Abstract
To predict transporter-mediated drug disposition using physiologically based pharmacokinetic models, one approach is to measure transport activity and relate it to protein expression levels in cell lines (overexpressing the transporter) and then scale these to via in vitro to in vivo extrapolation (IVIVE). This approach makes two major assumptions. First, that the expression of the transporter is predominantly in the plasma membrane. Second, that there is a linear correlation between expression level and activity of the transporter protein. The present study was conducted to test these two assumptions. We evaluated two commercially available kits that claimed to separate plasma membrane from other cell membranes. The Qiagen Qproteome kit yielded very little protein in the fraction purported to be the plasma membrane. The Abcam Phase Separation kit enriched the plasma membrane but did not separate it from other intracellular membranes. For the Abcam method, the expression level of organic anion-transporting polypeptides (OATP) 1B1/2B1 and breast cancer resistance protein (BCRP) proteins in all subcellular fractions isolated from cells or human liver tissue tracked that of Na+-K+ ATPase. Assuming that Na+-K+ ATPase is predominantly located in the plasma membrane, these data suggest that the transporters measured are also primarily located in the plasma membrane. Using short hairpin RNA, we created clones of cell lines with varying degrees of OATP1B1 or BCRP expression level. In these clones, transport activity of OATP1B1 or BCRP was highly correlated with protein expression level (r2 > 0.9). These data support the use of transporter expression level data and activity data from transporter overexpressing cell lines for IVIVE of transporter-mediated disposition of drugs.
Introduction
Physiologically based pharmacokinetic models are increasingly used in drug development to predict drug disposition in humans based on in vitro data. Such in vitro to in vivo extrapolation (IVIVE) has been most successful for drugs cleared predominantly by cytochrome P450 metabolism. This is because scaling factors relating the expression of the enzyme in vitro (in microsomes) to that in vivo are available. The increased focus on designing compounds that are metabolically stable has resulted in the development of drugs that are significantly cleared by transporters. However, until recently, the corresponding scaling factors for transporters were lacking primarily because the traditional method for measurement of transporter expression, namely, Western blotting, yields only relative expression data (because pure standards of membrane proteins are not available) that are semiquantitative at best. To overcome these deficiencies of Western blotting, proteomics, based on liquid chromatography with tandem mass spectrometry (LC-MS/MS) and surrogate unique peptide(s), is increasingly being used to quantify transporter protein expression (Prasad and Unadkat 2014). This method is both quantitative and independent of the availability of pure transport protein standards.
Irrespective of the method used, it should ideally measure transporter protein that is active and present in the plasma membrane. However, the current approaches used for transporter proteomics either do not isolate the plasma membrane from other intracellular membranes (Prasad et al., 2013, 2014; Qiu et al., 2013) or give poor yield (Suski et al., 2014). In addition, for IVIVE, the method assumes that expression and activity are directly proportional to each other. This makes it possible to extrapolate the activity and expression of the transporter measured in cell lines (overexpressing the transporter) to that in vivo where the expression of the transporter may be variable and likely lower. Therefore, our study determined whether 1) pure plasma membrane can be isolated from tissues and cells using commercially available kits and 2) the expression of transporter protein (in total membrane) is correlated with activity when the former is varied.
Materials and Methods
Short hairpin RNA (shRNA) or scrambled shRNA lentiviral particles (for ABCG2 and SLCO1B1), polybrene and puromycin dihydrochloride were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The ProteoExtract native membrane protein extraction kits were procured from Calbiochem (Temecula, CA). The plasma membrane protein extraction kits, ab65400 and Qproteome, were procured from Abcam (Cambridge, MA) and Qiagen (Hilden, Germany), respectively. The protein quantification BCA kit, dithiotreitol, iodoacetamide, and MS grade trypsin were purchased from Pierce Biotechnology (Rockford, IL). Synthetic signature peptides for organic anion-transporting polypeptide (OATP) 1B1/2B1 and breast cancer resistance protein (BCRP) (Prasad et al., 2013, 2014) were obtained from New England Peptides (Boston, MA). The corresponding stable isotope-labeled peptides were obtained from Thermo Fisher Scientific (Rockford, IL). Tritium-labeled estradiol 17β glucuronide ([3H]-EG) (specific activity: 50 Ci/mmol, concentration: 1 mCi/ml) was purchased from American Radiolabeled Chemicals (St. Louis, MO). ScintiVerse BD Cocktail liquid scintillant was purchased from Fisher Scientific. High-pressure liquid chromatography–grade acetonitrile was purchased from Fischer Scientific (Fair Lawn, NJ). Estrone 3-sulfate sodium salt and formic acid were purchased from Sigma-Aldrich (St. Louis, MO). All reagents were analytic grade.
