprazole clearance in all simulated trials for all ethnic groups are shown in Table 2. Summarized data for each ethnicity are given in Table 3. The ratios of predicted to observed GM of CL PO in Caucasian EMs and CLPO and CLIV in Chinese EMs were
0.88, 1.16, and 0.99, respectively. This showed good accuracy of predictions for Caucasian and Chinese EMs. However, the ratios of predicted to observed GM of CLIV and CLPO in Japanese EMs were 0.88 and 0.71, respectively, demonstrating a tendency to under-prediction for CLPO in Japanese EMs. Prediction of clearance in Caucasian, Chinese, and Japanese PMs Although CYP2C19 has no activity in PMs, the clearance of omeprazole in PMs was also simulated. The results can be seen in Online Resource 4. In Caucasians, the ratio of predicted to observed GM of CLPO was 1.53 for the studies reported by Andersson et al. [24]. In Chinese, the ratios of predicted to observed GM of CLIV and CLPO were 0.50 and 0.77 for the studies reported by Wang et al. [19] and Yin et al. [25], respectively. In Japanese, the ratios of predicted to observed GM of CLIV were 1.34 and 1.42 for the studies reported by Ishizawa et al. [20] and Uno et al. [26], respectively. The ratios of predicted to observed GM of CLPO were 1.36 and 1.31 for the studies reported by Uno et al. [26] and Shirai et al. [27], respectively. Caucasian, Chinese, and Japanese EM clearance differences The differences in observed and predicted GM omeprazole clearance in Caucasian, Chinese, and Japanese EMs when combining all data across studies can be seen in Table 3. The observed pooled CLPO in Caucasians was 2.03-fold and 2.05-fold higher than the values in Chinese and in Japanese, respectively. The ratios of observed pooled CL (i.v. and oral)
Eur J Clin Pharmacol (2015) 71:617–624
621
Fig. 1 Simulated (lines) and observed (circles) mean plasma concentration–time curves of omeprazole in healthy Caucasian (a, b), Chinese (c, d), and Japanese (e–h) CYP2C19 extensive metabolizers. The gray lines represent the mean of simulations of individual trials (ten trials). The black line represents the mean for the total virtual population. The dashed lines represent the 5th and 95th percentiles for the total virtual population
between Chinese and Japanese were 0.96 and 1.01, respectively. The pooled predicted CLPO in Caucasians was 1.79-fold higher than the value in Chinese, which was slightly lower than the observed difference (2.03-fold). The pooled predicted CLPO in Caucasians was 2.53-fold higher than the value in Japanese, which was higher than the observed difference (2.05-fold). The pooled predicted CLIV and CLPO in Chinese were 1.27- and 1.41-fold higher than the values in Japanese. These simulations suggested that the differences in CLPO between Chinese and Caucasian EMs were well predicted while the differences in CL PO between Japanese and Caucasian EMs were overpredicted due to under-prediction of CLPO in Japanese.
Discussion Known differences in physiological parameters between Caucasians, Chinese, and Japanese have been implemented in several commercial PBPK tools, and the application of physiologically based methods to compare CYP-mediated clearance between Caucasians and Chinese using SimCYP has been previously reported by Barter et al. [7]. They reported that the predicted omeprazole CLPO in Chinese was approximately 2-fold higher than the observed CLPO reported for the study by Hu et al. [28]. In our work, we have attempted to increase the number of studies in the clinical dataset while keeping consistency in study condition (e.g., formulation) and have also extended the dataset to include intravenous PK data
622
Eur J Clin Pharmacol (2015) 71:617–624
Table 3 Summary of observed and predicted geometric mean omeprazole clearances in healthy Caucasian, Chinese, and Japanese subjects (EM, pooled) Ethnicity
Phenotype
n
Route
Observed CLa (L/h)
Predicted CLa (L/h)
Ratiob
Caucasian Chinese
EM EM EM EM EM
32 16 19 31 25
p.o. i.v. p.o. i.v. p.o.
48.0 11.8 23.7 12.3 23.4
42.0 13.7 23.4 10.8 16.6
0.88 1.16 0.99 0.88 0.71
Japanese
EM extensive metabolizers, n number of the subjects, CL clearance (which is calculated as dose / AUC), p.o. oral, i.v. intravenous a
CLPO for oral administration, CLPO: apparent clearance after oral administration
b
Ratios of predicted to observed CL
using individual data kindly provided by the authors of two Chinese papers [19, 25]. Based on in vitro studies with human liver microsomes [29], the difference in intrinsic clearance between Caucasians and Chinese was larger for CYP2C19 than for other CYPs (CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2D6, CYP2E1, and CYP3A). The median CLint value for (S)-mephenytoin 4′-hydroxylation (a probe drug in vitro for CYP2C19) in Chinese liver microsomes was 0.43 μL/min/mg, which was 26 % of the corresponding value in Caucasians (1.64 μL/min/mg) [29]. Our work showed that this in vitro difference was reflected in vivo. Although comparison between different studies was difficult due to different conditions and formulations, the clearance of CYP2C19 substrates in Caucasian CYP2C19 EMs seemed to be approximately 2-fold higher than that in Asian EMs (Chinese, Japanese). The reported GM CLPO of omeprazole in Korean EMs (20.5 L/h (n=21) and 25.4 L/h (n=6)) [30, 31] was also very similar to those values in Chinese and Japanese EMs. For the Japanese population, the default value for the CYP2C19 abundance in the liver is 1 pmol/mL in SimCYP [8, 32]. Using this value, the ratios of predicted vs observed clearance were 0.41 (intravenous, Ishizawa et al. [20]), 0.57 (intravenous, Uno et al. [26]), 0.41 (oral, Uno et al. [26]), and 0.24 (oral, Shirai et al. [27]), which were significantly underpredicted. Inoue et al. [21] reported that the mean abundance values of CYP2C19 in Japanese and Caucasian EM liver microsomes were 1.521 and 4.465 pmol/mL, respectively. When this 3-fold difference in CYP2C19 abundance in the liver between Japanese and Caucasian EMs was considered in the Japanese model, the abundances could be adjusted to 4.77 and 0.511 pmol/mL in the liver and gastrointestinal tract, respectively, based on the default values in SimCYP for the Caucasian EMs which were 14 and 1.5 pmol/mL in the liver and gastrointestinal tract, respectively. Using these modified
CYP2C19 abundances, the recovery of clearance in Japanese was better with ratios of predicted to observed clearance of 0.76 (intravenous, Ishizawa et al. [20]), 1.05 (intravenous, Uno et al. [26]), 0.90 (oral, Uno et al. [26]), and 0.52 (oral, Shirai et al. [27]). This indicates that the abundance of CYP2C19 in Japanese requires more investigation. Different liver weight values in Chinese have been reported [33], and the reported mean value is only 9 % higher than the value currently used in SimCYP [34]. As regard the cardiac output and hepatic blood flow, the reported values are from the Sichuan province of China, which is known for small body size and is not representative of the whole Chinese population. Also, omeprazole has a low hepatic extraction ratio, and so, clearance in Chinese is not significantly affected by hepatic blood flow. EMs can be divided into homozygous EMs (wt/wt) and heterozygous EMs (wt/m1 or wt/m2) based on the genotypes. The difference in clearance of omeprazole was significant between homozygous EMs and heterozygous EMs. For example, the CLPO in homozygous EMs was almost 2-fold higher than in heterozygous EMs in Japanese, though the difference became smaller when omeprazole was given intravenously [26]. However, the models in SimCYP have not distinguished these two groups. One of the reasons may be that the difference between homozygous EMs and heterozygous EMs was small compared to the difference between EMs and PMs. Another reason may be due to sparse abundance data for each genotype. Therefore, homozygous EMs and heterozygous EMs were pooled as general EMs for the simulations in this study although they were separated into two groups in the observed studies. Although EMs data was used to investigate the CYP2C19 activity, the PMs were also simulated and showed ratios of predicted to observed GM of CL (i.v. and oral) all within 2-fold (Online Resource 4). Omeprazole is reported to exhibit a dose- and timedependent nonlinear pharmacokinetics [35]. Although omeprazole is a time-dependent inhibitor of CYP2C19 and CYP3A4, it is the metabolites of omeprazole which are believed to be the major cause for the nonlinear pharmacokinetics [36–41]. The clearance model in SimCYP was modified by inputting the omeprazole time-dependent inhibition parameters [27], and the effect of auto-inhibition of omeprazole was found to be very limited. Therefore, the modification was not implemented in the final omeprazole clearance model. In summary, the omeprazole clearance prediction using Caucasian, Chinese, and Japanese population models in SimCYP was within 2-fold of observations in Caucasian, Chinese, and Japanese after adjusting the CYP2C19 abundance in Japanese EMs. This indicated that the virtual Caucasian, Chinese, and Japanese models in SimCYP can be used to predict the clearance of substrates of CYP2C19. We also used the PBPK method to assess the differences in oral and intravenous clearance of omeprazole in Caucasians,
Eur J Clin Pharmacol (2015) 71:617–624
623
Chinese, and Japanese. The results showed that PBPK modeling could reasonably recover the ethnic sensitivity. This may lead to early identification of ethnic sensitivity in clearance and the need for different dosing regimens in a specific ethnic group for substrates of CYP2C19 which can support the rational design of bridging clinical trials.
