Abstract
The current studies assessed the utility of freshly plated hepatocytes, cryopreserved plated hepatocytes, and cryopreserved plated HepaRG cells for the estimation of inactivation parameters kinact and KI for CYP3A. This was achieved using a subset of CYP3A time-dependent inhibitors (fluoxetine, verapamil, clarithromycin, troleandomycin, and mibefradil) representing a range of potencies. The estimated kinact and KI values for each time-dependent inhibitor were compared with those obtained using human liver microsomes and used to estimate the magnitude of clinical pharmacokinetic drug-drug interaction (DDI). The inactivation kinetic parameter, kinact, was most consistent across systems tested for clarithromycin, verapamil, and troleandomycin, with a high kinact of 0.91 min−1 observed for mibefradil in HepaRG cells. The apparent KI estimates derived from the various systems displayed a range of variability from 3-fold for clarithromycin (5.4–17.7 μM) to 6-fold for verapamil (1.9–12.6 μM). In general, the inactivation kinetic parameters derived from the cell systems tested fairly replicated what was observed in time-dependent inhibition studies using human liver microsomes. Despite some of the observed differences in inactivation kinetic parameters, the estimated DDIs derived from each of the tested systems generally agreed with the clinically reported DDI within approximately 2-fold. In addition, a plated cell approach offered the ability to conduct longer primary incubations (greater than 30 min), which afforded improved ability to identify the weak time-dependent inhibitor fluoxetine. Overall, results from these studies suggest that in vitro inactivation parameters generated from plated cell systems may be a practical approach for identifying time-dependent inhibitors and for estimating the magnitude of clinical DDIs.
Introduction
Time-dependent inhibition (TDI) of cytochrome P450 (P450) enzymes is a major concern due to the potential for clinically significant drug-drug interactions (DDI) (Lin and Lu, 1998). Therefore, from a drug discovery perspective, it is not only critical to qualitatively identify but also to quantitatively estimate the magnitude of any potential interaction for new chemical entities before preclinical development (Mayhew et al., 2000).
In general, kinetic parameters of enzyme inactivation (kinact and KI) are measured in a two-staged assay with multiple concentrations and time points using human liver microsomes (HLM) (Jones et al., 1999; Grimm et al., 2009). The process of using in vitro inactivation data to estimate and rank order the DDI potential of compounds is well established (Mayhew et al., 2000; Obach et al., 2006, 2007), and the in vitro inactivation kinetic parameters along with the in vivo systemic and intestinal concentration of the time-dependent inhibitor, fraction metabolized, and intestinal extraction ratio of the victim probe drug, and enzyme degradation rates (Correia, 1991; Greenblatt et al., 2003; Ghanbari et al., 2006) are used to estimate the potential for a clinical DDI (Obach et al., 2006, 2007). However, precipitants of TDI are not always products of oxidative metabolism (Ogilvie et al., 2006; Baer et al., 2009; Xu et al., 2009; Honkalammi et al., 2011), and therefore it may be more relevant to assess TDI in a system containing the full complement of drug-metabolizing enzymes, such as hepatocytes (Li, 1997; Zhao et al., 2005; Zhao, 2008; Li and Doshi, 2011).
Many investigators have evaluated the use of hepatocyte suspensions (Zhao et al., 2005, 2007; McGinnity et al., 2006; Van et al., 2007; Lu et al., 2008a,b; Baer et al., 2009; Mao et al., 2011) for time-dependent inhibition studies. The use of hepatocytes for time-dependent inhibition holds distinct advantages over the use of microsomes, including being more physiologically relevant and offering the ability to conduct longer incubations. However, there are some potential difficulties associated with time-dependent inhibition studies using hepatocyte suspensions, including removal of time-dependent inhibitor while maintaining cell viability (Zhao, 2008).