Liver tissue samples from adult healthy donors were obtained from the human liver bank of the School of Pharmacy, University of Washington. Due to their anonymous nature, the use of these samples was classified as nonhuman subjects research by the University of Washington Human Subjects Division. BCRP-expressing Madin-Darby canine kidney II (MDCKII) cells were a gift from Dr. Qingcheng Mao, University of Washington. OATP1B1-expressing Chinese hamster ovary (CHO) cells were a gift from Dr. Bruno Stieger, University of Zurich. OATP1B1-expressing MDCKII cells were a gift from Dr. Yuichi Sugiyama, University of Tokyo. OATP2B1-expressing HEK293 cells were a gift from Dr. Markus Keiser, University of Greifswald.
Plasma Membrane Protein Extraction
Two different commercially available kits were used to isolate plasma membrane: Qproteome (Qiagen, Hilden, Germany) and ab65400 (Abcam, Cambridge, MA).
For the Qproteome protocol, 1 × 107 HEK293 cells expressing OATP2B1 were centrifuged for 5 minutes at 450g, and the cell pellet was washed with phosphate-buffered saline (PBS). The washed cell pellet was resuspended in 2 ml of Lysis Buffer PM without protease inhibitors and centrifuged again for 5 minutes at 450g. The resulting cell pellet was resuspended in 500 μl of Lysis Buffer PM with protease inhibitors and incubated for 15 minutes at 4°C with gentle vortexing. We added 2.5 μl of Lysis Solution PL to the cell suspension, mixed by briefly vortexing, and then incubated the suspension for 5 minutes at 4°C. Complete cell disruption was achieved by multiple aspiration through a needle (26 or 21 gauge), and the lysate was centrifuged at 12,000g and 4°C for 20 minutes (Pellet (12k × g); see Fig. 1A).
Protein expression in sequential steps used in isolation of plasma membrane with the (A) Qproteome (Qiagen) or (B) Abcam kit. OATP2B1 expression tracks expression of Na+-K+ ATPase, a plasma membrane marker, but not that of other markers, namely, GS28 (Golgi), calreticulin (ER), Tim23 (mitochondria), GAPDH (cytosol), or histone 3 (nucleus). In the Qproteome and Abcam kits, the eluent and fraction 4 are supposed to respectively contain only the plasma membrane.
The supernatant was transferred to a new microcentrifuge tube, and 20 μL of the reconstituted Binding Ligand PBL was added. The equilibrated Strep-Tactin magnetic beads were added to the reaction mix and incubated with gentle agitation for 60 minutes at 4°C. The tube was kept on a magnetic separator for 1 minute, and the supernatant was removed (Bead supernatant). We added 500 μl of Lysis Buffer PM of the kit with protease inhibitors to the pellet, and the beads were resuspended and incubated on ice for 5 minutes. The tube was again placed on a magnetic separator for 1 minute, and the supernatant was collected (washing). We added 500 μl of Elution buffer PME to the pellet, mixed well, and incubated the tube on ice for 5 minutes. The tube was placed on a magnetic separator for 1 minute, and the supernatant and beads were collected (Eluent and Beads, respectively). Samples from each of these steps were analyzed by LC-MS/MS for OATP2B1 and subcellular marker expression (see Fig. 1).
