Abstract
The ability to predict human liver-to-plasma unbound partition coefficient (Kpuu) is of great importance to estimate unbound liver concentration, develop PK/PD relationships, predict efficacy and toxicity in the liver, and model the drug-drug interaction potential for drugs that are asymmetrically distributed into the liver. A novel in vitro method has been developed to predict in vivo Kpuu with good accuracy using cryopreserved suspension hepatocytes in InVitroGRO HI media with 4% BSA. Validation was performed using six OATP substrates with rat in vivo Kpuu data from i.v. infusion studies where a steady state was achieved. Good in vitro-in vivo correlation (IVIVE) was observed as the in vitro Kpuu values were mostly within 2-fold of in vivo Kpuu. Good Kpuu IVIVE in human was also observed with in vivo Kpuu data of dehydropravastatin from positron emission tomography and in vivo Kpuu data from PK/PD modeling for pravastatin and rosuvastatin. Under the specific Kpuu assay conditions, the drug-metabolizing enzymes and influx/efflux transporters appear to function at physiologic levels. No scaling factors are necessary to predict in vivo Kpuu from in vitro data. The novel in vitro Kpuu method provides a useful tool in drug discovery to project in vivo Kpuu.
Introduction
Liver is an important organ for many disease targets, such as dyslipidemia, diabetes, obesity, and nonalcoholic steatohepatitis. It is critical to understand the unbound drug concentration in the liver, because it impacts pharmacological activity, metabolic and biliary clearance, and DDI (Smith et al., 2010). For compounds that are not actively transported and are not influenced by membrane potential or pH gradient (Scott et al., 2016), the unbound drug concentration in the liver will be the same as that in the plasma. In this case, the unbound partition coefficient (Kpuu) between liver and plasma is close to 1. When compounds are uptake transporter substrates (e.g., OATPs), Kpuu values can be greater than 1 due to active influx. Kpuu represents the distribution of unbound drugs between liver and plasma in vivo or between hepatocytes and media in vitro when multiple processes, including metabolism, uptake, efflux, and passive diffusion, have achieved a steady state. Kpuu can be described using the extended clearance equation incorporating the multiple mechanisms (Shitara et al., 2006; Watanabe et al., 2010; Yabe et al., 2011). It is important to be able to estimate in vivo Kpuu, since it is the link between unbound plasma concentration and unbound liver concentration. Because it is challenging to measure the unbound liver concentration directly in higher species (e.g., nonhuman primate) and humans, the ability to predict Kpuu will enable direct estimation of unbound liver concentrations from unbound plasma concentrations.
Currently, several in vitro methods (Riccardi et al., 2016) are available to estimate Kpuu, including the binding method (Mateus et al., 2013), the kinetic method (Yabe et al., 2011), and the temperature method (Shitara et al., 2013). However, validation of these methods with in vivo exposure/pharmacology or in vitro activity data are fairly limited (Shitara et al., 2013; Riccardi et al., 2016). IVIVE for Kpuu using hepatocyte systems for OATP substrates has not been established. Several studies have shown internalization or downregulation of transporters in the hepatocyte systems (Roelofsen et al., 1995; Bow et al., 2008; Kimoto et al., 2012; Kunze et al., 2014; Morse et al., 2015; Vildhede et al., 2015), although others have shown no significant difference in transporter abundance between hepatocytes and liver tissues (Prasad et al., 2014; Badee et al., 2015). It is uncertain whether a direct translation is possible without scaling factors from in vitro hepatocytes to in vivo Kpuu. In this study, we explored the IVIVE of Kpuu using cryopreserved suspension rat and human hepatocytes for OATP substrates. The ability to predict in vivo liver-to-plasma Kpuu from in vitro systems will provide a useful tool in drug discovery to predict unbound liver concentration as well as clearance and dose, to design drugs for liver targeting, to develop pharmacokinetic (PK)/pharmacodynamics (PD) relationships for disease targets residing in the liver, and model DDI due to inhibition/induction of liver enzymes when transporters are involved in the distribution processes.