The objectives of the current study were to investigate fresh plated primary hepatocytes, cryopreserved plated hepatocytes, and cryopreserved HepaRG-plated cell systems for their ability to generate in vitro-inactivation parameters comparable with those generated using HLM. The kinact and KI values obtained in the plated cell systems and human liver microsomes were used to predict clinical pharmacokinetic DDIs and were then compared with the observed clinical drug interaction. The plated methods used to conduct the time-dependent inhibition experiments overcome many of the limitations of a hepatocyte suspension assay. The advantages, disadvantages, and ability to quantitatively predict clinical pharmacokinetic DDIs for each of the tested systems, compared with the commonly used TDI assay in HLM, are discussed.
Materials and Methods
Chemicals.
Clarithromycin, mibefradil, verapamil, troleandomycin, and fluoxetine (Fig. 1), acetonitrile with 0.1% formic acid, water with 0.1% formic acid, 1.0 M monobasic potassium phosphate solution, and 1.0 M dibasic potassium phosphate solution were purchased from Sigma-Aldrich (St. Louis, MO). Midazolam and α-hydroxymidazolam-d4 were obtained from Cerilliant Corporation (Round Rock, TX). NADPH was obtained from EMD (Gibbstown, NJ). Primary plated human hepatocytes isolated from a 55-year-old male (donor H1014) were obtained in a 24-well format from XenoTech, LLC (Lenexa, KS). The following were acquired from Invitrogen (Carlsbad, CA): plateable, inducible cryopreserved human hepatocytes (lot 4205) isolated from a 42-year-old white male; HepaRG cells; Williams' medium E; collagen-coated, 24-well plates; Geltrex; cryopreserved hepatocyte recovery medium (CHRM); primary hepatocyte thawing, plating, and maintenance supplements; GlutaMAX; and HepaRG thawing, plating, and general purpose medium supplement. Human liver microsomes (BD Ultra pool 150) were purchased from BD Biosciences (Bedford, MA). All other reagents were of the highest purity available.
Chemical structures of CYP3A time-dependent inhibitors.
TDI in Human Liver Microsomes.
Incubations were performed in an automated fashion using a Tecan Evo (Durham, NC) in polypropylene tubes containing 100 mM phosphate buffer (pH 7.4), 0.1 mg/ml human liver microsomes, and 1 mM NADPH at 37°C (primary incubation). Final test compound concentrations were as follows: 0 to 50 μM fluoxetine and clarithromycin, 0 to 7.5 μM mibefradil, 0 to 5 μM troleandomycin, and 0 to 25 μM verapamil. After initiation of the primary incubation by the addition of NADPH, samples (10 μl) were taken at 0, 1, 3, 5, 15, and 30 min and added to a secondary incubation consisting of 190 μl of phosphate buffer containing 15 μM midazolam (saturating concentration) and 1 mM NADPH. The secondary reactions were quenched after 8 min using one volume of ice-cold acetonitrile containing α-hydroxymidazolam-d4 (100 nM) and centrifuged at 3000g for 10 min. Supernatants were transferred to a 96-well plate and analyzed by liquid chromatography tandem mass spectrometry (LC/MS/MS).
Cell Culture.
Upon receipt of plated primary human hepatocytes, cells were washed and incubated for 48 h with daily media changes using hepatocyte maintenance media. Plateable inducible cryopreserved human hepatocytes were thawed for 2 min at 37°C and resuspended in prewarmed CHRM media. Resuspended cells were centrifuged at 100g for 10 min. Supernatant was removed, and hepatocyte maintenance media was added to achieve a final cell density of 8.0 × 105 viable cells/ml. Cell viability was assessed by trypan blue exclusion and was greater than 70%. Collagen-coated, 24-well plates were seeded at a density of 4.0 × 105 cells/well. Plates were incubated for 6 h in a humidified incubator at 37°C and 5% CO2. After incubation, media was replaced with fresh ice-cold hepatocyte maintenance media containing 0.35 mg/ml Geltrex and incubated overnight.
