Advertisement

Improved Prediction of the Drug-Drug Interactions of Pemafibrate Caused by Cyclosporine A and Rifampicin via PBPK Modeling: Consideration of the Albumin-Mediated Hepatic Uptake of Pemafibrate and Inhibition Constants With Preincubation Against OATP1B

  • Ji Eun Park
    Affiliations
    Sugiyama Laboratory, RIKEN Baton Zone Program, RIKEN Cluster for Science, RIKEN, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama 230-0045, Japan

    Pharmacokinetics, Dynamics and Metabolism, Translational Medicine and Early Development, R&D, Sanofi K.K., 3 Chome-20-2, Nishishinjuku, Shinjuku-ku, Tokyo, 160-0023, Japan
    Search for articles by this author
  • Yoshihisa Shitara
    Affiliations
    Pharmacokinetics, Dynamics and Metabolism, Translational Medicine and Early Development, R&D, Sanofi K.K., 3 Chome-20-2, Nishishinjuku, Shinjuku-ku, Tokyo, 160-0023, Japan
    Search for articles by this author
  • Wooin Lee
    Affiliations
    College of Pharmacy and Research Institute of Pharmaceutical Sciences, Seoul National University, Bldg 21 Rm 309, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, S. Korea
    Search for articles by this author
  • Shigemichi Morita
    Affiliations
    Pharmacokinetics, Dynamics and Metabolism, Translational Medicine and Early Development, R&D, Sanofi K.K., 3 Chome-20-2, Nishishinjuku, Shinjuku-ku, Tokyo, 160-0023, Japan
    Search for articles by this author
  • Jasminder Sahi
    Affiliations
    Pharmacokinetics, Dynamics and Metabolism, Translational Medicine and Early Development, R&D, Sanofi China, 1228 Yan'an Middle Road, Jing'an District, Shanghai, China
    Search for articles by this author
  • Kota Toshimoto
    Affiliations
    Sugiyama Laboratory, RIKEN Baton Zone Program, RIKEN Cluster for Science, RIKEN, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama 230-0045, Japan
    Search for articles by this author
  • Yuichi Sugiyama
    Correspondence
    Corresponding author.
    Affiliations
    Sugiyama Laboratory, RIKEN Baton Zone Program, RIKEN Cluster for Science, RIKEN, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama 230-0045, Japan
    Search for articles by this author
Published:October 12, 2020DOI:https://doi.org/10.1016/j.xphs.2020.10.016

      Abstract

      Pemafibrate (PMF) is highly albumin-bound (>99.8%) and a substrate for hepatic uptake transporters (OATP1B) and CYP enzymes. Here, we developed a PBPK model of PMF to capture drug-drug interactions (DDI) incurred by cyclosporine (CsA) and rifampicin (RIF), the two OATP1B inhibitors. Initial PBPK modeling of PMF utilized in vitro hepatic uptake clearance (PSinf) obtained in the absence of albumin, but failed in capturing the blood PMF pharmacokinetic (PK) profiles. Based on the results that in vitro PSinf of unbound PMF was enhanced in the presence of albumin, we applied the albumin-facilitated dissociation model and the resulting PSinf parameters improved the prediction of the blood PMF PK profiles. In refining our PBPK model toward improved prediction of the observed DDI data (PMF co-administered with single dosing of CsA or RIF; PMF following multiple RIF dosing), we adjusted the previously obtained in vivo OATP1B inhibition constants (Ki,OATP1B) of CsA or RIF for pitavastatin by correcting for substrate-dependency. We also incorporated the induction of OATP1B and CYP enzymes after multiple RIF dosing. Sensitivity analysis informed that the higher gastrointestinal absorption rate constant could further improve capturing the observed DDI data, suggesting the possible inhibition of intestinal ABC transporter(s) by CsA or RIF.

      Keywords

      Introduction

      Pemafibrate (PMF, Parmodia®) belongs to the class of fibrates (selective peroxisome proliferator-activated receptor-α modulators) and it is approved in Japan for the hyperlipidemia therapy.
      • Fruchart J.C.
      Pemafibrate (K-877), a novel selective peroxisome proliferator-activated receptor alpha modulator for management of atherogenic dyslipidaemia.
      Encouraged by the findings that PMF displays efficacy and safety profiles superior to other conventional fibrates,
      • Ishibashi S.
      • Yamashita S.
      • Arai H.
      • et al.
      Effects of K-877, a novel selective PPARalpha modulator (SPPARMalpha), in dyslipidaemic patients: a randomized, double blind, active- and placebo-controlled, phase 2 trial.
      ,
      • Ishibashi S.
      • Arai H.
      • Yokote K.
      • et al.
      Efficacy and safety of pemafibrate (K-877), a selective peroxisome proliferator-activated receptor alpha modulator, in patients with dyslipidemia: results from a 24-week, randomized, double blind, active-controlled, phase 3 trial.
      the phase III trial (PROMINENT) is ongoing in approximately 10,000 participants with type 2 diabetes mellitus.
      • Pradhan A.D.
      • Paynter N.P.
      • Everett B.M.
      • et al.
      Rationale and design of the pemafibrate to reduce cardiovascular outcomes by reducing triglycerides in patients with diabetes (PROMINENT) study.
      There is growing interest in expanding the use of PMF in patients with cardiovascular disease, and it is thus important to enhance our understanding of the pharmacokinetics (PKs) of PMF as well as the potential drug-drug interactions (DDIs).
      PMF is currently approved only in Japan (June 2018), and the following information is available on the PKs of PMF. The PKs for PMF was proportional to the oral doses ranging from 0.1 to 0.8 mg.
      Pharmaceuticals and Medical Devices Agency.
      PMF is a substrate of organic anion transporting polypeptides 1B1 (OATP1B1), and metabolizing enzymes such as cytochrome P450 (CYP) 3A4/2C8/2C9.
      • Ogawa S.I.
      • Shimizu M.
      • Yamazaki H.
      Plasma concentrations of pemafibrate with co-administered drugs predicted by physiologically based pharmacokinetic modeling in virtual populations with renal/hepatic impairment.
      In clinical studies that assessed the potential for DDIs, the systemic exposure of PMF was increased by co-administration of a single dose of cyclosporine A (CsA; increasing the AUC and Cmax of PMF by 14- and 9-fold, respectively) or a single dose of rifampicin (RIF; increasing the AUC and Cmax of PMF by 11- and 9-fold, respectively).
      Pharmaceuticals and Medical Devices Agency.
      On the other hand, the systemic exposure of PMF was decreased by multiple doses of RIF (AUC, to 0.22-fold; Cmax, to 0.38-fold).
      Pharmaceuticals and Medical Devices Agency.
      CsA is an inhibitor of OATP1B, and RIF can cause complex DDIs via multiple mechanisms, including the induction of CYPs and OATP1B (by repeated dosing), and the inhibition of OATP1B (by co-administration).
      • Niemi M.
      • Backman J.T.
      • Fromm M.F.
      • Neuvonen P.J.
      • Kivisto K.T.
      Pharmacokinetic interactions with rifampicin : clinical relevance.
      ,
      • Kapetas A.J.
      • Sorich M.J.
      • Rodrigues A.D.
      • Rowland A.
      Guidance for rifampin and midazolam dosing protocols to study intestinal and hepatic cytochrome P450 (CYP) 3A4 induction and de-induction.
      The observed changes in the PMF exposure caused by CsA and RIF likely involve multiple mechanisms. Physiologically based pharmacokinetic (PBPK) modeling for PMF may offer mechanistic insights into the complex DDIs caused by CsA (incorporating the inhibition of OATP1B), and RIF (incorporating the inhibition of OATP1B and the induction of OATP1B and CYP3A4/2C8/2C9).
      In capturing the PK profiles of PMF using a PBPK model under control and DDI conditions, the following considerations are worthwhile, especially in improving the prediction accuracy: albumin-facilitated hepatic uptake and substrate-dependent inhibition potency. In the case of PMF (a substrate for OATP1B1), it is important to identify the rate-determining process of the hepatic elimination (according to the extended clearance concept). The in vitro-to-in vivo extrapolation (IVIVE) method (bottom-up approach) is commonly used to quantitatively predict in vivo hepatic uptake clearance using in vitro experimental systems (e.g., over-expression systems, human hepatocytes), but for highly protein-bound drugs, discrepancies tend to persist between bottom-up and top-down approaches.
      • Miyauchi S.
      • Masuda M.
      • Kim S.J.
      • et al.
      The phenomenon of albumin-mediated hepatic uptake of organic anion transport polypeptide substrates: prediction of the in vivo uptake clearance from the in vitro uptake by isolated hepatocytes using a facilitated-dissociation model.
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      • Poulin P.
      • Haddad S.
      Albumin and uptake of drugs in cells: additional validation exercises of a recently published equation that quantifies the albumin-facilitated uptake mechanism(s) in physiologically based pharmacokinetic and pharmacodynamic modeling Research.
      • Bteich M.
      • Poulin P.
      • Haddad S.
      The potential protein-mediated hepatic uptake: discussion on the molecular interactions between albumin and the hepatocyte cell surface and their implications for the in vitro-to-in vivo extrapolations of hepatic clearance of drugs.
      One of plausible mechanisms is the enhancing effect of albumin on the hepatic uptake. Several reports indicated that the in vivo hepatic clearance of OATP1B substrate drugs was better described using the in vitro uptake clearance measured in the presence of albumin (by incorporating a model that describes the albumin-mediated hepatic uptake),
      • Miyauchi S.
      • Masuda M.
      • Kim S.J.
      • et al.
      The phenomenon of albumin-mediated hepatic uptake of organic anion transport polypeptide substrates: prediction of the in vivo uptake clearance from the in vitro uptake by isolated hepatocytes using a facilitated-dissociation model.
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      • Poulin P.
      • Haddad S.
      Albumin and uptake of drugs in cells: additional validation exercises of a recently published equation that quantifies the albumin-facilitated uptake mechanism(s) in physiologically based pharmacokinetic and pharmacodynamic modeling Research.
      • Bteich M.
      • Poulin P.
      • Haddad S.
      The potential protein-mediated hepatic uptake: discussion on the molecular interactions between albumin and the hepatocyte cell surface and their implications for the in vitro-to-in vivo extrapolations of hepatic clearance of drugs.
      compared to the conventional free-drug hypothesis (in which the free fraction of a drug governs the hepatic uptake rate
      • Shand D.G.
      • Cotham R.H.
      • Wilkinson G.R.
      Perfusion-limited of plasma drug binding on hepatic drug extraction.
      ). Another consideration is the in vitro/in vivo discrepancy of inhibition potency. For the known inhibitors of OATP1B (CsA and RIF), the inhibition constants against OATP1B (Ki,OATP1B) display considerable discrepancies between in vitro and in vivo (in vitro Ki,OATP1B values being 5–20-fold larger than in vivo Ki,OATP1B values).
      • Yoshida K.
      • Maeda K.
      • Sugiyama Y.
      Hepatic and intestinal drug transporters: prediction of pharmacokinetic effects caused by drug-drug interactions and genetic polymorphisms.
      ,
      • Li R.
      • Barton H.A.
      • Varma M.V.
      Prediction of pharmacokinetics and drug-drug interactions when hepatic transporters are involved.
      Moreover, in vitro Ki,OATP1B values vary depending on the incubation times with inhibitors and the substrates tested.
      • Izumi S.
      • Nozaki Y.
      • Maeda K.
      • et al.
      Investigation of the impact of substrate selection on in vitro organic anion transporting polypeptide 1B1 inhibition profiles for the prediction of drug-drug interactions.
      ,
      • Noe J.
      • Portmann R.
      • Brun M.E.
      • Funk C.
      Substrate-dependent drug-drug interactions between gemfibrozil, fluvastatin and other organic anion-transporting peptide (OATP) substrates on OATP1B1, OATP2B1, and OATP1B3.
      By incorporating the considerations described above, the current study established a PBPK model of PMF. The PBPK model parameters for PMF were optimized via fitting to the blood concentration-time profile of PMF under control conditions (top-down approach). The analysis was also performed to measure the hepatic uptake clearance in the presence of albumin (bottom-up approach), subsequently compared with the top-down approach. Additionally, we retrospectively investigated the reported DDIs of PMF caused by CsA (co-administration) and RIF (co-administration and multiple RIF dosing followed by single PMF dosing) using the Ki,OATP1B values considering the pre-incubation effect of CsA/RIF on OATP1B inhibition and substrate-dependent Ki,OATP1B differences, following step-wise protocols proposed by Yoshikado et al.
      • Yoshikado T.
      • Yoshida K.
      • Kotani N.
      • et al.
      Quantitative analyses of hepatic OATP-mediated interactions between statins and inhibitors using PBPK modeling with a parameter optimization method.
      ,
      • Yoshikado T.
      • Toshimoto K.
      • Maeda K.
      • et al.
      PBPK modeling of coproporphyrin I as an endogenous biomarker for drug interactions involving inhibition of hepatic OATP1B1 and OATP1B3.