11.
Acknowledgments The study was supported by Roche Postdoc Fellowship Program.
13.
Conflict of interest All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare the following: S.F., Y.C., N.P., P.H., C.W., and J.S. had support from F. Hoffmann-La Roche for the submitted work; Y.C., N.P., C.W., and J.S. are full-time employees of F. Hoffmann-La Roche; P.H. is a full-time employee of Peking Union Medical College Hospital; S.F. is a Roche Postdoctoral Fellow; no financial relationships with any organization that might have an interest in the submitted work in the previous 3 years; and no other relationships or activities that could appear to have influenced the submitted work.
14.
Author contribution Sheng Feng and Yumi Cleary performed the analysis. All the authors contributed to the discussion of results. Sheng Feng wrote the first draft. Yongqing Wang and Ophelia Q. P. Yin kindly provided the individual demographic and concentration–time data for the two Chinese studies. All the authors reviewed the data and approved the final version for submission.
12.
15.
16.
17.
18.
19.
References 1. 2. 3. 4.
5.
6.
7.
8.
9.
10.
Thiers FA, Sinskey AJ, Berndt ER (2008) Trends in the globalization of clinical trials. Nat Rev Drug Discov 7:13–14 Karlberg JPE (2009) Globalization of sponsored clinical trials. Nat Rev Drug Discov 7:458–460 ICH Topic E 5 (R1) Ethnic factors in the acceptability of foreign clinical data. Zhao P, Rowland M, Huang SM (2012) Best practice in the use of physiologically based pharmacokinetic modeling and simulation to address clinical pharmacology regulatory questions. Clin Pharmacol Ther 92:17–20 Rowland M, Peck C, Tucker G (2011) Physiologically-based pharmacokinetics in drug development and regulatory science. Annu Rev Pharmacol Toxicol 51:45–73 Huang SM, Rowland M (2012) The role of physiologically based pharmacokinetic modeling in regulatory review. Clin Pharmacol Ther 91(3):542–549 Barter ZE, Tucker GT, Rowland-Yeo K (2013) Differences in cytochrome p450-mediated pharmacokinetics between Chinese and Caucasian populations predicted by mechanistic physiologically based pharmacokinetic modeling. Clin Pharmacokinet 52:1085– 1100 Inoue S, Howgate EM, Rowland-Yeo K, Shimada T, Yamazaki H, Tucker GT, Rostami-Hodjegan A (2006) Prediction of in vivo drug clearance from in vitro data. II: potential inter-ethnic differences. Xenobiotica 36(6):499–513 Shu Y, Wang LS, Xiao WM, Wang W, Huang SL, Zhou HH (2000) Probing CYP2C19 and CYP3A4 activities in Chinese liver microsomes by quantification of 5-hydroxyomeprazole and omeprazole sulphone. Acta Pharmacol Sin 21(8):753–758 FDA (2012) Drug interaction studies—study design, data analysis, implications for dosing, and labeling recommendations (draft).
20.
21.
22.
23.
24.
25.
26.
27.
Chang M, Dahl ML, Tybring G, Götharson E, Bertilsson L (1995) Use of omeprazole as a probe drug for CYP2C19 phenotype in Swedish Caucasians: comparison with S-mephenytoin hydroxylation phenotype and CYP2C19 genotype. Pharmacogenetics 5: 358–363 Chang M, Tybring G, Dahl ML, Götharson E, Sagar M, Seensalu R, Bertilsson L (1995) Interphenotype differences in disposition and effect on gastrin levels of omeprazole—suitability of omeprazole as a probe for CYP2C19. Br J Clin Pharmacol 39:511–518 Desta Z, Zhao X, Shin JG, Flockhart DA (2002) Clinical significance of the cytochrome P450 2C19 genetic polymorphism. Clin Pharmacokinet 4:913–958 Zanger UM, Schwab M (2013) Cytochrome P450 enzymes in drug metabolism: regulation of gene expression, enzyme activities, and impact of genetic variation. Pharmacol Ther 138:103–141 Nakamoto K, Kidd JR, Jenison RD, Klaassen CD, Wan YJ, Kidd KK, Zhong XB (2007) Genotyping and haplotyping of CYP2C19 functional alleles on thin-film biosensor chips. Pharmacogenet Genomics 17:103–114 Wu F, Gaohua L, Zhao P, Jamei M, Huang SM, Bashaw ED, Lee SC (2014) Predicting nonlinear pharmacokinetics of omeprazole enantiomers and racemic drug using physiologically based pharmacokinetic modeling and simulation: application to predict drug/ genetic interactions. Pharm Res 31(8):1919–1929 Jamei M, Turner D, Yang J, Neuhoff S, Polak S, Rostami-Hodjegan A, Tucker G (2009) Population-based mechanistic prediction of oral drug absorption. AAPS J 11:225–237 Regårdh CG, Andersson T, Lagerström PO, Lundborg P, Skånberg I (1990) The pharmacokinetics of omeprazole in humans—a study of single intravenous and oral doses. Ther Drug Monit 12:163–172 Wang Y, Zhang H, Meng L, Wang M, Yuan H, Ou N, Zhang H, Li Z, Shi R (2010) Influence of CYP2C19 on the relationship between pharmacokinetics and intragastric pH of omeprazole administered by successive intravenous infusions in Chinese healthy volunteers. Eur J Clin Pharmacol 66:563–569 Ishizawa Y, Yasui-Furukori N, Takahata T, Sasaki M, Tateishi T (2005) The effect of aging on the relationship between the cytochrome P450 2C19 genotype and omeprazole pharmacokinetics. Clin Pharmacokinet 44:1179–1189 Inoue K, Yamazaki H, Imiya K, Akasaka S, Guengerich FP, Shimada T (1997) Relationship between CYP2C9 and 2C19 genotypes and tolbutamide methyl hydroxylation and S-mephenytoin 4′-hydroxylation activities in livers of Japanese and Caucasian populations. Pharmacogenetics 7:103–113 Baldwin RM, Ohlsson S, Pedersen RS, Mwinyi J, IngelmanSundberg M, Eliasson E, Bertilsson L (2008) Increased omeprazole metabolism in carriers of the CYP2C19*17 allele; a pharmacokinetic study in healthy volunteers. Br J Clin Pharmacol 65:767–774 Shirasaka Y, Sager JE, Lutz JD, Davis C, Isoherranen N (2013) Inhibition of CYP2C19 and CYP3A4 by omeprazole metabolites and their contribution to drug-drug interactions. Drug Metab Dispos 41:1414–1424 Andersson T, Holmberg J, Röhss K, Walan A (1998) Pharmacokinetics and effect on caffeine metabolism of the proton pump inhibitors, omeprazole, lansoprazole, and pantoprazole. Br J Clin Pharmacol 45:369–375 Yin OQ, Tomlinson B, Chow AH, Waye MM, Chow MS (2004) Omeprazole as a CYP2C19 marker in Chinese subjects: assessment of its gene-dose effect and intrasubject variability. J Clin Pharmacol 44:582–589 Uno T, Niioka T, Hayakari M, Yasui-Furukori N, Sugawara K, Tateishi T (2007) Absolute bioavailability and metabolism of omeprazole in relation to CYP2C19 genotypes following single intravenous and oral administrations. Eur J Clin Pharmacol 63:143–149 Shirai N, Furuta T, Moriyama Y, Okochi H, Kobayashi K, Takashima M, Xiao F, Kosuge K, Nakagawa K, Hanai H, Chiba