HepaRG cells were thawed for 2 min at 37°C and resuspended in 50 ml of prewarmed media consisting of Williams' medium E with 1% GlutaMAX and HepaRG thawing supplement. Resuspended cells were centrifuged at 360g for 4 min. Supernatant was removed, and HepaRG plating media consisting of Williams' medium E with 1% GlutaMAX and HepaRG plating supplement was added to achieve a final cell density of 6.0 × 105 viable cells/ml. Cell viability was assessed by trypan blue exclusion and was greater than 70%. Collagen-coated, 24-well plates were seeded at a density of 3.0 × 105 cells/well and cultured in a humidified incubator at 37°C and 5% CO2 for 4 h.
TDI in Plated Cells.
Incubations were performed in a humidified incubator at 37°C and 5% CO2. Williams' medium E was removed from each collagen-coated, 24-well plate by vacuum aspiration and replaced with 0.5 ml of Williams' medium E containing time-dependent inhibitor to initiate the primary incubation. Williams' medium E containing time-dependent inhibitor was removed by vacuum aspiration at 0, 5, 15, and 30 min (0, 15, 45, and 90 min for fluoxetine). Cells were washed with 0.5 ml of Williams' medium E and removed using vacuum aspiration before the addition of 0.5 ml of Williams' medium E containing 15 μM midazolam to initiate the secondary incubation. Incubate (100 μl) was removed after an 8-min secondary incubation, and the reaction was immediately terminated by flash freezing in a dry ice/methanol bath. After flash freezing the samples, 0.3 ml of ice-cold acetonitrile containing α-hydroxymidazolam-d4 (100 nM) was added. Samples were centrifuged at 3000g for 10 min. Supernatant was transferred to a 96-well plate and analyzed by LC/MS/MS.
LC/MS/MS Analysis.
Quantitation of 1′-hydroxy midazolam (MRM 342 > 203) was performed using an AB Sciex (Foster City, CA) API-5000 triple quadrupole mass spectrometer equipped with an electrospray ionization interface in positive ion mode and connected in-line to a Waters (Milford, MA) ACQUITY UPLC system. Separation of 1′-hydroxy midazolam was performed using a Waters ACQUITY BEH C8 1.7 μm (2.1 × 50 mm) column maintained at 50°C. The mobile phase flow rate was 0.5 ml/min. Initial conditions were 55% A (0.1% formic acid in water) and 45% B (0.1% formic acid in acetonitrile). Initial conditions were held for 0.5 min, followed by a linear gradient to 90% B over 0.05 min (hold for 0.25 min), and then re-equilibration to initial conditions with a total analytical run time of 1.0 min.
Data Analysis.
Peak areas for 1′hydroxy midazolam and α-hydroxymidazolam-d4 were determined using Analyst software (version 1.4.2; AB Sciex). Analyte/internal standard peak area ratios were used to calculate the percentage of activity remaining based on comparison with time 0 samples. The apparent inactivation rate (kobs) for each time-dependent inhibitor concentration was determined from plots of the ln percentage activity remaining and primary incubation time. The rate equations for enzyme inactivation were determined using eq. 1:
where kobs represents the observed rate of inactivation, kinact (min−1) is the maximum rate of inactivation, [I] represents the time-dependent inhibitor concentration, and KI is the concentration of time-dependent inhibitor that achieves one-half the maximum rate of inactivation. Kinetic parameters of CYP3A inactivation (kinact and KI) were determined by nonlinear regression of kobs and nominal time-dependent inhibitor concentration using GraFit software (version 7.0.2; Erithacus Software, Horley, Surrey, UK). The reported apparent KI values are estimates because the intracellular concentrations for each drug were not determined.