Materials and Methods
Materials.
Test compounds were obtained from obtained from Sigma-Aldrich (St. Louis, MO) or Pfizer (Groton, CT). PF-04991532 (Compound 19 in the reference) (Pfefferkorn et al., 2012) and PF-05187965 (Compound 7 in the reference) (Stevens et al., 2013) were synthesized according to the methods reported in the referenced publications. Rat plasma (from 14 males and 14 females, pooled) and human plasma (from 6 males, pooled), cryopreserved human hepatocytes (Lot DCM, custom-pooled of both male and female, 10 donors), and Wistar-Han rat hepatocytes (Lot VSU, 35 male pooled donors) were purchased from BioreclamationIVT, LLC (Hicksville, NY). Human liver (1 male donor) was from Analytical Biologic Services Inc. (Wilmington, DE). Wistar-Han rat liver (four male donors) was obtained internally at Pfizer Global Research and Development (Groton, CT). Williams’ Medium E (catalog #C1984, custom formula number 91-5233EC; ThermoFisher Scientific, Waltham, MA) contained 26 mM sodium bicarbonate and 50 mM HEPES, InVitroGRO HI media, and M-PER buffer were purchased from ThermoFisher Scientific. BSA (free of fatty acid; catalog #A4612) and other reagents were from Sigma-Aldrich unless specified. The equilibrium dialysis device (96-well format) and cellulose membranes (molecular weight cutoff, 12,000–14,000) were obtained from HTDialysis, LLC (Gales Ferry, CT). Breathe Easy sealing membranes were obtained from Sigma-Aldrich.
Determination of Fraction Unbound.
In preparation of in vitro fraction unbound (fu) measurement, rat and human liver tissues were homogenized in PBS (1:5 tissue: PBS dilution) at RT with an Omni TH tissue homogenizer (Omni International, Kennesaw, GA). A probe (7 × 110 mm) was used for 30-second pulses at high speed. InVitroGRO HI media containing 4% BSA and plasma were used directly without any dilution for binding determination. Before an experiment, the dialysis membranes were soaked in water for 15 minutes, 30% ethanol/water for 15 minutes, and PBS for 15 minutes or overnight. The dialysis device was put together following the instructions from the manufacturer (http://www.htdialysis.com). Compound stock solutions were prepared at 200 μM in dimethylsulfoxide (DMSO), added to matrices (1:100 dilution), and mixed well with a multichannel pipettor (Eppendorf; VWR, Radnor, PA). The final test compound concentration for equilibrium dialysis was 2 μM with 1% DMSO. An aliquot (150 μL) of matrix (plasma, liver homogenate, or assay media) containing 2 μM test compound was added to the donor side of the membrane, and PBS (150 μL) was added to the receiver side of the membrane. The dialysis device was sealed with Breathe-Easy Membranes (Sigma-Aldrich). Quadruplicates were used for each compound in binding experiments. The dialysis device was incubated in a humidified incubator (75% relative humidity, 5% CO2/95% air) at 37°C for 6 hours at 200 rpm with an orbital shaker (VWR). Alternative binding methods (before saturation or dilution) were also used for highly bound compounds to ensure that equilibrium had been achieved (Riccardi et al., 2015). After the incubation was completed, matrix samples (15 μL) from the donor wells were added to 45 μL of PBS in a 96-well plate. Dialyzed PBS (45 μL) from the receiver wells was added to the blank matrix (15 μL). Matrix material (15 μL) containing 2 μM compound from both before and after incubation was taken and added to 45 μL PBS in a 96-well plate. They were used to assess the recovery and stability of the samples. Cold acetonitrile (200 μL) containing internal standard (IS; a cocktail of 5 ng/ml terfenadine and 0.5 ng/ml tolbutamide) was added to all the samples for protein precipitation. The samples were vortexed for 3 minutes (VWR) and centrifuged at 3000 rpm for 5 minutes (Allegra 6R; Beckman Coulter, Fullerton, CA) at RT. The supernatant was transferred, dried down, reconstituted with solvents, and analyzed by LC-MS/MS. Sertraline was used on every incubation plate for quality control. Calculation of fu has been discussed previously (Riccardi et al., 2015, 2016).