624
28.
29.
30.
31.
32.
33.
34.
Eur J Clin Pharmacol (2015) 71:617–624 K, Ohashi K, Ishizaki T (2001) Effects of CYP2C19 genotypic differences in the metabolism of omeprazole and rabeprazole on intragastric pH. Aliment Pharmacol Ther 15:1929–1937 Hu XP, Xu JM, Hu YM, Mei Q, Xu XH (2007) Effects of CYP2C19 genetic polymorphism on the pharmacokinetics and pharmacodynamics of omeprazole in Chinese people. J Clin Pharm Ther 32:517–524 Yang J, He MM, Niu W, Wrighton SA, Li L, Liu Y, Li C (2012) Metabolic capabilities of cytochrome P450 enzymes in Chinese liver microsomes compared with those in Caucasian liver microsomes. Br J Clin Pharmacol 73:268–284 Jin SK, Kang TS, Eom SO, Kim JI, Lee HJ, Roh J (2009) CYP2C19 haplotypes in Koreans as a marker of enzyme activity evaluated with omeprazole. J Clin Pharm Ther 34:437–446 Cho JY, Yu KS, Jang IJ, Yang BH, Shin SG, Yim DS (2002) Omeprazole hydroxylation is inhibited by a single dose of moclobemide in homozygotic EM genotype for CYP2C19. Br J Clin Pharmacol 53:393–397 Yasumori T, Murayama N, Yamazoe Y, Kato R (1990) Polymorphism in hydroxylation of mephenytoin and hexobarbital stereoisomers in relation to hepatic P-450 human-2. Clin Pharmacol Ther 47:313–322 Li GF, Yu G, Liu HX, Zheng QS (2014) Ethnic-specific in vitro-in vivo extrapolation and physiologically based pharmacokinetic approaches to predict cytochrome P450-mediated pharmacokinetics in the Chinese population: opportunities and challenges. Clin Pharmacokinet 53:197–202 Barter ZE, Tucker GT, Rowland-Yeo K (2014) Response to BEthnic-specific in vitro-in vivo extrapolation and physiologically based pharmacokinetic approaches to predict cytochrome P450mediated pharmacokinetics in Chinese population: opportunities and challenges^. Clin Pharmacokinet 53:203
35.
Regårdh CG (1986) Pharmacokinetics and metabolism of omeprazole in man. Scand J Gastroenterol Suppl 118:99–104 36. Soons PA, van den Berg G, Danhof M, van Brummelen P, Jansen JB, Lamers CB, Breimer DD (1992) Influence of single- and multiple-dose omeprazole treatment on nifedipine pharmacokinetics and effects in healthy subjects. Eur J Clin Pharmacol 42:319– 324 37. Funck-Brentano C, Becquemont L, Lenevu A, Roux A, Jaillon P, Beaune P (1997) Inhibition by omeprazole of proguanil metabolism: mechanism of the interaction in vitro and prediction of in vivo results from the in vitro experiments. J Pharmacol Exp Ther 280: 730–738 38. Yu KS, Yim DS, Cho JY, Park SS, Park JY, Lee KH, Jang IJ, Yi SY, Bae KS, Shin SG (2001) Effect of omeprazole on the pharmacokinetics of moclobemide according to the genetic polymorphism of CYP2C19. Clin Pharmacol Ther 69:266–273 39. Angiolillo DJ, Gibson CM, Cheng S, Ollier C, Nicolas O, Bergougnan L, Perrin L, LaCreta FP, Hurbin F, Dubar M (2011) Differential effects of omeprazole and pantoprazole on the pharmacodynamics and pharmacokinetics of clopidogrel in healthy subjects: randomized, placebo-controlled, crossover comparison studies. Clin Pharmacol Ther 89:65–74 40. Boulenc X, Djebli N, Shi J, Perrin L, Brian W, Van Horn R, Hurbin F (2012) Effects of omeprazole and genetic polymorphism of CYP2C19 on the clopidogrel active metabolite. Drug Metab Dispos 40:187–197 41. Ogilvie BW, Yerino P, Kazmi F, Buckley DB, Rostami-Hodjegan A, Paris BL, Toren P, Parkinson A (2011) The proton pump inhibitor, omeprazole, but not lansoprazole or pantoprazole, is a metabolism-dependent inhibitor of CYP2C19: implications for coadministration with clopidogrel. Drug Metab Dispos 39:2020–2033
PHARMACOKINETICS AND DISPOSITION
Evaluating a physiologically based pharmacokinetic model for prediction of omeprazole clearance and assessing ethnic sensitivity in CYP2C19 metabolic pathway Sheng Feng 1 & Yumi Cleary 2 & Neil Parrott 2 & Pei Hu 3 & Cornelia Weber 2 & Yongqing Wang 4 & Ophelia Q. P. Yin 5 & Jun Shi 1
Received: 30 November 2014 / Accepted: 10 March 2015 / Published online: 24 March 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose The purpose of this study is to evaluate the ethnicityspecific population models in the SimCYP Simulator® for prediction of omeprazole clearance with attention to differences in the CYP2C19 metabolic pathway. Methods The SimCYP® models incorporating Caucasian, Chinese, and Japanese population-specific demographic, physiological, and enzyme data were applied to simulate omeprazole pharmacokinetics. Published pharmacokinetic data of omeprazole after intravenous or oral administration in Caucasian, Chinese, and Japanese were used for the evaluation. Results Following oral administration, the ratio of the predicted to observed geometric mean of omeprazole clearance in Caucasian extensive metabolizers (EMs) was 0.88. The ratios in Chinese EMs were 1.16 and 0.99 after intravenous and oral administration, respectively. The ratios in Japanese EMs were 0.88 and 0.71 after intravenous and oral administration, respectively. Electronic supplementary material The online version of this article (doi:10.1007/s00228-015-1834-y) contains supplementary material, which is available to authorized users. * Jun Shi [email protected] 1
Roche Innovation Center Shanghai, Building 6, Lane 917, Ha Lei Road, Pudong, Shanghai, China
2
Roche Innovation Center Basel, Basel, Switzerland
3
Peking Union Medical College Hospital, Beijing, China
4
The First Affiliated Hospital, Nanjing Medical University, Nanjing, China
5
Chinese University of Hong Kong, Hong Kong, People’s Republic of China
Significant differences (2-fold) in the observed oral clearance of omeprazole were identified between Caucasian and Asian (Chinese and Japanese) EMs while the observed oral and intravenous clearances of omeprazole were similar between Chinese and Japanese EMs. Physiologically based pharmacokinetics (PBPK) models within SimCYP accurately predicted the difference in the observed oral clearance between Caucasian and Chinese EMs but overpredicted the difference between Caucasians and Japanese EMs due to underprediction of oral clearance in Japanese EMs. Conclusions The PBPK model within SimCYP adequately predicted omeprazole clearance in Caucasian, Chinese, and Japanese EMs and the 2-fold differences in clearance of omeprazole between Caucasian and Asian EMs. This may lead to early identification of ethnic sensitivity in clearance and the need for different dosing regimens in a specific ethnic group for substrates of CYP2C19 which can support the rational design of bridging clinical trials.