Clinical DDI predictions were determined for each of the time-dependent inhibitors using the input parameters from Table 1 and the corresponding in vitro inactivation kinetic parameters. Area under the curve (AUC) ratio (fold change in midazolam AUC) was estimated using eq. 2 (Obach et al., 2006):
where kinact is the maximal rate of inactivation and KI is the concentration of drug that yields ½-maximal rate of inactivation, fm(CYP3A) is the fraction metabolized of selected substrate by CYP3A (midazolam, 0.93), Fg is the contribution of intestinal extraction (midazolam, 0.57), [I]in vivo is the free systemic concentration in vivo, and kdeg is the first-order degradation rate constant of CYP3A hepatic (0.00032 min−1) and CYP3A gut (0.00048 min−1) (Greenblatt et al., 2003; Ghanbari et al., 2006).
Summary of published data used for the prediction of clinical drug-drug interaction for each of the tested CYP3A time-dependent inhibitors
The intestinal concentration of each of the time-dependent inhibitors was estimated using eq. 3 (Obach et al., 2006):
where D is the dose administered, ka is the absorption rate constant (0.03 min−1), Fa is the fraction of the time-dependent inhibitor absorbed through the gut unchanged, and Qg is human intestinal blood flow (248 ml/min). For a top limit estimate, each compound was assumed to be completely absorbed (Fa = 1). Human plasma protein binding values were used for estimating free systemic Cmax (Obach et al., 2006).
Results
Estimation of In Vitro Inactivation Parameters.
Clarithromycin, mibefradil, verapamil, troleandomycin, and fluoxetine were incubated in a 150- donor pool of human liver microsomes for up to 30 min and in freshly plated hepatocytes, cryopreserved plated hepatocytes, and cryopreserved plated HepaRG for up to 30 min (fluoxetine up to 90 min), to determine pseudo first-order rates of inactivation (kobs). The apparent KI and kinact were then estimated from nonlinear regression of the kobs determined for each time-dependent inhibitor concentration using eq. 1.
The apparent KI and kinact values obtained using HLM were within the range of those previously published (Table 2). Mibefradil and troleandomycin both exhibited moderate to strong inactivation potential with apparent KI values of 1.1 and 1.8 μM, respectively, and kinact values of 0.7 and 0.31 min−1, respectively. Clarithromycin and verapamil both exhibited weak to moderate inactivation potential with apparent KI values of 17.7 and 8.9 μM, respectively, and kinact values of 0.05 and 0.1 min−1, respectively. Fluoxetine exhibited weak inactivation potential with an apparent KI of 8.6 μM and a kinact of 0.005 min−1. Given the weak inactivation of fluoxetine in vitro, it has been our experience that this is often not reproducible (data not shown).
Kinetic parameters of CYP3A time-dependent inhibitors
The KI and kinact values obtained using freshly plated hepatocytes are reported in Table 2. Mibefradil and troleandomycin both exhibited moderate to strong inactivation potential with apparent KI values of 0.6 and 0.9 μM, respectively, and kinact values of 0.1 min−1. Clarithromycin and verapamil both exhibited weak to moderate inactivation potential with apparent KI values of 12.6 μM and kinact values of 0.06 and 0.14 min−1, respectively. Fluoxetine exhibited weak inactivation potential with an apparent KI value of 7.3 μM and a kinact value of 0.03 min−1.
The apparent KI and kinact values obtained using plated cryopreserved hepatocytes are reported in Table 2. Mibefradil and troleandomycin both exhibited moderate to strong inactivation potential with apparent KI values of 0.2 and 0.6 μM, respectively, and kinact values of 0.28 and 0.08 min−1, respectively. Clarithromycin and verapamil both exhibited weak to moderate inactivation potential with apparent KI values of 8.1 and 1.9 μM, respectively, and kinact values of 0.09 and 0.14 min−1, respectively. Fluoxetine exhibited weak inactivation potential with an apparent KI of 17.1 μM and a kinact of 0.03 min−1.