In Vitro Kpuu Measurement.
The cryopreserved hepatocytes were thawed and resuspended in William’s Medium E. The number of cells and viability were determined using the Trypan Blue exclusion method. Cell suspensions were centrifuged (Allegra 6R; Beckman Coulter) at 50g at RT for 3 minutes. Media were removed and cells were resuspended in InVitroGRO HI media supplemented with 4% BSA. Test compounds (1 mM) were prepared in DMSO, and 1 µl was added to the suspended hepatocytes at 0.5 million cells/ml in 1 ml. The final compound concentration is 1 µM with 0.1% DMSO. Two to four replicates were used for each compound. The suspension was incubated at 37°C in a humidified incubator (75% relative humidity, 5% CO2/95% air) with for 4 hours to ensure that the steady state was reached. At the end of the incubation, the cell suspension was centrifuged for 3 minutes at 500 rpm and supernatant was sampled to determine the medium concentration. The remaining medium was removed from the hepatocytes. The cells were washed with cold PBS three times (1 ml each time) and lysed with 75 μl M-PER buffer. The solution of the lysed cells was sampled for analysis. Cold acetonitrile with IS was added to both the supernatant and the cell-lysed solutions and mixed. The solution was centrifuged at 3000 rpm for 5 minutes at RT, and the supernatants were transferred for LC-MS/MS analysis using standard curves from both media and cells. Calculation of in vitro Kpuu has been discussed previously (Riccardi et al., 2016). Here, fu,cell is replaced with fu,liver since the two values are quite comparable, as liver comprised of 80% hepatocytes by volume (Bayliss and Skett, 1996).
In Vivo Rat Liver-to-Plasma Kpuu Determination.
The i.v. infusion experiments in rats were conducted at BioDuro (Shanghai, People’s Republic of China). Wistar-Han rats (male, n = 3, fed) were infused i.v. through a jugular vein cannula at a rate of 4–9 μl/min with test compounds using a programmable pump (Harvard 2000; Harvard Apparatus, Holliston, MA). The doses were selected based on the i.v. bolus data and the detection limit. Infusion time was determined using a duration greater than five times the terminal half-life that was predetermined from i.v. bolus data. Under this condition, Kpuu should be close to steady state, because a greater than 97% steady state is achieved at five times the half-life (Ito, 2011; Hedaya, 2012). Dose and formulation of the compounds are summarized in Table 1. At the end of infusion, blood samples were obtained from the carotid artery catheter, and livers were also collected. Liver samples were rinsed with saline and patted dry with a paper towel, and all the blood vessels attached were also removed to minimize the potential contamination from blood and bile. Since the total volume of the biliary tree is quite small compared with the liver for both rat (0.5%) and human (0.3%) (Casali et al., 1994; Masyuk et al., 2001), the impact of bile contamination on liver concentration is likely to be small. Concentrations were determined using LC-MS/MS. Both pravastatin and its isomer (van Haandel et al., 2016), 3′α-hydroxy-pravastatin, were included in the calculation of Kpuu. Free concentrations were calculated by multiplying the total plasma or total liver concentration by the fu values of plasma or liver. The in vivo unbound liver-to-plasma ratio, Kpuu, was obtained by dividing the steady-state unbound liver concentration by the unbound plasma concentration.
Rat i.v. infusion experimental conditions
LC-MS/MS Quantification.
A generic LC-MS/MS method is discussed here, and equivalent methods were also used based on compound properties. Two LC mobile phases were used: (A) 95% 2 mM ammonium acetate in water and 5% 50/50 methanol/acetonitrile and (B) 90% 50/50 methanol/acetonitrile and 10% 2 mM ammonium acetate in water, or (A) water with 0.1% formic acid and (B) acetonitrile with 0.1% formic acid. A flow rate of 0.5 ml/min was used with solvent gradient from 5% (B) to 95% (B) over 1.1 minutes to elute the compounds from the ultra-performance LC column (BEH C18, 1.7 μm, 50 × 2.1 mm; Waters, Milford, MA). The injection volume was 10 μl, and the cycle time was 2.5 minutes/injection. A CTC PAL autosampler (LEAP Technologies, Carrboro, NC), an model 1290 binary pump (Agilent, Santa Clara, CA) and an API 6500 triple quadrupole mass spectrometer with a TurboIonSpray source (AB Sciex, Foster City, CA) in MRM mode were used for sample analysis. Data collection processing and analysis were conducted with Analyst 1.6.1 software (Applied Biosystems, Foster City, CA).
Human in Vivo Kpuu Estimation Based on PK/PD Modeling.
The in vivo human Kpuu values for rosuvastatin and pravastatin were calculated as the ratio between in vitro and in vivo IC50 values for 3-hydroxy-3-methylglutaryl-CoA reductase inhibition. The in vitro IC50 values were obtained from the literature (McTaggart et al., 2001; Holdgate et al., 2003; Gazzerro et al., 2012). The in vivo IC50 values were estimated using PK/PD modeling (Supplemental eq. 1). The rosuvastatin model published previously has adapted to estimate the in vivo IC50 in humans (Aoyama et al., 2010). This is a two compartment PK model developed with an indirect PD response incorporating circadian rhythm of mevalonic acid (MVA) production (Aoyama et al., 2010). Pravastatin plasma PK (Pan et al., 1990; van Luin et al., 2010) and plasma MVA concentrations in response to pravastatin (Nozaki et al., 1996) were used in the PK/PD modeling with the same method as rosuvastatin. All modeling was performed in NONMEM 7.2 (ICON Plc, Dublin, Ireland).
Calculation Methods.
fu was calculated with eq. 1 based on compound concentrations or area ratios between the test compound and the IS. For samples with diluted matrices, fu was obtained with eqs. 2 and 3, where D is the dilution factor. The calculations of recovery and stability are shown in eqs. 4 and 5, respectively. In vitro Kpuu, unbound cell concentration, and unbound medium concentration were obtained using eqs. 6–8.
Eq. (1)
Eq. (2)
Eq. (3)
Eq. (4)
Eq. (5)
Eq. (6)
Eq. (7)
Eq. (8)Results
Development of IVIVE requires high-quality in vivo Kpuu data to verify the in vitro results. As in vivo rat Kpuu data can be obtained relatively easily by using i.v. infusion to ensure steady state has been achieved, Kpuu IVIVE was first developed using the rat animal model. The method was then extended to humans, where quality in vivo Kpuu data are quite limited. Four statins (cerivastatin, fluvastatin, rosuvastatin, and pravastatin) and two Pfizer internal liver-targeting compounds [PF-04991532 (Pfefferkorn et al., 2012) and PF-05187965 (Stevens et al., 2013)] were used for the study of rat IVIVE. All six compounds are OATP substrates consisting of a carboxylic acid functional group. Cerivastatin and pravastatin are OATP1B1 substrates; rosuvastatin and fluvastatin are substrates of OATP1B1, OATP1B3, and OATP2B1; and the two Pfizer compounds (PF-04991532 and PF-05187965) are substrates of OATP1B1 and OATP1B3 (University of Washington Drug-Drug Interaction Database, UCSF-FDA TransPortal) (Pfefferkorn et al., 2012; Stevens et al., 2013). The plasma and liver concentrations, measured at steady state, from in vivo rat i.v. infusion studies are shown in Table 2. Kpuu was calculated based on in vivo liver-to-plasma Kp and in vitro fu,p and fu,liver from the equilibrium dialysis assay. The in vivo Kpuu values of the compounds range from 2.2 to 57, covering a wide range of liver distribution properties. The in vitro Kpuu data using cryopreserved suspension rat hepatocytes are summarized in Table 3. In vitro Kpuu was calculated by multiplying in vitro Kp with fu,cell and dividing by fu,media. Since fu,cell is similar to fu,liver, as 80% of liver cells are hepatocytes (Bayliss and Skett, 1996), rat fu,liver data from Table 2 were used as fu,cell for the in vitro Kpuu calculation. From this and previous studies (Riccardi et al., 2016), a 4-hour incubation time is a conservative time point to achieve a steady state in cells for most compounds. For example, the Kpuu values for cerivastatin in rat hepatocytes were 39 ± 2.0, 34 ± 3.2, 24 ± 1.4, 31 ± 0.42, and 23 ± 2.3 at 1, 2, 3, 4 and 5 hours with 0.1 µM incubation, suggesting that a steady state has been achieved even after 1 hour of incubation. The comparison between in vitro Kpuu from suspension rat hepatocytes and in vivo rat Kpuu is shown in Table 4. A good correlation between in vitro and in vivo has been observed, and the Kpuu values are mostly within 2-fold of each other. The lower Kpuu value of pravastatin compared with the other three statins could potentially be due to higher biliary clearance, higher basolateral efflux, and lower active uptake. It appears that the drug-metabolizing enzymes and influx/efflux transporters are functioning at physiologic levels in the rat hepatocytes in suspension under the specific assay conditions. No scaling factors are necessary to predict in vivo Kpuu data from in vitro data.
In vivo rat i.v. infusion study for determination of liver-to-plasma Kpuu
In vitro Kp and Kpuu between cells and media in rat suspension hepatocytes
Correlation between in vivo rat liver-to-plasma Kpuu and in vitro suspension rat hepatocyte Kpuu
High-quality in vivo Kpuu data from humans are very scarce. Liver biopsy is invasive and positron emission tomography (PET) imaging has a number of limitations (e.g., interference from metabolites, nonspecific binding to tissues at low doses). Nevertheless, a few in vivo human Kpuu values are available to evaluate the in vitro Kpuu method. Human liver PET data has been reported for [11C]dehydropravastatin (DHP) (Shingaki et al., 2014). The Kpuu was estimated to be 2.0 using terminal-phase (after 15 minutes) DHP data and in-house–measured pravastatin human fu,p (0.64) and fu,liver (0.17) values. It has been shown that the transporter and dispositional properties of DHP and pravastatin are very similar using an in vitro sandwich-cultured human hepatocyte assay and in vivo rat studies (Y. Sugiyama, personal communication). Therefore, the Kpuu value of DHP can be used as a surrogate for the Kpuu value of pravastatin.
Because high-quality directly measured human in vivo Kpuu data are fairly limited, human Kpuu values of rosuvastatin and pravastatin were obtained indirectly using PK/PD modeling. Other statins were not included in the modeling due to the interference of active metabolites or insufficient literature data. The human in vivo liver-to-plasma Kpuu of rosuvastatin was estimated to be 10. The average value reported in the literature for in vitro IC50 was 7 nM, ranging from 2 to 10 nM (McTaggart et al., 2001; Holdgate et al., 2003; Gazzerro et al., 2012). The in vivo IC50 of rosuvastatin was fitted to be 2.1 ng/ml with a 95 confidence interval (CI) of 1.87–2.32 ng/ml. The corresponding unbound IC50 was 0.7 nM (95% CI, 0.62–0.78 nM) using an in-house–measured fu,p of 0.16. The PK/PD model fits the data well (Supplemental Fig. 1), and the parameter estimates agree with those published previously (Supplemental Table 1) (Aoyama et al., 2010). The human in vivo liver-to-plasma Kpuu of pravastatin was estimated to be 5.3. The average of the reported in vitro IC50 values was 48 nM ranging from 29 to 70 nM (Gazzerro et al., 2012). The in vivo IC50 of pravastatin was modeled to be 6.0 ng/ml with a 95% CI of 3.6–12.7 ng/ml, converting to an unbound in vivo IC50 of 9.0 nM (95% CI, 5.4–19.3 nM) based on an in-house–measured fu,p of 0.64. The modeling results are shown in Supplemental Table 2 and Supplemental Figure 2.
The comparison of human in vivo Kpuu from PET or PK/PD modeling and in vitro Kpuu from human hepatocytes in suspension is shown in Table 5. The in vitro assay predicted in vivo DHP Kpuu values from PET data well (2.3 versus 2.0). The Kpuu data from PK/PD modeling carry some uncertainties because it is an indirect in vivo measure of Kpuu from many in vitro and in vivo parameters. Certain assumptions were made to obtain the Kpuu values from PK/PD modeling. It is assumed that the in vitro assays fully capture the in vivo conditions and that the measured in vitro IC50 value can be directly used to quantitatively explain the PD (e.g., MVA) response to unbound liver drug concentrations. In addition, there are only eight and five time point measurements of plasma concentrations of MVA, respectively, in each group of the rosuvastatin and pravastatin studies, and only one PD study of each statin was available to be included in the analysis. Giving the uncertainty of the in vivo Kpuu values from PK/PD modeling, it is reasonable that the in vitro Kpuu data are within 2-fold and 4-fold of the in vivo Kpuu estimates from PK/PD modeling. The human Kpuu IVIVE is similar to that observed in the rat based on limited data, suggesting that the in vitro Kpuu method with suspension hepatocytes is a suitable in vitro tool to predict in vivo Kpuu under specific assay conditions. As more human in vivo Kpuu data become available, the performance of the in vitro Kpuu method will continue to be verified.
Correlation between in vitro and in vivo Kpuu for humans
Discussion
Determination of the unbound liver concentration is critical to understand the pharmacology of disease targets in the liver, develop PK/PD relationships, predict DDI potentials, and anticipate liver toxicity. The ability to predict in vivo liver-to-plasma Kpuu values from in vitro assays is highly desirable as there is no easy way to measure the human unbound liver concentration. With Kpuu values, unbound liver concentrations can be estimated from unbound plasma concentrations, which can be readily measured. As reported in the literature, a direct translation of in vitro Kpuu to in vivo Kpuu can be challenging because transporter protein levels and functions in the in vitro systems are quite different than those in the in vivo systems (Roelofsen et al., 1995; Bow et al., 2008; Kimoto et al., 2012; Kunze et al., 2014; Morse et al., 2015; Vildhede et al., 2015). Typically, empirical scaling factors are needed to predict in vivo drug disposition from in vitro data (Jones et al., 2012; Li et al., 2014). The scaling factors are system dependent and can vary with assay conditions, such as cell culture time/media, plated versus suspension cells, BSA versus no proteins, cell types (transfected cells versus hepatocytes), and medium composition. However, under this specific assay condition with cryopreserved hepatocyte suspension in InVitroGRO HI media with 4% BSA, the transporters and enzymes appear to be functioning at the physiologic level. Good Kpuu IVIVE has been observed in both rat and human without any scaling factors. Both the specific assay media and the physiologic amount of BSA (4%) in the assay are important to generate good Kpuu IVIVE. This is the first time that an in vitro assay shows good prediction of in vivo Kpuu for OATP substrates. Based on the extended clearance concept, Kpuu is affected by intrinsic clearance of passive diffusion, active hepatic uptake, sinusoidal efflux, biliary excretion, as well as metabolism. Based on some literature data, cryopreserved suspended hepatocytes did not retain proper functional activity of efflux transporters. This was likely due to the internalization or downregulation of some transporters (Bow et al., 2008). However, the information is controversial in the literature, as efflux activity has been reported in suspension hepatocytes of multiple species including rat, human, dog, and monkey (Li et al., 2008). The mechanistic understanding of why this particular assay condition seems to perform better others requires further investigation.