For the Abcam protocol, 3 × 108 HEK293 cells expressing OATP2B1 or 1 g of human liver tissue (HL-105, HL-115, or HL-129) were used. The cells were washed with 1 ml of ice-cold PBS, and the pellet was resuspended in 1 ml of the homogenization buffer mix of the kit in an ice-cold Dounce homogenizer. Homogenization was performed on ice for 30–50 times. The liver tissue was homogenized in 3 times volume of the 1X Homogenization Buffer Mix until it was completely lysed (40 times). The homogenate was transferred to a 1.5-ml tube and centrifuged at 700g for 10 minutes at 4°C. The supernatant (fraction 1; see Fig. 1B) was collected in a new vial and centrifuged at 8,000g for 10 minutes at 4°C. The pellet was collected (fraction 2), and the supernatant was transferred to a new vial and centrifuged at 20,000g for 30 minutes at 4°C. The supernatant (fraction 3) was collected, and the pellet (representing total cellular membrane protein) was resuspended in 200 μl of the Upper Phase Solution of the kit, to which 200 μl of warmed (30°C) Lower Phase Solution was added. The mix was vortexed and incubated on ice for 20 minutes. The tube was centrifuged at 1000g for 5 minutes, and the upper phase was carefully transferred to a new tube kept on ice.
The upper phase was then extracted by adding 100 μl of the warmed (30°C) lower phase solution. Then, after mixing well, it was incubated on ice for 20 minutes and centrifuged at 1000g for 5 minutes. Finally, the upper phase was carefully collected and diluted in 5 volumes of water and incubated overnight on ice. On day 2, the sample was centrifuged at 15,000g for 10 minutes at 4°C, and the pellet was recovered as the plasma membrane fraction (fraction 4). The lower phase was also collected (fraction 5). Each of these listed fractions was analyzed by LC-MS/MS for OATP2B1 and subcellular marker expression (see Fig. 1B).
Generation of Cells Expressing Varying Levels of OATP1B1 or BCRP Using shRNA
For these experiments, OATP1B1-expressing CHO cells were selected over OATP1B1-expressing MDCKII cells because of the higher expression of OATP1B1 in these cells. OATP1B1 (CHO)– or BCRP (MDCKII)–expressing cells were transduced with the shRNA or control (scrambled) lentiviral particles using the manufacturer’s protocol with some optimization. Briefly, before viral infection, cells were plated in a 12-well plate for 24 hours with 1 ml of Dulbecco’s modified Eagle’s medium including fetal bovine serum and penicillin/streptomycin. On day 2, the medium was removed and replaced with Dulbecco’s modified Eagle’s medium containing Polybrene (sc-134220) (final concentration, 5 μg/ml for MDCKII cells and 10 μg/ml for CHO cells). Then varying amounts (12.5–125 µl, each µl has about 5000 lentiviral particles carrying shRNA) of lentiviral particles were added to the cells. On day 3, the medium was removed and replaced by 1 ml of complete medium (without Polybrene), and the cells were grown for the next 24 hours. The cells were split on day 4 and incubated for 48 hours in complete medium. Stable clones expressing shRNA (or scrambled shRNA) were selected by adding an optimized concentration of puromycin dihydrochloride: 5 μg/ml for OATP1B1 (CHO cells) or 10 μg/ml for BCRP (MDCKII cells).
Transporter Activity Assay
OATP1B1.
[3H]-estradiol glucuronide ([3H]-EG) was used as a substrate to measure OATP1B1 transport activity. CHO cells grown in T75-flasks until 80%–90% confluent were harvested using trypsin, seeded into 12-well poly-d-lysine coated plates (density of approximately 75,000 cells per well), and allowed to grow for 2 days. Subsequently, the medium was removed and replaced with 10 mM sodium butyrate containing complete medium, and the cells were allowed to grow for the next 24 hours. Then the cells were washed with 1 ml/well of Dulbecco’s phosphate-buffered saline buffer followed by a 10 minutes before incubation with 300 µl of Hank’s balanced salt solution (HBSS) buffer with or without a specific inhibitor of OATP1B1 (1 µM estrone sulfate). Then the cells were incubated for 5 minutes with 10 nM [3H]-EG. The uptake was quenched by washing the cells 3 times with ice-cold Dulbecco’s phosphate-buffered saline buffer (1 ml each).
Next, the cells were lysed with 160 µl of 0.2% (w/w) aqueous sodium dodecyl sulfate solution. We used 30 µl of this lysate solution for total protein estimation using the BCA method, and 90 µl to analyze total radioactivity by Tri-Carb Liquid Scintillation Counters (PerkinElmer). Transporter and Na+-K+ ATPase expression was measured by LC-MS/MS as described herein.