The apparent KI and kinact values obtained using plated HepaRG are reported in Table 2. Mibefradil and troleandomycin both exhibited moderate to strong inactivation potential with apparent KI values of 1.1 and 2.2 μM, respectively, and kinact values of 0.91 and 0.17 min−1, respectively. Clarithromycin and verapamil both exhibited weak to moderate inactivation potential with apparent KI values of 5.4 and 5.5 μM, respectively, and kinact values of 0.08 and 0.12 min−1, respectively. Fluoxetine exhibited weak inactivation potential with an apparent KI of 35.7 μM and a kinact value of 0.01 min−1.
Prediction of DDIs.
Clinical DDI predictions were determined for each of the time-dependent inhibitors using the input parameters from Table 1 and the corresponding in vitro inactivation kinetic parameters from Table 2. Fold change in AUC of victim drug midazolam was calculated using eq. 2 (Obach et al., 2006).
Fluoxetine is a known weak CYP3A time-dependent inhibitor with generally a lack of reported clinical CYP3A DDI (Lam et al., 2003). After an oral dose of 20 to 60 mg q.d. for 12 days, the observed clinical AUCi/AUC ratio was 0.8. All plated cell systems and HLM agreed with the general lack of clinically observed DDI in humans after oral dosing with the object drug midazolam (Fig. 2; Table 3). However, only the plated cell systems exhibited measureable time-dependent loss of activity due to the longer preincubation times (Fig. 3).
Kinetic plots (kobs versus time-dependent inhibitor concentration) illustrating time-dependent loss of CYP3A activity with fluoxetine in human liver microsomes (A), freshly plated hepatocytes (B), cryopreserved plated hepatocytes (C), and cryopreserved plated HepaRG (D).
Comparison of predicted clinical drug-drug interactions (fold change in AUC with object drug midazolam) from in vitro kinetic parameters of CYP3A time-dependent inhibitors
Predicted fold change in AUC was calculated using eq. 2.
Semilogarithmic plots (ln percentage activity remaining versus incubation time) illustrating the time-dependent loss of enzymatic activity in human liver microsomes (A) and cryopreserved plated HepaRG (B) with the weak CYP3A time-dependent inhibitor fluoxetine.
Verapamil has a reported clinical CYP3A-mediated DDI where after an oral dose of 80 mg t.i.d. for 2 days, the observed clinical AUCi/AUC ratio was 3 to 4 (Backman et al., 1994). The predicted DDI for verapamil using inactivation kinetic parameters derived from HLM and freshly plated hepatocytes was favorable at ∼4-fold, whereas in cryopreserved plated hepatocytes and plated HepaRG, DDI predictions were 10- and 5-fold, respectively. Overall, with the exception of cryopreserved plated hepatocytes, the plated cell systems agreed with the HLM data and fairly closely predicted the clinically observed DDI for verapamil (Fig. 4; Table 3).
Kinetic plots (kobs versus time-dependent inhibitor concentration) illustrating time-dependent loss of CYP3A activity with verapamil in human liver microsomes (A), freshly plated hepatocytes (B), cryopreserved plated hepatocytes (C), and cryopreserved plated HepaRG (D).
A reported clinical CYP3A-mediated DDI has also been reported for clarithromycin (Gorski et al., 1998). After an oral-dosing regimen of 500 mg b.i.d. for 7 days, the observed clinical AUCi/AUC ratio was 5 to 8 (Table 3). The predicted DDIs were as follows: for HLM, 5-fold; in freshly plated hepatocytes, 7-fold; in cryopreserved plated hepatocytes, 11-fold; and in cryopreserved plated HepaRG, 12-fold. The plated system in closest agreement with the HLM predicted and actual reported DDI was freshly plated hepatocytes, with a predicted AUC ratio of 7 (Table 3). Meanwhile, DDI estimates derived from cryopreserved plated hepatocytes and cryopreserved plated HepaRG, which were ∼2-fold higher than predictions from HLM, were within approximately 2-fold of the reported clinical DDI (Fig. 5; Table 3).
