Utility of Chimeric Mice with Humanized Liver for Predicting Human Pharmacokinetics in Drug Discovery: Comparison with in Vitro–in Vivo Extrapolation and Allometric Scaling

Abstract

Predicting human pharmacokinetics (PK) such as clearance (CL) and volume of distribution (Vd) is a critical component of drug discovery. These predictions are mainly performed by in vitro–in vivo extrapolation (IVIVE) using human biological samples, such as hepatic microsomes and hepatocytes. However, some issues with this process have arisen, such as inconsistencies between in vitro and in vivo findings; the integration of predicted CYP, non-CYP and transporter-mediated human PK; and the difficulty of evaluating very metabolically stable compounds. Various approaches to solving these issues have been reported. Allometric scaling using experimental animals has also often been used. However, this method has also shown many problems due to interspecies differences, albeit that various correction methods have been proposed. Another approach involves the production of chimeric mice with humanized liver via the transplantation of human hepatocytes into mice. The livers of these mice are repopulated mostly with human hepatocytes and express human drug-metabolizing enzymes and drug transporters, suggesting that these mice are useful for solving the issues of IVIVE and allometric scaling, and more reliably predicting human PK. In this review, we summarize human PK prediction methods using IVIVE, allometric scaling and chimeric mice with humanized liver, and discuss the utility of predicting human PK in drug discovery by comparing these chimeric mice with IVIVE and allometric scaling.

1. INTRODUCTION

Predicting human efficacy, toxicity, and drug metabolism and pharmacokinetics (PK) is critical to drug discovery. Historically, poor drug metabolism and PK profiles have been the most significant cause of attrition of drug candidates.¹⁾ Pharmaceutical companies have therefore tried to predict human drug metabolism and PK based on in vitro studies using human biological samples and in vivo studies using experimental animals, and selected drug candidates which were considered to show reliable drug metabolism and PK profiles in humans. Although these strategies produced a dramatic reduction in attrition, drug metabolism and PK still remain as non-negligible cause of attrition.²⁾

To select drug candidates, it is necessary to predict human quantitative PK, metabolite profiles and drug–drug interactions (DDIs) caused by drug enzyme inductions and inhibitions. In particular, predicting human PK is important for drug discovery. Conventionally, this has been done by evaluating PK parameters such as clearance (CL) and volume of distribution (Vd). In addition, more detailed human PK predictions such as the time course of plasma and tissue drug concentrations have become to be used because these predictions are necessary to estimate quantitative efficacy and toxicity.³⁾

Prediction of human PK has mainly been conducted using compounds metabolized by CYP. However, recently, it has also become necessary to predict human PK for compounds metabolized by non-CYP, such as uridine 5′-diphosphate (UDP)-glucuronosyltransferase (UGT) and aldehyde oxidase (AO).^3,4) Further, attention has also focused on predicting human PK involving drug transporters.³⁾ It is also important to evaluate the integration of these prediction methods using CYP, non-CYP and drug transporters-mediated human PK.

Prediction of human PK is most commonly conducted using in vitro–in vivo extrapolation (IVIVE) and allometric scaling. IVIVE predicts in vivo CL using in vitro data for biological materials such as hepatic microsomes and hepatocytes, based on the CL concepts. This method is very useful for predicting human PK using human biological samples. However, differences between in vitro and in vivo findings have produced many difficulties with this prediction, for which various correction methods have been proposed.⁵⁾ As mentioned above, IVIVE has been mainly established for compounds metabolized by CYP. Presently, it is also necessary to establish prediction methods for compounds metabolized by non-CYP. Allometric scaling predicts CL and Vd based on the relationship between in vivo PK parameters and body weights in experimental animals such as mice, rats, dogs and monkeys. Prediction with allometric scaling has also often failed due to interspecies differences in drug metabolism and pharmacokinetics. However, this methodology has been reconsidered following proposals for various modified methods and its predictability has been revised.^5,6)

Despite these methods for improving IVIVE and allometric scaling, various issues concerning their accuracy and application remain unsolved. Accordingly, more accurate and useful prediction methods are required. Recently, human PK prediction methods using chimeric mice with humanized liver have been reported.⁷⁾ These mice are constructed by transplanting them with human hepatocytes.⁸⁾ The livers of these mice are then mostly repopulated with human hepatocytes and express human drug-metabolizing enzymes and drug transporters.⁹⁾ These chimeric mice with humanized liver are therefore considered useful animals for solving the issues of PK prediction using IVIVE and allometric scaling.

Chimeric mice with humanized liver have been used for PK, metabolite profiling and DDIs studies. There have been several reviews with respect to the studies, including our recent review.^7,10–13) In the present review, we particularly focus on human PK prediction using chimeric mice with humanized liver. First, we explain an overview of IVIVE and allometric scaling and problems with their use. Next, we discuss the utility of chimeric mice with humanized liver for predicting human PK in drug discovery by comparing this method with IVIVE and allometric scaling.

2. PREDICTION OF HUMAN PK USING IVIVE

In drug discovery, predicting human CL is essential for PK studies. IVIVE has been widely used for this purpose. Namely, mathematical models based on CL concepts were proposed for predicting quantitative hepatic CL from the 1970s to 1980s.^14–16) The reliability of the mathematical models was initially studied for compounds mainly metabolized by CYP using experimental animal biological samples such as hepatic microsomes and hepatocytes. Since the 1990s, prediction of human CL has progressed by the application of human biological samples. Hepatic CL (CL_h) is predicted using mathematical models. The equations of these models consist of intrinsic CL (CL_int), blood unbound fraction (f_u) for each compound, and hepatic blood flow (Q_h). Typical models include the well-stirred model¹⁴⁾ (Eq. 1), the parallel-tube model¹⁵⁾ (Eq. 2) and the dispersion model¹⁶⁾ (Eq. 3).

  

(1)

  

(2)

  

(3)

where D_n (the dispersion number) = 0.17, and R_n = f_u·CL_int/Q_h

CL_h is predicted by substituting in vitro CL_int (CL_{int, in vitro}) estimated from in vitro metabolic experiments using hepatic microsomes or hepatocytes into CL_int in the equations (Fig. 1). However, difficulties in predicting human CL have been often observed due to the inconsistency between in vitro and in vivo findings. Various correction methods have therefore been proposed.⁵⁾ Naritomi et al.^17,18) evaluated a prediction method using a compound-specific scaling factor, the ratio of animal CL_{int, in vitro} and in vivo CL_int (CL_{int, in vivo}). Animal and human CL_{int, in vitro} are first estimated using hepatic microsomes or hepatocytes. The animal CL_{int, in vivo} is then calculated from in vivo animal PK experiments using the mathematical models. Finally, human CL_h is predicted using human CL_{int, in vitro} corrected by the animal scaling factor and the mathematical models. Obach¹⁹⁾ evaluated the predictability of human CL using the mathematical models with or without plasma and microsomal protein binding. Poulin et al.²⁰⁾ proposed a novel method which considered the fraction unbound to hepatocytes in incubation and the unbound fraction in the liver. Despite these efforts, however, the discrepancy between in vitro and in vivo methods has still not necessarily been resolved. Resolution of this issue would require further improvement in IVIVE methods or completely different approaches.