      Materials and Methods

       Materials

      Pitavastatin (PTV) calcium and rosuvastatin (RSV) calcium were obtained from Wako Pure Chemical Industries (Osaka, Japan). PMF, and the deuterium-labeled compounds PTV-d5, and RSV-d6 were purchased from Toronto Research Chemicals (North York, Canada). CsA was from Santa Cruz (Dallas, TX, USA). RIF and human serum albumin (HSA) were from Merck Millipore (Burlington, MA, USA). Cryopreserved human hepatocytes [plated human hepatocytes (PHH); Lot HH1052 (male, 44-year-old, Caucasian) and HH1103 (female, 44-year-old, Caucasian/Hispanic)] were purchased from In Vitro ADMET Laboratories (Columbia, MD, USA). Other reagents and solvents were of analytical grade from Wako Pure Chemical Industries and Thermo Fisher Scientific (Waltham, MA, USA).

       Uptake Study in Human Embryonic Kidney (HEK) 293 Cells Stably Expressing OATP Isoforms

      The HEK293 cells expressing an OATP isoform and control cells were seeded in the poly-d-lysine-coated 48-well plate (1 × 105 cells/well; BD Biosciences, San Jose, CA) and grown in culture media for 48 h prior to the uptake study. After washing out the media and incubating in pre-warmed Krebs-Henseleit Buffer (KHB; pH 7.4) for 15 min, the dosing solution in pre-warmed KHB was added into the well at 37 °C, starting the uptake reaction. The medium was aliquoted from the well immediately before terminating the uptake reaction, by adding cold phosphate-buffered saline (PBS) at designated time points. Each well was washed with cold PBS three times. The attached cells were harvested with methanol containing an internal standard (IS) and lysed by sonication. After centrifuging the cell lysate and the medium mixed with acetonitrile containing IS (10,000 g, 10 min, 4 °C), the drug levels in the resulting supernatant were quantified by liquid chromatography with tandem mass spectrometry (LC−MS/MS). Detailed information on LC−MS/MS analysis and kinetic analysis of the uptake clearance was described Supplementary Text.

       Uptake Study in PHH

      The cryopreserved human hepatocytes were thawed out according to the manufacturer's instructions, and seeded in collagen I-coated 24-well plates (Corning, NY, USA) using Universal Primary Cell Plating Medium (In Vitro ADMET Laboratories) (4 × 105 cells/well, cell viability > 85%) for 5 h prior to the uptake study. The uptake assay in PHH was performed in the same way as the HEK293 cells stably expressing OATP isoforms, except the procedure that the uptake reaction was terminated by transferring the plate on ice and immediately washing with cold KHB three times at designated time points.
      To investigate the albumin-mediated uptake, HSA was added at varying concentrations (0, 0.125, 0.25, 0.5, 1, and 5% for Lot HH1052; 0, 0.25, 0.5, 1, 3, and 5% for Lot HH1103) in the cassette dosing solution containing PMF, PTV, and RSV (3 μM each, in the pre-warmed KHB buffer).

       Kinetic Analysis of Albumin-Mediated Uptake Study in PHH

      The uptake clearance for a drug (PSinf) was calculated as the uptake clearance for the unbound drug in the absence of albumin (PSu,inf(−)) multiplied with the fu value (unbound fraction) according to the free-drug hypothesis (i.e., PSinf = PSu,inf(−) × fu). The albumin-mediated uptake clearance via the drug-albumin complex was additionally considered by applying the facilitated-dissociation model as below
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      (graphic presentation shown in Supplementary Fig. S1).
      PSinf=PSu,inf()×KdKd+[Alb]+PSb,inf×BmaxKd,m+[Alb]×[Alb]Kd+[Alb]
      (1)


      where Kd, the dissociation constant between HSA and a drug, which calculated as fu × [Alb]/(1 - fu); [Alb], the molar concentration of HSA; PSb,inf, the hepatic uptake clearance for the unbound drug dissociated from the albumin-drug complex near the cell surface; Bmax, the clearance capacity of albumin-binding sites on the surface of hepatocyte; Kd,m, the dissociation constant between albumin and the surface of hepatocytes.
      The Kd values of PTV and RSV for HSA binding were from the previous report.
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      For statins, their fu values in the presence of 5% HSA (fu(5%HSA)) were similar to the unbound fraction in plasma (fp) values,
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      and we assumed that the fu(5%HSA) value of PMF is the same as the reported fp value of 0.0016
      Pharmaceuticals and Medical Devices Agency.
      and the Kd value for PMF was calculated from [Alb] and fu as described above. In the current study, the Kd,m value was fixed as 45.2 μM for all the drugs, based on the previous study,
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      and PSb,inf ×Bmax values for PMF, PTV, and PRV were obtained with Eq. (1) via nonlinear least-squares fitting.
      The unbound uptake clearance in the presence of albumin (PSu,inf(+)) was calculated using the following equation
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      ;
      PSu,inf(+)=PSinffu=PSu,inf()+PSb,inf×Bmax×[Alb]Kd,m+[Alb]×1Kd
      (2)


       In Vitro Ki,OATP1B Values by CsA or RIF Determined From the Uptake Study in HEK293 Cells Stably Expressing OATP1B1 (HEK/OATP1B1)

      The inhibitory potencies of CsA or RIF were assessed under co-incubation and pre-incubation conditions. Before starting the uptake reaction, HEK/OATP1B1 cells were equilibrated with the KHB buffer containing CsA or RIF (for the pre-incubation conditions) or lacking CsA or RIF (for the co-incubation conditions) for 30 min at 37 °C. After removing the buffer treated, HEK/OATP1B1 or control cells were incubated with PMF (0.1 μM), PTV (0.1 μM), and RSV (0.3 μM) as a cassette dosing for 1 min at 37 °C in the absence and presence of CsA (0.001, 0.01, 0.03, 0.1, 0.3, 1, or 10 μM) or RIF (0.01, 0.1, 0.3, 1, 3, 10, 30, or 100 μM).
      Half maximal inhibitory concentrations (IC50) of CsA or RIF were estimated using the following equation;
      PSinf(+inh)=PSact,inf1+[I]IC50+PSdif,inf
      (3)


      where PSinf(+inh) and PSact,inf represent the uptake clearance in the presence and absence of an inhibitor (CsA or RIF), respectively. [I] represents the concentration of the inhibitor. The PSdif,inf represents the passive diffusion clearance that is not inhibited at the maximum concentration of the inhibitor tested (set as a free parameter for fitting). The parameters were estimated by a nonlinear least-squares regression analysis. Since the drug concentrations were chosen to be much lower than their reported Km values for OATP1B1, the IC50 values obtained by fitting were used as in vitro Ki,OATP1B values.
      For correcting the Ki,OATP1B values considering the substrate-dependency, the in vivo Ki,OATP1B values of PMF were calculated by multiplying the in vitro Ki,OATP1B values of PMF obtained in pre-incubation with the ratio of in vivo to in vitro Ki,OATP1B of the reference Drug A (PTV in the current study
      • Yoshikado T.
      • Toshimoto K.
      • Maeda K.
      • et al.
      PBPK modeling of coproporphyrin I as an endogenous biomarker for drug interactions involving inhibition of hepatic OATP1B1 and OATP1B3.
      ).
      (invivo)Ki,OATP1B,PMF=(invitro)Ki,OATP1B,PMF×(invivo)Ki,OATP1B,DrugA(invitro)Ki,OATP1B,DrugA
      (4)


       Structures of PBPK Model for PMF

      The PBPK model for PMF (Fig. 1) was constructed by incorporating the PK characteristics of PMF (Supplementary Table S1). The PMF PBPK model included the organ compartments (muscle, skin, and adipose tissue) to which lipophilic compounds can readily distribute. OATP1B-mediated hepatic uptake of PMF was incorporated in the five-liver compartments (mimicking a dispersion model).
      • 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.
      Intestinal absorption of PMF was incorporated using a first-order absorption model. Since the bile excretion of PMF was reported in rats,
      Pharmaceuticals and Medical Devices Agency.
      enterohepatic circulation was included using three transit compartments. We also confirmed that PMF is a substrate for the bile canalicular transporters (MRP2, MDR1, and BCRP) using the transcellular transport assay (Supplementary Fig. S2).
      Figure thumbnail gr1
      Fig. 1(a) Structure of the PBPK model of pemafibrate (PMF). The PBPK model of PMF included the organ compartments (muscle, skin, and adipose tissue), where lipophilic compounds are readily distributed. OATP1B-mediated hepatic uptake of PMF was incorporated in the five-liver compartments. Intestinal absorption of PMF was described using a first-order absorption model. Enterohepatic circulation (EHC) was included using three transit compartments. (b) Scheme showing the possible mechanisms of DDI of PMF with cyclosporine A (CsA) and rifampin (RIF) in hepatocytes. CsA can lead to the inhibition of OATP1B-mediated uptake and CYP3A4-mediated metabolism of PMF. The single dose of RIF can inhibit the OATP1B-mediated uptake of PMF, whereas the multiple doses of RIF can lead to the induction of multiple CYP enzymes and OATP1B as well as the inhibition of OATP1B-mediated uptake of PMF. CLint, bile, intrinsic clearance of biliary excretion CLint,met, intrinsic clearance of hepatic metabolism; CLR, renal clearance; fB, unbound fraction in blood; fH, unbound fraction in hepatocytes; HC, hepatocytes; HE, hepatic extracellular space; ka, the absorption rate constant in first order absorption model; kbile, the transit rate constant for enterohepatic circulation (EHC); PSact,inf, active uptake intrinsic clearance on sinusoidal membrane; PSdif,eff, efflux intrinsic clearance by passive diffusion through sinusoidal membrane; PSdif,inf, influx intrinsic clearance by passive diffusion through sinusoidal membrane; Qtissue, blood flow rate in tissue.

       Parameter Optimization for the PBPK Model of PMF

      All nomenclature and Differential equations are described in Supplementary Text. The PBPK model parameters for PMF were optimized according to the step-wise protocol proposed by Yoshikado et al.
      • Yoshikado T.
      • Yoshida K.
      • Kotani N.
      • et al.
      Quantitative analyses of hepatic OATP-mediated interactions between statins and inhibitors using PBPK modeling with a parameter optimization method.
      Briefly, the physiological parameters and PK properties of PMF were collected from the previous reports, and set as fixed values (Supplementary Tables S1 and S3) or estimated as initial parameters (Supplementary Table S4). The PBPK model parameters (ka, Tlag, CLint,all, and fbile) for PMF were optimized using the blood concentration-time profile of a single oral dose of PMF (0.4 mg, control conditions in a reported DDI study with RIF
      Pharmaceuticals and Medical Devices Agency.
      ) (Supplementary Table S4).

       Simulation of DDI of PMF With CsA or RIF

      We adopted the previously reported PBPK model structures and parameters for CsA
      • Yoshikado T.
      • Yoshida K.
      • Kotani N.
      • et al.
      Quantitative analyses of hepatic OATP-mediated interactions between statins and inhibitors using PBPK modeling with a parameter optimization method.
      and RIF
      • Asaumi R.
      • Menzel K.
      • Lee W.
      • et al.
      Expanded physiologically-based pharmacokinetic model of rifampicin for predicting interactions with drugs and an endogenous biomarker via complex mechanisms including organic anion transporting polypeptide 1B induction.
      (Supplementary Fig. S3). The inhibitory effects of CsA and RIF on OATP1B-mediate hepatic uptake or the induction effects of RIF on OATP1B and CYPs were incorporated into the PMF PBPK model to simulate DDIs (Fig. 1b). All Supplementary Information (text, figures, and tables) is available as separate files.

       Software

      Numeric Analysis Program for Pharmacokinetics (Napp, ver. 2.31) was used for optimization and simulation in the current study. Nonlinear least-squares calculation was conducted with the weight for the square root of the original value.

      Results

       Uptake of PMF in HEK293 Cells Stably Expressing OATP-Expression Systems and PHH

      Cellular uptake of PMF was significantly enhanced in OATP1B1- and OATP1B3-expressing HEK293 cells compared to the control cells, but not in OATP2B1-expressing cells (Fig. 2a–c). The time-dependent uptake of PMF was also observed in PHH (Fig. 2d). When the relative contribution of OATP1B1 and OATP1B3 was calculated using the relative activity factor (RAF) or relative expression factor (REF) method, the contribution of OATP1B1 to the net hepatic uptake of PMF was found to be similar to (by the REF method) or two-fold larger (by the RAF method) than that of OATP1B3 (Supplementary Table S5). When nonlinear least-squares regression analysis was performed (based on Eq. S3) on the PMF uptake data, the Km and Vmax values of PMF were determined as 4.49 μM and 185 pmol/min/mg protein, respectively, in HEK293 cells expressing OATP1B1 (Fig. 2e and Supplementary Table S6). The kinetic parameters obtained in PHH were as follows: Km (3.1 μM), Vmax (80 pmol/min/mg protein), PSdif.inf (13.7 μL/min/mg protein), and PSact,inf (26.1 μL/min/mg protein) (Fig. 2f and Supplementary Table S6).
      Figure thumbnail gr2
      Fig. 2(a–d) Cellular uptake of PMF in HEK293 cells stably expressing OATP isoforms or plated human hepatocytes (PHH). Time-dependent uptake of PMF (0.1 μM) was assessed in HEK293 cells expressing OATP1B1 (a, HEK/OATP1B1), OATP1B3 (b, HEK/OATP1B3), or OATP2B1 (c, HEK/OATP2B1) (closed circles) and mock control cells (open circles). Time-dependent uptake of PMF (0.1 μM) was also assessed using PHH (d, lot HH1052). (e) Concentration-dependent uptake of PMF (0.03–100 μM) in HEK/OATP1B1. The uptake of PMF was calculated by the difference of uptake volume between HEK/OATP1B1 and control cells for 1 min (, ) and shown as Eadie-Hofstee plot. (f) Concentration-dependent uptake of PMF (0.1–100 μM) in PHH (lot HH1052). The uptake was calculated using the uptake volume between 0.25–2 min ( ) and shown as Eadie-Hofstee plot. ∗p < 0.05 compared with the uptake in control cells (by two-sided, paired Student's t-test). Solid lines in the panels (e) and (f) represent the fitting by a nonlinear least-squares regression analysis based on . The results are shown as mean ± SD (n = 3–4).