Kinetic plots (kobs versus time-dependent inhibitor concentration) illustrating time-dependent loss of CYP3A activity with clarithromycin in human liver microsomes (A), freshly plated hepatocytes (B), cryopreserved plated hepatocytes (C), and cryopreserved plated HepaRG (D).
The reported clinical CYP3A-mediated DDI for mibefradil is ∼9 after a single oral dose of 100 mg (Veronese et al., 2003). The predicted DDIs using inactivation parameters were as follows: for HLM, 12-fold; in freshly plated hepatocytes, 6-fold; in cryopreserved plated hepatocytes, 16-fold; and in cryopreserved plated HepaRG, 13-fold. The plated system in closest agreement with the HLM-derived DDI prediction of 12 was cryopreserved plated HepaRG (13-fold). Cryopreserved plated hepatocytes were within approximately 2-fold of the observed clinical DDI. Freshly plated hepatocytes underpredicted the observed clinical DDI, due to the especially low kinact observed at 0.1 min−1. Each of the plated cell systems tested agreed with the human liver microsomal predicted DDI and roughly agreed with the clinically observed DDI (Fig. 6; Table 3).
Kinetic plots (kobs versus time-dependent inhibitor concentration) illustrating time-dependent loss of CYP3A activity with mibefradil in human liver microsomes (A), freshly plated hepatocytes (B), cryopreserved plated hepatocytes (C), and cryopreserved plated HepaRG (D).
Troleandomycin has a reported clinical CYP3A DDI (Kharasch et al., 2004). After a single oral dose of 500 mg, the observed clinical AUCi/AUC ratio was 15. The predicted DDI for troleandomycin using inactivation kinetic parameters derived from HLM and each cell system was comparable, between 21- to 23-fold. Each of the plated cell systems tested agreed with the human liver microsomal predicted DDI and roughly agreed with the clinically reported DDI (Fig. 7; Table 3).
Kinetic plots (kobs versus time-dependent inhibitor concentration) illustrating time-dependent loss of CYP3A activity with troleandomycin in human liver microsomes (A), freshly plated hepatocytes (B), cryopreserved plated hepatocytes (C), and cryopreserved plated HepaRG (D).
Donor Variability.
To assess the effect of donor-to-donor variability on clinical DDI prediction, freshly plated primary hepatocytes from three different donors were incubated with mibefradil. The three donors exhibited measurable differences in the in vitro inactivation kinetics with apparent KI values of 0.54, 0.77, and 1.6 μM and kinact values of 0.1, 0.16, and 0.36 min−1, for donors H1014, FHVL, and MHVL, respectively (Fig. 8). However, the differences between in vitro kinetic parameters did not translate to a difference in the predicted clinical interaction between midazolam and mibefradil, because the estimated DDI values for all three donors ranged from 5.5- to 6.3-fold.
Kinetic plots (kobs versus time-dependent inhibitor concentration) illustrating time-dependent loss of CYP3A activity with mibefradil in freshly plated human hepatocytes with lot H1014 (A), lot FHVL (B), and lot MHVL (C) demonstrating the effect of donor variability.
Discussion
The U.S. Food and Drug Administration (FDA) has issued several guidance documents that outline the need to evaluate drug candidates for the potential to cause drug-drug interactions (U.S. Department of Health and Human Services, 2006, http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm292362.pdf). More importantly, these documents describe decision trees and recommendations for examining the propensity of a drug to cause time-dependent inhibition of P450. Despite improvements in the in vitro assays that are used to predict clinical DDI (Obach et al., 2006; Grimm et al., 2009), unexpected drug-drug interactions still occur, which may be due to limitations of in vitro systems. The ideal in vitro system for comprehensively evaluating drug-drug interactions should theoretically be human hepatocytes (Zhao, 2008), given the full complement of metabolic clearance mechanisms. Even though the ability to conduct in vitro DDI studies has been demonstrated in human hepatocyte suspensions (Mao et al., 2011), there are limitations with the long-term viability and enzymatic activity of the cells in suspension (Zhao, 2008). It has been demonstrated that human plated cell systems offer unique advantages over human cell suspensions for time-dependent inhibition studies (Li and Doshi, 2011). Thus, the primary goal of this study was to evaluate a number of different plated cell systems for their ability to generate inactivation kinetic parameters for CYP3A, the most important drug-metabolizing enzyme with respect to characterized mechanism for DDI. Furthermore, we investigated plated cell systems for their ability to identify drug candidates that exhibit weak, time-dependent inhibition of CYP3A by conducting experiments with preincubation times greater than 30 min with fluoxetine, a reported weak time-dependent inhibitor of CYP3A.
