Abstract
In recent years, some endogenous substrates of organic anion transporting polypeptide 1B (OATP1B) have been identified and characterized as potential biomarkers to assess OATP1B-mediated clinical drug–drug interactions (DDIs). However, quantitative determination of their selectivity to OATP1B is still limited. In this study, we developed a relative activity factor (RAF) method to determine the relative contribution of hepatic uptake transporters OATP1B1, OATP1B3, OATP2B1, and sodium-taurocholate co-transporting polypeptide (NTCP) on hepatic uptake of several OATP1B biomarkers, including coproporphyrin I (CPI), coproporphyrin I CPIII, and sulfate conjugates of bile acids: glycochenodeoxycholic acid sulfate (GCDCA-S), glycodeoxycholic acid sulfate (GDCA-S), and taurochenodeoxycholic acid sulfate (TCDCA-S). RAF values for OATP1B1, OATP1B3, OATP2B1, and NTCP were determined in cryopreserved human hepatocytes and transporter transfected cells using pitavastatin, cholecystokinin, resveratrol-3-O-β-D-glucuronide, and taurocholic acid (TCA) as reference compounds, respectively. OATP1B1-specific pitavastatin uptake in hepatocytes was measured in the absence and presence of 1 µM estropipate, whereas NTCP-specific TCA uptake was measured in the presence of 10 µM rifampin. Our studies suggested that CPI was a more selective biomarker for OATP1B1 than CPIII, whereas GCDCA-S and TCDCA-S were more selective to OATP1B3. OATP1B1 and OATP1B3 equally contributed to hepatic uptake of GDCA-S. The mechanistic static model, incorporating the fraction transported of CPI/III estimated by RAF and in vivo elimination data, predicted several perpetrator interactions with CPI/III. Overall, RAF method combined with pharmacogenomic and DDI studies is a useful tool to determine the selectivity of transporter biomarkers and facilitate the selection of appropriate biomarkers for DDI evaluation.
SIGNIFICANCE STATEMENT The authors developed a new relative activity factor (RAF) method to quantify the contribution of hepatic uptake transporters organic anion transporting polypeptide (OATP)1B1, OATP1B3, OATP2B1, and sodium taurocholate co-transporting polypeptide (NTCP) on several OATP1B biomarkers and evaluated their predictive value on drug-drug interactions (DDI). These studies suggest that the RAF method is a useful tool to determine the selectivity of transporter biomarkers. This method combined with pharmacogenomic and DDI studies will mechanistically facilitate the selection of appropriate biomarkers for DDI prediction.
Introduction
Inhibition of organic anion transporting polypeptide (OATP)1B1 and OATP1B3 (collectively referred to throughout as “OATP1B”), the major hepatic uptake transporters for many clinically used drugs, has been associated with serious adverse effects, including statin-induced myopathy and rhabdomyolysis (Neuvonen et al., 2006; Gupta et al., 2016; Hougaard Christensen et al., 2020). Given the importance of transporters in drug interactions, regulatory agencies have recommended evaluating transporter-mediated drug–drug interaction (DDI) for new molecular entities (https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-investigation-drug-interactions-revision-1_en.pdf and https://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-drug-interaction-studies-cytochrome-p450-enzyme-and-transporter-mediated-drug-interactions). However, there are still significant challenges to use in vitro data to predict transporter-related perpetrator DDIs due to the limitations of in vitro transporter inhibition assays and gaps in in vitro to in vivo extrapolation (Chu et al., 2017). Therefore, in recent years, there has been a tremendous interest and progress in identifying, characterizing, and applying endogenous biomarkers to access DDI potential for hepatic and renal transporters in early clinical trials (Chu et al., 2018; Rodrigues et al., 2018; Mochizuki et al., 2021; Arya et al., 2022). Among all, several classes of endogenous substrates of OATP1B demonstrated the potential to serve as surrogate clinical probes to assess DDIs. For instance, the suitability of coproporphyrins I and III (CPI and CPIII), the by-products of heme synthesis, as OATP1B biomarkers have been broadly evaluated in vitro and in vivo. The majority of CPs are synthesized in the liver (Yoshikado et al., 2018). They are metabolically stable and eliminated intact into bile and urine (Shen et al., 2016). CPI/III are substrates of OATP1B1, OATP1B3, and multidrug resistance proteins (MRP) 2/3/4. CPIII is also a substrate of OATP2B1 (Bednarczyk and Boiselle, 2016; Kunze et al., 2018; Chatterjee et al., 2021). In collective clinical studies, CPI has exhibited higher sensitivity to detect potent, moderate, and weak OATP1B inhibition (Kunze et al., 2018; Mori et al., 2020b; Zhang et al., 2020; Mochizuki et al., 2022; Tess et al., 2022). Recently, CPI model-based approaches have been developed and applied to quantitatively predict the magnitude of DDIs for OATP1B clinical substrates (Yoshikado et al., 2018; Takita et al., 2021; Kimoto et al., 2022; Tess et al., 2022).
Other than CPs, conjugated bile acids (BA) could also be sensitive biomarkers for OATP1B. Bile acids are first synthesized in liver from cholesterol, conjugated with sulfate or glucuronide, then actively secreted into bile and eventually reabsorbed into blood circulation. In several clinical DDI studies, the area under plasma concentration-time profile (AUC) of glycochenodeoxycholic acid sulfate (GCDCA-S), a substrate of OATP1B1, OATP1B3, sodium-taurocholate co-transporting polypeptide (NTCP), and MRP2, were significantly increased following the administration of rifampin [600 mg, single dose (SD), by mouth], a potent OATP1B inhibitor (Takehara et al., 2017, 2018; Tatosian et al., 2021). In fact, the response of GCDCA-S (the ratio of AUC [AUCR] ∼10- 5 to 20-fold) after rifampin administration was much greater than other biomarkers (e.g., CPI). Likewise, a recent clinical pharmacogenomic study suggested that glycochenodeoxycholate and glycodeoxycholate 3-O-glucuronides (GCDCA-3G and GDCA-3G) could also be sensitive and specific OATP1B1 biomarkers (Neuvonen et al., 2021).
Because multiple transporters are involved in hepatic elimination of OATP1B biomarkers, identifying the biomarkers with higher selectivity to the transporter(s) of the interest will enable mechanistic interpretation of biomarker data and quantitative prediction of DDIs (Chu et al., 2018). The selectivity of transporter biomarkers can be determined in vitro and in vivo. The in vitro relative activity factor (RAF) method has been established to quantitatively determine relative contribution of multiple uptake transporters to hepatic uptake of drugs (Hirano et al., 2004; Kunze et al., 2014; Izumi et al., 2018, 2020). It can be explored to study the selectivity of OATP1B biomarkers. Using this approach, functional uptake of a “relatively” selective probe of specific transporter in hepatocyte over that in transporter transfected cells can be measured to obtain the RAF value. However, previous RAF methods may over- or underestimate the contribution of selective uptake transporter(s), due to the lack of selective probe drugs.
The objectives of this work are to evaluate the in vitro selectivity of CPI, CPIII, GCDCA-S, glycodeoxycholic acid sulfate (GDCA-S), and taurochenodeoxycholic acid sulfate (TCDCA-S) by determining the relative contribution of major hepatic uptake transporters (OATP1B1, OATP1B3, OATP2B1, and NTCP) on uptake of these biomarkers. In this study, we have developed a new RAF method to determine transporter specific uptake by employing selective probe substrates and inhibitors, which will improve the accuracy of determining in vitro selectivity of OATP1B biomarkers. The data generated will be useful to improve quantitative prediction of OATP1B-mediated DDIs using biomarkers. It is worth noting that organic anion transporter OAT2 and organic cation transporter OCT1 may also contribute to hepatic uptake of endogenous compounds and substrate drugs. Because the OATP1B biomarkers we studied are not substrates of these transporters (Table 1), RAF values of OAT2 and OCT1 are not evaluated.
Kinetic parameters of OATP1B1-, OATP1B3-, OATP2B1-, and NTCP-mediated uptake of several OATP1B biomarkers
Each value represents the best-fit value ± S.E. CPI, CPIII, and GCDCA-S are not substrates of human organic anion transporter OAT2 (Shen et al., 2017; Willemin et al., 2021) and organic cation transporter OCT1 (Bednarczyk and Boiselle, 2016; Kunze et al., 2018).