       Effect of Albumin on Uptake Clearance of PMF in PHH

      To examine whether the high protein-bound PMF (fp = 0.0016) displays the albumin-mediated hepatic uptake, we measured the uptake clearance in the presence of varying concentrations of HSA in two lots of PHH (HH1052 and HH1103). PTV (fp = 0.0054; high protein binding) and RSV (fp = 0.13; relatively low protein binding) were used for comparison. When the experimentally observed PSinf values were plotted on a logarithmic scale (closed symbols, Fig. 3), the PSinf values of PMF and PTV showed a decreasing trend as the HSA concentrations increased: The decreasing trend was more pronounced in the lot HH1103 than in the lot HH1052. For all three drugs, the observed PSinf values were much larger than the predictions based on the free-drug hypothesis (shown in blue, Fig. 3). When the PSinf values were predicted based on the facilitated-dissociation model (Eq. (1)) describing the albumin-mediated uptake, the observed data were well captured (shown in red, Fig. 3).
      Figure thumbnail gr3
      Fig. 3Effects of varying concentrations of human serum albumin on the uptake clearances (PSinf) of PMF (a, d), PTV (b, e), and RSV (c, f). Two different lots of plated human hepatocytes were used: (a–c) lot HH1052; (d–e) lot HH1103. The uptake of PMF, PTV, and RSV (3 μM each) was assessed in the presence of the concentrations of human serum albumin ranging from 0 to 5%. The uptake clearance was calculated using the uptake volume between 0.25 and 1.25 min. The closed circles represent the observed data (n = 2). The PSinf values based on the free-drug hypothesis are shown in blue, while the values calculated according to the facilitated-dissociation model (Eq. ) are shown in red. The nominal values for the related parameters are listed in .
      Table 1 summarizes the nominal values of the hepatic uptake clearances for the unbound drug in the presence of 5% albumin (PSu,inf(+5%HSA)) via different calculation methods. When the PSu,inf(+5%HSA) values for PMF, PTV, and RSV were calculated from the experimentally observed clearance in the presence of 5% HSA divided by the unbound fraction [i.e., PSinf(+5%HSA)/fu(5%HSA); denoted as PSu,inf(+5%HSA,obs)], the resulting values were much greater than those observed in the absence of albumin (PSu,inf(−)). The PSu,inf(+5%HSA,obs) values displayed considerable differences between the lots of HH1052 and HH1103, but the corresponding values for the highly protein-bound PMF and PTV were consistently larger than those of RSV in both lots. The largest inter-lot difference was nearly 15-fold for the PSu,inf(+5%HSA,obs) values of PMF: 2656 and 181 μL/min/mg protein for the lots HH1052 and HH1103, respectively. Another calculation method employed the fitting to the facilitated-dissociation model [the obtained parameter denoted as PSu,inf(+5%HSA,extrapolated)]. The PSu,inf(+5%HSA,extrapolated) values were in the same rank order: PMF ≧ PTV > RSV as PSu,inf(+5%HSA,obs). The hepatic clearance values of the unbound drug (PSu,inf) were calculated using the values for individual hepatocyte lots and different methods and scaled up per kg body weight for comparison (Table 1b). The PSu,inf values from the calculation considering the albumin-mediated hepatic uptake (either from the calculation using the experimentally observed values or extrapolation) were larger (by more than 1500-fold) than the values based on the free-drug hypothesis. These values were used for the subsequent bottom-up simulations for the blood PMF concentration-time profiles.
      Table 1Summary of the Uptake Clearance-Related Parameters for Pemafibrate (PMF), Pitavastatin (PTV), and Rosuvastatin (RSV), Calculated From the Experimental Observations (Using the Two Different Hepatocyte Lots HH1052 and HH1103), and Model-Based Analysis (From the Free-Drug Hypothesis or the Facilitated-Dissociation Model).
      (A) Uptake Clearance-Related Parameters for PMF, PTV, and RSV
      Drugfu(5%HSA)Kd (μM)Experimentally observedCalculation Based on Observed PSinf(+5%HSA)Extrapolation Based on the Facilitated-Dissociation Model
      PSu,inf(−) (μL/min/mg protein)PSinf(+5%HSA) (μL/min/mg protein)PSu,inf(+5%HSA,obs)
      Calculated, PSinf(+5%HSA)/fu(5%HSA).
      (μL/min/mg protein)
      Kd,m (μM)[PSb,inf × Bmax]
      Obtained by nonlinear least-squares fitting using Eq. (1) (same as VB,max value (the uptake clearance for the drug amount) in reported kinetic analysis10).
      (μM)
      PSu,inf(+5%HSA,extrapolated)
      Calculated using Eq. (2) in the presence of 5% HSA, PSu,inf(−) + (PSb,inf × Bmax × [Alb]5%HSA)/(Kd,m + [Alb]5%HSA)/Kd.
      (μL/min/mg protein)
      Lot HH1052Lot HH1103Lot HH1052Lot HH1103Lot HH1052Lot HH1103Lot HH1052Lot HH1103Lot HH1052Lot HH1103
      PMF0.0016
      assumed that fu(5%HSA) value was the same as the fp value of PMF.
      1.21
      Calculated, fu × [Alb]/(1 - fu).
      53
      Mean of observed values (n = 2).
      (54, 52)
      42
      Mean of observed values (n = 2).
      (43, 41)
      4.3
      Mean of observed values (n = 2).
      (5.3, 3.2)
      0.29
      Mean of observed values (n = 2).
      (0.33, 0.25)
      265618145.2
      Reported.10.
      520 ± 119363 ± 53458325
      PTV0.0056
      Reported.10.
      4.44
      Reported.10.
      19.7
      Mean of observed values (n = 2).
      (21.1, 18.3)
      27.4
      Mean of observed values (n = 2).
      (25.5, 29.3)
      5.4
      Mean of observed values (n = 2).
      (6.0, 4.7)
      1.2
      Mean of observed values (n = 2).
      (1.2, 1.2)
      955207661 ± 191890 ± 51160217
      RSV0.099
      Reported.10.
      94.5
      Reported.10.
      3.6
      Mean of observed values (n = 2).
      (4.0, 3.2)
      4.9
      Mean of observed values (n = 2).
      (4.4, 5.3)
      5.3
      Mean of observed values (n = 2).
      (5.9, 4.6)
      1.8
      Mean of observed values (n = 2).
      (2.0, 1.7)
      53.018.61293 ± 4521383 ± 1301619
      (B) PSu,inf of PMF scaled up per kg body weight
      Calculation MethodPSu,inf (L/h/kg body weight)
      Lot HH1052Lot HH1103
      Bottom-up approachScaled up from PSu,inf(−), based on the free-drug hypothesis (= fp × PSu,inf(−)/Rb)0.0310.024
      Scaled up from PSu,inf(+5%HSA,obs), experimentally observed in the presence of 5% HSA (= fu(5%HSA) × PSu,inf(+5%HSA,obs)/Rb)53036.2
      Scaled up from PSu,inf(+5%HSA,extrapolated), extrapolated from the facilitated-dissociation model92.765.7
      Top-down approachin vivo (fitted) (Supplementary Table S4)1652 (β = 0.11, optimized)

      300 (β = 0.65, calculated based on in vitro data)
      [Allb]5%HSA, the concentration of 5% HSA (= 752 μM); β, the ratio of overall hepatic intrinsic clearance to hepatic uptake clearance ((= CLint,all/(PSact,inf + PSdif,inf)); Bmax, the clearance capacity of albumin-binding sites on the surface of hepatocyte; fP, unbound fraction in plasma; fu(5%HSA), the unbound fraction under 5% HSA; Kd, the dissociation constant between drug and HSA; Kd,m, the dissociation constant between albumin and the surface of hepatocytes; PSb,inf, the hepatic uptake clearance for the unbound drug dissociated from the albumin-drug complex near the cell surface; PSu,inf(−), the hepatic uptake clearances in the absence of albumin; PSinf(+5%HSA), the hepatic uptake clearances under 5% HSA; PSu,inf(+5%HSA,obs), the observed hepatic uptake clearances for the unbound drug under 5% HSA; PSu,inf(+5%HSA,extrapolated), the hepatic uptake clearances for unbound form under 5% HSA, where estimated from fitted parameters in facilitated-dissociation model; Rb, the ratio of blood to plasma concentration.
      a assumed that fu(5%HSA) value was the same as the fp value of PMF.
      b Calculated, fu × [Alb]/(1 - fu).
      c Reported.
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      .
      d Mean of observed values (n = 2).
      e Calculated, PSinf(+5%HSA)/fu(5%HSA).
      f Obtained by nonlinear least-squares fitting using Eq. 1 (same as VB,max value (the uptake clearance for the drug amount) in reported kinetic analysis
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      ).
      g Calculated using Eq. 2 in the presence of 5% HSA, PSu,inf(−) + (PSb,inf × Bmax × [Alb]5%HSA)/(Kd,m + [Alb]5%HSA)/Kd.

       Simulation of the Blood PMF Concentration-Time Profiles: Impact of Considering the Albumin-Mediated Uptake Mechanism (Versus the Free-Drug Hypothesis) and Comparison of Top-Down and Bottom-up Approaches

      The values for the four PBPK model parameters (ka, Tlag, CLint,all, and fbile) were not attainable by estimation from in vitro data. Thus, these parameters were fitted to the blood concentration-time profile of a single-dose of PMF (0.4 mg, control conditions). As it is often difficult to experimentally estimate the β value (a hybrid parameter that can reflect the rate-determining step in the overall intrinsic hepatic clearance), the PBPK modeling was performed with the β value either fixed as 0.65 (calculated based on in vitro data) or set as a free parameter (i.e., optimized parameter). The PBPK model of PMF successfully captured the PK profiles of PMF under control conditions: the fitted profiles (via top-down approach) were closer to the observed data with the β value of 0.11 (optimized) than with the β value of 0.65 (fixed, based on the in vitro data) (Fig. 4, solid black lines and Supplementary Table S4). From this top-down approach, the in vivo hepatic uptake clearances were 1652 and 300 L/h/kg body weight for the β values of 0.11 (optimized) and 0.65 (fixed, based on the in vitro data), respectively.
      Figure thumbnail gr4
      Fig. 4Comparison of bottom-up simulations of the blood PMF concentration-time profiles (under control conditions) using the hepatic uptake clearances based on the free-drug hypothesis, and those considering the albumin-mediated uptake (a, b: using the values from lot HH1052; c, d: using the values from lot HH1103). The open circles represent the observed data following a single dose of PMF (0.4 mg, control conditions in a reported DDI study
      Pharmaceuticals and Medical Devices Agency.
      ). The solid black lines are the fitted blood PMF concentration-time profiles. The dotted blue lines represent the bottom-up simulations using the hepatic uptake clearances based on the free-drug hypothesis. The red solid and broken lines represent the bottom-up simulations using the hepatic uptake clearances observed in the presence of 5% human serum albumin (PSu,inf(+5%HSA,obs)) and the extrapolated hepatic uptake clearances (PSu,inf(+5%HSA,extrapolated)), respectively. Description on the use of the two β values (0.11 and 0.65) is provided in the main text.
      Bottom-up simulation was also performed using the hepatic uptake clearance of PMF based on the free-drug hypothesis or the albumin-mediated uptake. For the in vitro data using the lot HH1052 (Fig. 4a and 4b; Table 1b), the consideration of the albumin-mediated uptake yielded the simulated profiles closer to the observed blood PMF PK profiles, compared to the simulation based on the free-drug hypothesis. For the in vitro data using the lot HH1103 (Fig. 4c and 4d; Table 1b), the consideration of the albumin-mediated uptake also reduced the differences between the simulated and observed data, but the extent of improvement was less than the simulation using the in vitro data using the lot HH1052. Although the improvement by considering the albumin-mediated uptake was observed in either β value used, the simulated profiles using the β value of 0.65 (the case in which the uptake clearance is rate-determining for intrinsic clearance) were closer to the observed data than those using the β value of 0.11 (Fig. 4b and 4d versus 4a and 4c).

       Inhibitory Effects of CsA and RIF on OATP1B1-Mediated Uptake of PMF: Comparison of Pre-incubation and Co-incubation Conditions

      CsA and RIF inhibited the uptake of PMF, PTV, and RSV in a concentration-dependent manner for both pre-incubation and co-incubation conditions (Fig. 5; Supplementary Table S7). Under pre-incubation conditions, the inhibitory potencies of CsA against the OATP1B1-mediated uptake of three substrates (PMF, PTV, and RSV) were much enhanced: the in vitro Ki,OATP1B values of CsA (Ki,OATP1B_CsA) under pre-incubation conditions were 0.041, 0.048, and 0.095 μM for PMF, PTV, and RSV, respectively, being 6–60-fold lower than those obtained under co-incubation conditions (Fig. 5a and Supplementary Table S7a). After pre-incubation, the in vitro Ki,OATP1B values of RIF (Ki,OATP1B_RIF) were 0.39, 1.25, and 1.05 μM for PMF, PTV, and RSV, respectively, being were 3~9-fold lower than those obtained under co-incubation conditions (Fig. 5b and Supplementary Table S7b).
      Figure thumbnail gr5
      Fig. 5Comparison of the inhibitory potencies of CsA (a) and RIF (b) on OATP1B1-mediated uptake of PMF (0.1 μM), PTV (0.1 μM), and RSV (0.3 μM) between co-incubation and pre-incubation conditions. For the co-incubation condition (Co; black squares), the uptake of substrates was measured for 1 min at 37 °C by incubating the inhibitor (CsA or RIF) and substrates simultaneously. For the pre-incubation condition (Pre; blue circles), HEK/OATP1B1 cells were pre-incubated with the inhibitor (CsA or RIF) for 30 min, and subsequently the uptake of substrates was measured for 1 min at 37 °C by incubating the inhibitor and substrates simultaneously. Solid lines represent the fitted results by a nonlinear least-squares regression analysis based on Eq. . Mean ± SD (n = 3).

       Simulation of the Blood PMF Concentration-Time Profiles Under DDI Conditions Using the Ki,OATP1B Values of CsA and RIF Obtained in vitro (With Pre-Incubation) or Corrected for Substrate-Dependent Variation

      The blood PMF concentration-time profiles under DDI conditions were simulated using the in vitro Ki,OATP1B values of CsA and RIF (with pre-incubation), but this bottom-up simulation yielded the PK profiles which considerably deviated from the observed data (Fig. 6, dotted pink lines). In order to improve the DDI prediction accuracy via PBPK modeling, the simulations were performed using the in vivo Ki,OATP1B values of CsA and RIF corrected for substrate-dependent variation. The previous PBPK modeling
      • Yoshikado T.
      • Yoshida K.
      • Kotani N.
      • et al.
      Quantitative analyses of hepatic OATP-mediated interactions between statins and inhibitors using PBPK modeling with a parameter optimization method.
      reported the in vivo Ki,OATP1B value of CsA or RIF by fitting to the clinical DDI data between PTV and CsA or RIF (in vivo Ki,OATP1B_CsA, 0.012 μM; in vivo Ki,OATP1B_RIF, 0.23 μM). To correct for the substrate-dependent variation in the Ki,OATP1B values of CsA and RIF, as reported previously,
      • Li R.
      • Barton H.A.
      • Varma M.V.
      Prediction of pharmacokinetics and drug-drug interactions when hepatic transporters are involved.
      ,
      • Izumi S.
      • Nozaki Y.
      • Maeda K.
      • et al.
      Investigation of the impact of substrate selection on in vitro organic anion transporting polypeptide 1B1 inhibition profiles for the prediction of drug-drug interactions.
      the in vitro Ki,OATP1B values of CsA and RIF in inhibiting the uptake of PMF and PTV (with pre-incubation) were used for the calculation using Eq. (4). The in vivo Ki,OATP1B_CsA and Ki,OATP1B_RIF values corrected for PMF were as follows: in vivo Ki,OATP1B_CsA, 0.010 μM (=0.04 × 0.012/0.048); in vivo Ki,OATP1B_RIF, 0.072 μM (=0.39 × 0.23/1.25). The use of these corrected in vivo Ki,OATP1B_CsA and Ki,OATP1B_RIF values improved the DDI simulation (Fig. 6, solid red lines), capturing the observed fold-differences in AUC (CsA, by 13-fold; RIF, by 9-fold). However, the simulations underpredicted the observed changes in Cmax (observed versus simulated; CsA, 9- vs 6-fold; RIF, 10- vs 4-fold). The use of a higher ka value (of 2/h) improved in capturing the observed Cmax values, but deviations occurred at later time points (Fig. 6b, broken green lines, upper panels). When the Ki,OATP1B_RIF value was further decreased to 0.036 μM (half value of 0.072 μM) with the ka value of 2/h, the DDI simulation tended to improve in capturing the blood PMF concentration-time profiles with RIF co-administration (Fig. 6b, lower panels).
      Figure thumbnail gr6
      Fig. 6Simulation of the blood PMF concentration-time profiles under the DDI conditions caused by co-administration of a single dose of CsA (a) or RIF (b) using the in vitro Ki,OATP1B values from pre-incubation and the in vivo Ki,OATP1B values corrected for substrate-dependency. The blue circles and lines represent the observed data and the fitted results under the control conditions (a single oral dose of PMF, 0.4 mg
      Pharmaceuticals and Medical Devices Agency.
      ), respectively. The red circles represent the observed data under the DDI conditions [a single oral dose of PMF, 0.4 mg; an oral dose of CsA (600 mg) or RIF (600 mg)].
      Pharmaceuticals and Medical Devices Agency.
      The lines of different colors represent the simulation results using the parameters indicated in the right-hand side.

       Simulation of the Blood PMF Concentration-time Profiles Under DDI Conditions Caused by Multiple RIF Dosing

      To simulate the PK profiles of PMF under DDI conditions caused by repeated RIF dosing, we used the induction parameters obtained by typical probe substrates for individual CYP isoforms and OATP1B from our previous report (Supplementary Table S9).
      • Asaumi R.
      • Menzel K.
      • Lee W.
      • et al.
      Expanded physiologically-based pharmacokinetic model of rifampicin for predicting interactions with drugs and an endogenous biomarker via complex mechanisms including organic anion transporting polypeptide 1B induction.
      ,
      • Asaumi R.
      • Toshimoto K.
      • Tobe Y.
      • et al.
      Comprehensive PBPK model of rifampicin for quantitative prediction of complex drug-drug interactions: CYP3A/2C9 induction and OATP inhibition effects.
      Since the PMF is a substrate for CYP3A4/CYP2C8/CYP2C9,
      • Ogawa S.I.
      • Shimizu M.
      • Yamazaki H.
      Plasma concentrations of pemafibrate with co-administered drugs predicted by physiologically based pharmacokinetic modeling in virtual populations with renal/hepatic impairment.
      the PMF PBPK model incorporated the components for the induction of OATP1B as well as CYP3A4/CYP2C8/CYP2C9 with reported fm,CYP values
      • Ogawa S.I.
      • Shimizu M.
      • Yamazaki H.
      Plasma concentrations of pemafibrate with co-administered drugs predicted by physiologically based pharmacokinetic modeling in virtual populations with renal/hepatic impairment.
      and the inhibition of OATP1B by RIF. Simulation was performed for the dosing regimen, in which RIF was administered for 10 days (p.o., 600 mg QD), and PMF (p.o., 0.4 mg) was administered 24 h after the last dose of RIF. When the induction effects were considered only for CYP3A/CYP2C8/CYP2C9, the simulation results tended to overpredict the observed AUC (0.24-fold) and Cmax (0.39-fold): AUC (simulated: 0.67- and 0.73-fold for β = 0.11 and 0.65, respectively) and Cmax (simulated: 0.65- and 0.63-fold for β = 0.11 and 0.65, respectively). When the induction effects were considered for both CYP enzymes (CYP3A/CYP2C8/CYP2C9) and OATP1B, the DDI prediction was improved: simulated AUC, 0.22- and 0.25-fold; simulated Cmax, 0.22- and 0.23-fold for β = 0.11 and 0.65, respectively) (Fig. 7 and Supplementary Table S8c).
      Figure thumbnail gr7
      Fig. 7Simulation of the blood PMF concentration-time profiles under the DDI conditions caused by multiple RIF dosing, considering the induction of OATP1B and CYPs (CYP3A/CYP2C8/CYP2C9) under two β values (0.11 (a) and 0.65 (b). The blue circles and lines represent the observed data and the fitted results under the control conditions (a single oral dose of PMF, 0.4 mg
      Pharmaceuticals and Medical Devices Agency.
      ), respectively. The red circles represent the observed data under the DDI conditions [multiple oral doses of RIF (600 mg/day for 10 days); oral dosing of PMF alone (0.4 mg) on day 11 (i.e., 24 h after the last-dose of RIF)].
      Pharmaceuticals and Medical Devices Agency.
      The red dotted and solid lines represent the simulated profiles under the DDI conditions considering the metabolic enzyme induction only (CYP3A/CYP2C8/CYP2C9) or the induction of both metabolic enzymes and OATP1B, respectively.

      Discussion

      With the enhanced understanding of hepatic drug elimination mechanisms, the cases of complex DDIs involving the interplay of hepatic enzymes and transporters have been increasingly recognized. However, a considerable gap remains in the quantitative bottom-up prediction of the complex DDIs using the parameters obtained in vitro. For drugs that are highly protein-bound and substrates for the hepatic uptake transporters such as OATP1B, doubt has been cast over the prediction accuracy of the hepatic uptake clearance using the IVIVE method based on the conventional free-drug hypothesis.
      • Miyauchi S.
      • Masuda M.
      • Kim S.J.
      • et al.
      The phenomenon of albumin-mediated hepatic uptake of organic anion transport polypeptide substrates: prediction of the in vivo uptake clearance from the in vitro uptake by isolated hepatocytes using a facilitated-dissociation model.
      ,
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      The inhibitory potency of perpetrator drugs may also differ between in vitro and in vivo settings. In the current study, we developed and refined a PBPK model for PMF in order to accurately capture the PK changes by the DDI caused by CsA and RIF. To improve the prediction accuracy of the hepatic uptake clearance by IVIVE, we considered the albumin-mediated hepatic uptake mechanism.
      • Miyauchi S.
      • Masuda M.
      • Kim S.J.
      • et al.
      The phenomenon of albumin-mediated hepatic uptake of organic anion transport polypeptide substrates: prediction of the in vivo uptake clearance from the in vitro uptake by isolated hepatocytes using a facilitated-dissociation model.
      ,
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      For the DDI prediction, we adjusted the OATP1B1 inhibition potency with PMF as a substrate by performing the in vitro experiments under the pre-incubation conditions together with correcting for substrate-dependent Ki,OATP1B differences. These strategies led to an overall improvement in capturing the observed DDI data between PMF and CsA/RIF.
      Recently, Ogawa et al.
      • Ogawa S.I.
      • Shimizu M.
      • Yamazaki H.
      Plasma concentrations of pemafibrate with co-administered drugs predicted by physiologically based pharmacokinetic modeling in virtual populations with renal/hepatic impairment.
      reported the quantitative analyses of DDIs of PMF caused by the co-administration of OATP1B inhibitors (a single dose of RIF or sacubitril). The authors reported, but could not resolve the discrepancy in the hepatic uptake clearances between in vitro and in vivo. As the in vitro/in vivo differences of intrinsic hepatic clearance tend to increase for highly protein-bound drugs,
      • Kim S.J.
      • Lee K.R.
      • Miyauchi S.
      • Sugiyama Y.
      Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
      we hypothesized that the highly (>99.8%) protein-bound PMF may undergo the albumin-mediated hepatic uptake. The presence of the albumin-mediated uptake mechanism for PMF was confirmed by our experimental results, which were well captured by applying the facilitated-dissociation model (Fig. 3). As expected, the hepatic uptake of PMF was more efficient than estimated based on the free-drug hypothesis (Fig. 3a and 3d). The extent by which the hepatic uptake was enhanced in the presence of 5% HSA varied depending on the drugs and the hepatocyte lots. The fold differences of the PSu,inf(+5%HSA,obs) values compared to the PSu,inf(−) values were large for the highly protein-bound PMF (50- and 4-fold for the lots HH1052 and HH1103, respectively) and PTV (48- and 8-fold for the lots HH1052 and HH1103, respectively), but considerably small for RSV (relatively low protein binding; 15- and 4-fold for HH1052 and HH1103, respectively). The inter-lot differences (lots HH1052 versus HH1103) in the PSu,inf(+5%HSA,obs) values also varied depending on the drugs: PMF, 15-fold; PTV, 5-fold; RSV, 3-fold. The inter-lot differences of PSu,inf(5%HSA) values were noted between lots HH1052 and HH1103 and the possible reason(s) for the inter-lot differences of PSu,inf(+5%HSA) remain(s) to be clarified yet. One possibility is that the Kd,m value in Eq. (1) representing the interaction between albumin and the surface of hepatocytes may exhibit large inter-individual/inter-lot differences, but further in-depth studies are warranted. Nevertheless, the bottom-up prediction based on the in vitro hepatic uptake clearance of PMF measured in the presence of 5% HSA was better than that in the absence of HSA (based on the free-drug hypothesis) (Fig. 4). These findings may also be applicable for other highly albumin-bound OATP1B substrates, for which in vitro uptake clearances measured in the presence of 5% albumin may improve the quantitative prediction by IVIVE. In addition to the facilitated-dissociation model used in the current study,
      • Miyauchi S.
      • Masuda M.
      • Kim S.J.
      • et al.
      The phenomenon of albumin-mediated hepatic uptake of organic anion transport polypeptide substrates: prediction of the in vivo uptake clearance from the in vitro uptake by isolated hepatocytes using a facilitated-dissociation model.
      ,
      • Forker E.L.
      • Luxon B.A.
      Albumin-mediated transport of rose bengal by perfused rat liver. Kinetics of the reaction at the cell surface.
      other hypotheses and models (e.g., the albumin-receptor hypothesis,
      • Weisiger R.
      • Gollan J.
      • Ockner R.
      Receptor for albumin on the liver cell surface may mediate uptake of fatty acids and other albumin-bound substances.
      the fu,p-adjusted model,
      • Poulin P.
      • Haddad S.
      Albumin and uptake of drugs in cells: additional validation exercises of a recently published equation that quantifies the albumin-facilitated uptake mechanism(s) in physiologically based pharmacokinetic and pharmacodynamic modeling Research.
      the transporter-induced protein-binding shift model
      • Bowman C.M.
      • Benet L.Z.
      An examination of protein binding and protein-facilitated uptake relating to in vitro-in vivo extrapolation.
      ) have been proposed for the mechanisms of the albumin-mediated uptake phenomenon. Further investigation in this area is critical in narrowing the gap between in vitro and in vivo.
      It is now well-recognized that the inhibition potency of CsA on OATP1B1 and OATP1B3 can enhance greatly by pre-incubation.
      • Izumi S.
      • Nozaki Y.
      • Maeda K.
      • et al.
      Investigation of the impact of substrate selection on in vitro organic anion transporting polypeptide 1B1 inhibition profiles for the prediction of drug-drug interactions.
      ,
      • Amundsen R.
      • Christensen H.
      • Zabihyan B.
      • Asberg A.
      Cyclosporine A, but not tacrolimus, shows relevant inhibition of organic anion-transporting protein 1B1-mediated transport of atorvastatin.
      • Shitara Y.
      • Takeuchi K.
      • Nagamatsu Y.
      • Wada S.
      • Sugiyama Y.
      • Horie T.
      Long-lasting inhibitory effects of cyclosporin A, but not tacrolimus, on OATP1B1- and OATP1B3-mediated uptake.
      • Gertz M.
      • Cartwright C.M.
      • Hobbs M.J.
      • et al.
      Cyclosporine inhibition of hepatic and intestinal CYP3A4, uptake and efflux transporters: application of PBPK modeling in the assessment of drug-drug interaction potential.
      For example, the pre-incubation of CsA lowered its Ki,OATP1B values in inhibiting the uptake of PTV and atorvastatin (by 4–22-fold), compared to those obtained under the co-incubation conditions. The DDI guidance for the US Food and Drug Administration recommended using the Ki,OATP1B values of CsA under the pre-incubation conditions for DDI prediction (https://www.fda.gov/downloads/Drugs/Guidances/UCM581965.pdf). The pronounced pre-incubation effect of CsA on the inhibition potency against OATP1B1 was also observed in the current study (Fig. 5a). The fold-differences in the in vitro Ki,OATP1B values of CsA under pre-incubation vs co-incubation were as follows: for PMF, 60-fold; for PTV, 27-fold; for RSV, 6-fold) (Supplementary Table S7a), similar to the previous reports by others (16, 25–27).
      In addition to CsA, other OATP1B inhibitors (e.g., simeprevir, asunaprevir) also exhibited enhanced inhibition potency by pre-incubation.
      • Furihata T.
      • Matsumoto S.
      • Fu Z.
      • et al.
      Different interaction profiles of direct-acting anti-hepatitis C virus agents with human organic anion transporting polypeptides.
      • Shitara Y.
      • Sugiyama Y.
      Preincubation-dependent and long-lasting inhibition of organic anion transporting polypeptide (OATP) and its impact on drug-drug interactions.
      • Pahwa S.
      • Alam K.
      • Crowe A.
      • et al.
      Pretreatment with rifampicin and tyrosine kinase inhibitor dasatinib potentiates the inhibitory effects toward OATP1B1- and OATP1B3-mediated transport.
      However, there is a limited understanding of the clinical significance of pre-incubation for other OATP1B inhibitors. For verifying the pre-incubation effect of RIF, we also compared the Ki,OATP1B values of RIF under pre- and co-incubation conditions. Compared to CsA, the pre-incubation effect of RIF was modest for the compounds tested (PMF, 9-fold; PTV, 3-fold; RSV, 3-fold) (Fig. 5b and Supplementary Table S7b). Other groups also reported the marginal pre-incubation effect (by 1.2–2 fold) of RIF on the inhibition of OATP-mediated PTV and RSV uptake, in monkey hepatocytes
      • Ufuk A.
      • Kosa R.E.
      • Gao H.
      • et al.
      In vitro-in vivo extrapolation of OATP1B-mediated drug-drug interactions in cynomolgus monkey.
      and human hepatocytes.
      • Barnett S.
      • Ogungbenro K.
      • Menochet K.
      • et al.
      Gaining mechanistic insight into coproporphyrin I as endogenous biomarker for OATP1B-mediated drug-drug interactions using population pharmacokinetic modeling and simulation.
      Our understanding is currently limited regarding the mechanisms for varying degrees of the pre-incubation effects depending on OATP1B inhibitors. For the pre-incubation effect of CsA, the trans-inhibition mechanism was previously proposed to explain the changes in the Ki values depending on the pre-incubation time.
      • Shitara Y.
      • Sugiyama Y.
      Preincubation-dependent and long-lasting inhibition of organic anion transporting polypeptide (OATP) and its impact on drug-drug interactions.
      Further mechanistic investigations are necessary to better understand the pre-incubation effect and its inhibitor-dependent aspects.
      Our current study implemented step-wise adjustment of the inhibition potency of CsA or RIF against OATP1B to capture the observed DDI data of PMF by CsA or RIF. The use of the in vitro Ki,OATP1B values obtained by pre-incubation underestimated the PK profiles of PMF under the DDI conditions caused by CsA/RIF (Fig. 6). Upon further correction for substrate-dependency as outlined by Yoshikado et al.,
      • Yoshikado T.
      • Toshimoto K.
      • Maeda K.
      • et al.
      PBPK modeling of coproporphyrin I as an endogenous biomarker for drug interactions involving inhibition of hepatic OATP1B1 and OATP1B3.
      the simulated profiles were in better agreement with the observed data, regarding the extent of the increase for the AUC, but not for Cmax (Fig. 6, Supplementary Tables S8a, b). The use of a higher ka value (approximately 3~4-fold to the fitted values) captured the Cmax value and reproduced the blood PMF concentration-time profiles in the presence of CsA/RIF. CsA and RIF are known to inhibit BCRP (Ki value; CsA, 1.5 μM; RIF, 14–18 μM) and MDR1 (Ki value; CsA, 1.7 μM; RIF, 4.3 μM).
      • Yoshida K.
      • Maeda K.
      • Sugiyama Y.
      Transporter-mediated drug--drug interactions involving OATP substrates: predictions based on in vitro inhibition studies.
      • Prueksaritanont T.
      • Chu X.
      • Evers R.
      • et al.
      Pitavastatin is a more sensitive and selective organic anion-transporting polypeptide 1B clinical probe than rosuvastatin.
      • Pedersen J.M.
      • Khan E.K.
      • Bergstrom C.A.S.
      • Palm J.
      • Hoogstraate J.
      • Artursson P.
      Substrate and method dependent inhibition of three ABC-transporters (MDR1, BCRP, and MRP2).
      PMF is likely a substrate of BCRP and MDR1 (Supplementary Fig. S2). The intestinal efflux transporters, BCRP and MDR1, may be involved in the clinically observed DDI between PMF and CsA/RIF. PMF is reported to have the FaFg value of 0.96 (Supplementary Table S1), and the effect of CsA/RIF on the extent of intestinal absorption fraction (FaFg) of PMF is likely to be negligible. However, PMF may stay longer in the intestinal lumen, possibly due to the enteric recycling of PMF by efflux transporters.
      • Estudante M.
      • Morais J.G.
      • Soveral G.
      • Benet L.Z.
      Intestinal drug transporters: an overview.
      The altered Cmax values of PMF under the DDI conditions may be attributable to a decreased enteric recycling, thereby shortening the residence time of PMF at the gut lumen. Potential inhibition effects of CsA/RIF on intestinal efflux transporters may be incorporated to further improve the prediction accuracy of DDIs of PMF via PBPK modeling. Another consideration may include the absorption compartments that can model the intestinal transporter-enzyme interplay by incorporating the regional variations in the expression levels of CYP3A, MDR1, and BCRP from the upper to the lower region of the intestine.
      • Harwood M.D.
      • Neuhoff S.
      • Carlson G.L.
      • Warhurst G.
      • Rostami-Hodjegan A.
      Absolute abundance and function of intestinal drug transporters: a prerequisite for fully mechanistic in vitro-in vivo extrapolation of oral drug absorption.
      The recently reported PBPK model of RIF incorporating the induction of CYP3A/CYP2C8/CYP2C9/OATP1B by multiple RIF dosing successfully explained the several cases of complex DDIs by RIF.
      • Asaumi R.
      • Menzel K.
      • Lee W.
      • et al.
      Expanded physiologically-based pharmacokinetic model of rifampicin for predicting interactions with drugs and an endogenous biomarker via complex mechanisms including organic anion transporting polypeptide 1B induction.
      For example, the developed PBPK model of RIF well captured the reported DDI data of repaglinide (a substrate of OATP1B/CYP3A/CYP2C8) or glibenclamide (a substrate of OATP1B/CYP3A/CYP2C9) with varying dosing regimens (in terms of the duration and time interval of administration).
      • Asaumi R.
      • Menzel K.
      • Lee W.
      • et al.
      Expanded physiologically-based pharmacokinetic model of rifampicin for predicting interactions with drugs and an endogenous biomarker via complex mechanisms including organic anion transporting polypeptide 1B induction.
      Being a substrate of OATP1B and three metabolizing enzymes (i.e., CYP3A4, CYP2C8, CYP2C9), PMF is a representative example to which all possible induction components by RIF apply. The decreases in the AUC and Cmax of PMF upon repeated RIF dosing were explained at least partly by the induction of CYP3A/CYP2C8/CYP2C9, and the DDI prediction further improved by consideration of additional induction component for OATP1B (Fig. 7 and Supplementary Table S8c). When the hepatic uptake clearance is rate-determining for the overall intrinsic clearance (e.g., a large β value of 0.65) (Fig. 7b), the contribution of enzyme induction was less pronounced than the case with a small β value is of 0.11 (Fig. 7a). The effect of the repeated exposure to RIF on OATP1B induction is still controversial,
      • Zamek-Gliszczynski M.J.
      • Patel M.
      • Yang X.
      • et al.
      Intestinal P-gp and putative hepatic OATP1B induction: international transporter consortium perspective on drug development implications.
      ,
      • Rodrigues A.D.
      • Lai Y.
      • Shen H.
      • Varma M.V.S.
      • Rowland A.
      • Oswald S.
      Induction of human intestinal and hepatic organic anion transporting polypeptides: where is the evidence for its relevance in drug-drug interactions?.
      but the current results support the importance of considering OATP1B induction by RIF in quantitatively predicting the DDI between PMF and RIF, and possibly for other OATP1B substrates.
      In conclusion, we verified that the hepatic uptake clearance of PMF in the presence of 5% HSA was better than those in the absence of HSA to improve the prediction accuracy of IVIVE, which may be applied to other highly protein-bound OATP1B substrate drugs. We also developed a PBPK model of PMF that can capture the observed DDI data caused by CsA and RIF implementing step-wise adjustment (pre-incubation and further correcting for substrate-dependency) of the inhibition potency of CsA/RIF. For further improvement of the PBPK model of PMF, additional strategies may include the consideration of the inhibition of the intestinal efflux transporters.

      Conflict of Interest

      The authors declare no conflicts of interest.

      Appendix A. Supplementary Data

      References

        • Fruchart J.C.
        Pemafibrate (K-877), a novel selective peroxisome proliferator-activated receptor alpha modulator for management of atherogenic dyslipidaemia.
        Cardiovasc Diabetol. 2017; 16: 124
        • Ishibashi S.
        • Yamashita S.
        • Arai H.
        • et al.
        Effects of K-877, a novel selective PPARalpha modulator (SPPARMalpha), in dyslipidaemic patients: a randomized, double blind, active- and placebo-controlled, phase 2 trial.
        Atherosclerosis. 2016; 249: 36-43
        • Ishibashi S.
        • Arai H.
        • Yokote K.
        • et al.
        Efficacy and safety of pemafibrate (K-877), a selective peroxisome proliferator-activated receptor alpha modulator, in patients with dyslipidemia: results from a 24-week, randomized, double blind, active-controlled, phase 3 trial.
        J Clin Lipidol. 2018; 12: 173-184
        • Pradhan A.D.
        • Paynter N.P.
        • Everett B.M.
        • et al.
        Rationale and design of the pemafibrate to reduce cardiovascular outcomes by reducing triglycerides in patients with diabetes (PROMINENT) study.
        Am Heart J. 2018; 206: 80-93
      1. Pharmaceuticals and Medical Devices Agency.
        (Available at:)
        • Ogawa S.I.
        • Shimizu M.
        • Yamazaki H.
        Plasma concentrations of pemafibrate with co-administered drugs predicted by physiologically based pharmacokinetic modeling in virtual populations with renal/hepatic impairment.
        Xenobiotica. 2020; : 1-9
        • Niemi M.
        • Backman J.T.
        • Fromm M.F.
        • Neuvonen P.J.
        • Kivisto K.T.
        Pharmacokinetic interactions with rifampicin : clinical relevance.
        Clin Pharmacokinet. 2003; 42: 819-850
        • Kapetas A.J.
        • Sorich M.J.
        • Rodrigues A.D.
        • Rowland A.
        Guidance for rifampin and midazolam dosing protocols to study intestinal and hepatic cytochrome P450 (CYP) 3A4 induction and de-induction.
        AAPS J. 2019; 21: 78
        • Miyauchi S.
        • Masuda M.
        • Kim S.J.
        • et al.
        The phenomenon of albumin-mediated hepatic uptake of organic anion transport polypeptide substrates: prediction of the in vivo uptake clearance from the in vitro uptake by isolated hepatocytes using a facilitated-dissociation model.
        Drug Metab Dispos. 2018; 46: 259-267
        • Kim S.J.
        • Lee K.R.
        • Miyauchi S.
        • Sugiyama Y.
        Extrapolation of in vivo hepatic clearance from in vitro uptake clearance by suspended human hepatocytes for anionic drugs with high binding to human albumin: improvement of in vitro-to-in vivo extrapolation by considering the "Albumin-Mediated" hepatic uptake mechanism on the basis of the "Facilitated-Dissociation model".
        Drug Metab Dispos. 2019; 47: 94-103
        • Poulin P.
        • Haddad S.
        Albumin and uptake of drugs in cells: additional validation exercises of a recently published equation that quantifies the albumin-facilitated uptake mechanism(s) in physiologically based pharmacokinetic and pharmacodynamic modeling Research.
        J Pharm Sci. 2015; 104: 4448-4458
        • Bteich M.
        • Poulin P.
        • Haddad S.
        The potential protein-mediated hepatic uptake: discussion on the molecular interactions between albumin and the hepatocyte cell surface and their implications for the in vitro-to-in vivo extrapolations of hepatic clearance of drugs.
        Expert Opin Drug Metab Toxicol. 2019; 15: 633-658
        • Shand D.G.
        • Cotham R.H.
        • Wilkinson G.R.
        Perfusion-limited of plasma drug binding on hepatic drug extraction.
        Life Sci. 1976; 19: 125-130
        • Yoshida K.
        • Maeda K.
        • Sugiyama Y.
        Hepatic and intestinal drug transporters: prediction of pharmacokinetic effects caused by drug-drug interactions and genetic polymorphisms.
        Annu Rev Pharmacol Toxicol. 2013; 53: 581-612
        • Li R.
        • Barton H.A.
        • Varma M.V.
        Prediction of pharmacokinetics and drug-drug interactions when hepatic transporters are involved.
        Clin Pharmacokinet. 2014; 53: 659-678
        • Izumi S.
        • Nozaki Y.
        • Maeda K.
        • et al.
        Investigation of the impact of substrate selection on in vitro organic anion transporting polypeptide 1B1 inhibition profiles for the prediction of drug-drug interactions.
        Drug Metab Dispos. 2015; 43: 235-247
        • Noe J.
        • Portmann R.
        • Brun M.E.
        • Funk C.
        Substrate-dependent drug-drug interactions between gemfibrozil, fluvastatin and other organic anion-transporting peptide (OATP) substrates on OATP1B1, OATP2B1, and OATP1B3.
        Drug Metab Dispos. 2007; 35: 1308-1314
        • Yoshikado T.
        • Yoshida K.
        • Kotani N.
        • et al.
        Quantitative analyses of hepatic OATP-mediated interactions between statins and inhibitors using PBPK modeling with a parameter optimization method.
        Clin Pharmacol Ther. 2016; 100: 513-523
        • Yoshikado T.
        • Toshimoto K.
        • Maeda K.
        • et al.
        PBPK modeling of coproporphyrin I as an endogenous biomarker for drug interactions involving inhibition of hepatic OATP1B1 and OATP1B3.
        CPT Pharmacometrics Syst Pharmacol. 2018; 7: 739-747
        • 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. 2009; 328: 652-662
        • Asaumi R.
        • Menzel K.
        • Lee W.
        • et al.
        Expanded physiologically-based pharmacokinetic model of rifampicin for predicting interactions with drugs and an endogenous biomarker via complex mechanisms including organic anion transporting polypeptide 1B induction.
        CPT Pharmacometrics Syst Pharmacol. 2019; 8: 845-857
        • Asaumi R.
        • Toshimoto K.
        • Tobe Y.
        • et al.
        Comprehensive PBPK model of rifampicin for quantitative prediction of complex drug-drug interactions: CYP3A/2C9 induction and OATP inhibition effects.
        CPT Pharmacometrics Syst Pharmacol. 2018; 7: 186-196
        • Forker E.L.
        • Luxon B.A.
        Albumin-mediated transport of rose bengal by perfused rat liver. Kinetics of the reaction at the cell surface.
        J Clin Invest. 1983; 72: 1764-1771
        • Weisiger R.
        • Gollan J.
        • Ockner R.
        Receptor for albumin on the liver cell surface may mediate uptake of fatty acids and other albumin-bound substances.
        Science. 1981; 211: 1048-1051
        • Bowman C.M.
        • Benet L.Z.
        An examination of protein binding and protein-facilitated uptake relating to in vitro-in vivo extrapolation.
        Eur J Pharm Sci. 2018; 123: 502-514
        • Amundsen R.
        • Christensen H.
        • Zabihyan B.
        • Asberg A.
        Cyclosporine A, but not tacrolimus, shows relevant inhibition of organic anion-transporting protein 1B1-mediated transport of atorvastatin.
        Drug Metab Dispos. 2010; 38: 1499-1504
        • Shitara Y.
        • Takeuchi K.
        • Nagamatsu Y.
        • Wada S.
        • Sugiyama Y.
        • Horie T.
        Long-lasting inhibitory effects of cyclosporin A, but not tacrolimus, on OATP1B1- and OATP1B3-mediated uptake.
        Drug Metab Pharmacokinet. 2012; 27: 368-378
        • Gertz M.
        • Cartwright C.M.
        • Hobbs M.J.
        • et al.
        Cyclosporine inhibition of hepatic and intestinal CYP3A4, uptake and efflux transporters: application of PBPK modeling in the assessment of drug-drug interaction potential.
        Pharm Res (N Y). 2013; 30: 761-780
        • Furihata T.
        • Matsumoto S.
        • Fu Z.
        • et al.
        Different interaction profiles of direct-acting anti-hepatitis C virus agents with human organic anion transporting polypeptides.
        Antimicrobial Agents Chemother. 2014; 58: 4555-4564
        • Shitara Y.
        • Sugiyama Y.
        Preincubation-dependent and long-lasting inhibition of organic anion transporting polypeptide (OATP) and its impact on drug-drug interactions.
        Pharmacol Ther. 2017; 177: 67-80
        • Pahwa S.
        • Alam K.
        • Crowe A.
        • et al.
        Pretreatment with rifampicin and tyrosine kinase inhibitor dasatinib potentiates the inhibitory effects toward OATP1B1- and OATP1B3-mediated transport.
        J Pharm Sci. 2017; 106: 2123-2135
        • Ufuk A.
        • Kosa R.E.
        • Gao H.
        • et al.
        In vitro-in vivo extrapolation of OATP1B-mediated drug-drug interactions in cynomolgus monkey.
        J Pharmacol Exp Ther. 2018; 365: 688-699
        • Barnett S.
        • Ogungbenro K.
        • Menochet K.
        • et al.
        Gaining mechanistic insight into coproporphyrin I as endogenous biomarker for OATP1B-mediated drug-drug interactions using population pharmacokinetic modeling and simulation.
        Clin Pharmacol Ther. 2018; 104: 564-574
        • Yoshida K.
        • Maeda K.
        • Sugiyama Y.
        Transporter-mediated drug--drug interactions involving OATP substrates: predictions based on in vitro inhibition studies.
        Clin Pharmacol Ther. 2012; 91: 1053-1064
        • Prueksaritanont T.
        • Chu X.
        • Evers R.
        • et al.
        Pitavastatin is a more sensitive and selective organic anion-transporting polypeptide 1B clinical probe than rosuvastatin.
        Br J Clin Pharmacol. 2014; 78: 587-598
        • Pedersen J.M.
        • Khan E.K.
        • Bergstrom C.A.S.
        • Palm J.
        • Hoogstraate J.
        • Artursson P.
        Substrate and method dependent inhibition of three ABC-transporters (MDR1, BCRP, and MRP2).
        Eur J Pharm Sci. 2017; 103: 70-76
        • Estudante M.
        • Morais J.G.
        • Soveral G.
        • Benet L.Z.
        Intestinal drug transporters: an overview.
        Adv Drug Deliv Rev. 2013; 65: 1340-1356
        • Harwood M.D.
        • Neuhoff S.
        • Carlson G.L.
        • Warhurst G.
        • Rostami-Hodjegan A.
        Absolute abundance and function of intestinal drug transporters: a prerequisite for fully mechanistic in vitro-in vivo extrapolation of oral drug absorption.
        Biopharm Drug Dispos. 2013; 34: 2-28
        • Zamek-Gliszczynski M.J.
        • Patel M.
        • Yang X.
        • et al.
        Intestinal P-gp and putative hepatic OATP1B induction: international transporter consortium perspective on drug development implications.
        Clin Pharmacol Ther. 2020; https://doi.org/10.1002/cpt.1916
        • Rodrigues A.D.
        • Lai Y.
        • Shen H.
        • Varma M.V.S.
        • Rowland A.
        • Oswald S.
        Induction of human intestinal and hepatic organic anion transporting polypeptides: where is the evidence for its relevance in drug-drug interactions?.
        Drug Metab Dispos. 2020; 48: 205-216