Fig. 1. Procedure for Predicting CL_h Using IVIVE

As mentioned above, IVIVE has been primarily evaluated for compounds mainly metabolized by CYP. In contrast, non-CYP such as UGT and AO are involved in the metabolism of many compounds. Testa et al.²¹⁾ reported that although CYP is involved in 40% of drug metabolism, non-CYP is involved in the remaining part. Cerny²²⁾ reported the contributions of CYP and non-CYP in the formation of major metabolites for U.S. Food and Drug Administration (FDA)-approved intravenous and orally administered small-molecule drugs (2006–2015). He showed that although CYP contributes the largest percentage of metabolism, non-CYP or a mix of both CYP and non-CYP represent a considerable portion. Thus, it is also necessary to consider prediction methods for non-CYP metabolism.

When predicting CL using IVIVE for a compound, it is necessary to select in vitro reaction conditions which accord with the characteristics of the drug metabolizing enzymes involved in the metabolism. So which enzymes metabolize the compounds should be clarified. Drug metabolizing enzymes involved in the metabolism of test compounds are usually predicted by detecting comprehensive metabolites, the quantitative relative amount of the metabolites, and structural elucidation of the metabolites in hepatocytes. The metabolites produced are generally identified using LC-MS. Radio-labelled test compounds are useful for detecting comprehensive metabolite. However, the use of radio-labelled test compounds is limited in drug discovery. One method for identifying enzymes involved in the metabolism of a compound is to use a specific inhibitor for each drug-metabolizing enzyme in in vitro metabolic studies using hepatocytes. Jinno et al.²³⁾ evaluated the contribution of CYP and UGT to the metabolism of various carboxylic acid compounds by adding 1-aminobenzotriazole, a CYP inhibitor, or (−)-borneol, a UGT inhibitor, in in vitro human hepatocytes reaction mixtures. As an alternative method, Di et al.²⁴⁾ reported an approach for providing mechanistic insights into metabolic pathways by comparing CL_{int, in vitro} in hepatic microsomes (CYP-mediated reaction) and hepatocytes: for compounds mainly metabolized by CYP, the two CL_{int, in vitro} values are comparable. On the other hand, for compounds mainly metabolized by non-CYP, CL_{int, in vitro} in hepatocytes is larger than in hepatic microsomes. When uptake is rate-limiting in hepatocytes, CL_{int, in vitro} in hepatocytes is smaller than in hepatic microsomes.

Unlike the case with CYP, it is necessary to establish an IVIVE method for each non-CYP enzyme. It has been reported that CL_{int, in vitro} in hepatic microsomes underpredicts CL_{int, in vivo} for UGT metabolism.^25,26) One reason is that unsaturated fatty acids released from microsomal membranes inhibit the metabolic activities of UGT isoforms such as UGT1A9 and UGT2B7.²⁷⁾ The predictability of human CL_h was improved by the addition of bovine serum albumin (BSA), which can sequester unsaturated fatty acids in in vitro reaction mixtures. For AO metabolism, CL_{int, in vitro} in hepatic cytosol and S9 and hepatocytes underpredicted human CL_{int, in vivo}.²⁸⁾ Akabane et al.²⁹⁾ compared human CL_{int, in vivo} and CL_{int, in vitro} in pooled hepatocytes for eight AO substrates. CL_{int, in vitro} showed an approximately 10-fold underestimation compared to CL_{int, in vivo}. By taking account of this result, they were able to predict CL_h quantitatively. Sulfotransferase (SULT), carboxylesterase (CES), flavin-containing monooxygenase (FMO) and monoamine oxidase (MAO) are also important non-CYP enzymes. However, reports on predicting human CL with these drug-metabolizing enzymes are limited. Cubit et al.³⁰⁾ tried to predict human CL_{int, in vivo} based on CL_{int, in vitro} of hepatic and intestinal CYP, UGT and SULT for four compounds. They used hepatic and intestinal cytosol for SULT metabolism. Nishimuta et al.³¹⁾ evaluated the prediction of human CL for eight CES 1 substrates using hepatocytes and S9. The mean value of predicted CL_{int,in vivo} based on hepatocyte CL_{int, in vitro} showed only 20% of the observed value. On the other hand, prediction bias was reduced with hepatic S9. Jones et al.³²⁾ studied the correlation between CL_{int, in vivo} and CL_{int, in vitro} in humans for 10 FMO substrates. They found that CL_{int, in vitro} in hepatocytes can predict CL_{int, in vivo} and CL_h well. For MAO metabolism, PK prediction methods using human hepatic mitochondria and microsomes have been reported.^33,34) However, further improvement in predictability appears to require additional evaluation.

Recently, predicting drug transporter-mediated human PK using IVIVE has attracted attention.³⁵⁾ Watanabe et al.³⁶⁾ predicted human PK for pravastatin, which is a substrate of organic anion-transporting polypeptide (OATP) 1B1 and multidrug resistance-associated protein (MRP) 2, using a physiologically based pharmacokinetic (PBPK) model with human cryopreserved hepatocytes. They showed the excellent predictability of plasma concentration profiles and distribution of pravastatin by incorporating empirical scaling factors which were calculated from in vivo PK data and in vitro data determined in hepatocytes in rats. De Bruyn et al.³⁷⁾ reported that the use of the empirical scaling factors determined from monkey data improved the prediction of human PK for nine OATP substrates. The empirical value of the scaling factors is thought to differ among compounds, donors of human hepatocytes, in vitro conditions and animal species.^38–41)

As described above, various IVIVE methods for predicting human PK participants for each drug-metabolizing enzyme or drug transporter have been evaluated. However, it is difficult to clarify which of the drug-metabolizing enzymes and transporters contributes to the PK of test compounds in early drug discovery. It should be also considered that multiple drug-metabolizing enzymes and transporters are involved in PK. Therefore, it is necessary to establish unified PK prediction methods for various drug-metabolizing enzymes and drug transporters.

When test compounds are metabolically very stable, it is difficult to estimate CL_{int, in vitro} accurately since incubation times are limited in in vitro experiments with microsomes and hepatocytes.⁴²⁾ Various approaches to this issue have been reported. Di et al.⁴³⁾ proposed the relay method, in which human cryopreserved hepatocytes are transferred in the supernatant of an in vitro incubation mixture to another in vitro incubation mixture with freshly thawed hepatocytes to prolong the reaction time. They reported that CL_{int, in vitro} estimated using the relay method showed excellent predictions of human CL_{int, in vivo}. Further, novel in vitro systems such as HepatoPac^44,45) and HμREL,^46,47) which are micropatterned hepatocyte coculture systems, and cell systems such as upcyte,⁴⁸⁾ which is derived from human hepatocytes, have been used to estimate low CL_{int, in vitro} and predict in vivo CL in humans.

Prediction of human CL_h has been mainly investigated using IVIVE methods_. However, predicting extrahepatic CL is also important. To predict intestinal clearance or intestinal availability, various mathematical models^4,49) such as the Q_gut model⁵⁰⁾ and the simplified intestinal availability model⁵¹⁾ have been proposed. To predict renal clearance, IVIVE methods have been also evaluated.^52–55) Further, future studies will require human PK prediction methods which unify hepatic and extrahepatic clearances.

3. PREDICTION OF HUMAN PK USING ALLOMETRIC SCALING

Allometric scaling is a PK prediction method based on the relationship between in vivo PK parameters such as total CL (CL_t) and Vd, and body weights in experimental animals such as mice, rats, dogs and monkeys. Simple allometry is a basic method. To predict human CL_t, the following equation (Eq. 4) obtained from in vivo PK studies in three or more species is used.

  

(4)

where a and b are the coefficient and scaling exponent, respectively, and BW is the body weight.