To confirm that the decrease in OATP1B1 transport activity in the knockdown cells was due to a change in Vmax and not in Km, these parameters were estimated in control (transduced with scrambled shRNA) and maximal knockdown cells. A simplified two-point approach was used to determine the Vmax and Km. Preliminary data indicated that [3H]-EG uptake (Vmax) was achieved in the presence of 120 µM EG. Therefore, the velocity of [3H]-EG uptake by the cells was determined in the presence of 120 µM EG (Vmax) or 20 nM EG (Vmax/Km). The reported values are mean ± S.D. of triplicates.
BCRP.
The cellular accumulation of Hoechst 33342 dye was used to indirectly measure BCRP efflux activity as described elsewhere (Kim et al., 2002). Hoechst 33342 emits blue fluorescence when bound to double-strand DNA. MDCKII cells grown in T75-flasks until 80%–90% confluent were harvested using trypsin, seeded into black 96-well plates (density of approximately 40,000 cells per well), and allowed to adhere overnight. Then the culture medium was removed, and the cell monolayer was washed twice with Krebs-Hepes buffer and resuspended in 100 µl buffer. LY335984 ((2R)-1-{4-[(1aR,10bS)-1,1-Dichloro-1,1a,6,10b-tetrahydrodibenzo[a,e] cyclopropa[c][7]annulen-6-yl}-3-(quinolin-5-yloxy)propan-2-ol) (0.3 μM) and Ko143 ((3S,6S,12aS)-1,2,3,4,6,7,12,12a-Octahydro-9-methoxy-6-(2-methylpropyl)-1,4-dioxopyrazino[1′,2′:1,6]pyrido[3,4-b]indole-3-propanoic acid 1,1-dimethylethyl ester) (1 μM) were used to selectively inhibit MDR1 (endogenous) and BCRP activity, respectively. After 30 minutes of preincubation with inhibitors at 37°C, the cells were incubated with varying concentrations of Hoechst 33342 solution (0.5 μM, 1 μM, 2.5 μM, 5 μM, 10 μM, and 30 μM). The fluorescence (excitation at 355 nm and an emission at 460 nm) in each well was immediately measured at 4-minute intervals (at 37°C) up to 40 minutes.
After correcting for background fluorescence, the rate of cellular uptake of Hoechst 33342 was assessed by taking the linear slope of the difference between the fluorescence signal in the presence of both the LY335984 and Ko143 inhibitors (passive diffusion) and the LY335984 inhibitor (BCRP efflux plus passive diffusion). The Vmax and Km values were estimated as described earlier for OAPT1B1. The reported values are mean ± S.D. of triplicates.
For expression-activity correlation, various BCRP knockdown clones were incubated with 2.5 μM Hoechst 33342 for 40 minutes, and activity was assessed as the fluorescence Au (arbitrary fluorescence unit) difference between passive diffusion (Ko14 + LY335984 inhibitor) and BCRP efflux (LY335984 inhibitor). Transporter and Na+-K+ ATPase expression was measured by LC-MS/MS as described herein.
LC-MS/MS Quantification of Transporters and Membrane Markers
The protein expression of BCRP, OATP1B1/2B1, and subcellular markers (e.g., Na+-K+ ATPase) in cell lines or liver tissues was quantified using LC-MS/MS and surrogate peptides as described previously elsewhere (Prasad et al., 2013, 2014). The corresponding heavy labeled peptide was used as an internal standard. For the expression-activity correlation study, the total membrane protein was extracted using the Calbiochem kit as described elsewhere (Prasad et al., 2013, 2014). The multiple reaction monitoring parameters for quantification of the subcellular marker proteins are shown in Supplemental Table 1. Protein concentrations in the individual samples were determined using the BCA protein assay kit (Thermo Fisher Scientific).
Results
Plasma Membrane Isolation.
