Abstract
The levels of metabolizing enzymes and transporters expressed in hepatocytes are decisive factors for hepatobiliary disposition of most drugs. Induction via nuclear receptor activation can significantly alter those levels, with the coregulation of multiple enzymes and transporters occurring to different extents. Here, we report the use of a targeted liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) method for concurrent quantification of multiple cytochrome P450 (P450), UDP-glucuronosyltransferase (UGT), and transporter proteins in cultured primary human hepatocytes. The effects of culture format (i.e., sandwich culture versus conventional culture) and of dexamethasone (DEX) media concentrations on mRNA, protein, and activity levels were determined for three donors, and protein expression was compared with that in liver. In general, P450 and UGT expression was lower in hepatocyte cultures than that in liver, and CYP2C9 was found to be the most abundant P450 isoform expressed in cultured hepatocytes. The sandwich culture format and 0.1 μM DEX in media retained the protein expression in the hepatocytes closest to the levels found in liver. However, higher in vitro expression was observed for drug transporters, especially for multidrug resistance protein 1 and breast cancer resistance protein. Direct protein quantification was applied successfully to study in vitro induction in sandwich cultured primary hepatocytes in a 24-well format using the prototypical inducers rifampicin, omeprazole, and phenobarbital. We conclude that targeted absolute LC-MS/MS quantification of drug-metabolizing enzymes and transporters can broaden the scope and significantly increase the impact of in vitro drug metabolism studies, such as induction, as an important supplement or future alternative to mRNA and activity data.
Introduction
The assessment of potential drug-drug interactions by in vitro methods has become an indispensable part of the preclinical development process of new drug candidates. In addition to investigations on reversible or irreversible inhibition of drug-metabolizing enzymes, studies on the effects of enzyme induction and on the interaction with drug transporters have to be performed and are considered as mandatory by current regulatory guidelines and draft guidance documents [U.S. Food and Drug Administration, Draft Guidance for Industry: Drug Interaction Studies-Study Design, Data Analysis, and Implications for Dosing and Labeling (http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm072101.pdf); European Medicine Agency, Draft Guideline on the Investigation of Drug Interactions (http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2010/05/WC500090112.pdf]. Primary human hepatocytes are the recommended and broadly used experimental system for these types of studies (Hewitt et al., 2007; Sahi et al., 2010; Swift et al., 2010). The use of hepatocytes provides, among others, the advantage of including active drug transport mechanisms as well as effects of drug metabolites that are formed during the experimental incubation period by phase I and II metabolism.
Cultured hepatocytes offer the possibility to assess long-term effects, such as enzyme induction (i.e., changes of the expression levels of molecular entities of the hepatocytes). Common culture formats used for in vitro enzyme induction studies comprise conventional and sandwich cultures, the former referring to plated hepatocytes maintained on rigid substratum (e.g., collagen), whereas in the latter cells are “sandwiched” by an additional layer of either collagen or a composite of extracellular matrix proteins, such as Matrigel (LeCluyse, 2001). Experimental protocols for in vitro enzyme induction studies vary substantially between laboratories across the industry, and no standard criteria for induction assays are applied currently (Sinz et al., 2008). Assay conditions, such as culture format and duration, media supplements etc., were reported previously to alter the phenotype and induction response of cultured hepatocytes (LeCluyse, 2001). Therefore, selection of the appropriate culture format and conditions for a specific in vitro study is crucial with regard to the in vivo predictability of the data generated. Although the quantification of enzymatic activities using suitable probe substrates has become a standard procedure to measure the changes in expression levels and activities of cytochrome P450 (P450) enzymes, other approaches are feasible but are not necessarily performed on a routine basis. The assessment of changes in expression at the mRNA or protein level is such an approach. Quantification of mRNA by reverse transcription polymerase chain reaction (RT-PCR) methodologies is relatively straightforward and often conducted to supplement data on enzyme activity. However, changes in mRNA levels may not necessarily reflect protein expression, especially for drug transporters (Ohtsuki et al., 2012).
Measuring changes of protein amounts is performed far less frequently (Sinz et al., 2008; Chu et al., 2009) and not required by current regulatory guidelines. This is because common experimental procedures are relatively tedious, require the availability of specific antibodies, and typically lack the necessary level of accuracy (e.g., compared with enzymatic activity assays). Yet, quantitative data on protein levels would be highly valuable for several reasons: first, to confirm the suitability of the experimental system for the respective study by comparing relevant in vitro protein expression (P450s, drug transporters, etc.) with in vivo data; second, because results for enzyme activity from in vitro induction studies can be confounded by concomitant enzymatic inhibition by the test compound, especially irreversible inhibition, therefore leading to false negative results. Such effects have been described for troleandomycin, mifepristone, verapamil, and several HIV protease inhibitors, specifically ritonavir (Hsu et al., 1998; Luo et al., 2002; McGinnity et al., 2006; Kenny et al., 2008). Even if enzyme inhibition by the test compound may be recognized beforehand during separate in vitro studies, effects of inhibitory metabolites formed by the hepatocytes can occur and may remain unrecognized (Ogilvie et al., 2006). However, using mRNA as the sole parameter to measure enzyme induction is prone to misinterpretation, most often by overestimating the induction potential. To avoid such problems, routine quantification of expression levels for a panel of important proteins during in vitro enzyme induction studies would be highly desirable. Liquid chromatography coupled tandem mass spectrometry (LC-MS/MS)-based techniques recently have been developed successfully for absolute quantification of membrane-bound enzymes and transporter proteins (Kamiie et al., 2008; Kawakami et al., 2011).
The purpose of the present study is to demonstrate the suitability of LC-MS/MS-based absolute protein quantification for sensitive and specific determination of molecular entities in primary human hepatocyte cultures that are involved in hepatic drug metabolism. This includes the effects from culture and induction not only for drug-metabolizing enzymes but also for drug transporters, which have been recognized as a decisive factor for the understanding of hepatobiliary drug disposition (Shitara et al., 2006; International Transporter Consortium et al., 2010). To our knowledge, this is the first report comparing enzyme activity and mRNA levels with absolute quantitative data on protein concentration in primary hepatocyte cultures measured by a targeted LC-MS/MS approach.
Materials and Methods
Materials.
Amodiaquine, dexamethasone (DEX), diclofenac, dimethyl sulfoxide, extracellular matrix (ECM) gel from Engelbreth-Holm-Swarm murine sarcoma, omeprazole (OME), penicillin-streptomycin (100×), phenobarbital (PB), rifampicin (RIF), testosterone, trypan blue solution, and Williams' medium E (WME) were purchased from Sigma-Aldrich (St. Louis, MO). Bupropion was from Promochem (Wesel, Germany), S-mephenytoin was from Enzo Life Sciences GmbH (Lörrach, Germany), and phenacetin was from Acros Organics (Fairlawn, NJ). Rat tail collagen I was from Roche Diagnostics (Mannheim, Germany), fetal calf serum, phosphate-buffered saline, and stabilized l-glutamine (GlutaMAX, 100×) was from Invitrogen (Carlsbad, CA), and insulin-transferrin-selenous acid and bovine serum albumin premix (100×) was from BD Biosciences (San Jose, CA). Dulbecco's modified Eagle's medium (10×) was purchased from Biozol (Eching, Germany).
Hepatocyte Culture.
Fresh human hepatocytes from Caucasian donors were provided by the Human Tissue and Cell Research Foundation/Hepacult GmbH (Regensburg, Germany). Cells were prepared from surplus liver tissue obtained from tumor resections using the two-step collagenase perfusion technique as described previously (LeCluyse et al., 2005) and shipped as a suspension. Donor demographics, a list of latest medications, and informed consent were provided. WME supplemented with 6.25 μg/ml insulin, 6.25 μg/ml transferrin, 6.25 ng/ml selenous acid, 5.35 μg/ml linoleic acid, 1.25 mg/ml bovine serum albumin, 0.01 μM DEX, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine was used for human hepatocyte culture. Upon arrival, hepatocytes were sedimented for 5 min at 72g and 4°C. Cell viability was assessed by trypan blue exclusion. Hepatocyte suspensions were adjusted to 750,000 viable cells per milliliter in WME containing 6% fetal calf serum, and cells were plated on BD BioCoat collagen I-coated plates (BD Biosciences). Seeding density was 375,000 cells per well of a 24-well plate for enzyme activity and RNA expression analysis and 10 million (mio) cells per 100-mm dish for microsomal and plasma membrane preparations, respectively. After an incubation period of 4 h at 37°C, 5% CO2, and 95% humidity, nonattached cells were removed, and fresh WME without serum was supplied. One hundred-millimeter dishes were used to prepare microsomal and plasma membrane fractions for absolute protein quantification of P450s, UGT enzymes, and transporters by LC-MS/MS, whereas 24-well plates were used for P450 in situ enzyme activity assessment and subsequent mRNA expression analysis for P450s, UGTs, and drug transporters. Cultures were assessed daily by light microscopy to ensure morphological normalcy (i.e., polygonal shape and distinct bile canaliculi) as well as confluence adequate for experiments (Supplemental Table 1).
Sandwich Culture Using ECM Gel.
Twenty-four hours after plating, medium was removed, and hepatocytes were overlaid with 0.5 ml per 24-well or 9 ml per 100-mm dish, respectively, of ice-cold WME containing 0.25 mg/ml ECM gel and incubated for 24 h at 37°C, 5% CO2, and 95% humidity to obtain the sandwich cultures. For the high DEX cultures, medium containing 0.1 μM DEX was used starting from 24 h after plating.
Sandwich Culture Using Collagen I.
Lyophilized rat tail collagen I was dissolved overnight in 0.2% acetic acid to give a 0.83 mg/ml stock solution. A total of 0.55 ml of 10× Dulbecco's modified Eagle's medium was added to 4.95 ml of collagen I stock solution on ice, and the pH was adjusted to 7.2 to 7.4 with 1 N NaOH, resulting in a final collagen I solution of 0.75 mg/ml. Medium was removed 24 h after initial plating, and 75 μl of ice-cold collagen I solution was added per well in a 24-well plate or 1.5 ml per 100-mm dish, respectively. Culture plates were incubated for 45 min at 37°C, 5% CO2, and 95% humidity for collagen gelling before fresh WME was supplied.
Single Gel Monolayer Culture.
Hepatocytes were maintained in culture after initial seeding without any matrix overlay. All of the cultures were kept in total for 4 days in 24-well format before in situ enzyme activity assessment and RNA isolation and for 5 days in 100-mm format, respectively, before cell harvest for protein quantification. Fresh medium was supplied daily. Maintenance medium for all of the cultures was identical, containing 0.01 μM DEX, except for the high DEX condition where 0.1 μM DEX was used.
The research protocols for the present study were approved by the Ethics Committee of the Graduate School of Pharmaceutical Sciences of Tohoku University. The study has been carried out in accordance with the Declaration of Helsinki.
In Situ P450 Enzyme Activity Assessment.
P450 enzyme activities were determined in situ by monitoring metabolite formation in culture medium over time (0–45 or 0–90 min) with LC-MS/MS after incubation with specific probe substrates. Validated enzyme activity assays comprised separate probe reactions performed in 24-well format. Probe substrates and assay details are provided as supplemental information (Supplemental Table 3). The following equipment was used for LC-MS/MS analysis of incubation samples: HTS PAL autosampler (CTC Analytics, Zwingen, Switzerland), 1200 series binary pump and degasser (Agilent Technologies, Santa Clara, CA), column switch thermostat Spark 880 (May Lab, Vienna, Austria), and API 4000 QTRAP (Applied Biosystems/MDS Sciex, Foster City, CA). Chromatographic separation of incubation samples was performed on YMC-Pack Pro C18 (50 × 2.1 mm, 5 μm; YMC Europe, Dinslaken, Germany) and XTerra MS C18 (50 × 2.1 mm, 5 μm; Waters, Milford, MA) analytical columns. Twenty-five microliters of the incubation sample were diluted with 480 μl of water containing 5 to 10% acetonitrile, 0.1% formic acid, and the respective stable isotope-labeled metabolite as an internal standard. Five microliters was injected into the LC-MS/MS system operated with an atmospheric pressure electrospray ionization source. Details on analytical assays are provided in Supplemental Table 3.
Microsomal and Plasma Membrane Preparation.
Hepatocytes were scraped from culture dishes using a rubber policeman, washed with phosphate-buffered saline, sedimented at 800g for 10 min at 4°C, and snap-frozen in liquid nitrogen. After being thawed, cells were lysed by two-step nitrogen cavitation at 450 and 800 psi each for 15 min at 4°C using a pressure vessel (Parr, Moline, IL). The homogenates were centrifuged at 10,800g for 20 min at 4°C, and the supernatants were collected and ultracentrifuged at 100,000g for 60 min at 4°C. The microsomal pellet was suspended in buffer A [0.1 M KCl-phosphate buffer (pH 7.4) containing protease inhibitor cocktail] and ultracentrifuged at 100,000g for 60 min at 4°C. The resulting pellet was suspended buffer B (20 mM Tris-HCl buffer containing 0.25 M sucrose and 5.4 mM EDTA), and an aliquot of the suspension was stored as a microsomal fraction at −80°C. An enriched plasma membrane fraction was prepared from the microsomal fraction for the quantification of lower abundance transporter proteins. Therefore, the remaining volume of the microsomal fraction was layered on top of a 38% (w/v) sucrose solution and centrifuged at 100,000g for 30 min at 4°C. The turbid layer at the interface was recovered, suspended in buffer B, and centrifuged at 100,000g for 30 min at 4°C. The plasma membrane fraction, containing both canalicular and sinusoidal membranes, was obtained from the resulting pellet by suspension in buffer B. Levels of Na/K-ATPase and γ-glutamyl transpeptidase (γ-GTP), plasma membrane marker proteins used for data normalization, typically were increased in plasma membrane factions compared with those in microsomal fractions (Table 1). Protein concentrations were measured by the Lowry method using the detergent compatible protein assay reagent (Bio-Rad Laboratories, Hercules, CA).
Absolute protein levels and enzyme activities in human hepatocytes from three donors kept in sandwich culture (ECM overlay, 0.01 μM DEX)
P450 and UGT proteins were determined from microsomal fractions, and transporter proteins were determined from plasma membrane preparations. Concentrations are calculated as the total amount of microsomal and plasma membrane protein, respectively. P450 enzyme activities were determined in situ by measuring phenacetin O-deethylation (CYP1A2), bupropion 2-hydroxylation (CYP2B6), amodiaquine N-deethylation (CYP2C8), diclofenac 4′-hydroxylation (CYP2C9), S-mephenytoin 4′-hydroxylation (CYP2C19), and testosterone-6β hydroxylation (CYP3A4).
TaqMan RT-PCR Gene Expression Analysis.
Total RNA from hepatocytes cultured in 24-well plates was isolated using the RNeasy protocol (QIAGEN GmbH, Hilden, Germany). The yield of extracted RNA was analyzed spectrophotometrically at 260 nm. Purity and integrity of RNA were assessed using RNA 6000 LabChips and 2100 Bioanalyzer (Agilent Technologies). For first-strand cDNA synthesis, 200 ng of total RNA was reverse-transcribed in a final volume of 20 μl following the protocol for the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). cDNA samples were diluted to 60 μl with water and stored at −20°C. One microliter of diluted cDNA was used for quantitative real-time PCR. P450 and UGT analyses were performed on a 7500 Fast Real-Time PCR System (Applied Biosystems), and transporter analyses were performed on an iQ5 Real-Time PCR System (Bio-Rad Laboratories). Assays were performed as technical triplicates. Information on inventoried assays and primer and probe sequences of custom-made assays are provided in Supplemental Tables 4 and 5.
Data Normalization and Analysis.
Gene expression data were normalized to β-actin as an endogenous control quantified in the same sample (duplexed RT-PCR) or as a separate sample on the same assay plate for inventoried assays. Relative expression differences were calculated using the comparative ΔΔCt method, accounting for the actual PCR efficiency for each target and run.
P450, UGT, and Transporter Protein Quantification.
Absolute quantification of P450, UGT, and transporter proteins was conducted by LC-MS/MS measurement of tryptic protein fragments similar to the procedures described previously (Kamiie et al., 2008; Kawakami et al., 2011). In brief, protein samples were diluted in suspension buffer containing 7 M guanidine-HCl, 500 mM Tris-HCl (pH 8.0), and 10 mM EDTA. Samples were reduced by the addition of 1 mM dithiothreitol and alkylated under a nitrogen atmosphere with 10 mM iodoacetamide. The S-carbamoylmethylated proteins were precipitated by the addition of a mixture of methanol and chloroform. Precipitates were dissolved in 6 M urea, diluted with 100 mM Tris-HCl (pH 8.0), and digested with tosyl phenylalanyl chloromethyl ketone-treated trypsin at an enzyme/substrate ratio of 1:100 at 37°C for 16 h. The tryptic digests were spiked with stable isotope-labeled peptides of identical amino acid sequences as the corresponding analyte peptides as internal standards and acidified with formic acid for analysis with the nano-LC system or high-performance liquid chromatography system, which was connected to an electrospray ionization triple-quadrupole mass spectrometer [API 5000 or 4000 QTRAP (for CYP2B6 and 2C19); Applied Biosystems/MDS Sciex]. LC was performed with C18 capillary columns. Linear gradients of 1 to 45% acetonitrile in 0.1% formic acid were applied to elute the peptides at a flow rate of 100 nl/min for nano-LC and 50 μl/min for high-performance liquid chromatography. The mass spectrometer was prepared to run multiple reaction monitoring (MRM) experiments for peptides. Peak integration of ion count chromatograms was performed using Analyst software (Applied Biosystems/MDS Sciex). A specific tryptic fragment peptide was selected for the quantification of each target protein and monitored in four different MRM transitions. The amount of each peptide was determined as an average of three to four MRM transitions from one sample. In the cases where signal peaks were detected for less than three transitions, the amount of peptide was defined as below the limit of quantification. Quantification of CYP3A4 was performed indirectly by the subtraction of the values determined for the peptide specific for CYP3A43 and the one detecting both CYP3A4 and CYP3A43: CYP3A4 = CYP3A4/CYP3A43 − CYP3A43.
Results
Absolute Protein Expression of P450, UGT, and Drug Transporter Proteins in Sandwich Cultured Human Hepatocytes.
The absolute amounts of P450, UGT, and transporter proteins were determined in primary human hepatocytes from three donors cultured for 5 days on 100-mm BD BioCoat collagen I-coated culture dishes, overlaid with a film of ECM protein gel and with 0.01 μM DEX present in culture medium (Fig. 1; Table 1). The expression levels of P450 and UGT enzymes were quantified in the microsomal fractions, and drug transporters were quantified in the plasma membrane fractions containing both sinusoidal and canalicular membranes, respectively. CYP2C9 and UGT2B7 showed the highest protein expression levels of the measured P450 and UGT enzymes (Fig. 1), whereas the levels of CYP2B6 and CYP3A5 were the lowest and even below the limit of quantification (LOQ) for two donors (Table 1). CYP2A6 concentrations varied the most among hepatocyte cultures from three donors (approximately 7.8-fold). Multidrug resistance protein 1 (MDR1) and multidrug resistance associated protein 2 (MRP2) showed the highest expression levels of canalicular transporters, and organic cation transporter 1 (OCT1) showed the highest expression level among the sinusoidal transporters (Fig. 1). The highest variability in expression levels was observed for ABCA8 (7.7-fold), ABCG8 (7.5-fold), and MRP1 (10.7-fold). This was predominantly due to higher concentrations of these transporters found in the plasma membrane fractions from donor HuHep10261, exhibiting up to 10.6-fold higher expression compared with the average of donors HuHep9966 and HuHep10901, whereas a 1.74-fold difference in expression was observed at most between these two donors (data not shown). MRP3 and MRP4 were below the LOQ for both of the latter two donors.
Absolute expression levels of P450, UGT (A), and transporter proteins (B) in sandwich cultured human hepatocytes. Expression levels of P450 and UGT enzymes were determined in microsomal fractions of human hepatocytes from three donors cultured for 5 days under standard conditions (i.e., ECM gel and 0.01 μM DEX). Expression levels of transporter proteins were determined in purified plasma membrane fractions. Data represent mean ± S.D. (n = 2 or 3, excluding data below LOQ). Open squares represent data determined from only one donor. *, amounts of CYP3A4 were determined indirectly as described under Materials and Methods.
Effects of Culture Conditions on P450 Protein Expression, mRNA Levels, and Enzymatic Activity in Human Hepatocytes.
Hepatocytes from the identical three donor preparations described above were cultured in parallel to control for three additional conditions (see Materials and Methods). The effects were assessed on enzyme activity and mRNA levels for CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, and CYP3A4 after 4 days of culture in 24-well plates and on protein expression after 5 days in 100-mm dishes, respectively (Table 2). Enzyme activities were determined in situ from 24-well plates by measuring the metabolite formation in culture supernatants after the addition of specific probe substrates, whereas P450 protein was quantified from the microsomal preparations obtained from 100-mm dishes. Hepatocytes plated on BD BioCoat collagen I-coated culture dishes with ECM overlay and 0.01 μM DEX in culture medium served as a control. Compared with these culture conditions, hepatocytes also were cultured either using the same sandwich format with a 10-fold higher media concentration of DEX (0.1 μM DEX) or with an overlay of gelled collagen I (Collagen I) instead of ECM gel or using a conventional culture on BD BioCoat collagen I-coated culture dishes without any additional overlay (single gel). Mean relative changes (n = 3) in protein levels, enzyme activity, and mRNA expression of the selected P450 isoforms CYP1A2, CYP2B6, and CYP3A4 for the different culture conditions are depicted in Fig. 2. Table 2 summarizes the mean relative changes of protein levels, mRNA, and enzyme activity for the different culture conditions compared with the control condition (0.01 μM DEX). No consistently significant effect on P450 expression or activity was seen (Fig. 2), because donor variability superseded the effects caused by different culture conditions. However, the high DEX condition generally increased P450 expression, particularly CYP2A6, CYP2E1, and CYP3A7, mRNA levels of CYP1A2, CYP2A6, and CYP3A4, and enzymatic activity of CYP3A4. In contrast, the conventional single gel condition exhibited the lowest levels in P450 protein, mRNA, and enzyme activity. Collagen I cultures showed the lowest enzyme activity for CYP2C9 (0.49-fold) but the highest protein levels (1.53-fold, compared with control) for all of the tested culture conditions. Because this appears to be an obvious contradiction, it has to be taken into account that, unlike P450 protein that was measured in microsomal preparations, P450 activities were determined in situ; therefore, additional factors such as active transport and/or properties of matrix overlay may affect metabolite kinetics in the supernatant.
Mean relative change in mRNA expression, protein levels, and in situ enzyme activity in hepatocytes cultured at different conditions compared with control (ECM overlay, 0.01 μM DEX)
The same donor preparations (n = 3) were used for all of the tested conditions. An analysis of mRNA expression and enzyme activity was performed after 4 days in culture and protein after 5 days, respectively. P450 and UGT proteins were determined from microsomal fractions, and transporter proteins were determined from plasma membrane preparations of hepatocyte cultures.
Relative changes in mRNA, enzymatic activity, and protein concentration of CYP1A2 (A), CYP2B6 (B), and CYP3A4 (C) in human hepatocyte cultures in response to culture conditions. The effects of 0.1 μM DEX, collagen I, and single gel conditions were compared relative to that of control (ECM gel sandwich and 0.01 μM DEX). Expression of mRNA was determined via RT-PCR, and in situ enzyme activity was determined by probe substrate reactions after 4 days of culture in 24-well format. P450 protein was quantified in microsomal preparations from the cultures of the same donors on 100-mm plates after 5 days. Data represent mean ± S.D. of three donors. Significance was determined applying a two-tailed, two-sample t test. *, p < 0.05; **, p < 0.01.
Effects of Culture Conditions on UGT Protein Expression and mRNA Levels.
Microsomal preparations from hepatocytes were analyzed for protein expression of UGT1A1 and UGT2B7 by LC-MS/MS (Table 2). Semiquantitative RT-PCR was applied to determine the mRNA expression of UGT1A1, UGT1A4, UGT1A6, UGT1A9, UGT2B7, and UGT2B15 relative to that of endogenous control β-actin. Results for relative mean changes under the different culture conditions are summarized in Table 2. Both mRNA and protein levels were within a 2-fold range for all of the conditions tested, showing less sensitivity for UGT enzyme expression toward external conditions compared with P450s. Single gel conditions reduced mRNA expression of UGT1A1 approximately 0.64-fold compared with that of control, and UGT1A1 protein likewise was reduced in the single gel cultures. However, the decrease was not statistically significant. UGT2B7 mRNA expression was the highest under control conditions, whereas protein levels were maximal in single gel culture (1.46-fold, compared with control).
Effects of Culture Conditions on Drug Transporter Protein Expression and mRNA Levels.
Mean expression levels of transporter proteins in plasma membrane fractions of hepatocytes from three donors cultured under the described conditions were within a 0.3- to 3.6-fold range compared with those of the control conditions (Table 2), and the control conditions showed overall the highest protein levels for both ATP-binding cassette (ABC) and solute carrier transporters. Except for MRP1, notable differences in the expression of the transporters among culture conditions were not observed. Protein expression of MRP1 under the collagen I condition was found to be on average 2.29-fold higher compared with that of the control. Mean MRP5 expression was increased by 3.6-fold in single gel culture; however, there was no statistically significant difference in the expression under control conditions. For mRNA, bile salt export pump (BSEP) and OCT1 showed the highest interculture differences in expression with approximately 3-fold higher levels in control condition compared with those in the single gel. MRP2 and OCT1 transcripts were the most abundant among the tested drug transporters; MRP1, breast cancer resistance protein (BCRP), and OATP1B3 mRNA was found to be the least abundant (data not shown).
Comparison of Human P450, UGT, and Transporter Protein Expression between Cultured Hepatocytes and Liver.
The expression of P450, UGT, and transporter proteins in hepatocytes under the tested culture conditions was compared with corresponding protein levels determined in liver tissue (Ohtsuki et al., 2012). Figure 3 shows the correlation between protein expression of P450 and UGT enzymes in human liver (n = 17) and that in hepatocytes from three donors kept over 5 days under four different culture conditions. Corresponding data for transporter protein expression are available in Supplemental Fig. 6.
Comparison of protein expression for P450 and UGT enzymes between human liver tissue and hepatocytes cultured under different conditions. Protein concentrations were determined from microsomal preparations. Data represent mean ± S.D. for human liver tissue (n = 17) and mean ± S.D. for the same human hepatocyte preparations (n = 3) cultured under different conditions. Values below LOQ were set to 0 fmol/μg protein for mean calculation. Dashed lines indicate 5-fold variation. Selected data are shown for clarity reasons.
As stated previously, relatively high interdonor variability in expression levels was observed among the three hepatocyte preparations, most markedly for CYP2C8, CYP2C19, and CYP2E1 (>10-fold) and irrespective of the culture condition applied. Variability of the lower abundance CYP2B6 and CYP3A5 protein in hepatocyte cultures was partly due to analytical sensitivity. Liver tissue samples obtained from 17 donors (donor demographics are given in Supplemental Table 2) showed the largest variability in CYP3A4 expression (>1000-fold), followed by CYP2C19 (approximately 100-fold) and CYP3A5 (>10-fold). Among the drug-metabolizing enzymes quantified, CYP2C9 and UGT2B7 were found to be the most abundant in both cultured hepatocytes and liver tissue. Mean expression of CYP2B6, CYP3A5, and CYP2E1 in cultures was the lowest relative to liver. For the other P450 and UGT proteins analyzed, expression was found to be within a 5-fold range compared with liver expression. Overall, relative effects of culture conditions on P450 expression were similar among the three donors, with 0.1 μM DEX retaining levels closest to liver. However, these effects were superseded altogether by donor variability in P450 expression.
In contrast to the P450 and UGT enzymes, transporter protein expression in cultured hepatocytes was generally higher than that in human liver (Supplemental Fig. 6), and different culture formats and DEX concentrations showed no relevant effects on protein levels. It is noteworthy that MRP3 was quantified in all of the liver samples but could not be detected in cultured hepatocytes, whereas the opposite was true for ABCG8. Mean expression levels of the canalicular transporter BCRP in cultured hepatocytes were >10-fold higher than those in the samples from liver biopsies, likely reflecting extensive bile canalicular network formation under in vitro conditions.
Induction of P450 and Drug Transporters by Prototypical Inducers in Sandwich Cultured Human Hepatocytes.
Fresh human hepatocytes from a separate donor preparation were cultured in 24-well format, applying our laboratory standard condition for in vitro induction studies as described above. Cells were exposed for 48 h to prototypical inducers RIF (at 25 μM), OME (at 50 μM), PB (at 1 mM), and solvent control. In situ enzyme activities and mRNA expression were determined. Cells from a parallel incubation were scraped from the 24 wells to prepare microsomal and plasma membrane fractions for protein quantification. Table 3 summarizes the results for absolute concentrations of P450, UGT, and transporter proteins and in situ P450 activities. Relative induction of CYP1A2, CYP2B6, and CYP3A4 mRNA, enzyme activity, and protein by RIF, OME, and PB is shown in Fig. 4. Marked response to the selective inducers was observed for P450 protein, yet exceeding the response of mRNA and activity for CYP1A2 and CYP3A4.
Absolute protein levels and enzyme activities in sandwich cultures from donor preparation HuHep10536 after 48 h of treatment with solvent control and prototypical inducers OME, PB, and RIF, respectively
Protein concentrations (except for Na/K-ATPase) were normalized to γ-GTP for each incubation condition. P450 enzyme activities were determined in situ by measuring phenacetin O-deethylation (CYP1A2), bupropion 2-hydroxylation (CYP2B6), amodiaquine N-deethylation (CYP2C8), and testosterone-6β hydroxylation (CYP3A4).
Induction of mRNA, enzyme activity, and protein of CYP1A2 (A), CYP2B6 (B), and CYP3A4 (C) in sandwich cultured human hepatocytes after 48 h of treatment with prototypical inducers. Hepatocytes from donor HuHep10536 were cultured in 24-well format under standard conditions and treated for 48 h with solvent control, 25 μM RIF, 50 μM OME, and 1 mM PB, respectively. In situ enzyme activity and mRNA were determined from the same culture plate. P450 protein was quantified in microsomal preparations from a separate 24-well plate per condition and concentration normalized to γ-GTP expression. Data represent mean ± S.D.
Figure 5 shows the effects of prototypical inducers on the relative expression levels of the canalicular transporters MDR1 and BCRP, sinusoidal transporters OATP1B1 and OCT1, and phase II enzyme UGT1A1. Clear induction of MDR1 protein by all three inducers RIF (1.78-fold), OME (2.46-fold), and PB (2.82-fold) was observed. A significant increase in BCRP protein expression was seen after treatment with OME (1.43-fold) and PB (1.37-fold). It is noteworthy that RIF induced the concentration of the sinusoidal transporter OATP1B1 (1.50-fold), whereas OCT1 expression was suppressed significantly (0.74-fold).
Induction of transporter and UGT proteins in sandwich cultured human hepatocytes after 48 h of treatment with prototypical inducers. Hepatocytes of donor HuHep10536 were cultured in 24-well format under standard conditions and treated for 48 h with solvent control, 25 μM RIF, 50 μM OME, and 1 mM PB, respectively. Protein concentrations of UGT1A1 were determined from microsomal preparations, and transporter proteins were quantified from purified plasma membrane preparations. Data were normalized to γ-GTP expression and represent mean ± S.D. Significance was determined by applying a two-tailed, two-sample t test. *, p < 0.05; **, p < 0.01.
Discussion
LC-MS/MS assays were used in the present study to quantify P450, UGT, and transporter proteins in cultured primary human hepatocytes, and the data were compared with mRNA expression and enzyme activities. In addition, protein concentrations in cultured hepatocytes were compared with those from human liver tissues. We further demonstrated that targeted absolute protein quantification can be applied to standard in vitro induction studies.
MRM analysis of diagnostic peptides with stable isotope-labeled internal standards and reference peptides allows for sensitive and selective absolute quantification of protein concentrations with a linear dynamic range similar to that of small molecules (Elliott et al., 2009; Ohtsuki et al., 2011). Integral membrane proteins yet pose an analytical challenge due to their hydrophobic nature and low abundance, as seen for many drug transporters (Barnidge et al., 2003; Kamiie et al., 2008). Being a relatively new technique, LC-MS/MS-based absolute protein quantification has not yet been introduced widely as a standard method for the assessment of drug-metabolizing enzymes and drug transporter proteins; therefore, data on assay development and validation still are sparse. The validity of the quantification assays used in this study was demonstrated recently (Sakamoto et al., 2011) with regard to the acceptance criteria for robustness and reproducibility of bioanalytical assays that are used commonly in drug development. This underlines the potential and applicability for nonclinical studies on drug metabolism and pharmacokinetics, where validated analytical assays are highly desired—if not mandatory—that meet quality criteria required by regulatory guidance. Besides adding a valuable end point to standard in vitro studies, such as enzyme induction, these methods can be used to quantify transporter protein expression in cellular systems used for studies on drug transport, which may help the interpretation of results obtained from, for example, Caco-2 cultures and recombinant cells.
Primary hepatocytes are the preferred system to study in vitro hepatic metabolism and induction because they express the full complement of regulating nuclear cofactors, metabolic enzymes, and transporters. In vitro culture conditions, such as culture format and duration, media supplements, etc., can affect the phenotype of cultured hepatocytes (LeCluyse 2001; Pascussi et al., 2001). We found that, among the tested culture conditions, expression and activity of P450 proteins were the highest under high DEX and sandwich culture conditions, which is in line with previous reports (Pascussi et al., 2001). The synthetic glucocorticoid DEX is used commonly as a media supplement to compensate, at least to some extent, for reduced expression of P450 and activity that occur over time in primary culture, especially for CYP3A (Madan et al., 2003). We observed that sensitivity toward induction for CYP3A4 by prototypical inducer RIF in human hepatocyte sandwich cultures decreased markedly at near micromolar DEX (0.1–1 μM) media concentrations (unpublished data), reflecting substantially increased basal CYP3A4 activity caused by pregnane X receptor activation of DEX at these concentrations (McCune et al., 2000). Therefore, DEX at 0.01 μM medium concentration is used in our laboratory for induction studies to provide for adequate sensitivity of induction. Basal P450 activities are monitored routinely to ensure suitability of the culture. In contrast to P450, the highest levels of transporter expression were obtained under the lower DEX control condition. The single gel cultures showed lowest levels in mRNA, protein, and activity among the tested conditions, thus confirming reports that hepatocyte dedifferentiation occurs more rapidly in cultures lacking surrounding matrix protein (LeCluyse, 2001).
P450 and UGT protein levels in cultured hepatocytes from the three donors were generally lower compared with the expression in human liver tissues determined for 17 biopsy samples (Ohtsuki et al., 2012). Expression profiles under 0.1 μM DEX and collagen I showed the smallest differences compared with that of liver tissue. However, the impact of culture conditions on P450 expression was superseded by donor variability. Transporter protein expression was similar to that of human liver, yet showing higher expression of canalicular transporters such as BCRP and MDR1. This could possibly be explained by the extensive formation of bile canalicular networks in sandwich culture (LeCluyse et al., 1994), resulting in the larger formation of canalicular membranes, and consequently higher amounts of these canalicular transporters are found in the plasma membrane fractions prepared from hepatocytes than those in liver tissue.
In the course of in vitro studies on enzyme induction, coregulation of multiple enzymes and transporters by a single inducer can occur simultaneously but to a different extent. Inhibition, direct inhibition, or indirect inhibition by a formed metabolite can be a confounding factor when determining induction by enzymatic activity assessment (McGinnity et al., 2006). Because mRNA expression is a direct and very sensitive parameter to measure nuclear receptor-mediated transactivation, it is discussed and even recommended as the preferred end point for a quantitative assessment of the induction potential of drug candidates, because it allows detection of the intrinsic induction potential of a drug candidate independent of concurrent inhibition (Fahmi et al., 2010). Then again, an increase in mRNA and nuclear activation in response to the induction can be significantly greater than increases in protein and activity (LeCluyse, 2001; Almond et al., 2009), indicating a degree of redundancy in the amount of mRNA synthesized, and therefore such effects on mRNA may lack relevance to the changes of drug metabolism and disposition with respect to pharmacokinetics, drug safety, and efficacy. In addition, there is an even more remote correlation between mRNA expression and protein levels for drug transporters (Ohtsuki et al., 2012). LC-MS/MS-based protein quantification would certainly allow for more accurate evaluation of the intrinsic induction potential of drugs, including those that simultaneously inactivate drug-metabolizing enzymes.
The judgment whether a drug candidate may potentially cause enzyme induction in vivo depends on the application of thresholds that are proposed by regulatory guidance [U.S. Food and Drug Administration, Draft Guidance for Industry: Drug Interaction Studies-Study Design, Data Analysis, and Implications for Dosing and Labeling (http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm072101.pdf); European Medicine Agency, Draft Guideline on the Investigation of Drug Interactions (http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2010/05/WC500090112.pdf]. Exceeding a predefined cut-off value at pharmacological concentrations compared with a positive control in vitro qualifies a compound as an inducer and usually triggers a clinical drug interaction study. Recently, more quantitative approaches to the assessment of in vitro induction data are discussed, applying mechanistic in silico models instead of empirical threshold values (Almond et al., 2009) to better predict the in vivo induction potential of a drug candidate. These in silico approaches include physiologically based pharmacokinetics and a combination of in vitro and in vivo pharmacokinetic data. For such approaches, however, absolute data on in vitro protein expression of a set of the most relevant drug-metabolizing enzymes and transporters in response to treatment with a drug candidate would certainly be of great benefit. This would be even more the case if such protein concentration data of drug-metabolizing enzymes and drug transporters also were available for the various organs involved in drug metabolism (e.g., kidney and intestine). Therefore, “pharmacoproteomics” (Ohtsuki et al., 2011) is a promising new approach adding to the established methodologies in drug research and development. Here, we demonstrated that LC-MS/MS quantification of major proteins involved in hepatic metabolism disposition can be performed in the course of common induction studies. Considering the data of this study together with the results of the accompanying study on protein expression in human liver tissue (Ohtsuki et al., 2012), valuable information is provided for a future better understanding and quantitative prediction of in vivo induction effects by mechanistic and physiologically based pharmacokinetics modeling.
In summary, we used novel LC-MS/MS assays, in conjunction with common enzyme activity and mRNA techniques, to quantify major drug-metabolizing enzymes and transporters in cultured primary human hepatocytes to elucidate differences in expression profiles in response to various culture conditions and in comparison to human liver. The methods described herein are suitable for absolute quantification of the expression of these molecular entities during in vitro induction studies.
Authorship Contributions
Participated in research design: Schaefer, Ohtsuki, Ishiguro, Matsumaru, Terasaki, and Ebner.
Conducted experiments: Schaefer, Kawakami, Inoue, Liehner, Saito, and Sakamoto.
Performed data analysis: Schaefer, Ohtsuki, Kawakami, Inoue, Liehner, Saito, Sakamoto, and Ishiguro.
Wrote or contributed to the writing of the manuscript: Schaefer, Ohtsuki, Terasaki, and Ebner.
Acknowledgments
We thank Michelle Werner, Britta Koch, and Ikumi Washio for assistance with hepatocyte culture, Ikue Kitahara for assistance with mRNA analysis, and Takashi Igarashi for collaborative support.
Footnotes
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
ABBREVIATIONS:
- P450
- cytochrome P450
- ABC
- ATP-binding cassette protein
- BCRP
- breast cancer resistance protein
- BSEP
- bile salt export pump
- CNT
- concentrative nucleoside transporter
- DEX
- dexamethasone
- ECM
- extracellular matrix protein
- ENT
- equilibrative nucleoside transporter
- γ-GTP
- γ-glutamyl transpeptidase
- HuHep
- human hepatocyte preparation
- LC-MS/MS
- liquid chromatography coupled to tandem mass spectrometry
- LOQ
- limit of quantification
- MATE
- multidrug and toxin extrusion protein
- MCT
- monocarboxylate transporter
- mio
- million
- MDR
- multidrug resistance protein
- MRP
- multidrug resistance-associated protein
- MRM
- multiple reaction monitoring
- NTCP
- sodium/taurocholate cotransporting polypeptide
- OATP
- organic anion-transporting polypeptide
- OCT
- organic cation transporter
- OME
- omeprazole
- PB
- phenobarbital
- RIF
- rifampicin
- RT-PCR
- reverse transcription-polymerase chain reaction
- UGT
- UDP-glucuronosyltransferase
- WME
- Williams' medium E
- Ct
- cycle threshold.
- Received August 8, 2011.
- Accepted October 5, 2011.
- Copyright © 2012 by The American Society for Pharmacology and Experimental Therapeutics