Figure 2 shows an illustration of simple allometry for predicting human CL_t. However, this method has often failed at predictions. Therefore, various modified allometric scaling methods aimed at improving the prediction of human pharmacokinetics have been proposed.

Fig. 2. Illustration of Simple Allometry for Predicting Human CL_t

Mahmood and Balian⁵⁶⁾ proposed the rule of exponents (ROE). This method predicts human CL_t based on the scaling exponents of simple allometry as follows:

  

(5)

  

(6)

  

(7)

  

(8)

  

(9)

where MLP represents the maximum lifespan potential and BrW indicates brain weight.

Ward and Smith⁵⁷⁾ proposed the liver blood flow (LBF) method. This method predicts human CL_t using animal CL_t and the ratio of human to animal LBF rates. Predicted human CL_t can be calculated using the following equation (Eq. 10):

  

(10)

The LBF values of 85, 30, 45 and 21 mL/min/kg for rats, dogs, monkeys and humans, respectively, were used.^58,59) Prediction accuracy of the LBF method was better with monkey clearance data than with rat and dog data. Nagilla and Ward,⁶⁰⁾ Ward et al.,⁶¹⁾ and Tang et al.⁶²⁾ reported similar results. Tang and Mayersohn⁶³⁾ developed the unbound fraction in plasma (f_u) corrected intercept method (FCIM), which incorporates rat and human plasma protein binding data with allometric scaling. The equation (Eq. 11) for predicting human CL_t is as follows:

  

(11)

where a is the coefficient of simple allometric scaling, and Rf_u is the ratio of f_u in rats to humans. This method showed better predictability than the ROE.

Tang et al.⁶²⁾ proposed alternative approaches based on CL_t data sets from one or two animal species. The one animal species method uses rat, dog, or monkey CL_t data with a fixed coefficient as follows:

  

(12)

  

(13)

  

(14)

The two animal species method uses rat and dog, or rat and monkey CL data with a fixed coefficient and exponent as follows:

  

(15)

  

(16)

where a is the coefficient calculated from allometric scaling using the two species. They reported that the one species method using monkey data shows better predictability than that using rat or dog data, and that two species methods show comparable predictability with that of the ROE.

Hosea et al.⁶⁴⁾ reported that, from retrospective analysis, single species scaling (SSS) with rat, dog or monkey data shows similar or improved accuracy compared to multiple species allometric scaling methods such as simple allometric scaling and ROE. This method for predicting human CL uses a fixed scaling exponent of 0.75 as follows:

  

(17)

Deguchi et al.⁶⁵⁾ evaluated the accuracy of predicting human CL_t using the ROE and SSS with mouse, rat, monkey and dog PK data for UGT substrates. They reported that SSS using monkeys showed the most accurate prediction of human CL_t. Recently, Yoshimatsu et al.⁶⁶⁾ evaluated the prediction of human CL_t using SSS with minipigs. SSS with minipigs showed higher predictability compared with mice and rats.

In their comprehensive assessment of human CL_t prediction methods, Lombardo et al.⁶⁷⁾ compared the accuracy of 37 different human CL_t prediction methods by analyzing intravenous PK data from rats, dogs and monkeys for approximately 400 compounds. They estimated human CL_t and calculated the geometric mean-fold error, % of <2-fold error and median bias as the predictability indices for each method. They evaluated allometric scaling methods, including the one animal species method and LBF method using rat, dog and monkey data, and FCIM using rat, dog and monkey data. The one animal species method and LBF method using monkey data, and FCIM showed good predictability. Ring et al.⁶⁸⁾ evaluated the accuracy of 29 different human CL_t prediction methods for 108 compounds. They reported that several allometric scaling methods, including ROE, FCIM, the one animal species method using dog data, and the two animal species method using rat and dog data performed relatively well. Liu et al.⁶⁹⁾ collected rat, dog and monkey PK data for 446 compounds and evaluated the predictabilities of human CL_t using various allometric scaling methods including SSS, the two animal species method, ROE, FCIM and LBF methods. They classified reliable CL_t prediction methods based on their disposition pathway (renal excretion or non-renal excretion (metabolism, biliary excretion or minor renal excretion)), organ extraction ratio (hepatic extraction ratio or ratio of unbound CL to renal glomerular filtration rate) and Log P of the compounds.

Volume of distribution at steady state (Vd_ss) is an important PK parameter. In general, Vd_ss can be predicted using allometric scaling. It has been also reported that correction with unbound fraction in plasma improves human Vd_ss prediction.^6,70) Lombardo et al.⁷¹⁾ comprehensively estimated the predictabilities of prediction methods of Vd_ss in humans using rat, dog, and monkey PK data for approximately 400 compounds. They suggested that SSS using dogs may be a reasonable method for allometric scaling.

As described above, various modified methods have been proposed to improve the predictability of human PK parameters, especially CL. However, as long as allometric scaling methods use typical experimental animals such as mice, rats, dogs and monkeys, failures in predicting human PK will occur due to interspecies differences.

4. PREDICTION OF HUMAN PK USING CHIMERIC MICE WITH HUMANIZED LIVER

As described in Sections 2 and 3, IVIVE and allometric scaling have been widely used to predict human PK. However, the various issues mentioned above remain. One reason for these issues is that in vivo situations are not reflected in IVIVE methods. In addition, predictions with allometric scaling fail due to interspecies differences in PK. To solve these issues, the application of humanized mouse models has been considered. For human PK prediction methods, genetically humanized mice and chimeric mice with humanized liver are often investigated.⁸⁾ Genetically humanized mice result from the introduction of specific human genes into the mouse genome. The mice then express only the selected human drug metabolizing enzymes and transporters. Accordingly, the use if these mice in predicting human PK is limited to those compounds whose metabolism and disposition involve human drug metabolizing enzymes and transporters expressed in the mice. It is also necessary to consider the effects of drug metabolizing enzymes and transporters originally expressed in the host mice. Mitsui et al.⁷²⁾ predicted human CL for compounds mainly metabolized by CYP3A4 using transgenic mice carrying the human CYP3A4 gene and lacking their mouse Cyp3a gene. They reported an excellent correlation of CL_{int, in vivo} between the mice and humans for six CYP3A4 substrates. Further, they showed that the correlation could be improved by considering the relative contribution of CYP3A4 to overall metabolism in the mice.

In contrast, chimeric mice with humanized liver are generated by transplanting human cells into host mice which have immunodeficiencies and impaired hepatic function. The mice express human drug metabolizing enzymes and transporters in the liver since the mouse hepatocytes are mostly replaced with human hepatocytes.^7–12) Thus, the mice are useful for predicting human PK which is mainly determined by the liver. To produce such chimeric mice with humanized liver, various host mice have been developed, such as uPA/SCID, FRG, TK-NOG, AFC8, Alb-TRECK/SCID and FDG/NOD mice.⁸⁾ Among the models of chimeric mice with humanized liver, the most frequently used are humanized liver uPA/SCID mice, humanized liver TK-NOG mice and humanized liver FRG mice.^7,8,12) A number of researchers reported that levels of mRNA and protein expression and activities of drug metabolizing enzymes and transporters in the livers of chimeric mice with humanized liver were similar to those in human liver,^73–78) suggesting that chimeric mice with humanized liver would be useful in quantitatively predicting human PK.

However, it is only recently that chimeric mice with humanized liver have been applied to the prediction of human PK. In a previous review, we summarized human PK studies using chimeric mice with humanized liver.⁷⁾ In the present review, we add a number of novel research reports and discuss points to consider when applying chimeric mice with humanized liver to predict human PK.

As basic research into predicting human CL, Sanoh et al.⁷⁹⁾ evaluated the relationship of CL_int in humanized liver uPA/SCID mice and humans for 13 compounds metabolized by CYP and/or non-CYP. First, they compared CL_{int, in vitro} estimated from human hepatocytes isolated from the liver of the mice and CL_{int, in vivo} in humans. Human CL_{int, in vivo} was calculated from in vivo human CL_t using the well-stirred model. These CL_int data showed a poor correlation (Fig. 3), suggesting the need for a specific scaling factor for each compound, the ratio of CL_{int, in vitro} and CL_{int, in vivo}, as was also reported by Naritomi et al.^17,18) Next, they compared CL_{int, in vivo} in humanized liver uPA/SCID mice and humans. These CL_{int, in vivo} showed a good correlation except for a few compounds (Fig. 4), suggesting that it would be possible to quantitatively predict human CL based on in vivo PK studies using chimeric mice with humanized liver. However, absolute values of CL_{int, in vivo} in the mice were not consistent with those in humans, and the same scaling factor, the ratio of CL_{int, in vivo} between the mice and humans, were observed among the compounds. To predict human CL_{int, in vivo} using chimeric mice with humanized liver, it is necessary to consider the compound common scaling factor.

Fig. 3. Comparison of CL_{int, in vitro} in Human Hepatocytes Isolated from the Liver of Humanized Liver uPA/SCID Mice and CL_{int, in vivo} in Humans

Solid line represents unity. Dotted lines represent 3-fold differences. Adapted from Sanoh et al.,⁷⁹⁾ with permission from the American Society for Pharmacology and Experimental Therapeutics.

Fig. 4. Comparison of CL_{int, in vivo} in PXB Mice, a Humanized Liver uPA/SCID Mouse Model and Humans

Solid line represents unity. Dotted lines represent 3-fold differences. Adapted from Sanoh et al.,⁷⁹⁾ with permission from the American Society for Pharmacology and Experimental Therapeutics.

Sanoh et al.⁸⁰⁾ reported a human PK prediction method based on allometric scaling using humanized liver uPA/SCID mice. First, they predicted human CL_t and Vd_ss for various compounds metabolized by CYP and/or non-CYP using SSS. 82.4% and 100% of predicted human CL_t and Vd_ss were within 3-fold of the observed values, respectively, indicating that SSS using chimeric mice with humanized liver is useful for quantitatively predicting human PK. However, this method failed the prediction of CL_t for diazepam. In general, the liver of chimeric mice with humanized liver is not perfectly replaced with human hepatocytes.^9–11) The replacement index (RI), the replacement rate of mouse hepatocytes with human hepatocytes, of the mice used in the study was approximately 80%. The authors suggested that the reason for this failure to predict human CL_t for diazepam was metabolism by residual mouse hepatocytes. Next, they evaluated the prediction of intravenous plasma concentration–time curves in humans. They used the complex Dedrick plot,⁸¹⁾ a prediction method for human plasma concentration–time profiles based on allometric scaling. Results also showed good predictability for most of the compounds evaluated. However, this method failed to predict the plasma concentration–time curve for diazepam, and also failed to predict CL_t.

Miyamoto et al.⁸²⁾ evaluated the utility of SSS using chimeric mice with humanized liver by comparing the predictability of human CL_t and Vd_ss using humanized liver uPA/SCID mice, rats and monkeys. When using the mice, 83.3% and 79.3% of the predicted values of CL_t and Vd_ss were within 3-fold of the observed values, respectively. They also showed the excellent predictability of human PK parameters using the mice compared with those with rats and monkeys. These results indicate that chimeric mice with humanized liver are useful for predicting human PK based on SSS. However, when using mice, human CL_t for diazepam was overpredicted, as was also found by Sanoh et al.⁸⁰⁾ In addition, human CL_t was also overpredicted for benzydamine, which is metabolized by FMO, a non-CYP enzyme. It was considered that these overpredictions were due to metabolism of the compound by residual mouse hepatocytes in the liver of the mice. On the other hand, for mycophenolic acid and indomethacin, human CL_t were underpredicted using the mice, estimated due to differences in extra-hepatic metabolism between the mice and humans.

Physiologically based pharmacokinetic (PBPK) modeling is a mathematical model which uses a pharmacokinetic model consisting of multiple compartments. This model mimics human physiology to quantitatively predict concentration–time profiles of compounds in the blood or other biological fluids, and tissues.⁸³⁾ Human PK prediction methods based on PBPK models using chimeric mice with humanized liver have been also reported. Utoh et al.⁸⁴⁾ reported the suitable predictability of human plasma concentration profiles of CYP probes based on a simplified PBPK model consisting of three compartments using humanized liver TK-NOG mice. However, Yamazaki-Nishioka et al.⁸⁵⁾ reported the overprediction of human CL_h for benzydamine, a FMO substrate, based on a simplified PBPK model using humanized liver TK-NOG mice, as was also reported by Miyamoto et al.⁸²⁾ They suggested that the failure of prediction might be due to the effect of oxidation in the kidney, extra-hepatic metabolism. Nakayama et al.⁸⁶⁾ predicted human CL_t and Vd_ss from in vivo PK data of humanized liver uPA/SCID mice using the well-stirred model and the Rodgers equation for 16 CYP and non-CYP substrates, and two OATP1B1 or OATP1B3 substrates. Results showed excellent predictability. Further, they estimated plasma concentration-time profiles in humans using a more complicated PBPK model composed of 11 tissue compartments for 17 compounds. They observed generally successful prediction of human plasma concentration–time profiles.

Here, we consider some points for the application of chimeric mice with humanized liver for predicting human PK. When generating chimeric mice with humanized liver, human hepatocytes transplanted into host mice are usually prepared from a specific donor. Therefore, it is necessary to evaluate the differences in PK among chimeric mice with humanized liver transplanted with human hepatocytes obtained from different donors. As described above, Sanoh et al.⁸⁰⁾ and Miyamoto et al.⁸²⁾ reported the excellent predictability of human CL_t and Vd_ss using chimeric mice with humanized liver based on SSS. These reports used humanized liver uPA/SCID mice transplanted with commercially available human hepatocytes of different lots. Figure 5 compares CL_t and Vd_ss between humanized liver uPA/SCID mice transplanted with human hepatocytes of Lots. BD85 and BD195. PK parameters in the mice transplanted with Lot. BD85 were reported by Sanoh et al,⁸⁰⁾ whereas those transplanted with human hepatocytes of Lot. BD195 were reported by Miyamoto et al.⁸²⁾ Although the CL_t values obtained from Lot. BD195 tended to be slightly higher than those from Lot. BD85, an excellent correlation was observed. Moreover, Vd_ss were similar between mice obtained from Lot. BD85 and 195 for most of the compounds, showing an excellent correlation. In the future, comparison of chimeric mouse with humanized liver PK parameters should be evaluated for additional human hepatocyte lots.

Fig. 5. Comparison of CL_t and Vd_ss between Humanized Liver uPA/SCID Mice Transplanted with Human Hepatocytes of Lot. BD85 and BD195

Solid line represents unity. Data from Sanoh et al.⁸⁰⁾ and Miyamoto et al.⁸²⁾

By considering the relationship of mice and human PK, it is possible to reliably predict human PK. Therefore, estimation of PK in mice for test compounds should incorporate the evaluation of PK for a reference compound whose human PK has already been clarified. For example, prediction of human CL_t and Vd_ss using SSS should include application of the fixed values of a scaling exponent calculated from the relationship between chimeric mouse with humanized liver and human PK for a reference compound (Fig. 6(A)).

Fig. 6. Procedure of SSS (A) and Administration of Compounds (B) in in Vivo PK Studies Using Chimeric Mice with Humanized Liver

A–L: test compounds.

Since chimeric mice with humanized liver are expensive, cost effectiveness can be improved by performing in vivo PK studies by administering test compounds at a certain interval and sequentially estimating the PK, and using the cassette dosing method and simultaneously estimating the PK to reduce the number of mice used (Fig. 6(B)).^82,84,86)

As described above, drug metabolism by residual mouse hepatocytes in the liver of chimeric mice with humanized liver affects the predictability of human PK. To improve predictability, correction methods using CL_{int, in vitro} in humans and animals have been reported.^87,88) To improve prediction accuracy, use of in vitro data using hepatic microsomes in chimeric mice with humanized liver and humans might be effective. Kamimura et al.⁸⁹⁾ estimated CL_t in humanized TK-NOG mice with 100% RI by extrapolating the values of CL_t in humanized TK-NOG mice which showed different RIs for PF-04937319, a glucokinase activator. Further, they predicted human PK based on the estimated CL_t in the mice. To suppress the metabolism of residual mouse hepatocytes, novel types of chimeric mice with humanized liver have been reported. Examples include Cyp3a-knockout chimeric mice with humanized liver,⁹⁰⁾ which avoid drug metabolism by mouse Cyp3a, and mouse reduced nicotinamide adenine dinucleotide phosphate (NADPH)-CYP oxidoreductase-knockout chimeric mice with humanized liver,⁹¹⁾ which avoid mouse CYP metabolism.

It is also important to consider the influence of extra-hepatic metabolism of chimeric mice with humanized liver. Although a difficult issue to solve, incorporation of in vitro metabolic studies using chimeric mouse with humanized liver and human biological materials may be effective.

For IVIVE, it is difficult to estimate CL_{int, in vitro} and predict human PK for very metabolically stable compounds. On the contrary, it is easy to predict human PK using chimeric mice with humanized liver because the in vivo PK data of mice are directly used in the predictions.

Human PK prediction methods using chimeric mice with humanized liver have mainly evaluated CYP and non-CYP substrates. Results have suggested that it is possible to simultaneously evaluated both types of substrate.^80,82,86) However, the prediction of drug transporter-mediated human PK remains limited; future studies may need to evaluate the prediction of human PK for various compounds involved in transport and integrate results for prediction of CYP-, non-CYP- and transporter-mediated human PK.

5. CONCLUSION

This review summarizes the prediction of human PK using IVIVE, allometric scaling and chimeric mice with humanized liver, and discusses the utility of predicting human PK using these mice in drug discovery by comparing results with those obtained with IVIVE and allometric scaling. IVIVE is a typical method for predicting human PK and has been applied to numerous compounds. However, inconsistency between predicted and actual CL has often been observed, and various correction methods have been reported. For IVIVE, it is necessary to establish separate methods for CYP, non-CYP enzyme and drug transporter metabolism and then integrate these methods in predicting human PK. It is difficult to evaluate very metabolically stable compounds, although various approaches such as the relay method, HepatoPac, HμREL and upcyte have been proposed. Allometric scaling using experimental animals has been also often used. Since there have been many failures in predicting human PK due to interspecies differences, various correction methods have been proposed. However, these difficulties in prediction still remain. In contrast, it is considered that PK in chimeric mice with humanized liver mimics that in humans, and that human PK prediction methods such as allometric scaling and PBPK modeling using in vivo PK data of these mice may solve the issues of IVIVE and allometric scaling. Nevertheless, a number of issues in predicting human PK remain, such as the influence of residual mouse hepatocytes and extra-hepatic metabolism. Solving these issues would make human PK prediction methods using chimeric mice with humanized liver more effective and useful.

Conflict of Interest

S.S. and S.O. have performed a collaborative study with PhoenixBio Co., Ltd. The authors report no other conflict of interest.

REFERENCES

• 1) Kola I, Landis J. Can the pharmaceutical industry reduce attrition rates? Nat. Rev. Drug Discov., 3, 711–715 (2004).
• 2) Waring MJ, Arrowsmith J, Leach AR, Leeson PD, Mandrell S, Owen RM, Pairaudeau G, Pennie WD, Pickett SD, Wang J, Wallace O, Weir A. An analysis of the attrition of drug candidates from four major pharmaceutical companies. Nat. Rev. Drug Discov., 14, 475–486 (2015).
• 3) Naritomi Y. Future perspectives of predicting human pharmacokinetics. Drug Metab. Pharmacokinet., 29, 287–288 (2014).
• 4) Naritomi Y, Nakamori F, Furukawa T, Tabata K. Prediction of hepatic and intestinal glucuronidation using in vitro–in vivo extrapolation. Drug Metab. Pharmacokinet., 30, 21–29 (2015).
• 5) Zou P, Yu Y, Zheng N, Yang Y, Paholak HJ, Yu LX, Sun D. Applications of human pharmacokinetic prediction in first-in-human dose estimation. AAPS J., 14, 262–281 (2012).
• 6) Zou P, Zheng N, Yang Y, Yu LX, Sun D. Prediction of volume of distribution at steady state in humans: comparison of different approaches. Expert Opin. Drug Metab. Toxicol., 8, 855–872 (2012).
• 7) Naritomi Y, Sanoh S, Ohta S. Chimeric mice with humanized liver: application in drug metabolism and pharmacokinetics studies for drug discovery. Drug Metab. Pharmacokinet., 33, 31–39 (2018).
• 8) Scheer N, Wilson ID. A comparison between genetically humanized and chimeric liver humanized mouse models for studies in drug metabolism and toxicity. Drug Discov. Today, 21, 250–263 (2016).
• 9) Yoshizato K, Tateno C. In vivo modeling of human liver for pharmacological study using humanized mouse. Expert Opin. Drug Metab. Toxicol., 5, 1435–1446 (2009).
• 10) Strom SC, Davila J, Grompe M. Chimeric mice with humanized liver: tools for the study of drug metabolism, excretion, and toxicity. Methods Mol. Biol., 640, 491–509 (2010).
• 11) Sanoh S, Ohta S. Chimeric mice transplanted with human hepatocytes as a model for prediction of human drug metabolism and pharmacokinetics. Biopharm. Drug Dispos., 35, 71–86 (2014).
• 12) Xu D, Peltz G. Can humanized mice predict drug “behavior” in humans? Annu. Rev. Pharmacol. Toxicol., 56, 323–338 (2016).
• 13) Kamimura H, Ito S. Assessment of chimeric mice with humanized livers in new drug development: generation of pharmacokinetics, metabolism and toxicity data for selecting the final candidate compound. Xenobiotica, 46, 557–569 (2016).
• 14) Rane A, Wilkinson GR, Shand DG. Prediction of hepatic extraction ratio from in vitro measurement of intrinsic clearance. J. Pharmacol. Exp. Ther., 200, 420–424 (1977).
• 15) Pang KS, Rowland M. Hepatic clearance of drugs. I. Theoretical considerations of a “well-stirred” model and a “parallel tube” model. Influence of hepatic blood flow, plasma and blood cell binding, and the hepatocellular enzymatic activity on hepatic drug clearance. J. Pharmacokinet. Biopharm., 5, 625–653 (1977).
• 16) Roberts MS, Rowland M. Correlation between in-vitro microsomal enzyme activity and whole organ hepatic elimination kinetics: analysis with a dispersion model. J. Pharm. Pharmacol., 38, 177–181 (1986).
• 17) Naritomi Y, Terashita S, Kimura S, Suzuki A, Kagayama A, Sugiyama Y. Prediction of human hepatic clearance from in vivo animal experiments and in vitro metabolic studies with liver microsomes from animals and humans. Drug Metab. Dispos., 29, 1316–1324 (2001).
• 18) Naritomi Y, Terashita S, Kagayama A, Sugiyama Y. Utility of hepatocytes in predicting drug metabolism: comparison of hepatic intrinsic clearance in rats and humans in vivo and in vitro. Drug Metab. Dispos., 31, 580–588 (2003).
• 19) Obach RS. Prediction of human clearance of twenty-nine drugs from hepatic microsomal intrinsic clearance data: an examination of in vitro half-life approach and nonspecific binding to microsomes. Drug Metab. Dispos., 27, 1350–1359 (1999).
• 20) Poulin P, Kenny JR, Hop CE, Haddad S. In vitro–in vivo extrapolation of clearance: modeling hepatic metabolic clearance of highly bound drugs and comparative assessment with existing calculation methods. J. Pharm. Sci., 101, 838–851 (2012).
• 21) Testa B, Pedretti A, Vistoli G. Reactions and enzymes in the metabolism of drugs and other xenobiotics. Drug Discov. Today, 17, 549–560 (2012).
• 22) Cerny MA. Prevalence of non-cytochrome P450-mediated metabolism in food and drug administration-approved oral and intravenous drugs: 2006–2015. Drug Metab. Dispos., 44, 1246–1252 (2016).
• 23) Jinno N, Tagashira M, Tsurui K, Yamada S. Contribution of cytochrome P450 and UDT-glucuronosyltransferase to the metabolism of drugs containing carboxylic acid groups: risk assessment of acylglucuronides using human hepatocytes. Xenobiotica, 44, 677–686 (2014).
• 24) Di L, Keefer C, Scott DO, Strelevitz TJ, Chang G, Bi YA, Lai Y, Duckworth J, Fenner K, Troutman MD, Obach RS. Mechanistic insights from comparing intrinsic clearance values between human liver microsomes and hepatocytes to guide drug design. Eur. J. Med. Chem., 57, 441–448 (2012).
• 25) Boase S, Miners JO. In vitro–in vivo correlations for drugs eliminated by glucuronidation: investigations with the model substrate zidovudine. Br. J. Clin. Pharmacol., 54, 493–503 (2002).
• 26) Soars MG, Burchell B, Riley RJ. In vitro analysis of human drug glucuronidation and prediction of in vivo metabolic clearance. J. Pharmacol. Exp. Ther., 301, 382–390 (2002).
• 27) Miners JO, Mackenzie PI, Knights KM. The prediction of drug-glucuronidation parameters in humans: UDP-glucuronosyltransferase enzyme-selective substrate and inhibitor probes for reaction phenotyping and in vitro–in vivo extrapolation of drug clearance and drug–drug interaction potential. Drug Metab. Rev., 42, 196–208 (2010).
• 28) Zientek M, Jiang Y, Youdim K, Obach RS. In vitro–in vivo correlation for intrinsic clearance for drugs metabolized by human aldehyde oxidase. Drug Metab. Dispos., 38, 1322–1327 (2010).
• 29) Akabane T, Gerst N, Masters JN, Tamura K. A quantitative approach to hepatic clearance prediction of metabolism by aldehyde oxidase using custom pooled hepatocytes. Xenobiotica, 42, 863–871 (2012).
• 30) Cubitt HE, Houston JB, Galetin A. Prediction of human drug clearance by multiple metabolic pathways: integration of hepatic and intestinal microsomal and cytosolic data. Drug Metab. Dispos., 39, 864–873 (2011).
• 31) Nishimuta H, Houston JB, Galetin A. Hepatic, intestinal, renal, and plasma hydrolysis of prodrugs in human, cynomolgus monkey, dog, and rat: implications for in vitro–in vivo extrapolation of clearance of prodrugs. Drug Metab. Dispos., 42, 1522–1531 (2014).
• 32) Jones BC, Srivastava A, Colclough N, Wilson J, Reddy VP, Amberntsson S, Li D. An investigation into the prediction of in vivo clearance for a range of flavin-containing monooxygenase substrates. Drug Metab. Dispos., 45, 1060–1067 (2017).
• 33) Kamel A, Colizza K, Gunduz M, Harriman S, Obach RS. In vitro–in vivo correlation for intrinsic clearance for CP-409,092 and sumatriptan: a case study to predict the in vivo clearance for compounds metabolized by monoamine oxidase. Xenobiotica, 42, 355–362 (2012).
• 34) Masuo Y, Nagamori S, Hasegawa A, Hayashi K, Isozumi N, Nakamichi N, Kanai Y, Kato Y. Utilization of liver microsomes to estimate hepatic intrinsic clearance of monoamine oxidase substrate drugs in humans. Pharm. Res., 34, 1233–1243 (2017).
• 35) Kusuhara H, Sugiyama Y. In vitro–in vivo extrapolation of transporter-mediated clearance in the liver and kidney. Drug Metab. Pharmacokinet., 24, 37–52 (2009).
• 36) Watanabe T, Kusuhara H, Maeda K, Shitara Y, Sugiyama Y. Physiologically based pharmacokinetic modeling to predict transporter-mediated clearance and distribution of pravastatin in humans. J. Pharmacol. Exp. Ther., 328, 652–662 (2009).
• 37) De Bruyn T, Ufuk A, Cantrill C, Kosa RE, Bi YA, Niosi M, Modi S, Rodrigues AD, Tremaine LM, Varma MVS, Galetin A, Houston JB. Predicting human clearance of organic anion transporting polypeptide substrates using cynomolgus monkey: in vitro–in vivo scaling of hepatic uptake clearance. Drug Metab. Dispos., 46, 989–1000 (2018).
• 38) Jones HM, Barton HA, Lai Y, Bi YA, Kimoto E, Kempshall S, Tate SC, El-Kattan A, Houston JB, Galetin A, Fenner KS. Mechanistic pharmacokinetic modeling for the prediction of transporter-mediated disposition in humans from sandwich culture human hepatocyte data. Drug Metab. Dispos., 40, 1007–1017 (2012).
• 39) Ménochet K, Kenworthy KE, Houston JB, Galetin A. Use of mechanistic modeling to assess interindividual variability and interspecies differences in active uptake in human and rat hepatocytes. Drug Metab. Dispos., 40, 1744–1756 (2012).
• 40) Feng B, Varma MV, Costales C, Zhang H, Tremaine L. In vitro and in vivo approaches to characterize transporter-mediated disposition in drug discovery. Expert Opin. Drug Discov., 9, 873–890 (2014).
• 41) Izumi S, Nozaki Y, Komori T, Takenaka O, Maeda K, Kusuhara H, Sugiyama Y. Comparison of the predictability of human hepatic clearance for organic anion transporting polypeptide substrate drugs between different in vitro–in vivo extrapolation approaches. J. Pharm. Sci., 106, 2678–2687 (2017).
• 42) Hutzler JM, Ring BJ, Anderson SR. Low-turnover drug molecules: a current challenge for drug metabolism scientists. Drug Metab. Dispos., 43, 1917–1928 (2015).
• 43) Di L, Trapa P, Obach RS, Atkinson K, Bi YA, Wolford AC, Tan B, McDonald TS, Lai Y, Tremaine LM. A novel relay method for determining low-clearance values. Drug Metab. Dispos., 40, 1860–1865 (2012).
• 44) Chan TS, Yu H, Moore A, Khetani SR, Tweedie D. Meeting the challenge of predicting hepatic clearance of compounds slowly metabolized by cytochrome P450 using a novel hepatocyte model, HepatoPac. Drug Metab. Dispos., 41, 2024–2032 (2013).
• 45) Da-Silva F, Boulenc X, Vermet H, Compigne P, Gerbal-Chaloin S, Daujat-Chavanieu M, Klieber S, Poulin P. Improving prediction of metabolic clearance using quantitative extrapolation of results obtained from human hepatic micropatterned cocultures model and by considering the impact of albumin binding. J. Pharm. Sci., 107, 1957–1972 (2018).
• 46) Bonn B, Svanberg P, Janefeldt A, Hultman Ia, Grime K. Determination of human hepatocyte intrinsic clearance for slowly metabolized compounds: comparison of a primary hepatocyte/stromal cell co-culture with plated primary hepatocytes and HepaRG. Drug Metab. Dispos., 44, 527–533 (2016).
• 47) Hultman Ia, Vedin C, Abrahamsson A, Winiwarter S, Darnell M. Use of HμREL human coculture system for prediction of intrinsic clearance and metabolite formation for slowly metabolized compounds. Mol. Pharm., 13, 2796–2807 (2016).
• 48) Schaefer M, Schänzle G, Bischoff D, Süssmuth RD. Upcyte human hepatocytes: a potent in vitro tool for the prediction of hepatic clearance of metabolically stable compounds. Drug Metab. Dispos., 44, 435–444 (2016).
• 49) Peters SA, Jones CR, Ungell AL, Hatley OJ. Predicting drug extraction in the human gut wall: assessing contributions from drug metabolizing enzymes and transporter proteins using preclinical models. Clin. Pharmacokinet., 55, 673–696 (2016).
• 50) Yang J, Jamei M, Yeo KR, Tucker GT, Rostami-Hodjegan A. Prediction of intestinal first-pass drug metabolism. Curr. Drug Metab., 8, 676–684 (2007).
• 51) Kadono K, Akabane T, Tabata K, Gato K, Terashita S, Teramura T. Quantitative prediction of intestinal metabolism in humans from a simplified intestinal availability model and empirical scaling factor. Drug Metab. Dispos., 38, 1230–1237 (2010).
• 52) Gill KL, Houston JB, Galetin A. Characterization of in vitro glucuronidation clearance of a range of drugs in human kidney microsomes: comparison with liver and intestinal glucuronidation and impact of albumin. Drug Metab. Dispos., 40, 825–835 (2012).
• 53) Scotcher D, Jones C, Posada M, Galetin A, Rostami-Hodjegan A. Key to opening kidney for in vitro–in vivo extrapolation entrance in health and disease: Part II: mechanistic models and in vitro–in vivo extrapolation. AAPS J., 18, 1082–1094 (2016).
• 54) Mathialagan S, Piotrowski MA, Tess DA, Feng B, Litchfield J, Varma MV. Quantitative prediction of human renal clearance and drug–drug interactions of organic anion transporter substrates using in vitro transport data: a relative activity factor approach. Drug Metab. Dispos., 45, 409–417 (2017).
• 55) Scotcher D, Billington S, Brown J, Jones CR, Brown CDA, Rostami-Hodjegan A, Galetin A. Microsomal and cytosolic scaling factors in dog and human kidney cortex and application for in vitro–in vivo extrapolation of renal metabolic clearance. Drug Metab. Dispos., 45, 556–568 (2017).
• 56) Mahmood I, Balian JD. Interspecies scaling: predicting clearance of drugs in humans. Three different approaches. Xenobiotica, 26, 887–895 (1996).
• 57) Ward KW, Smith BR. A comprehensive quantitative and qualitative evaluation of extrapolation of intravenous pharmacokinetic parameters from rat, dog, and monkey to humans. I. clearance. Drug Metab. Dispos., 32, 603–611 (2004).
• 58) Davies B, Morris T. Physiological parameters in laboratory animals and humans. Pharm. Res., 10, 1093–1395 (1993).
• 59) Brown RP, Delp MD, Lindstedt SL, Rhomberg LR, Beliles RP. Physiological parameter values for physiologically based pharmacokinetic models. Toxicol. Ind. Health, 13, 407–484 (1997).
• 60) Nagilla R, Ward KW. A comprehensive analysis of the role of correction factors in the allometric predictivity of clearance from rat, dog, and monkey to humans. J. Pharm. Sci., 93, 2522–2534 (2004).
• 61) Ward KW, Nagilla R, Jolivette LJ. Comparative evaluation of oral systemic exposure of 56 xenobiotics in rat, dog, monkey and human. Xenobiotica, 35, 191–210 (2005).
• 62) Tang H, Hussain A, Leal M, Mayersohn M, Fluhler E. Interspecies prediction of human drug clearance based on scaling data from one or two animal species. Drug Metab. Dispos., 35, 1886–1893 (2007).
• 63) Tang H, Mayersohn M. A novel model for prediction of human drug clearance by allometric scaling. Drug Metab. Dispos., 33, 1297–1303 (2005).
• 64) Hosea NA, Collard WT, Cole S, Maurer TS, Fang RX, Jones H, Kakar SM, Nakai Y, Smith BJ, Webster R, Beaumont K. Prediction of human pharmacokinetics from preclinical information: comparative accuracy of quantitative prediction approaches. J. Clin. Pharmacol., 49, 513–533 (2009).
• 65) Deguchi T, Watanabe N, Kurihara A, Igeta K, Ikenaga H, Fusegawa K, Suzuki N, Murata S, Hirouchi M, Furuta Y, Iwasaki M, Okazaki O, Izumi T. Human pharmacokinetic prediction of UDP-glucuronosyltransferase substrates with an animal scale-up approach. Drug Metab. Dispos., 39, 820–829 (2011).
• 66) Yoshimatsu H, Konno Y, Ishii K, Satsukawa M, Yamashita S. Usefulness of minipigs for predicting human pharmacokinetics: prediction of distribution volume and plasma clearance. Drug Metab. Pharmacokinet., 31, 73–81 (2016).
• 67) Lombardo F, Waters NJ, Argikar UA, Dennehy MK, Zhan J, Gunduz M, Harriman SP, Berellini G, Liric Rajlic I, Obach RS. Comprehensive assessment of human pharmacokinetic prediction based on in vivo animal pharmacokinetic data, part 2: clearance. J. Clin. Pharmacol., 53, 178–191 (2013).
• 68) Ring BJ, Chien JY, Adkison KK, Jones HM, Rowland M, Jones RD, Yates JW, Ku MS, Gibson CR, He H, Vuppugalla R, Marathe P, Fischer V, Dutta S, Sinha VK, Björnsson T, Lavé T, Poulin PPRMA. CPCDC initiative on predictive models of human pharmacokinetics, part 3: comparative assessement of prediction methods of human clearance. J. Pharm. Sci., 100, 4090–4110 (2011).
• 69) Liu D, Song H, Song L, Liu Y, Cao Y, Jiang J, Hu P. A unified strategy in selection of the best allometric scaling methods to predict human clearance based on drug disposition pathway. Xenobiotica, 46, 1105–1111 (2016).
• 70) Kang HE, Lee MG. Approaches for predicting human pharmacokinetics using interspecies pharmacokinetic scaling. Arch. Pharm. Res., 34, 1779–1788 (2011).
• 71) Lombardo F, Waters NJ, Argikar UA, Dennehy MK, Zhan J, Gunduz M, Harriman SP, Berellini G, Rajlic IL, Obach RS. Comprehensive assessment of human pharmacokinetic prediction based on in vivo animal pharmacokinetic data, part 1: volume of distribution at steady state. J. Clin. Pharmacol., 53, 167–177 (2013).
• 72) Mitsui T, Nemoto T, Miyake T, Nagao S, Ogawa K, Kato M, Ishigai M, Yamada H. A useful model capable of predicting the clearance of cytochrome 3A4 (CYP3A4) substrates in humans: validity of CYP3A4 transgenic mice lacking their own Cyp3a enzymes. Drug Metab. Dispos., 42, 1540–1547 (2014).
• 73) Tateno C, Yoshizane Y, Saito N, Kataoka M, Utoh R, Yamasaki C, Tachibana A, Soeno Y, Asahina K, Hino H, Asahara T, Yokoi T, Furukawa T, Yoshizato K. Near completely humanized liver in mice shows human-type metabolic responses to drugs. Am. J. Pathol., 165, 901–912 (2004).
• 74) Katoh M, Matsui T, Nakajima M, Tateno C, Kataoka M, Soeno Y, Horie T, Iwasaki K, Yoshizato K, Yokoi T. Expression of human cytochromes P450 in chimeric mice with humanized liver. Drug Metab. Dispos., 32, 1402–1410 (2004).
• 75) Katoh M, Matsui T, Okumura H, Nakajima M, Nishimura M, Naito S, Tateno C, Yoshizato K, Yokoi T. Expression of human phase II enzymes in chimeric mice with humanized liver. Drug Metab. Dispos., 33, 1333–1340 (2005).
• 76) Yamasaki C, Kataoka M, Kato Y, Kakuni M, Usuda S, Ohzone Y, Matsuda S, Adachi Y, Ninomiya S, Itamoto T, Asahara T, Yoshizato K, Tateno C. In vitro evaluation of cytochrome P450 and glucuronidation activities in hepatocytes isolated from liver-humanized mice. Drug Metab. Pharmacokinet., 25, 539–550 (2010).
• 77) Hasegawa M, Kawai K, Mitsui T, Taniguchi K, Monnai M, Wakui M, Ito M, Suematsu M, Peltz G, Nakamura M, Suemizu H. The reconstituted ‘humanized liver’ in TK-NOG mice is mature and functional. Biochem. Biophys. Res. Commun., 405, 405–410 (2011).
• 78) Ohtsuki S, Kawakami H, Inoue T, Nakamura K, Tateno C, Katsukura Y, Obuchi W, Uchida Y, Kamiie J, Horie T, Terasaki T. Validation of uPA/SCID mouse with humanized liver as a human liver model: protein quantification of transporters, cytochromes P450, and UDP-glucuronosyltransferases by LC-MS/MS. Drug Metab. Dispos., 42, 1039–1043 (2014).
• 79) Sanoh S, Horiguchi A, Sugihara K, Kotake Y, Tayama Y, Ohshita H, Tateno C, Horie T, Kitamura S, Ohta S. Prediction of in vivo hepatic clearance and half-life of drug candidates in human using chimeric mice with humanized liver. Drug Metab. Dispos., 40, 322–328 (2012).
• 80) Sanoh S, Naritomi Y, Fujimoto M, Sato K, Kawamura A, Horiguchi A, Sugihara K, Kotake Y, Ohshita H, Tateno C, Horie T, Kitamura S, Ohta S. Predictability of plasma concentration–time curves in humans using single-species allometric scaling of chimeric mice with humanized liver. Xenobiotica, 45, 605–614 (2015).
• 81) Boxenbaum H, Ronfeld R. Interspecies pharmacokinetic scaling and the Dedrick plots. Am. J. Physiol., 245, R768–R775 (1983).
• 82) Miyamoto M, Iwasaki S, Chisaki I, Nakagawa S, Amano N, Hirabayashi H. Comparison of predictability for human pharmacokinetics parameters among monkeys, rats, and chimeric mice with humanised liver. Xenobiotica, 47, 1052–1063 (2017).
• 83) Saeheng T, Na-Bangchang K, Karbwang J. Utility of physiologically based pharmacokinetic (PBPK) modeling in oncology drug development and its accuracy: a systematic review. Eur. J. Clin. Pharmacol., 74, 1365–1376 (2018).
• 84) Utoh M, Suemizu H, Mitsui M, Kawano M, Toda A, Uehara S, Uno Y, Shimizu M, Sasaki E, Yamazaki H. Human plasma concentrations of cytochrome P450 probe cocktails extrapolated from pharmacokinetics in mice transplanted with human hepatocytes and from pharmacokinetics in common marmosets using physiologically based pharmacokinetic modeling. Xenobiotica, 46, 1049–1055 (2016).
• 85) Yamazaki-Nishioka M, Shimizu M, Suemizu H, Nishiwaki M, Mitsui M, Yamazaki H. Human plasma metabolic profiles of benzydamine, a flavin-containing monooxygenase probe substrate, simulated with pharmacokinetic data from control and humanized-liver mice. Xenobiotica, 48, 117–123 (2018).
• 86) Nakayama K, Ito S, Suzuki M, Takubo H, Yamazaki H, Nomura Y. Prediction of human pharmacokinetics of typical compounds by a physiologically based method using chimeric mice with humanized liver. Xenobiotica, 48, 1460516 (2018).
• 87) Lave T, Dupin S, Schmitt C, Chou RC, Jaeck D, Coassolo P. Integration of in vitro data into allometric scaling to predict hepatic metabolic clearance in man: application to 10 extensively metabolized drugs. J. Pharm. Sci., 86, 584–590 (1997).
• 88) Sayama H, Komura H, Kogayu M. Application of hybrid approach based on empirical and physiological concept for predicting pharmacokinetics in humans-usefulness of exponent on prospective evaluation of predictability. Drug Metab. Dispos., 41, 498–507 (2013).
• 89) Kamimura H, Ito S, Chijiwa H, Okuzono T, Ishiguro T, Yamamoto Y, Nishinoaki S, Ninomiya SI, Mitsui M, Kalgutkar AS, Yamazaki H, Suemizu H. Simulation of human plasma concentration–time profiles of the partial glucokinase activator PF-04937319 and its disproportionate N-demethylated metabolite using humanized chimeric mice and semi-physiological pharmacokinetic modeling. Xenobiotica, 47, 382–393 (2017).
• 90) Kato K, Ohbuchi M, Hamamura S, Ohshita H, Kazuki Y, Oshimura M, Sato K, Nakada N, Kawamura A, Usui T, Kamimura H, Tateno C. Development of murine Cyp3a knockout chimeric mice with humanized liver. Drug Metab. Dispos., 43, 1208–1217 (2015).
• 91) Barzi M, Pankowicz FP, Zorman B, Liu X, Legras X, Yang D, Borowiak M, Bissig-Choisat B, Sumazin P, Li F, Bissig KD. A novel humanized mouse lacking murine P450 oxidoreductase for studying human drug metabolism. Nat. Commun., 8, 39 (2017).