Materials and Methods
Chemicals and Reagents
[3H]Pitavastatin and unlabeled pitavastatin were purchased from American Radiolabeled Chemicals, Inc. (St. Louis, MO). [3H]estradiol-17β-D-glucuronide (E217βG), [3H]cholecystokinin (CCK8), [3H]estrone 3-sulfate (ES), and [3H]taurocholic acid (TCA) were purchased from PerkinElmer Life Sciences (Boston, MA). CPI, E217βG, ES, cyclosporine A (CsA), rifampin, rifamycin SV, quinidine, sodium taurochenodeoxycholate (TCDC), estropipate (EPP), and labetalol were purchased from Millipore Sigma (St. Louis, MO). CPIII was purchased from Chemscene (Monmouth Junction, NJ). GCDCA-S, GDCA-S, and TCDCA-S, unlabeled, as well as the deuterium-labeled version (GDCA-d4-S, GCDCA-d4-S, and TCDCA-d4-S), were purchased from Isosciences LLC (Ambler, PA). Resveratrol-3-O-β-D-glucuronide (R3G) was purchased from Cayman Chemical (Ann Arbor, MI). Bromosulfophthalein was purchased from MP Biomedicals (Solon, OH).
Cell Culture and Reagents
Human embryonic kidney 293 (HEK293) cells, HEK293 cells stably expressing OATP1B1 (SLCO1B1) or OATP1B3 (SLCO1B3) (HEK293-OATP1B1 or HEK293-OATP1B3, respectively) were provided by Solvo Biotechnology (Budapest, Hungary) and were used under license agreement; HEK293, HEK293-OATP1B1, and HEK293-OATP1B3 cells were cultured and cryopreserved by Evotec (Hamburg, Germany) and used for OATP1B1 and OATP1B3 experiments.
HEK293 cells transiently expressing OATP2B1 (SLCO2B1) or NTCP (SLC10A1) (HEK293-OATP2B1 or HEK293-NTCP, respectively) and control HECK293 cells were purchased from Corning Life Sciences (Manassas, VA).
Cryopreserved human hepatocytes, mixed gender, lot HCE (20 donors) and lot MMN (10 donors) were purchased from BioIVT (Baltimore, MD). All cell culture media and reagents were obtained from Thermo Fisher Scientific (Carlsbad, CA). All other reagents were commercially obtained with the highest analytical purity grade.
In Vitro Biomarker Uptake and Inhibition Studies in OATP1B1-, OATP1B3-, OATP2B1-, and NTCP-Transfected Cells
Uptake and inhibition studies of several OATP1B biomarkers were conducted in OATP1B1 or OATP1B3 stably transfected HEK293 cells, as well as OATP2B1 or NTCP transiently transfected HEK293 cells. Briefly, cryopreserved HEK293, HEK293-OATP1B1, or HEK293-OATP1B3 aliquots were thawed at 37°C, recovered in Dulbecco’s modified Eagle’s medium containing 10% FBS, and washed with Hanks’ balanced salt solution with 10 mM HEPES, pH 7.4. Cells were then re-suspended in 96-well glass coated plates at a density of 0.125 × 106 cells/well. For cryopreserved HEK293, HEK293-OATP2B1, or HEK293-NTCP aliquots, cells were thawed at 37°C, recovered in Dulbecco’s modified Eagle’s medium containing 10% FBS and cultured for 24 hours. For NTCP-transfected cells, cells were treated with 2 mM sodium butyrate (Millipore Sigma, Burlington, MA) for 24 hours to increase transporter expression prior to the experiment, and uptake/inhibition studies were conducted at a cell density of 0.2 × 106 cells/well.
For uptake time course studies, uptake was initiated by the addition of biomarker or positive control substrates (0.1 µM [3H]E217βG, 5 nM [3H]CCK8, 0.1 µM [3H]ES, and 1 µM [3H]TCA for OATP1B1, OATP1B3, OATP2B1, and NTCP, respectively), with or without positive control inhibitor (10 µM CsA for OATP1B1, OATP1B3, and NTCP and 100 µM bromosulfophthalein for OATP2B1), into control or transporter transfected cells at different time points.
For uptake kinetic studies, uptake was initiated by the addition of various concentrations of the biomarker at indicated time point based on the linear uptake of the biomarker in the respective transporter transfected cell lines.
Inhibitory effects of several compounds on uptake of biomarkers were also determined in transporter transfected and control cells. For inhibition studies in OATP1B1-, OATP1B3-, and OATP2B1-transfected cells, cells were preincubated with various concentrations of inhibitors for 30 minutes at 37°C, and uptake was then initiated by the addition of the biomarker and incubated at 37°C for indicated time. For inhibition studies in NTCP-transfected cells, uptake was initiated by the addition of the biomarker containing various concentrations of inhibitors into cells and coincubated at 37°C for indicated time.
For all studies, uptake was stopped by the addition of ice-cold PBS, followed by immediate centrifugation (Model 5180R; Eppendorf, Hamburg, Germany) for 1 minute at 3,000 rpm at 4°C and washing of the cell pellets with PBS three times.
For radiolabeled positive control samples, cell pellets were resuspended in 50% acetonitrile, and scintillation fluid (Ultima Gold; PerkinElmer, Waltham, MA) was added. Radioactivity was then determined by liquid scintillation counting in a Tri-Carb 4910TR liquid scintillation counter (PerkinElmer).
For non-radiolabeled biomarker samples, cell pellets were resuspended in 80% acetonitrile with 0.2 µM internal standard labetalol, followed by immediate centrifugation for 5 minutes at 3000 rpm at 4°C. The supernatant was then transferred to 96-well plate and analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS).
Determination of RAF of OATP1B1, OATP1B3, OATP2B1, and NTCP Using Reference Compounds
To estimate the contribution of individual transporter to hepatic uptake of biomarker, RAF of OATP1B1, OATP1B3, OATP2B1, and NTCP was measured by determining transporter-specific uptake kinetics in transporter transfected cells and hepatocytes using reference compounds. In this study, pitavastatin, CCK8, R3G, and TCA were selected as reference compounds for OATP1B1, OATP1B3, OATP2B1, and NTCP, respectively.
Uptake time course and kinetic studies of pitavastatin, CCK8, R3G, and TCA were conducted in HEK293-OATP1B1, HEK293-OATP1B3, HEK293-OATP2B1, and HEK293-NTCP, respectively, following the procedure described above.
Except for cholecystokinin (CCK-8) (specific substrate of OATP1B3) (Gui and Hagenbuch, 2008) and R3G (specific substrate of OATP2B1) (Bi et al., 2019) (Supplemental Fig. 1), pitavastatin (substrate of OATP1B1, OATP1B3, OATP2B1, and NTCP) (Prueksaritanont et al., 2014) and TCA (substrate of OATP1B1, OATP1B3, and NTCP) (Maeda et al., 2006) were not selective probe substrate for OATP1B1 and NTCP, respectively. For pitavastatin, OATP2B1 and NTCP was reported to only exhibit low to negligible roles on its hepatic uptake (Hirano et al., 2006; Vildhede et al., 2016; Bi et al., 2017, 2019). To measure OATP1B1 (pitavastatin)- and NTCP (TCA)-specific uptake kinetics of these reference compounds in human hepatocytes, we evaluated differential inhibitory effects of EPP (reported selective OATP1B1 inhibitor) (Zhang et al., 2019) and rifampin (not a potent inhibitor of NTCP) (Prueksaritanont et al., 2014) in transporter transfected cells. The results will help in selecting an appropriate inhibitor and inhibitor concentration to selectively inhibit relevant transporters (see below).
Inhibitory Effect of EPP on OATP1B1-Mediated Pitavastatin Uptake and OATP1B3-Mediated CCK8 Uptake
To determine the concentration of EPP that can only inhibit OATP1B1 in human hepatocytes, inhibitory effect of EPP on OATP1B1-mediated pitavastatin uptake and OATP1B3-mediated CCK8 uptake were assessed in transporter transfected cells, respectively. In these studies, we did not preincubate EPP with OATP1B1- and OATP1B3-transfected cells, because EPP will be used in uptake kinetic studies of pitavastatin in suspended human hepatocytes, where 30 minutes preincubation of hepatocytes at 37°C will reduce cell viability and subsequently the transport activity of substrate drugs (unpublished observation). Uptake was initiated by the addition of pitavastatin or CCK8 containing various concentration of EPP into control and OATP1B1- or OATP1B3-transfected cells and incubated at 37°C for indicated time. Other procedures were conducted as described above. Due to the very low transport activity of pitavastatin by OATP2B1 (data not shown), we are not able to estimate inhibitory effect of EPP on OATP2B1-mediated pitavastatin uptake.
Inhibitory Effect of Rifampin on OATP1B1-, OATP1B3-, and NTCP-Mediated TCA Uptake
TCA is a known substrate of OATP1B1, OATP1B3, and NTCP (Maeda et al., 2006). To determine the concentration of rifampin to selectively inhibit OATP1B1 and OATP1B3, but not NTCP, in human hepatocytes, inhibitory effect of rifampin on OATP1B1-, OATP1B3-, and NTCP-mediated TCA uptake were assessed. The studies were conducted without preincubation of rifampin with transfected cells based on the reasons discussed above for EPP inhibition study and the lack of preincubation-dependent inhibitory effects of rifampin to these transporters (unpublished observations). Uptake was initiated by the addition of TCA containing various concentration of rifampin into OATP1B1-, OATP1B3-, or NTCP-transfected and control cells and incubated at 37°C for indicated time. Other procedures were conducted as described above.
Uptake Kinetic Studies of Reference Compounds in Cryopreserved Human Hepatocytes
Uptake kinetic studies of reference compounds were conducted in two lots of cryopreserved human hepatocyte (lots HCE and MMN). Briefly, hepatocytes were thawed and resuspended in Krebs-Henseleit modified buffer (Sigma-Aldrich, St. Louis, MO) (pH 7.4) at a density of 0.5 × 106 cells/tube. After 3 minutes of preincubation at 37°C, uptake was initiated by the addition of various concentrations of reference compound into hepatocytes and incubated at 37°C at 0.5 and 1.0 or 1.5 minutes (within the linear range of uptake time course, data not shown) under the following conditions: 1) pitavastatin in the absence and presence of EPP (1 µM); 2) CCK8; 3) R3G; and 4) TCA in the presence of rifampin (10 µM). The reaction was terminated by spinning down the hepatocyte suspension through an oil layer (a mixture of silicone oil and mineral oil, density = 1.008) at 13,000 rpm for 1 minute. The supernatant (top portion above the oil layer) was collected for recovery determination. After aspirating the oil layer, the hepatocyte pellet was then lysed in 50% acetonitrile and mixed well. The radioactivity of pitavastatin, CCK8, and TCA in both supernatant and cell pellets was then determined by Tri-Carb liquid scintillation counter. For R3G (non-radioactive) samples, the hepatocyte pellets were resuspended in 80% acetonitrile with 0.2 µM internal standard labetalol, followed by centrifugation for 1 minute at 13,000 rpm. The supernatant was then transferred to 96-well plate and analyzed by LC-MS/MS.
Uptake of Several OATP1B Biomarkers into Cryopreserved Human Hepatocytes
To assess the contribution of active hepatic uptake versus passive diffusion of several OATP1B biomarkers, uptake of unlabeled CPI (0.01 µM) and CPIII (0.05 µM), and of deuterium (d4)-labeled GCDCA-S (0.1 µM), GDCA-S (0.1 µM), and TCDCA-S (0.1 µM), were conducted in cryopreserved human hepatocytes (lot HCE) at 0.5 and 1.5 minutes in absence and presence of a transporter inhibitor cocktail (CsA, 10 µM; rifamycin SV, 10 µM; rifampin, 100 µM; quinidine, 50 µM; and TCDC, 50 µM). CsA and TCDC are OATP1B and NTCP inhibitors (Schroeder et al., 1998), whereas rifampin and rifamycin SV are OATP1B and OATP2B1 inhibitors and quinidine is organic cation transporter 1 (OCT1) inhibitor. This transporter inhibitor cocktail was known to fully inhibit transporter-mediated uptake in human hepatocytes by OATP1B1, OATP1B3, OATP2B1, NTCP, and OCT1 (Houle et al., manuscript in preparation). The reaction was stopped according to the method described above and samples were then analyzed by LC-MS/MS.
LC-MS/MS Quantification of Biomarkers and R3G in In Vitro Transport Studies
The sample analysis was conducted on a SCIEX 4500 tandem mass spectrometer (Applied Biosystems/MDS SCIEX, Toronto, Canada) coupled to an ultra-high performance liquid chromatography Dionex Ultimate 3000 RS pump and an HTS PAL autosampler (CTC Analytics, Zwingen, Switzerland). Samples (5 µl) were injected onto an Acquity ultra high performance liquid chromatography HSS T3 2.1 50 mm UHPLC column, particle size 1.7 µm (Waters, Ireland). For quantification of biomarkers, the samples were eluted by a gradient program of 5% B, held for 15 seconds, 5% B to 95% B in 75 seconds, 95% B for 25 seconds, and held 95% B for 60 seconds. For quantification of R3G, the samples were eluted by a gradient program of 5% B, held for 15 seconds, 5% B to 95% B in 90 seconds, 95% B for 25 seconds, and held 95% B for 60 seconds. The mobile phase was a mixture of 0.1% formic acid in water (A) and acetonitrile containing 0.1% formic acid (B). The column temperature was maintained at ambient, and the flow rate was 0.75 ml/min. The spectrometry and multiple reaction monitoring precursor/product ion transitions of each biomarker were referenced in Supplemental Table 1.
Data Analysis
Inhibition Studies
Transporter-mediated uptake is calculated by subtracting the uptake rate in control cells from that in transporter-expressing cells (within the linear range of uptake time course of a probe substrate). Data are normalized to percent control (% control) and the IC50 values for inhibition of transporter-mediated uptake were obtained by fitting the data to the formula below by nonlinear regression analysis.
where “% Control” represents transporter-mediated uptake measured in the presence of various concentrations of inhibitor to that in the absence of inhibitor; “I” represents nominal inhibitor concentration (µM), and “γ” represents the Hill slope. The IC50 is calculated using GraphPad Prism (San Diego, CA) and reported as best-fit value ± S.E. from three independent measurements.
Uptake Kinetic Analysis of Biomarkers and Reference Compounds in Transporter Transfected Cell Lines
The initial uptake rate, determined within the linear uptake range, for transporter-mediated uptake of biomarkers or reference compounds was obtained by subtracting the uptake rate in control cells from that in the transporter-transfected cells, and the kinetic parameters were estimated by fitting the data to eq. 2 or eq. 3 (wherever appropriate) using nonlinear least squares method by GraphPad Prism.
where v is the initial uptake rate of test compound (pmol/min/106 cells), S is substrate concentration (µM), Km is the Michaelis-Menten constant (µM), Vmax is the maximum uptake rate (pmol/min/106 cells), and CLns is the nonsaturable uptake clearance via passive diffusion or other mechanisms (µl/min/106 cells).
CLint, cells, intrinsic uptake clearance in transfected cells mediated by the transporter, is the sum of Vmax/Km and CLns.
Uptake Kinetic Analysis of Reference Compounds in Human Hepatocytes
Pitavastatin
To estimate the hepatic uptake clearance of pitavastatin mediated by OATP1B1 in human hepatocytes, the initial uptake rate of pitavastatin (0.03–30 µM) in the absence or presence of EPP (1 µM) was first determined using the eq. 4 and eq. 5, respectively. The OATP1B1-mediated hepatic clearance of pitavastatin was then obtained by subtracting the uptake rate of pitavastatin in the presence of EPP from that in the absence of EPP (eq. 6) and the kinetic parameters were obtained by fitting the data to eq. 2 or eq. 3.
where vpita, vpita+EPP, and vpita, OATP1B1 represent the initial uptake rate of pitavastatin in the absence of EPP (1 µM) (uptake by OATP1B1 and OATP1B3), the initial uptake rate of pitavastatin in the presence of EPP (1 µM) (uptake only by OATP1B3, as 1 µM EPP inhibited OATP1B1-mediated uptake), and the OATP1B1-mediated hepatic uptake clearance of pitavastatin respectively.
CCK8
To estimate the hepatic uptake clearance of CCK8 mediated by OATP1B3 in human hepatocytes, the initial uptake rate of CCK8 (0.005–15 µM) was determined according to eq. 7 and the kinetic parameters were estimated by fitting the data to eq. 2 or eq. 3
R3G
To estimate the hepatic uptake clearance of R3G mediated by OATP2B1 in human hepatocytes, the initial uptake rate of R3G (0.1–15 µM) was determined according to eq. 8 and the kinetic parameters were estimated by fitting the data to eq. 2 or eq. 3
TCA
To estimate the hepatic uptake clearance of TCA mediated by NTCP in human hepatocytes, the initial uptake rate of TCA (0.1–50 µM) in the presence of rifampin (10 µM) (NTCP uptake only because 10 µM rifampin only inhibited OATP1B-mediated TCA uptake without impacting NTCP-mediated uptake) was determined according to eq. 9, and the kinetic parameters were estimated by fitting the data to eq. 2 or eq. 3
Total and active uptake clearance of reference compounds were subsequently estimated as follows:
Determination of RAF
RAF is defined as the ratio of active and transporter specific uptake clearance of respective reference compound in hepatocyte to that in transporter transfected cells and is determined as below
Estimation of Hepatic Uptake Clearance and Relative Contribution of Transporters to Hepatic Uptake of Biomarkers
Predicted Active Hepatic Uptake Clearance of Biomarkers (CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-S)
For CPI:
For CPIII:
For GCDCA-S, GDCA-S, and TCDCA-S:
Determination of the Fraction Transported (fth) of Biomarkers to Human Hepatocytes In Vitro
The fraction transported (fth) of the biomarkers by in vitro active hepatic uptake (fth,active) and passive diffusion (fth,passive) was estimated by uptake studies of the biomarker in the absence and present of transporter inhibitor cocktails and calculated using eqs. 18–21. The fraction transported by specific transporter (ft h, specific transporter) was calculated by eq. 22.
Uptake clearance = V/S, where S is the substrate concentration of the biomarker
Estimation of the Fraction Transported (ft) to Overall Elimination of Several Biomarkers and Prediction of Perpetrator–Biomarker Interactions Using Mechanistic Static Models
Estimation of ft Using Clinical Pharmacogenomic and DDI Data
ft of a biomarker by a transporter-mediated pathway (e.g., OATP1B1) can be calculated by eq. 23 or eq. 24.
AUCSLCO1B1 T/T or *1b/*1b and AUCSLCO1B1 C/C or *15/*15 represent plasma AUC of a biomarker in the subjects carry SLCO1B1 T/T or *1b/*1b genotype/haplotype (wild type) and SLCO1B1 C/C or *15/*15 (poor OATP1B1 function), respectively (Cooper-DeHoff et al., 2022). AUCcontrol and AUCinhibitor represent plasma AUC of a biomarker in the absence and presence of relatively selective and potent inhibitor to the transporter of the interest.
Estimation of ft for Several Biomarkers and Prediction of Perpetrator–Biomarker Interaction Using Mechanistic Static Models
CPI.
CPI is ∼85% and ∼15% eliminated via liver and kidney, respectively with minimal metabolism (Barnett et al., 2018; Takita et al., 2022), because OATP1B1 and OATP1B3 are major transporters for its hepatic uptake. The fold change of AUC (AUCR) for CPI with/without an OATP1B (OATP1B1 and OATP1B3) inhibitor drug can be calculated by eq. 25 (Chu et al., 2022):
The fraction transported by active hepatic uptake (ft, active hepatic uptake) of CPI, including the fraction transported by OATP1B1 (ft, OATP1B1) and OATP1B3 (ft, OATP1B3), were estimated to be 0.57 and 0.17, respectively, based on the fact that CPI was ∼85% eliminated via liver, and in vitro ft, h, OATP1B1 and ft, h, OATP1B3 were 0.67 and 0.20, respectively (Table 4). It was assumed that OATP1B-mediated hepatic uptake was the rate-determining step for hepatic elimination of CPI, and inhibition of hepatic efflux transporters (e.g., MRP2) has limited effect on its AUCR (Yoshikado et al., 2018). It was also assumed that OATP1B inhibitor drugs did not inhibit renal elimination and synthesis of CPI. Iin,max, u represents unbound inhibitor concentration at the inlet to the liver and was calculated via eq. 26 (https://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-drug-interaction-studies-cytochrome- p450-enzyme-and-transporter-mediated-drug-interactions).
where fu represents the plasma unbound fraction of the inhibitor drug, Imax is the maximum plasma concentration of the inhibitor drug, Fa is the fraction absorbed, Fg is the fraction escaping intestinal metabolism, ka is the absorption rate constant of the inhibitor, QH is the hepatic blood flow rate in humans (1500 ml/min), and Rb is the blood-to-plasma concentration ratio. To estimate the Iin,max, u value, Fa and Fg were set at 1, ka was set at 0.1 minute−1 as the worst-case scenario, and Rb was assumed to be 1; fu was set at 0.01 if fu of the inhibitor was determined to be less than 0.01 (https//www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-drug-interaction-studies-cytochrome-p450-enzyme-and-transporter- mediated-drug-interactions).
CPIII.
Likewise, AUCR of CPIII with or without OATP1B (OATP1B1, OATP1B3) and OATP2B1 inhibitor drug can be calculated by eq. 25. CPIII was also eliminated via hepatic and renal elimination. Based on the amount of CPIII and CPIII excreted into urine and the ratio of CPIII/CPI measured in feces in human (Kuhnel et al., 1999; Feng et al., 2021), as well as the fraction eliminated by liver and kidney for CPI (see above), the fraction of CPIII eliminated via liver and kidney were estimated to be 0.46 and 0.54, respectively. Given that CPIII in vitro ft, h, OATP1B1, ft, h, OATP1B3, and ft, h, OATP2B1 were 0.45, 0.22, and 0.07, respectively (averaged from hepatocyte lots HCE and MMN) (Table 4), the fraction transported of CPIII by OATP1B1 (ft, OATP1B1), OATP1B3 (ft, OATP1B3), and OATP2B1 (ft, OATP2B1) were estimated to be 0.21, 0.10, and 0.03, respectively. Again, it was under the assumption that OATP1B inhibitor drugs did not inhibit renal elimination and synthesis of CPIII.
Results
Uptake Time Course and Kinetics of Several OATP1B Biomarkers in Transporter Transfected Cells
Uptake time course of CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-S were conducted in OATP1B1-, OATP1B3- (Fig. 1), OATP2B1- (CPIII only), and NTCP (GCDCA-S, GDCA-S, and TCDCA-S only)-transfected and control cells (Fig. 2). Uptake of these biomarkers in transporter transfected cells was time-dependent and much higher than that in control cells, indicating that these biomarkers are substrates of OATP1B1 and OATP1B3. In addition, CPIII was transported by OATP2B1, whereas BA sulfate conjugates GCDCA-S, GDCA-S, and TCDCA-S were transported by NTCP. Uptake of positive control substrates (E217βG, CCK8, ES, and TCA for OATP1B1, OATP1B3, OATP2B1, and NTCP, respectively) were significantly higher in transporter transfected cells than control cells and were completely inhibited by respective positive control inhibitors, confirming the functionality of the assays (Figs. 1 and 2).
Uptake time course of several OATP1B biomarkers into human OATP1B1 and OATP1B3 stably transfected cells. (A–F) Time-dependent uptake of CPI (0.01 µM) (A), CPIII (0.1 µM) (B), GCDCA-S (0.1 µM) (C), GDCA-S (0.1 µM) (D), and TCDCA-S (0.1 µM) (E) into HEK293 and HEK293-OATP1B1 cells. (F) Positive control; uptake of [3H]E217βG (0.1 µM) into HEK293 and HEK293-OATP1B1 cells at 5 minutes in the presence or absence of CsA (10 µM). (G–L) Time-dependent uptake of CPI (0.01 µM) (G), CPIII (0.1 µM) (H), GCDCA-S (0.1 µM) (I), GDCA-S (0.1 µM) (J), and TCDCA-S (0.1 µM) (K) into HEK293 and HEK293-OATP1B3 cells. (L) Positive control; uptake of [3H]CCK8 (5 nM) into HEK293 and HEK293-OATP1B3 cells at 5 minutes in the presence or absence of CsA (10 µM). Open and closed squares/bars represent HEK293 and HEK293-OATP1B1 or HEK-OATP1B3, respectively. Each point represents the mean ± the S.D. of experiments performed in triplicate (n = 3).
Uptake time course of several OATP1B biomarkers into human NTCP and OATP2B1 stably transfected cells. (A–D) Time-dependent uptake of GCDCA-S (0.1 µM) (A), GDCA-S (0.1 µM) (B), and TCDCA-S (0.1 µM) (C) into HEK293 and HEK293-NTCP cells. (D) Positive control; uptake of [3H]TCA (1 µM) into HEK293 and HEK293-NTCP cells at 5 minutes in the presence or absence of CsA (10 µM). (E and F) Time-dependent uptake of CPIII (0.1 µM) (E) into HEK293 and HEK293-OATP2B1 cells; (F) positive control; uptake of [3H]ES (0.1 µM) into HEK293 and HEK293-OATP2B1 at 10 minutes in the presence or absence of bromosulfophthalein (100 µM). Open and closed squares/bars represent HEK293 and transporter (NTCP or OATP2B1) transfected HEK293 cells, respectively. Each point represents the mean ± the S.D. of experiments performed in triplicate (n = 3).
Transporter-mediated uptake kinetics of CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-S were conducted in OATP1B1-, OATP1B3-, OATP2B1- (CPIII), and NTCP (GCDCA-S, GDCA-S, and TCDCA-S)-transfected and control cells. Concentration- and transporter-mediated uptake kinetic data were shown in Fig. 3, and corresponding kinetic parameters were summarized in Table 1.
Concentration-dependent uptake of several OATP1B biomarkers by OATP1B1 (A–E), OATP1B3 (F–J), NTCP (K–M), and OATP2B1 (N). OATP1B1-mediated uptake of various concentrations of biomarkers at 37°C: CPI at 5 minutes (A), CPIII at 5 minutes (B), GCDCA-S at 5 minutes (C), GDCA-S at 2 minutes (D) and TCDCA-S at 5 minutes (E). OATP1B3-mediated uptake of various concentrations of biomarkers at 37°C: CPI at 5 minutes (F), CPIII at 5 minutes (G), GCDCA-S at 2 minutes (H), GDCA-S at 2 minutes (I) and TCDCA-S at 5 minutes (J). NTCP-mediated uptake of various concentrations of biomarkers at 37°C: GCDCA-S at 5 minutes (K), GDCA-S at 5 minutes (L) and TCDCA-S at 5 minutes (M). OATP2B1-mediated uptake of various concentrations of CPIII at 37°C at 5 minutes (N). Each point represents the mean ± the S.D. of experiments performed in triplicate (n = 3). The solid line depicts the fitted line obtained by nonlinear regression analysis as described in Materials and Methods.
Determination of RAF Values for OATP1B1, OATP1B3, OATP2B1, and NTCP
Transporter-Mediated Uptake of Reference Compounds in Transfected Cells
Pitavastatin, CCK-8, R3G, and TCA were selected as the reference compounds to determine RAF value for OATP1B1, OATP1B3, OATP2B1, and NTCP, respectively. Uptake time course and transporter-mediated uptake kinetics of pitavastatin, CCK-8, R3G, and TCA were conducted in OATP1B1-, OATP1B3-, OATP2B1-, and NTCP-transfected and control cells and the results were shown in Fig. 4, and corresponding kinetic parameters were summarized in Table 2.
Uptake time course and kinetics of transporter reference compounds into OATP1B1, OATP1B3, OATP2B1, and NTCP stably transfected cells. (A, D, G, and J) Time-dependent uptake of pitavastatin (0.1 µM) (A), CCK8 (5 nM) (D), resveratrol-3-O-β-D-glucuronide (R3G, 1 µM) (G), and TCA (0.1 µM) (J) into HEK293 and HEK293-OATP1B1, HEK293-OATP1B3, HEK293-OATP2B1, or HEK293-NTCP cells, respectively. (B, E, H, and K) Transporter-mediated uptake kinetics of pitavastatin (OATP1B1, B), CCK8 (OATP1B3, E), R3G (OATP2B1, H), TCA (NTCP, K), respectively. The solid line in these panels depicts the fitted line obtained by nonlinear regression analysis as described in Materials and Methods. All uptake rates were measured within linear range of transporter-mediated uptake. Eadie-Hofstee plots for the compounds that show biphasic kinetic profiles are also inserted. (C, F, I, and L) Uptake of pitavastatin, CCK8, R3G, and TCA at 5 minutes in the presence or absence of CsA (10 µM) for OATP1B1 (C), OATP1B3 (F), and NTCP (L), or rifamycin SV (20 µM) for OATP2B1 (I). Open and closed squares/bars represent parental HEK293 cells and transporter transfected HEK293 cells, respectively. Each point represents the mean ± the S.D. of experiments performed in triplicate (n = 3).
Kinetic parameters of reference compounds in transporter-transfected cells and human hepatocytes and their respective RAF values
Each value represents the mean ± S.E.
Selection of In Vitro Inhibitor Test Concentrations to Determine OATP1B1- and NTCP-Specific Uptake of Pitavastatin and TCA in Human Hepatocytes
Unlike CCK8 and R3G (OATP1B3 and OATP2B1 selective substrate, respectively) (Gui and Hagenbuch, 2008; Bi et al., 2019) (Supplemental Fig. 1), pitavastatin is not an OATP1B1-specific probe substrate because OATP1B3 also contributed to its hepatic uptake. To determine OATP1B1-specific uptake of pitavastatin, we evaluated the inhibitory effect of EPP, a relatively selective in vitro inhibitor of OATP1B1 (Zhang et al., 2019), on OATP1B1- and OATP1B3-mediated pitavastatin uptake in OATP1B1- and OATP1B3-transfected cells, respectively (Fig. 5A). EPP demonstrated differential inhibitory effects on OATP1B1- and OATP1B3-mediated pitavastatin uptake with obtained IC50 value of 0.08 ±0.02 µM and 12.32 ±1.54 µM, respectively. As such, 1 µM EPP, which inhibits OATP1B1, but not OATP1B3, was selected in subsequent uptake kinetic studies of pitavastatin to determine OATP1B1-specific uptake in human hepatocytes. Likewise, TCA is not a selective in vitro probe for NTCP, because it is also transported by OATP1B1 and OATP1B3 (Maeda et al., 2006). To determine NTCP-specific uptake of TCA, we measured inhibitory effects of rifampin on NTCP-, OATP1B1-, and OATP1B3-mediated TCA uptake in corresponding transporter transfected cells (Fig. 5B). Rifampin inhibited OATP1B1-, OATP1B3-, and NTCP-mediated TCA uptake with obtained IC50 value of 0.40 ± 0.10 µM, 0.15 ± 0.02 µM, and 140.3 ± 34.1 µM, respectively. Therefore, 10 µM rifampin, which completely inhibits OATP1B1- and OATP1B3-mediated TCA uptake without inhibiting NTCP-mediated uptake, was used in subsequent TCA uptake kinetic study to determine NTCP-specific TCA uptake in human hepatocytes.
Effects of EPP and rifampin on OATP1B1- and OATP1B3-mediated uptake (EPP and rifampin) and NTCP-mediated uptake (rifampin) in transporter transfected cells. (A) The inhibitory effect of EPP on OATP1B1-mediated pitavastatin (0.1 µM) uptake (open square), and OATP1B3-mediated CCK8 (5 nM) uptake (closed square) was conducted in HEK293 cells and HEK293-OATP1B1 or HEK293-OATP1B3 cells at 2 minutes, respectively. Obtained IC50 values for OATP1B1 and OATP1B3 were 0.08 ± 0.02 µM and 12.32 ± 1.54 µM, respectively. (B) The inhibitory effect of rifampin on OATP1B1-, OATP1B3-, and NTCP-mediated TCA uptake. The inhibitory effect of rifampin on OATP1B1-mediated (open square), OATP1B3-mediated (closed square), and NTCP-mediated (closed triangle) TCA (1 µM) uptake was conducted in HEK293 cells and HEK293-OATP1B1 at 4 minutes, HEK293-OATP1B3 at 2 minutes, or HEK293-NTCP cells at 2 minutes, respectively. Obtained IC50 values for OATP1B1, OATP1B3, and NTCP were 0.40 ± 0.10 µM, 0.15 ± 0.02 µM, and 140.3 ± 34.1 µM, respectively. Transporter-mediated uptake was determined by subtracting the uptake rate in control cells from that in transporter-transfected cells. Each point represents the mean ± the S.D. of experiments performed in triplicate (n = 3). The solid line depicts the fitted line obtained by nonlinear regression analysis as described in Materials and Methods.
Kinetics of OATP1B1-, OATP1B3-, OATP2B1-, and NTCP-Specific Uptake of the Reference Compound in Cryopreserved Human Hepatocytes and RAF Values
Kinetics of OATP1B1-, OATP1B3- OATP2B1-, and NTCP-specific uptake of pitavastatin, CCK8, R3G, and TCA was evaluated in two batches of cryopreserved human hepatocytes, lots MMN and HCE, respectively (Fig. 6). Concentration-dependent initial uptake rate of pitavastatin was measured in the absence and presence of 1 µM EPP. OATP1B1-specific pitavastatin uptake kinetics was obtained by subtracting its initial uptake rate in the presence of 1 µM EPP from that in its absence. NTCP-specific uptake kinetics of TCA was measured in the presence of 10 µM rifampin. OATP1B3-specific uptake kinetics of CCK8- and OATP2B1-specific uptake kinetics of R3G was measured directly because CCK8 and R3G are selective substrates of these transporters. Obtained kinetic parameters were summarized in Table 2. The RAF values for OATP1B1, OATP1B3, OATP2B1, and NTCP were 0.15, 0.50, 0.31, and 0.02 for hepatocytes lot HCE, and 0.28, 0.84, 0.37, and 0.01 for lot MMN, respectively (Table 2).
Uptake kinetics of reference compounds in cryopreserved human hepatocytes lot MMN (A–D) and lot HCE (E–H). Pitavastatin, CCK8, R3G, and TCA are reference compounds for OATP1B1, OATP1B3, OATP2B1, and NTCP, respectively. To estimate the hepatic uptake clearance of reference compounds mediated by specific transporter, uptake rate was determined as follows: for OATP1B1-specific uptake clearance of pitavastatin (A and E), OATP1B1-specific initial uptake rate of pitavastatin was calculated by subtracting the initial uptake rate of pitavastatin in the presence of estropipate (EPP, 1 µM) from that in its absence over a range of pitavastatin concentration at 0.5 minutes and 1.5 minutes; for OATP1B3-specific uptake clearance of CCK8 (B and F), initial uptake rate of CCK8 was obtained by measuring uptake of CCK8 over a range of concentration at 0.5 minutes and 1.5 minutes; for OATP2B1-specific uptake clearance of R3G (C and G), initial uptake rate of R3G was obtained by measuring uptake of R3G over a range of concentration at 0.5 minutes and 1.5 minutes; for NTCP-specific uptake clearance of TCA (D and H), initial uptake rate of TCA was obtained by the uptake rate of TCA in the presence of rifampin (10 µM) over a range of TCA concentration at 0.5 minutes and 1 minutes. Eadie-Hofstee plots for the compounds which show biphasic kinetic profiles are also inserted. Each point represents the mean ± the S.D. of experiments performed in triplicate (n = 3). The solid line depicts the fitted line obtained by nonlinear regression analysis as described in Materials and Methods.
Relative Contribution of OATP1B1, OATP1B3, OATP2B1, NTCP, and Passive Diffusion to Hepatic Uptake of Several OATP1B Biomarkers
Prediction of Relative Contribution of Each Transporter to Active Hepatic Uptake of Several OATP1B Biomarkers Using RAF Method
RAF-based uptake clearance of several OATP1B biomarkers mediated by OATP1B1 (RAF1B1 × CLuptake,1B1), OATP1B3 (RAF1B3 × CLuptake,1B3), OATP2B1 (RAF2B1 × CLuptake,2B1), and NTCP (RAFNTCP × CLuptake,NTCP) was summarized in Table 3. The relative contribution of these transporters to active hepatic uptake of CPI (OATP1B1 vs. OATP1B3), CPIII (OATP1B1 vs. OATP1B3 vs. OATP2B1), GCDCA-G, GDCA-S, and TCDCA-S (OATP1B1 vs. OATP1B3 vs. NTCP) was determined and the results were also shown in Table 3. OATP1B1 demonstrated higher contribution to active uptake of CPI (76.2%–78.6% in lot HCE and MMN) than that for CPIII (58.8%–63.7%), whereas OATP1B3 was a primary transporter for active hepatic uptake of GCDCA-S (94.3%–94.8%) and TCDCA-S (96.2%–96.5%). OATP1B1 and OATP1B3 equally contributed to hepatic uptake of GDCA-S (46.9%–50.5% for OATP1B1 and 49.3%–52.4% for OATP1B3). OATP2B1 only contributed to 7.5%–10.9% of CPIII active uptake. NTCP contributed negligibly to active hepatic uptake of GCDCA-S, GDCA-S, and TCDCA-S (0.1%–1.6%). Under the conditions tested, hepatocyte interlot variability between HCE and MMN was low.
Prediction of relative contribution of each transporter to active hepatic uptake of several OATP1B biomarkers using RAF method
Prediction of the Fraction Transported for Hepatic Uptake of Several OATP1B Biomarkers via Uptake Transporters and Passive Diffusion
To understand relative contribution of passive diffusion and active hepatic uptake, uptake rate of CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-S were measured in human hepatocytes (lot HCE) in the absence and presence of transporter inhibitor cocktails, which can fully inhibit active uptake by major human hepatic uptake transporters (OATP1B1, OATP1B3, OATP2B1, NTCP, and OCT1) (Houle et al., manuscript in preparation). Fraction transported for hepatic uptake by passive diffusion and active uptake by individual transporters was summarized in Table 4 and relative contribution (%) of each pathway on hepatic uptake of these biomarkers was shown in Fig. 7. Passive diffusion contributed to 13.4%, 27.4%, 10.8%, 22.7%, and 34.1% of hepatic uptake of CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-G, respectively. In terms of transporter-mediated uptake, OATP1B1 contributed to 66%–68%, 42.7%–46.2%, 3.6%–4.1%, 36.3%–39%, and 2.1%–2.4%, whereas OATP1B3 contributed to 18.6%–20.6%, 20.9%–22%, 84.2%–84.6%, 38.1%–40.5%, and 63.4%–63.6% of hepatic uptake of CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-G, respectively. In addition, OATP2B1 contributed to 5.4%–7.9% of CPIII uptake and NTCP only accounted for 0.5%–1.5%, 0.2%–0.5%, and 0.04%–0.1% of hepatic uptake of GCDCA-S, GDCA-S, and TCDCA-S, respectively.
Prediction of the fraction transported for hepatic uptake of several OATP1B biomarkers via hepatic uptake transporters and passive diffusion
Relative contribution of multiple hepatic uptake transporters and passive diffusion on hepatic uptake of several OATP1B biomarkers in cryopreserved human hepatocytes. Relative contribution of hepatic uptake transporters OATP1B1 (red bar), OATP1B3 (blue bar), OATP2B1 (yellow bar), NTCP (green bar), and passive diffusion (gray bar) on hepatic uptake of CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-S were determined using RAF method in cryopreserved human hepatocytes Lot MMN (A) and Lot HCE (B).
Inhibitory Effects of CsA, Rifampin, Simeprevir, and Itraconazole on Transporter-Mediated Uptake of CPI, CPIII, and GCDCA-S
Inhibitory effects of CsA, rifampin, simeprevir, and itraconazole on transporter-mediated uptake of CPI (OATP1B1 and OATP1B3), CPIII (OATP1B1, OATP1B3, and OATP2B1), and GCDCA-S (OATP1B1, OATP1B3, and NTCP) were conducted in transporter transfected cells, respectively. Obtained IC50 values were summarized in Table 5, and inhibitory profiles were shown in Supplemental Fig. 2. CsA and simeprevir showed potent inhibition (IC50 <1 µM) to uptake of CPI, CPIII, and GCDCA-S for all transporters tested. As expected, rifampin inhibited OATP1B1- and OATP1B3-mediated CPI, CPIII, and GCDCA-S uptake but showed moderate and weak inhibition to OATP2B1-mediated uptake of CPIII and NTCP-mediated uptake of GCDCA-S, respectively. Interestingly, inhibitory effects of itraconazole to OATP1B1 and OATP1B3 were substrate-dependent. It showed more potent inhibition to GCDCA-S (<1 µM) than CPI and CPIII (3–48 µM) (Table 5).
Summary of IC50 values of CsA, rifampin, simeprevir, and itraconazole on transporter-mediated uptake of several OATP1B biomarkers
Each value represents the best fit value ± S.E.
Prediction of Perpetrator-Biomarker Interaction of CPI and CPIII with Several OATP1B Inhibitor and Noninhibitor Drugs Using Mechanistic Static Model
To further understand the translation of in vitro RAF data on predicting perpetrator–biomarker interaction, we estimated ft values of CPI (OATP1B1 and OATP1B3) and CPIII (OATP1B1, OATP1B3, and OATP2B1) based on their in vivo elimination and in vitro relative contribution to hepatic uptake. Using mechanistic static modeling, we predicted AUCR of CPI and CPIII following the administration of several perpetrator drugs for which in vitro inhibition data using CPI and CPIII as probe substrates were determined in this study or reported in literature (Table 6). The predicted AUCRs were compared with those observed AUCRs of CPI and CPIII, as well as pitavastatin (relatively selective clinical probe drug for OATP1B) (Supplemental Fig. 3). Based on this limited dataset, mechanistic static models combined with ft values estimated by RAF method as well as in vitro inhibition data provided a reasonably good prediction of the perpetrator interaction with CPI and CPIII (predicted vs. observed AUCR <2-fold, except for one DDI study with 600 mg rifampin SD by mouth for CPIII: 2.3-fold). Furthermore, CPI AUCR estimated by ft and RAF method showed a good prediction of AUCR for pitavastatin (predicted vs. observed AUCR <2-fold). In contrast, predicted and observed AUCR of CPIII was less sensitive to detect weak to moderate clinical DDIs of pitavastatin (Supplemental Fig. 3).
Prediction of several perpetrator interaction with CPI and CPIII using mechanistic static model and comparison with pitavastatin DDIs
Cmax and fu values of perpetrator drugs and observed clinical perpetrator-biomarker AUCR and clinical DDIs with pitavastatin (the ratio of AUC with/without perpetrator drug; AUCR) were obtained from the University of Washington DDI database (http://www.druginteractioninfo.org) or cited references. Iin max, u and predicted CPI and CPIII AUCR using mechanistic static model were obtained as described in the Materials and Methods.
The ft, OATP1B1 of CPI, CPIII, and GCDCA-S was also estimated based on clinical pharmacogenomic data in healthy subjects carrying variant SLCO1B1 c.521C/C or SLCO1B1*15/15 genotypes, which have a poor but not complete deficiency of OATP1B1 function, as compared with those carrying wild SLCO1B1 c.521T/T or SLCO1B1*1b/1b (Supplemental Table 2). The estimated ft, OATP1B1 was 0.48 (geometric mean of four independent studies), 0.26 (geometric mean of two independent studies), and 0.41 (one study) for CPI, CPIII, and GCDCA-S, respectively. The ft, OATP1B1 of CPI and CPIII based on pharmacogenomic data are comparable to those estimated by RAF method (0.57 and 0.21 for CPI and CPIII).
Because of the lack of quantitative data on overall elimination of GCDCA-S via liver and kidney, the ft values of OATP1B1, OATP1B3, and NTCP could not be determined using the RAF approach. Despite that, our studies indicate that OATP1B3 is a major contributor than OATP1B1 and NTCP on hepatic uptake of GCDCA-S. Assuming that rifampin (600 mg SD, by mouth) completely inhibited OATP1B (OATP1B1 + OATP1B3)-mediated GCDCA-S uptake without impacting the uptake by NTCP, ftOATP1B1 + 1B3 was estimated to be 0.92 based on GCDCA-S AUCR of 13.2 (geometric mean of five rifampin clinical DDI studies) (Supplemental Table 3). As such, estimated ft, OATP1B1, ft, OATP1B3, and ft, NTCP was 0.04, 0.87, and 0.01, respectively, in which ft, OATP1B1 was much lower than the value estimated by clinical pharmacogenomic study (0.04 vs. 0.41).
Discussion
RAF method has been developed to determine relative contribution of OATP1B1 and OATP1B3 to hepatic uptake of several OATP1B substrate drugs (Hirano et al., 2004; Kunze et al., 2014; Izumi et al., 2020). In these studies, ES and CCK8 were selected as OATP1B1- and OATP1B3-specific substrates. ES is not specific to OATP1B1, however, as it is also transported by OATP1B3 and OATP2B1. Furthermore, the contribution of other hepatic uptake transporters, including OATP2B1 and NTCP, were not evaluated. As such, over- and/or underestimation of the contribution of each pathway and their hepatic uptake clearance is possible (Izumi et al., 2018, 2020). In our studies, we developed a new RAF method to determine transporter specific uptake of OATP1B1 (pitavastatin without or with 1 µM EPP), OATP1B3 (CCK8), OATP2B1 (R3G), and NTCP (TCA with 10 µM rifampin). This method has overcome the limitations of overlapping substrate specificity of reference compounds and provided a more accurate estimation of relative contribution of these transporters to hepatic uptake of transporter substrates. For the first time, we applied this approach to quantitatively determine the contribution of hepatic uptake transporters on several OATP1B biomarkers (CPI, CPIII, GCDCA-S, GDCA-S, and TCDCA-S) and evaluated their predictivity on perpetrator–biomarker interactions.
Our studies have indicated that CPI is a more selective biomarker for OATP1B1 than CPIII (ft, OATP1B1 was 0.57 and 0.21 for CPI and CPIII, respectively). Estimated ft, OATP1B1 values were generally in line with those obtained from pharmacogenomic studies (CPI, 0.48; CPIII, 0.26) (Supplemental Table 2). To further understand the predictivity of RAF method, we predicted AUCR of CPI/III with and/or without the administration of several OATP1B inhibitor or noninhibitor drugs. Within this limited dataset, predicted AUCR was within 2-fold of observed values for CPI, with one outlier for CPIII, suggesting that ft values and in vitro inhibition data generally predicted in vivo interaction with CPI/III (Table 6). Interestingly, predicted and observed AUCR of CPI was comparable to AUCR of pitavastatin, a relatively selective OATP1B substrate (Chu et al., 2018). However, CPIII trended toward underpredicting pitavastatin DDIs. CPIII has higher renal and passive hepatic diffusion clearance than CPI. This may explain the poor predictive performance of CPIII on pitavastatin DDIs.
In a clinical DDI study, several sulfated BAs demonstrated differential responses to rifampin (600 mg SD, by mouth) (Tatosian et al., 2021), suggesting that these biomarkers may have different selectivity to OATP1B. In fact, our studies suggest that OATP1B3 is a major contributor for hepatic uptake of GCDCA-S (84%) and TCDCA-S (64%), whereas OATP1B1 (38%) and OATP1B3 (39%) equally contribute to hepatic uptake of GDCA-S. Furthermore, % contribution of passive hepatic diffusion clearance varied among these biomarkers (GCDCA-S, 11%; GDCA-S, 22%; and TCDCA-S, 34%). These data highlight the need to quantitatively determine the selectivity of biomarkers to mechanistically interpret and translate observed data. Among these biomarkers, GCDCA-S is the best investigated. Besides being the substrate of OATP1B1, OATP1B3, NTCP, and MRP2, GCDCA-S undergoes active renal secretion via OAT3 (Tsuruya et al., 2016). Probenecid, a known inhibitor of organic anion transporters, decreased renal clearance of GCDCA-S dose-dependently with only minimal-to-weak increase of plasma AUC of GCDCA-S, which is attributed to inhibition of renal OAT3 (Tsuruya et al., 2016; Willemin et al., 2021). As such, GCDCA-S has also been considered as a sensitive biomarker for OAT3. It is worth noting that probenecid at high doses (1000 mg by mouth) weakly inhibited OATP1B (Zhang et al., 2020). Consequently, monitoring both renal and plasma exposure of GCDCA-S is recommended to deconvolute the inhibition between OATP1B and OAT3.
Unlike CPI/III, quantitative contribution of hepatic and renal clearance to elimination of GCDCA-S in human remains unknown, so we were unable to estimate ft values of respective transporters, despite our data suggesting that OATP1B3, not OATP1B1 and NTCP, is a primary transporter for hepatic uptake of GCDCA-S. Following rifampin administration (600 mg SD, by mouth), AUCR of GCDCA-S in healthy subjects was ∼13-fold in five independent studies, which was much higher than that for CPI/III, and some OATP1B probe substrates (e.g., pitavastatin) measured in the same study (Table 6, Supplemental Tables 3–5). Assuming that rifampin (600 mg SD, by mouth) completely inhibits OATP1B in vivo without affecting NTCP and OAT3 (IC50, NTCP = 48 µM) (Table 5) (IC50, OAT3 >300 µM) (Prueksaritanont et al., 2017), ft, OATP1B1 + 1B3 was 0.92. Based on our RAF data, ft, OATP1B1 and ft, OATP1B3 was estimated to be 0.04 and 0.77, respectively. In contrast, ft, OATP1B1 estimated by a clinical pharmacogenomic study (Supplemental Table 2) was 0.41 (AUCR SLCO1B1*15/15/*1b/*1b was 1.7) (Mori et al., 2019), which was 10-fold higher than the value obtained by rifampin DDI and RAF studies. It is worth noting that the reported pharmacogenomic study was based only on two subjects carrying OATP1B1*15/*15. Additional studies are needed to confirm this finding. Despite that, high AUCR of GCDCA-S following rifampin treatment cannot be fully explained by inhibition of OATP1B1/1B3 (predicted AUCR based on RAF and in vitro inhibition data were only 4.5). The possible explanations for such discrepancy are 1) rifampin inhibition of GCDCA-S is largely attributed to OATP1B3 and in vitro IC50 value of OATP1B3 were higher than in vivo Ki; 2) rifampin inhibited other elimination mechanisms of GCDCA-S (e.g., canalicular efflux by MRP2) (Takashima et al., 2012); and 3) rifampin increased synthesis rate of GCDCA-S from bile acid (Mori et al., 2019).
In this study, we observed a substrate-dependent inhibition of itraconazole to OATP1B biomarkers (Table 5). Consistent with clinical data (Sane et al., 2021), our studies indicated that itraconazole, at clinically relevant exposure, did not inhibit CPI/III uptake by OATP1B1/1B3. However, it showed potent in vitro inhibition to OATP1B1/1B3 by GCDCA-S. Future itraconazole clinical studies with GCDCA-S will confirm this in vitro finding. Nevertheless, caution is required to translate in vitro biomarker inhibition data from one to another.
Characterization of biomarker selectivity is an important element for DDI risk assessment. Currently, the selectivity of OATP1B biomarkers is investigated via 1) in vitro transporter phenotyping studies; 2) clinical DDI studies with potent OATP1B inhibitors (e.g., rifampin); and 3) pharmacogenomic/pharmacokinetic studies in subjects with poor transporter function (e.g., SLCO1B1 c.521C/C or SLCO1B1*15/15). Although these approaches have greatly improved our understanding on biomarker selectivity, they have some limitations. In vitro transporter phenotyping screening can only identify substrate specificity of biomarkers without quantitatively determine their contribution to disposition/elimination of biomarkers. Clinical DDIs with rifampin have provided a useful dataset to identify OATP1B biomarkers, but rifampin is neither a selective inhibitor to OATP1B nor is it able to differentiate the inhibition between OATP1B1 and OATP1B3. Pharmacogenomic studies are powerful tools to identify OATP1B1-specific biomarkers, but these studies require a large sample size and recruitment of the subjects with certain genotypes (e.g., SLCO1B1 C521T>C), which is resource- and time-consuming. Importantly, despite a poor function (SLCO1B1 C/C or *15/*15), OATP1B1 expression/activity is not completely abolished. As such, the contribution of OATP1B1 may be underestimated. In this regard, the RAF method has the advantage of quantitatively determining biomarker selectivity in vitro. Alternatively, relative expression factor method based on quantification of transporter expression in human hepatocytes and transfected cells using LC-MS/MS, is another useful tool to study relative contribution of multiple transporters (Kumar et al., 2021). However, the data could be confounded if transporter protein abundance did not correlate with functional activity. Future studies are needed to directly compare the RAF and relative expression factor approach to study biomarker selectivity.
Although physiologically based pharmacokinetic (PBPK) models are promising tools to predict transporter-mediated DDIs, underprediction of DDIs have been reported when in vitro transporter inhibition data are directly used in PBPK modeling (Taskar et al., 2020). To improve DDI prediction, Yoshikado et al. (2018) has proposed to use biomarkers as a translational tool to bridge in vitro and in vivo biomarker/probe drugs transporter inhibition. They successfully predicted several perpetrator DDIs with statins using CPI biomarker data and PBPK modeling (Yoshikado et al., 2018; Kimoto et al., 2022). However, in vivo Ki obtained from this approach is still an apparent parameter without deconvoluting the effects of perpetrator drugs to different OATP1B isoforms and other transporters. Understanding the contribution of transporters on disposition/elimination of biomarkers may help to accurately determine in vivo Kis and improve DDI prediction. Furthermore, the RAF method developed in this study can be extended to predict relative contribution of multiple hepatic uptake transporters on uptake of substrate drugs and therefore may improve DDI prediction for drugs that their hepatic uptake involves multiple transporters, especially for inhibitors with differential inhibition potency to multiple transporters (e.g., paclitaxel) (Mori et al., 2020a).
It is worth noting that the current RAF method can only determine the relative contribution of hepatic uptake transporters, but not the impact of canalicular (MRP2) and sinusoidal (MRP3/4) efflux transporters, or the net effects of all these transporters on hepatic disposition/elimination of OATP1B biomarkers. Although inhibition of hepatic uptake transporters will predominately impact plasma exposure of OATP1B biomarkers, inhibition of MRP2, to a less extent, may also increase plasma AUC of OATP1B biomarkers (Yoshikado et al., 2018). The roles of MRP2/3/4 on the disposition of these OATP1B biomarkers need further evaluation. Moreover, pharmacokinetic characterization of biomarker elimination clearance is required to extrapolate RAF data to in vivo.
In summary, we have developed a RAF method to quantitatively determine the selectivity of several OATP1B biomarkers. Our studies suggest that CPI is a more selective biomarker for OATP1B1 than CPIII, whereas GCDCA-S and TCDCA-S are more selective to OATP1B3. OATP1B1 and OATP1B3 equally contribute to hepatic uptake of GDCA-S. RAF combined with clinical pharmacogenomic and DDI studies will help to determine the selectivity of transporter biomarkers and provide valuable information on selection of appropriate biomarkers for DDI evaluation.
Acknowledgments
The authors thank Rosa Sanchez and Dan Rock for their support and review of the manuscript.
Authorship Contributions
Participated in research design: Chan, Houle, Zhang, Chu.
Conducted experiments: Chan, Houle, Zhang, Katwaru.
Performed data analysis: Chan, Houle, Zhang, Li, Chu.
Wrote or contributed to the writing of the manuscript: Chan, Chu.
Footnotes
- Received May 30, 2022.
- Accepted May 1, 2023.
All studies conducted in this manuscript were sponsored by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ. This work received no external funding.
No author has an actual or perceived conflict of interest with the contents of this article.
↵
This article has supplemental material available at dmd.aspetjournals.org.
Abbreviations
- AUC
- area under the plasma concentration−time curve
- AUCR
- area under the curve ratio
- CPI
- coproporphyrin I
- CPIII
- coproporphyrin III
- CCK-8
- cholecystokinin
- CsA
- cyclosporine A
- CLint,hepatocytes,
- intrinsic uptake clearance in hepatocytes
- CLint, cells
- intrinsic uptake clearance in transfected cells mediated by the transporter
- DDI
- drug–drug interaction
- EPP
- estropipate
- ES
- [3H]estrone 3-sulfate
- E217βG
- [3H]estradiol-17β-D-glucuronide
- GCDCA-S
- glycochenodeoxycholic acid sulfate
- GDCA-S
- glycodeoxycholic acid sulfate
- HEK293
- human embryonic kidney 293
- ft
- fraction transported
- LC-MS/MS
- liquid chromatography–tandem mass spectrometry
- MRP
- multidrug resistance protein
- NTCP
- sodiumtaurocholate co-transporting polypeptide
- OATP1B
- organic anion transporting polypeptide 1B
- OAT3
- organic anion transporter 3
- PBK
- physiologically based pharmacokinetic
- PBPK
- physiologically based pharmacokinetic
- RAF
- relative activity factor
- R3G
- resveratrol-3-O-glucuronide
- TCA
- taurocholic acid
- TCDCA-S
- taurochenodeoxycholic acid sulfate
- Copyright © 2023 by The American Society for Pharmacology and Experimental Therapeutics