Pooled HLM are commonly used for measuring the potential of a compound to cause time-dependent inhibition of P450 enzymes (Grimm et al., 2009). Based on the general acceptance of using human liver microsomes for time-dependent inhibition studies and the importance of evaluating CYP3A inactivation, this system was used as the standard for comparison when evaluating the ability of each of the plated cell systems to predict clinical DDI for CYP3A. The advantages of using HLM include availability, low cost, ease of use, and literature support for using time-dependent inhibition data derived from this system to estimate the magnitude of clinical DDIs with measurable accuracy (Obach et al., 2006, 2007). However, limitations with using HLM include inconsistent results for weak, time-dependent inhibitors such as fluoxetine (Fig. 3). Our laboratory routinely encounters weak, time-dependent inhibitors in the discovery setting, and a preincubation time of greater than 30 min may enable improved rank ordering of compounds due to greater dynamic range. However, due to enzyme thermal degradation, the preincubation time for HLM is generally limited to 30 min (Foti and Fisher, 2004). In addition, the inability to assess non-P450-mediated mechanisms (unless costly cofactors are added) is a limitation associated with the use of the HLM fraction.
Sandwich-cultured, freshly plated hepatocytes are commonly used for P450 induction assessment. As such, this system was chosen and evaluated for its ability to assess TDI of CYP3A. One of the main challenges associated with the use of freshly plated primary hepatocytes is the donor-to-donor variability and limited availability. Several donors were tested over a time span of approximately 1 year to determine the inactivation kinetic parameters of the five tested time-dependent inhibitors. The longer duration of the testing was due to the limited availability of cells from human donors with acceptable demographics and CYP3A activity. During the course of this study, we observed substantial donor-to-donor variability in CYP3A activity (data not shown), consistent with previous literature (Moore and Gould, 1984). In addition, the cost of purchasing the plated hepatocytes is substantially higher than using human liver microsomes. The advantages of using freshly plated hepatocytes are general ease of use due to the ability to purchase cells preplated in sandwich culture and the ability to assess non-P450-mediated mechanism due to the presence of a full complement of drug-metabolizing enzymes. Although the inactivation kinetic parameters derived from freshly plated human hepatocytes generally agreed with human liver microsomal parameters (Table 2) and the reported clinical DDI (Table 3), taking into consideration the limited availability of fresh primary human hepatocytes, this in vitro system may not be suitable for routine screening of compounds for CYP3A TDI, but it may serve as a suitable in vitro system for a more definitive assessment of CYP3A TDI, similar to the current approaches for assessing P450 induction liabilities.
Cryopreserved hepatocytes are commonly used to assess the in vitro metabolic clearance of pharmaceutical compounds. FDA guidance also establishes plateable cryopreserved hepatocytes as an acceptable model for assessment of human P450 induction. The advantages of using cryopreserved plated hepatocytes include availability of cells from multiple donors and the ability to assess non-P450 clearance mechanisms due to the presence of a full complement of hepatic drug-metabolizing enzymes. The apparent KI for mibefradil and verapamil tested in this system was generally 5- to 6-fold lower than the apparent KI derived in HLM and freshly plated systems (Table 2). This decrease in apparent KI translates to an overprediction compared with the reported pharmacokinetic CYP3A DDI (Table 3). It is unclear why this difference was observed. The main limitations in using cryopreserved hepatocytes are the overall cost for a plateable lot of cells and the plating process, which requires an overlay.
Based on previously published studies (Le Vee et al., 2006; Kanebratt and Andersson, 2008b), the HepaRG cell line has been reported to be useful for assessing P450 induction and other human drug metabolism studies. The advantages of using cryopreserved plated HepaRG are general ease of use due to good cell viability, simple plating procedures, acceptable expression levels of CYP3A (Kanebratt and Andersson, 2008a), and an unlimited supply of the same cell line, which would circumvent the need to recharacterize new lots of cryopreserved hepatocytes for CYP3A activity. Unlike human liver microsomes, where weak fluoxetine inactivation kinetics may be difficult to consistently reproduce due to limitations with the length of preincubation time, we were easily able to detect the weak CYP3A inactivation using cryopreserved plated HepaRG cells (Fig. 3), which suggests that this in vitro system may be suitable for this challenging task. However, one should understand that the expression levels of certain drug-metabolizing enzymes, such as CYP2D6, are lower than desired, which may limit the utility of the cell line for evaluating TDI of additional P450 enzymes (Kanebratt and Andersson, 2008a).
In summary, freshly plated hepatocytes seem to provide clinical DDI estimates that are roughly equivalent to human liver microsomes and approximate those observed clinically. However, given the evidence for potential donor-to-donor variability and infrequent availability, this system is not a viable choice for routine assessment of time-dependent inhibition in a discovery setting. In addition, use of cryopreserved hepatocytes, although also demonstrated as useful for characterizing CYP3A time-dependent inhibitors, is fraught with potential issues such as cost and the need to re-characterize new lots of cells. Overall, the HepaRG cell model appears to be a suitable system for an early assessment of time-dependent inhibition of CYP3A in the discovery setting, although further characterization for their utility in assessing TDI of additional P450 enzymes needs to be conducted. Data reported here should provide a benchmark for determined inactivation kinetic parameters and predicted clinical DDI for CYP3A time-dependent inhibitors using the cryopreserved plated HepaRG cell system. Now that we have conducted a baseline assessment for the ability to characterize well known time-dependent inhibitors in these cell models, our future studies will focus on evaluating non-P450-mediated mechanisms of time-dependent inhibition using plated cell systems, such as those reported with gemfibrozil (Ogilvie et al., 2006).
Authorship Contributions
Participated in research design: Albaugh, Fisher, Fullenwider, and Hutzler.
Conducted experiments: Albaugh and Fullenwider.
Contributed new reagents or analytic tools: Albaugh and Fullenwider.
Performed data analysis: Albaugh, Fullenwider, and Hutzler.
Wrote or contributed to the writing of the manuscript: Albaugh, Fisher, Fullenwider, and Hutzler.
Acknowledgments
We thank Dr. Dustin Smith (Boehringer-Ingelheim Pharmaceuticals, Ridgefield, CT) for carefully reviewing the manuscript and providing valuable comments and David Joseph (Boehringer-Ingelheim Pharmaceuticals, Ridgefield, CT) for valuable advice and suggestions regarding assessment of TDI in freshly plated hepatocytes. Finally, we thank Dr. Jonathon Jackson (Life Technologies, Durham, NC) for expert advice regarding plating and maintenance of HepaRG cells and cryopreserved hepatocytes.
Footnotes
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
ABBREVIATIONS:
- TDI
- time-dependent inhibition
- P450
- cytochrome P450
- DDI
- drug-drug interaction
- HLM
- human liver microsomes
- CHRM
- cryopreserved hepatocyte recovery medium
- LC/MS/MS
- liquid chromatography tandem mass spectrometry
- AUC
- area under the curve
- AUCi
- AUC in the presence of the inhibitor.
- Received January 19, 2012.
- Accepted April 9, 2012.
- Copyright © 2012 by The American Society for Pharmacology and Experimental Therapeutics