Keywords PBPK . CYP2C19 . Omeprazole . Ethnic sensitivity
Introduction In recent years, there has been a dramatic increase in the number of global clinical trials [1, 2]. This trend offers opportunities for cost saving and recruitment acceleration as well as minimizing duplication of clinical data and shortening the drug approval gap among regions. To ensure successful global drug development, ethnicity-related factors on PK/PD, medical practice, and disease biology must be taken into consideration [3]. Physiologically based pharmacokinetics (PBPK) modeling is one of the key components of model-based drug development and is increasingly embraced within the industry and regulatory
618
authorities [4–6]. PBPK modeling can be a useful tool to simulate PK in various ethnic groups and may be used to assess ethnic sensitivity. Currently, commercially available PBPK modeling software tools use parameters gathered mainly in Caucasians while the population data in Chinese and Japanese is less complete [7, 8]. For instance, some physiological data, such as milligrams of microsomal protein per gram of liver (MPPGL), in Chinese and Japanese are not available, and so, the Caucasian data has to be used for simulations in these populations. Even the existing physiological data of Chinese and Japanese are often from limited sources, for example, the hepatic CYP2C19 abundance in the liver was only from 20 Chinese livers and 31 Japanese livers, respectively [8, 9]. Therefore, to facilitate prospective application of PBPK models for drug development and bridging studies, more model system data are needed and the models must be evaluated for the Chinese and Japanese populations. Omeprazole is a proton pump inhibitor (PPI) used in the treatment of gastroesophageal reflux and other diseases, which are caused by the excess secretion of stomach acid. Due to its safety and sensitivity, omeprazole is recommended as a probe drug of CYP2C19 by the FDA and others [10–12]. CYP2C19 is well known to manifest polymorphism, and while the differences in frequency of extensive metabolizers (EMs) and poor metabolizers (PMs) in the Caucasian and Asian population are well documented, the differences in activity still need to be investigated [13, 14]. CYP2C19*2 and CYP2C19*3 are the most common variants. Both are null alleles and have no activity in vivo, thus no activity in PMs [14, 15]. Therefore, we used data from EMs to investigate the CYP2C19 activity differences in the three ethnicity groups. The objective of the study was to compare observed and SimCYP predicted clearance of omeprazole in Caucasian, Chinese, and Japanese EMs and examine if ethnic differences in the CYP2C19 metabolic pathway were well captured.
Methods
Eur J Clin Pharmacol (2015) 71:617–624
inactivated by gastric acid, enteric-coated or delayed release formations are often used. To decrease the effect of formulation differences, the default omeprazole model for an entericcoated tablet was selected for model evaluation and was compared to clinical studies with this formulation. For distribution, a minimal PBPK model with one compartment was applied. The default value of volume of distribution at steady state (Vss) for omeprazole model in SimCYP was 0.15 L/kg. However, the Vss was reset to 0.23, 0.249, and 0.219 L/kg for Caucasians, Chinese, and Japanese, respectively [18–20]. The Vss for Caucasians and Japanese was the values reported in the literature while the Vss for Chinese was the mean of parameter estimates based on the individual concentration data. The SimCYP omeprazole model default values of intrinsic clearance for hepatic metabolic clearance due to CYP2C19 and CYP3A4 were used, and all other drug-specific parameter models were identical across the simulated populations.
Virtual population The main differences in the inputs for the three virtual populations are summarized in Table 1 (see also Online Resource 2). The demographic databases of Caucasians, Chinese, and Japanese were as provided in SimCYP. The intrinsic catalytic activity of CYPs 2C19 and 3A4 per unit amount of enzyme variant and tissue composition were assumed to be the same in Caucasians, Chinese, and Japanese. The default mean abundances of CYP2C19 in the liver in Caucasian and Chinese and Japanese EMs were 14, 8, and 1 pmol/mg in SimCYP, respectively. The abundance in Japanese EMs was adjusted from 1 to 4.77 pmol/mg based on the reported 2.9-fold difference in CYP2C19 abundance between Caucasian and Japanese EMs [21]. Table 1 Summary of main differences in the inputs for the three virtual populations for the omeprazole clearance prediction (default values in SimCYP) Parameters
General Simulations were performed using the SimCYP Populationbased simulator®, version 14 (SimCYP Ltd, Sheffield, UK). The omeprazole model file and virtual populations for Caucasians, Chinese, and Japanese as provided in the software were used. Omeprazole model The input parameters for the omeprazole model [16] are summarized in the Online Resource 1. The omeprazole model used the advanced dissolution, absorption, and metabolism (ADAM) model for absorption [17]. As omeprazole is partly
Caucasians
Chinese
CYP2C19 abundance in gastrointestinal tract (pmol/mg) Extensive metabolizers 1.5 0.85 Poor metabolizers 0 0 CYP3A4 abundance in 66.2 58 gastrointestinal tract (pmol/mg) CYP2C19 abundance in liver (pmol/mg) Extensive metabolizers 14 8 Poor metabolizers 0 0 CYP3A4 abundance in liver 137 120 (pmol/mg) Average liver weight (L) 1.65056 1.402976 a
Japanese
0.511a 0 59
4.77a 0 122 1.65056
It was adjusted from the default value based on the study reported by Inoue et al. [21]
Eur J Clin Pharmacol (2015) 71:617–624
619
Simulation Simulations were conducted using virtual populations generated by matching distributions of age, sex, dosage, and phenotype (EM or PM) to the actual study data. The simulations followed the reported study designs and conditions as closely as possible including numbers of subjects and dosing regimens, and ten trial simulations were performed for each study.