The expressions of OATP2B1 and signature proteins representing various intracellular organelles (e.g., Na+-K+ ATPase for plasma membrane, GS28 for Golgi, calreticulin for endoplasmic reticulum, TIM23 for mitochondria, GAPDH for cytosol, and histone 3 for nucleus) in all subcellular fractions showed that neither method could separate plasma membrane from other cellular membranes (Fig. 1, A and B).
The Qproteome method yielded very little protein and no enrichment in the fraction purported to be the plasma membrane (eluent fraction, Fig. 1A). In contrast, the plasma membrane fraction (fraction 4, Fig. 1B) using the Abcam method showed enrichment (∼2.5-fold) of this membrane but was significantly contaminated by the Golgi and mitochondria. Interestingly, for both methods, the expression of OATP2B1 tracked the expression of Na+-K+ ATPase (Fig. 1, A and B) in all cellular fractions. Because Na+-K+ ATPase is predominantly expressed in the plasma membrane (Padilla-Benavides et al., 2010), these data suggest that OATP2B1 is also primarily localized in the plasma membrane.
Because of the higher protein yield, we used the Abcam method to determine whether the previous observation could be replicated for cells expressing OATP1B1 (MDCKII cells), OATP2B1 (HEK293 cells), BCRP (MDCKII cells), or for human liver tissues (n = 3). Indeed, the expression of OATP1B1, OATP2B1, and BCRP tracked the expression of Na+-K+ ATPase in all subcellular fractions isolated from overexpressed cells and the three individual human livers (Fig. 2).
Transporter protein expression (normalized to the expression of Na+-K+ ATPase) for sequential steps used in the isolation of plasma membrane using the Abcam kit. The y-axis represents the ratio (%) of peak areas of transporter versus membrane marker (Na+-K+ ATPase) in each fraction after each is normalized to the homogenate. Cells expressing BCRP or OATP1B1/2B1 as well as human liver tissues (HL 1–3) were used. Experiments 1 and 2 represent two independent experiments. Transporter protein expression tracked that of Na+-K+ ATPase. Fraction 4 is supposed to contain only the plasma membrane. BCRP expression was not detectable in fraction 3 or 5.
Expression-Activity Correlation in Cells Expressing OATP1B1 and BCRP.
OATP1B1 and BCRP transporter activity was highly correlated with the expression of these transporters in knockdown cells (r2 > 0.9) (Fig. 3). As expected, the decrease in transporter activity in the knockdown cells was explained by a decrease in transporter Vmax and not a change in Km (Table 1). The Km values were comparable to those reported by others (Yamazaki et al., 2005; Gui et al., 2008).
Protein expression and transport activity were highly correlated in various clones of cells expressing OATP1B1 (A) or BCRP (B) created by gene knockdown. OATP1B1 and BCRP transport activity were measured by [3H]-EG uptake and intracellular accumulation of Hoechst 33342 dye, respectively. Transport activity was expressed as a percentage of that observed in control cells transduced with scrambled shRNA.
Gene knockdown of OATP1B1 or BCRP significantly reduces the Vmax of the transporter without affecting its Km
Discussion
Most of the reports on transporter quantification (including from our laboratory) are based on measurements of total transporter protein (plasma membrane plus intracellular) and not just that present in the plasma membrane (Kawakami et al., 2011; Prasad et al., 2013, 2014; Qiu et al., 2013). Theoretically, the latter is more relevant because it is the transporter expressed in the plasma membrane that is responsible for compound cellular efflux or uptake and therefore needed in extrapolating the in vitro transporter activity in cell lines to that in vivo. Therefore, we investigated methods that could be used to routinely separate plasma membranes from other intracellular membranes. Though sucrose gradients (Schaefer et al., 2012) and biotinylation have been used for this purpose, both methods have limitations, as the former requires a large quantity of cells or tissue and the latter is best applied to intact cells but not tissues (Elia, 2012).
Therefore, we evaluated two kits that claimed to separate plasma membrane (in tissues and cells) from intracellular membranes: Qiagen Qproteome and Abcam ab65400. Neither method was able to separate the plasma membrane from intracellular membranes. The Qproteome kit gave poor yield of protein purported to be the plasma membrane. In contrast, when we used the Abcam kit, the expression of OATP1B1, OATP2B1, and BCRP tracked that of Na+-K+ ATPase in overexpressed cell lines as well as in HL tissue. As Na+-K+ ATPase has been shown to be predominantly localized in the plasma membrane (Padilla-Benavides et al., 2010), our data suggest that the expression of OATP1B1/2B1 and BCRP is also predominantly in the plasma membrane. Immunolocalization data of OATP and BCRP in cell lines support this conclusion (Kopplow et al., 2005; Xia et al., 2005).