The in vivo Kpuu measures the unbound drug concentration between liver and arterial blood rather than liver blood. The liver blood has lower drug concentrations than arterial blood for high-extraction drugs. Theoretically, for compounds with high liver extraction, the in vivo measured Kpuu value will be lower than the in vitro experimental Kpuu value. However, in practice, these differences have not been observed. The in vitro Kpuu values are shown to be both higher and lower than the in vivo Kpuu values. This could potentially be due to experimental variability from both in vitro and in vivo assays, making it difficult to detect the differences. This theoretical difference between in vivo and in vitro Kpuu values appears to be inconsequential for Kpuu IVIVE.
In this study, we observed good correlation between in vivo Kpuu values from PK/PD modeling and in vitro Kpuu. The in vivo Kpuu was calculated as the IC50 ratio of 3-hydroxy-3-methylglutaryl-CoA inhibition between in vitro and in vivo values. The in vitro IC50 was measured using human liver microsomes. Assuming that the in vitro IC50 fully captures in vivo conditions (e.g., enzyme activity, substrate concentration, pH, and temperature), has no impact of transporters, and represents the intrinsic inhibitory activity, this measured in vitro IC50 should explain the in vivo effect directly if the intracellular unbound drug concentration in the liver is known. However, in the absence of liver drug concentration, the human in vivo IC50 was estimated using PK/PD modeling based on unbound plasma concentration. The unbound statin concentration in the liver can be higher than unbound plasma concentrations due to OATP active uptake. Therefore, the IC50 ratio between in vitro and in vivo values reflects the unbound drug concentration difference between liver and plasma and can be used as a surrogate for Kpuu. This novel in vitro Kpuu method provides a new tool to assess in vivo Kpuu in drug discovery. The information is useful to estimate human unbound liver drug concentrations, predict efficacy, and model DDI risks for drugs that have active influx/efflux in the liver by transporters.
Acknowledgments
We thank Rui Li and Hugh Barton for useful discussion, and Larry Tremaine and Tess Wilson for their leadership and support.
Authorship Contributions
Participated in research design: Riccardi, Lin, Li, Niosi, Ryu, Hua, Atkinson, Kosa, Litchfield, Di.
Conducted experiments: Riccardi, Ryu, Hua.
Performed data analysis: Riccardi, Lin, Li, Niosi, Ryu, Hua, Atkinson, Kosa, Litchfield, Di.
Wrote or contributed to the writing of the manuscript: Li, Atkinson, Kosa, Di.
Footnotes
- Received December 8, 2016.
- Accepted March 1, 2017.
↵
This article has supplemental material available at dmd.aspetjournals.org.
Abbreviations
- BSA
- bovine serum albumin
- CI
- confidence interval
- DDI
- drug-drug interaction
- DHP
- dehydropravastatin
- DMSO
- dimethylsulfoxide
- fu
- fraction unbound
- fu,cell
- fraction unbound of cells
- fu,liver
- fraction unbound of liver
- fu,media
- fraction unbound of media
- fu,p
- fraction unbound of plasma
- IS
- internal standard
- IVIVE
- In vitro-in vivo extrapolation/correlation
- Kp
- partition coefficient
- Kpuu,
- unbound partition coefficient
- LC-MS/MS
- liquid chromatography-tandem mass spectrometry
- MVA
- mevalonic acid
- OATP
- organic anion-transporting polypeptide
- PBS
- phosphate-buffered saline
- PD
- pharmacodynamics
- PET
- positron emission tomography
- PK
- pharmacokinetics
- RT
- room temperature
- Copyright © 2017 by The American Society for Pharmacology and Experimental Therapeutics