Clinical data Clinical pharmacokinetic data for omeprazole following intravenous and oral administration in Caucasian, Chinese, and Japanese subjects were searched in the PubMed online database using omeprazole and pharmacokinetics as the search terms. Reports were excluded if they did not give phenotype information or reported clearance and area under the plasma concentration-time curve (AUC) that could not be used to calculate the geometric mean of clearance. Furthermore, only studies using enteric-coated or delayed release formulations were considered. Eight papers (three for Caucasians, two for Chinese, and three for Japanese) were considered eligible for the study [19, 20, 22–27]. Whereas both intravenous and oral studies were found for Chinese and Japanese, only studies with an oral administration of omeprazole were found for Caucasians. Lastly, the authors of the two Chinese papers [19, 25] kindly provided us with the individual data.
Population PK parameters are generally considered to be log normally distributed. Hence, values of geometric mean (GM) and 95 % confidence intervals (CIs) were used to compare the clearance. GMs and 95 % CIs were calculated from population mean of μ and standard deviation of σ, the natural logs of GM and geometric standard deviation, using Eqs. 1–4. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi σ ¼ Ln 1 þ CV2 ð1Þ μ ¼ LnðMeanÞ−0:5 σ2 ð2Þ
95% CIs ¼ exp
pffiffiffiffiffiσ μ 1:96 n−1
When data from more than one study or group were available for the same route of administration in the same ethnicity, the pooled geometric mean (PGM) was calculated using Eq. 7. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑n nj j¼1 PGM ¼ GM1n1 GM2n2 ⋯ GMj n j ð7Þ For those papers only reporting the mean and SD values of AUC0–inf [20, 23, 24] or AUC0–24 h [27], firstly, the GM of AUC was calculated using the Eqs. 1–3. Then, the GM of clearance was calculated by dose / AUC. The clearance was calculated from a non-compartment analysis using Phoenix software (WinNonlin models version 6.2, Pharsight Corp, Cary, NC, USA) if the individual data could be obtained, including observed data [19, 25] and SimCYP simulated data. The figures were drawn using R software (version 3.1.0). The mean concentration–time data was extracted with the Plot Digitizer software (version 2.6.3).
Results Subject characteristics The characteristics of the observed subjects including ethnicity, phenotype, the number of the observed subjects, and demographic data can be seen in Table 2.
Statistical analysis
GM ¼ expμ
When the mean and 95 % CI were reported, SD was calculated using the Eq. 6. pffiffiffi ðUpper Limit−Lower LimitÞ n ð6Þ SD ¼ 2 1:96
ð3Þ ð4Þ
where n=the number of the subjects in the study. Coefficient of variation (CV) was calculated from reported mean and standard deviation (SD) values in the published studies using Eq. 5. Standard deviation CV ¼ Mean
ð5Þ
Prediction of clearance in Caucasian EMs Plasma concentration–time profiles of omeprazole after 20 mg of oral administration in Caucasian EMs were available from two studies [23, 24]. Simulated and observed plasma concentration–time curves were compared for the ten virtual trials in Fig. 1a, b (n=9 and n=12, respectively). The plasma concentration–time profiles of omeprazole were generally well recovered for both studies. The ratios of predicted to the observed GM of CLPO in the entire virtual population of each study were 1.13, 0.79, and 0.75 (Table 2) and within 0.8–1.25 in 5, 4, and 5 out of the 10 trials for the studies reported by Baldwin et al. [22], Shirasaka et al. [23], and Andersson et al. [24], respectively (as shown in Online Resource 3). Prediction of clearance in Chinese EMs Simulated and observed plasma concentration–time curves after i.v. and oral administration of 40 mg omeprazole in Chinese EMs were compared for the ten virtual trials in
620 Table 2 Ethnicity
Eur J Clin Pharmacol (2015) 71:617–624 Details of the in vivo studies of observed and predicted geometric mean clearances in healthy Caucasian, Chinese, and Japanese subjects (EM) Phenotype n
Caucasian EM EM EM Chinese Japanese
Weight (kg)
Age (years)
11 69 (60–109) 33 (21–60) 9 Not given 12 68–86
Dose (mg) Route Observed CLa (L/h) Predicted CLa (L/h) Ratiob Reference 40
p.o.
35.1 (21.5–57.3)
39.6 (33.6–46.6)
1.13
Baldwin et al. [22]
18–50
20
p.o.
53.9 (38.9–74.8)
42.6 (35.7–50.9)
0.79
Shirasaka et al. [23]
22–32
20
p.o.
58.5 (44.2–78.2)
43.8 (38.0–50.4)
0.75
Andersson et al. [24]
EM
16 61.9 (51–72) 26 (22–31)
40
i.v.
11.8 (9.69–14.3)
13.7 (13.0–14.3)
1.16
Wang et al. [19]
EM
19 61.7 (54–71) 21 (20–24)
40
p.o.
23.7 (18.5–30.5)
23.4 (22.1–24.9)
0.99
Yin et al. [25]
EM
17 57.7
20
i.v.
13.8
10.5 (9.64–11.3)
0.76
Ishizawa et al. [20]
25
EM
14 63.0 (42–90) 28.6 (21–44) 20
i.v.
10.7
11.2 (10.2–12.2)
1.05
Uno et al. [26]
EM EM
14 63.0 (42–90) 28.6 (21–44) 40 11 43–72 22–38 20
p.o. p.o.
19.3 30.0
17.4 (15.4–19.7) 15.7 (13.6–18.2)
0.90 0.52
Shirai et al. [27]
Data are shown as geometric mean values (95 % CI in brackets) for clearance and as arithmetic mean values (ranges in brackets, if reported) for weight and age EM extensive metabolizers, n number of the subjects, CL clearance (which is calculated as dose / AUC0–inf; in the study of Shirai, it is calculated as dose / AUC0–24 h), p.o. oral, i.v. intravenous a
CLPO for oral administration, CLPO: apparent clearance after oral administration
b
Ratios of predicted to observed CL
Fig. 1c, d, respectively. The simulated plasma concentration– time curves were in a good agreement with the observed data. The ratios of predicted to observed GM CLIV and CLPO were 1.16 and 0.99 (Table 2), and 7 and 10 out of 10 trials had the ratio within 0.8–1.25. These simulations demonstrated that the Chinese population model was able to recover CLIV and CLPO of omeprazole very well.
Prediction of clearance in Japanese EMs For Japanese, two papers were found for each intravenous and oral administration. Simulated and observed plasma concentration–time curves after 20-mg omeprazole intravenous administration were compared for the ten virtual trials in Fig. 1e, f (n=17 and n=14, respectively). The ratios of the simulated to the observed GM of CLIV in the entire virtual population were 0.76 and 1.05 (Table 2) and within 0.8–1.25 in 3 and 9 out of the 10 trials for the studies reported by Ishizawa et al. [20] and Uno et al. [26], respectively. Simulated and observed plasma concentration–time curves after oral administration of omeprazole in Japanese EMs were compared for the ten virtual trials in Fig. 1g, h (40 mg, n=14 and 20 mg, n=12, respectively). The ratios of the simulated to the observed GM of omeprazole CLPO in the entire virtual population were 0.90 and 0.52 (Table 2) and within 0.8–1.25 in 7 and 0 out of the 10 trials for the studies reported by Uno et al. [26] and Shirai et al. [27], respectively. The predicted values of omeprazole clearance in all simulated trials for all ethnic groups are shown in Table 2. Summarized data for each ethnicity are given in Table 3. The ratios of predicted to observed GM of CL PO in Caucasian EMs and CLPO and CLIV in Chinese EMs were
0.88, 1.16, and 0.99, respectively. This showed good accuracy of predictions for Caucasian and Chinese EMs. However, the ratios of predicted to observed GM of CLIV and CLPO in Japanese EMs were 0.88 and 0.71, respectively, demonstrating a tendency to under-prediction for CLPO in Japanese EMs. Prediction of clearance in Caucasian, Chinese, and Japanese PMs Although CYP2C19 has no activity in PMs, the clearance of omeprazole in PMs was also simulated. The results can be seen in Online Resource 4. In Caucasians, the ratio of predicted to observed GM of CLPO was 1.53 for the studies reported by Andersson et al. [24]. In Chinese, the ratios of predicted to observed GM of CLIV and CLPO were 0.50 and 0.77 for the studies reported by Wang et al. [19] and Yin et al. [25], respectively. In Japanese, the ratios of predicted to observed GM of CLIV were 1.34 and 1.42 for the studies reported by Ishizawa et al. [20] and Uno et al. [26], respectively. The ratios of predicted to observed GM of CLPO were 1.36 and 1.31 for the studies reported by Uno et al. [26] and Shirai et al. [27], respectively. Caucasian, Chinese, and Japanese EM clearance differences The differences in observed and predicted GM omeprazole clearance in Caucasian, Chinese, and Japanese EMs when combining all data across studies can be seen in Table 3. The observed pooled CLPO in Caucasians was 2.03-fold and 2.05-fold higher than the values in Chinese and in Japanese, respectively. The ratios of observed pooled CL (i.v. and oral)
Eur J Clin Pharmacol (2015) 71:617–624
621
Fig. 1 Simulated (lines) and observed (circles) mean plasma concentration–time curves of omeprazole in healthy Caucasian (a, b), Chinese (c, d), and Japanese (e–h) CYP2C19 extensive metabolizers. The gray lines represent the mean of simulations of individual trials (ten trials). The black line represents the mean for the total virtual population. The dashed lines represent the 5th and 95th percentiles for the total virtual population
between Chinese and Japanese were 0.96 and 1.01, respectively. The pooled predicted CLPO in Caucasians was 1.79-fold higher than the value in Chinese, which was slightly lower than the observed difference (2.03-fold). The pooled predicted CLPO in Caucasians was 2.53-fold higher than the value in Japanese, which was higher than the observed difference (2.05-fold). The pooled predicted CLIV and CLPO in Chinese were 1.27- and 1.41-fold higher than the values in Japanese. These simulations suggested that the differences in CLPO between Chinese and Caucasian EMs were well predicted while the differences in CL PO between Japanese and Caucasian EMs were overpredicted due to under-prediction of CLPO in Japanese.
Discussion Known differences in physiological parameters between Caucasians, Chinese, and Japanese have been implemented in several commercial PBPK tools, and the application of physiologically based methods to compare CYP-mediated clearance between Caucasians and Chinese using SimCYP has been previously reported by Barter et al. [7]. They reported that the predicted omeprazole CLPO in Chinese was approximately 2-fold higher than the observed CLPO reported for the study by Hu et al. [28]. In our work, we have attempted to increase the number of studies in the clinical dataset while keeping consistency in study condition (e.g., formulation) and have also extended the dataset to include intravenous PK data
622
Eur J Clin Pharmacol (2015) 71:617–624
Table 3 Summary of observed and predicted geometric mean omeprazole clearances in healthy Caucasian, Chinese, and Japanese subjects (EM, pooled) Ethnicity
Phenotype
n
Route
Observed CLa (L/h)
Predicted CLa (L/h)
Ratiob
Caucasian Chinese
EM EM EM EM EM
32 16 19 31 25
p.o. i.v. p.o. i.v. p.o.
48.0 11.8 23.7 12.3 23.4
42.0 13.7 23.4 10.8 16.6
0.88 1.16 0.99 0.88 0.71
Japanese
EM extensive metabolizers, n number of the subjects, CL clearance (which is calculated as dose / AUC), p.o. oral, i.v. intravenous a
CLPO for oral administration, CLPO: apparent clearance after oral administration
b
Ratios of predicted to observed CL
using individual data kindly provided by the authors of two Chinese papers [19, 25]. Based on in vitro studies with human liver microsomes [29], the difference in intrinsic clearance between Caucasians and Chinese was larger for CYP2C19 than for other CYPs (CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2D6, CYP2E1, and CYP3A). The median CLint value for (S)-mephenytoin 4′-hydroxylation (a probe drug in vitro for CYP2C19) in Chinese liver microsomes was 0.43 μL/min/mg, which was 26 % of the corresponding value in Caucasians (1.64 μL/min/mg) [29]. Our work showed that this in vitro difference was reflected in vivo. Although comparison between different studies was difficult due to different conditions and formulations, the clearance of CYP2C19 substrates in Caucasian CYP2C19 EMs seemed to be approximately 2-fold higher than that in Asian EMs (Chinese, Japanese). The reported GM CLPO of omeprazole in Korean EMs (20.5 L/h (n=21) and 25.4 L/h (n=6)) [30, 31] was also very similar to those values in Chinese and Japanese EMs. For the Japanese population, the default value for the CYP2C19 abundance in the liver is 1 pmol/mL in SimCYP [8, 32]. Using this value, the ratios of predicted vs observed clearance were 0.41 (intravenous, Ishizawa et al. [20]), 0.57 (intravenous, Uno et al. [26]), 0.41 (oral, Uno et al. [26]), and 0.24 (oral, Shirai et al. [27]), which were significantly underpredicted. Inoue et al. [21] reported that the mean abundance values of CYP2C19 in Japanese and Caucasian EM liver microsomes were 1.521 and 4.465 pmol/mL, respectively. When this 3-fold difference in CYP2C19 abundance in the liver between Japanese and Caucasian EMs was considered in the Japanese model, the abundances could be adjusted to 4.77 and 0.511 pmol/mL in the liver and gastrointestinal tract, respectively, based on the default values in SimCYP for the Caucasian EMs which were 14 and 1.5 pmol/mL in the liver and gastrointestinal tract, respectively. Using these modified
CYP2C19 abundances, the recovery of clearance in Japanese was better with ratios of predicted to observed clearance of 0.76 (intravenous, Ishizawa et al. [20]), 1.05 (intravenous, Uno et al. [26]), 0.90 (oral, Uno et al. [26]), and 0.52 (oral, Shirai et al. [27]). This indicates that the abundance of CYP2C19 in Japanese requires more investigation. Different liver weight values in Chinese have been reported [33], and the reported mean value is only 9 % higher than the value currently used in SimCYP [34]. As regard the cardiac output and hepatic blood flow, the reported values are from the Sichuan province of China, which is known for small body size and is not representative of the whole Chinese population. Also, omeprazole has a low hepatic extraction ratio, and so, clearance in Chinese is not significantly affected by hepatic blood flow. EMs can be divided into homozygous EMs (wt/wt) and heterozygous EMs (wt/m1 or wt/m2) based on the genotypes. The difference in clearance of omeprazole was significant between homozygous EMs and heterozygous EMs. For example, the CLPO in homozygous EMs was almost 2-fold higher than in heterozygous EMs in Japanese, though the difference became smaller when omeprazole was given intravenously [26]. However, the models in SimCYP have not distinguished these two groups. One of the reasons may be that the difference between homozygous EMs and heterozygous EMs was small compared to the difference between EMs and PMs. Another reason may be due to sparse abundance data for each genotype. Therefore, homozygous EMs and heterozygous EMs were pooled as general EMs for the simulations in this study although they were separated into two groups in the observed studies. Although EMs data was used to investigate the CYP2C19 activity, the PMs were also simulated and showed ratios of predicted to observed GM of CL (i.v. and oral) all within 2-fold (Online Resource 4). Omeprazole is reported to exhibit a dose- and timedependent nonlinear pharmacokinetics [35]. Although omeprazole is a time-dependent inhibitor of CYP2C19 and CYP3A4, it is the metabolites of omeprazole which are believed to be the major cause for the nonlinear pharmacokinetics [36–41]. The clearance model in SimCYP was modified by inputting the omeprazole time-dependent inhibition parameters [27], and the effect of auto-inhibition of omeprazole was found to be very limited. Therefore, the modification was not implemented in the final omeprazole clearance model. In summary, the omeprazole clearance prediction using Caucasian, Chinese, and Japanese population models in SimCYP was within 2-fold of observations in Caucasian, Chinese, and Japanese after adjusting the CYP2C19 abundance in Japanese EMs. This indicated that the virtual Caucasian, Chinese, and Japanese models in SimCYP can be used to predict the clearance of substrates of CYP2C19. We also used the PBPK method to assess the differences in oral and intravenous clearance of omeprazole in Caucasians,
Eur J Clin Pharmacol (2015) 71:617–624
623
Chinese, and Japanese. The results showed that PBPK modeling could reasonably recover the ethnic sensitivity. This may lead to early identification of ethnic sensitivity in clearance and the need for different dosing regimens in a specific ethnic group for substrates of CYP2C19 which can support the rational design of bridging clinical trials.
11.
Acknowledgments The study was supported by Roche Postdoc Fellowship Program.
13.
Conflict of interest All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare the following: S.F., Y.C., N.P., P.H., C.W., and J.S. had support from F. Hoffmann-La Roche for the submitted work; Y.C., N.P., C.W., and J.S. are full-time employees of F. Hoffmann-La Roche; P.H. is a full-time employee of Peking Union Medical College Hospital; S.F. is a Roche Postdoctoral Fellow; no financial relationships with any organization that might have an interest in the submitted work in the previous 3 years; and no other relationships or activities that could appear to have influenced the submitted work.
14.
Author contribution Sheng Feng and Yumi Cleary performed the analysis. All the authors contributed to the discussion of results. Sheng Feng wrote the first draft. Yongqing Wang and Ophelia Q. P. Yin kindly provided the individual demographic and concentration–time data for the two Chinese studies. All the authors reviewed the data and approved the final version for submission.
12.
15.
16.
17.
18.
19.
References 1. 2. 3. 4.
5.
6.
7.
8.
9.
10.
Thiers FA, Sinskey AJ, Berndt ER (2008) Trends in the globalization of clinical trials. Nat Rev Drug Discov 7:13–14 Karlberg JPE (2009) Globalization of sponsored clinical trials. Nat Rev Drug Discov 7:458–460 ICH Topic E 5 (R1) Ethnic factors in the acceptability of foreign clinical data. Zhao P, Rowland M, Huang SM (2012) Best practice in the use of physiologically based pharmacokinetic modeling and simulation to address clinical pharmacology regulatory questions. Clin Pharmacol Ther 92:17–20 Rowland M, Peck C, Tucker G (2011) Physiologically-based pharmacokinetics in drug development and regulatory science. Annu Rev Pharmacol Toxicol 51:45–73 Huang SM, Rowland M (2012) The role of physiologically based pharmacokinetic modeling in regulatory review. Clin Pharmacol Ther 91(3):542–549 Barter ZE, Tucker GT, Rowland-Yeo K (2013) Differences in cytochrome p450-mediated pharmacokinetics between Chinese and Caucasian populations predicted by mechanistic physiologically based pharmacokinetic modeling. Clin Pharmacokinet 52:1085– 1100 Inoue S, Howgate EM, Rowland-Yeo K, Shimada T, Yamazaki H, Tucker GT, Rostami-Hodjegan A (2006) Prediction of in vivo drug clearance from in vitro data. II: potential inter-ethnic differences. Xenobiotica 36(6):499–513 Shu Y, Wang LS, Xiao WM, Wang W, Huang SL, Zhou HH (2000) Probing CYP2C19 and CYP3A4 activities in Chinese liver microsomes by quantification of 5-hydroxyomeprazole and omeprazole sulphone. Acta Pharmacol Sin 21(8):753–758 FDA (2012) Drug interaction studies—study design, data analysis, implications for dosing, and labeling recommendations (draft).
20.
21.
22.
23.
24.
25.
26.
27.
Chang M, Dahl ML, Tybring G, Götharson E, Bertilsson L (1995) Use of omeprazole as a probe drug for CYP2C19 phenotype in Swedish Caucasians: comparison with S-mephenytoin hydroxylation phenotype and CYP2C19 genotype. Pharmacogenetics 5: 358–363 Chang M, Tybring G, Dahl ML, Götharson E, Sagar M, Seensalu R, Bertilsson L (1995) Interphenotype differences in disposition and effect on gastrin levels of omeprazole—suitability of omeprazole as a probe for CYP2C19. Br J Clin Pharmacol 39:511–518 Desta Z, Zhao X, Shin JG, Flockhart DA (2002) Clinical significance of the cytochrome P450 2C19 genetic polymorphism. Clin Pharmacokinet 4:913–958 Zanger UM, Schwab M (2013) Cytochrome P450 enzymes in drug metabolism: regulation of gene expression, enzyme activities, and impact of genetic variation. Pharmacol Ther 138:103–141 Nakamoto K, Kidd JR, Jenison RD, Klaassen CD, Wan YJ, Kidd KK, Zhong XB (2007) Genotyping and haplotyping of CYP2C19 functional alleles on thin-film biosensor chips. Pharmacogenet Genomics 17:103–114 Wu F, Gaohua L, Zhao P, Jamei M, Huang SM, Bashaw ED, Lee SC (2014) Predicting nonlinear pharmacokinetics of omeprazole enantiomers and racemic drug using physiologically based pharmacokinetic modeling and simulation: application to predict drug/ genetic interactions. Pharm Res 31(8):1919–1929 Jamei M, Turner D, Yang J, Neuhoff S, Polak S, Rostami-Hodjegan A, Tucker G (2009) Population-based mechanistic prediction of oral drug absorption. AAPS J 11:225–237 Regårdh CG, Andersson T, Lagerström PO, Lundborg P, Skånberg I (1990) The pharmacokinetics of omeprazole in humans—a study of single intravenous and oral doses. Ther Drug Monit 12:163–172 Wang Y, Zhang H, Meng L, Wang M, Yuan H, Ou N, Zhang H, Li Z, Shi R (2010) Influence of CYP2C19 on the relationship between pharmacokinetics and intragastric pH of omeprazole administered by successive intravenous infusions in Chinese healthy volunteers. Eur J Clin Pharmacol 66:563–569 Ishizawa Y, Yasui-Furukori N, Takahata T, Sasaki M, Tateishi T (2005) The effect of aging on the relationship between the cytochrome P450 2C19 genotype and omeprazole pharmacokinetics. Clin Pharmacokinet 44:1179–1189 Inoue K, Yamazaki H, Imiya K, Akasaka S, Guengerich FP, Shimada T (1997) Relationship between CYP2C9 and 2C19 genotypes and tolbutamide methyl hydroxylation and S-mephenytoin 4′-hydroxylation activities in livers of Japanese and Caucasian populations. Pharmacogenetics 7:103–113 Baldwin RM, Ohlsson S, Pedersen RS, Mwinyi J, IngelmanSundberg M, Eliasson E, Bertilsson L (2008) Increased omeprazole metabolism in carriers of the CYP2C19*17 allele; a pharmacokinetic study in healthy volunteers. Br J Clin Pharmacol 65:767–774 Shirasaka Y, Sager JE, Lutz JD, Davis C, Isoherranen N (2013) Inhibition of CYP2C19 and CYP3A4 by omeprazole metabolites and their contribution to drug-drug interactions. Drug Metab Dispos 41:1414–1424 Andersson T, Holmberg J, Röhss K, Walan A (1998) Pharmacokinetics and effect on caffeine metabolism of the proton pump inhibitors, omeprazole, lansoprazole, and pantoprazole. Br J Clin Pharmacol 45:369–375 Yin OQ, Tomlinson B, Chow AH, Waye MM, Chow MS (2004) Omeprazole as a CYP2C19 marker in Chinese subjects: assessment of its gene-dose effect and intrasubject variability. J Clin Pharmacol 44:582–589 Uno T, Niioka T, Hayakari M, Yasui-Furukori N, Sugawara K, Tateishi T (2007) Absolute bioavailability and metabolism of omeprazole in relation to CYP2C19 genotypes following single intravenous and oral administrations. Eur J Clin Pharmacol 63:143–149 Shirai N, Furuta T, Moriyama Y, Okochi H, Kobayashi K, Takashima M, Xiao F, Kosuge K, Nakagawa K, Hanai H, Chiba
624
28.
29.
30.
31.
32.
33.
34.
Eur J Clin Pharmacol (2015) 71:617–624 K, Ohashi K, Ishizaki T (2001) Effects of CYP2C19 genotypic differences in the metabolism of omeprazole and rabeprazole on intragastric pH. Aliment Pharmacol Ther 15:1929–1937 Hu XP, Xu JM, Hu YM, Mei Q, Xu XH (2007) Effects of CYP2C19 genetic polymorphism on the pharmacokinetics and pharmacodynamics of omeprazole in Chinese people. J Clin Pharm Ther 32:517–524 Yang J, He MM, Niu W, Wrighton SA, Li L, Liu Y, Li C (2012) Metabolic capabilities of cytochrome P450 enzymes in Chinese liver microsomes compared with those in Caucasian liver microsomes. Br J Clin Pharmacol 73:268–284 Jin SK, Kang TS, Eom SO, Kim JI, Lee HJ, Roh J (2009) CYP2C19 haplotypes in Koreans as a marker of enzyme activity evaluated with omeprazole. J Clin Pharm Ther 34:437–446 Cho JY, Yu KS, Jang IJ, Yang BH, Shin SG, Yim DS (2002) Omeprazole hydroxylation is inhibited by a single dose of moclobemide in homozygotic EM genotype for CYP2C19. Br J Clin Pharmacol 53:393–397 Yasumori T, Murayama N, Yamazoe Y, Kato R (1990) Polymorphism in hydroxylation of mephenytoin and hexobarbital stereoisomers in relation to hepatic P-450 human-2. Clin Pharmacol Ther 47:313–322 Li GF, Yu G, Liu HX, Zheng QS (2014) Ethnic-specific in vitro-in vivo extrapolation and physiologically based pharmacokinetic approaches to predict cytochrome P450-mediated pharmacokinetics in the Chinese population: opportunities and challenges. Clin Pharmacokinet 53:197–202 Barter ZE, Tucker GT, Rowland-Yeo K (2014) Response to BEthnic-specific in vitro-in vivo extrapolation and physiologically based pharmacokinetic approaches to predict cytochrome P450mediated pharmacokinetics in Chinese population: opportunities and challenges^. Clin Pharmacokinet 53:203
35.
Regårdh CG (1986) Pharmacokinetics and metabolism of omeprazole in man. Scand J Gastroenterol Suppl 118:99–104 36. Soons PA, van den Berg G, Danhof M, van Brummelen P, Jansen JB, Lamers CB, Breimer DD (1992) Influence of single- and multiple-dose omeprazole treatment on nifedipine pharmacokinetics and effects in healthy subjects. Eur J Clin Pharmacol 42:319– 324 37. Funck-Brentano C, Becquemont L, Lenevu A, Roux A, Jaillon P, Beaune P (1997) Inhibition by omeprazole of proguanil metabolism: mechanism of the interaction in vitro and prediction of in vivo results from the in vitro experiments. J Pharmacol Exp Ther 280: 730–738 38. Yu KS, Yim DS, Cho JY, Park SS, Park JY, Lee KH, Jang IJ, Yi SY, Bae KS, Shin SG (2001) Effect of omeprazole on the pharmacokinetics of moclobemide according to the genetic polymorphism of CYP2C19. Clin Pharmacol Ther 69:266–273 39. Angiolillo DJ, Gibson CM, Cheng S, Ollier C, Nicolas O, Bergougnan L, Perrin L, LaCreta FP, Hurbin F, Dubar M (2011) Differential effects of omeprazole and pantoprazole on the pharmacodynamics and pharmacokinetics of clopidogrel in healthy subjects: randomized, placebo-controlled, crossover comparison studies. Clin Pharmacol Ther 89:65–74 40. Boulenc X, Djebli N, Shi J, Perrin L, Brian W, Van Horn R, Hurbin F (2012) Effects of omeprazole and genetic polymorphism of CYP2C19 on the clopidogrel active metabolite. Drug Metab Dispos 40:187–197 41. Ogilvie BW, Yerino P, Kazmi F, Buckley DB, Rostami-Hodjegan A, Paris BL, Toren P, Parkinson A (2011) The proton pump inhibitor, omeprazole, but not lansoprazole or pantoprazole, is a metabolism-dependent inhibitor of CYP2C19: implications for coadministration with clopidogrel. Drug Metab Dispos 39:2020–2033