Because the two kits failed to isolate pure plasma membrane and we are not aware of any other routine method that can do so, an alternative and relevant question would be, is the plasma membrane expression of the transporter as a percentage of total cellular expression of the transporter in overexpressed cell lines comparable to that in liver tissue? If it is, we can use with confidence the activity and expression of the transporter measured in cell lines for IVIVE. Assuming that Na+-K+ ATPase is predominately located in the plasma membrane, our data suggest that this assumption is valid. Therefore, we used Na+-K+ ATPase as a normalizing factor in analyzing the expression-activity correlation described herein.
Another assumption made when conducting IVIVE of transporter activity and expression in cell lines is that transporter expression and activity are linearly correlated. Using knockdown with lentiviral shRNA, we showed that transporter expression was highly correlated with activity (r2>0.9) for cells expressing OATP1B1 and BCRP. Moreover, this was due to a reduction in Vmax of the transporter and not due to a change in Km.
Collectively, our data support the use of cell lines overexpressing transporters for IVIVE of transporter-based disposition of drugs. Whether such IVIVE will be accurate needs to be verified.
Acknowledgments
The authors thank Dr. Qingcheng Mao, Department of Pharmaceutics, University of Washington, for BCRP-expressing MD CKII cells, Dr. Bruno Stieger, University of Zurich, for OATP1B1-expressing CHO cells, Dr. Yuichi Sugiyama, University of Tokyo, for OATP1B1-expressing MD CKII cells, and Dr. Markus Keiser, University of Greifswald for OATP2B1-expressing HEK293 cells.
Authorship Contributions
Participated in research design: Kumar, Prasad, Patilea, Evers, Gupta, Salphati, Hop, Unadkat.
Conducted experiments: Kumar, Prasad, Patilea.
Performed data analysis: Kumar, Prasad, Patilea, Unadkat.
Wrote or contributed to the writing of the manuscript: Kumar, Prasad, Patilea, Evers, Gupta, Salphati, Hop, Unadkat.
Footnotes
- Received October 17, 2014.
- Accepted December 8, 2014.
A.G., L.S., R.E., and C.E.C.A.H. contributed equally to this work.
This study was supported by the University of Washington Research Affiliate Program on Transporters (UWRAPT) sponsored by AstraZeneca, Genentech, and Merck & Co., Inc. (http://sop.washington.edu/department-of-pharmaceutics/research-affiliate-program-on-transporters-uwrapt/).
↵
This article has supplemental material available at dmd.aspetjournals.org.
Abbreviations
- AU
- arbitrary fluorescence unit
- BCRP
- breast cancer resistance protein
- CHO
- Chinese hamster ovary cells
- EG
- estradiol 17β glucuronide
- [3H]-EG
- tritium-labeled estradiol 17-β glucuronide
- HEK293
- human embryonic liver 293 cells
- HL
- human liver cells
- IVIVE
- in vitro to in vivo extrapolation
- Ko143
- (3S,6S,12aS)-1,2,3,4,6,7,12,12a-Octahydro-9-methoxy-6-(2-methylpropyl)-1,4-dioxopyrazino[1′,2′:1,6]pyrido[3,4-b]indole-3-propanoic acid 1,1-dimethylethyl ester
- LC-MS/MS
- liquid chromatography with tandem mass spectrometry
- LY335984
- (2R)-1-{4-[(1aR,10bS)-1,1-Dichloro-1,1a,6,10b-tetrahydrodibenzo[a,e] cyclopropa[c][7]annulen-6-yl}-3-(quinolin-5-yloxy)propan-2-ol
- MDCKII
- Madin-Darby canine kidney II cells
- OATP
- organic anion-transporting polypeptide
- PBS
- phosphate-buffered saline
- shRNA
- short hairpin RNA
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics