1521-009X/45/6/604–611$25.00 https://doi.org/10.1124/dmd.116.074740DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 45:604–611, June 2017Copyright ª 2017 by The American Society for Pharmacology and Experimental Therapeutics

Coproporphyrin-I: A Fluorescent, Endogenous Optimal ProbeSubstrate for ABCC2 (MRP2) Suitable for Vesicle-Based MRP2

Inhibition Assay s

Ravindranath Reddy Gilibili, Sagnik Chatterjee, Pravin Bagul, Kathleen W. Mosure,Bokka Venkata Murali, T. Thanga Mariappan, Sandhya Mandlekar, and Yurong Lai1

Pharmaceutical Candidate Optimization, Biocon Bristol-Myers Squibb R&D Center (BBRC), Syngene International Ltd., Bangalore,India (R.R.G., S.C., P.B., B.V.M., T.T.M.); Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Company, Princeton, NewJersey (K.W.M., Y.L.); and Pharmaceutical Candidate Optimization, Biocon Bristol-Myers Squibb R&D Center (BBRC), Bangalore,

India (S.M.)

Received December 21, 2016; accepted March 17, 2017

ABSTRACT

Inside-out–oriented membrane vesicles are useful tools to investi-gate whether a compound can be an inhibitor of efflux transporterssuch as multidrug resistance–associated protein 2 (MRP2). How-ever, because of technical limitations of substrate diffusion and lowdynamic uptake windows for interacting drugs used in the clinic,estradiol-17b-glucuronide (E17bG) remains the probe substrate thatis frequently used in MRP2 inhibition assays. Here we recapitulatedthe sigmoidal kinetics of MRP2-mediated transport of E17bG,with apparent Michaelis-Menten constant (Km) and Vmax values of170 617 mM and 1447 6 137 pmol/mg protein/min, respectively.The Hill coefficient (2.05 6 0.1) suggests multiple substrate bindingsites for E17bG transport with cooperative interactions. UsingE17bG as a probe substrate, 51 of 97 compounds tested (53%)showed up to 6-fold stimulatory effects. Here, we demonstrate for

the first time that coproporphyrin-I (CP-I) is a MRP2 substrate inmembrane vesicles. The uptake of CP-I followed a hyperbolicrelationship, adequately described by the standard Michaelis-Menten equation (apparent Km and Vmax values were 7.7 6 0.7 mMand 48 6 11 pmol/mg protein/min, respectively), suggesting theinvolvement of a single binding site. Of the 47 compounds tested,30 compounds were inhibitors of human MRP2 and 8 compounds(17%) stimulated MRP2-mediated CP-I transport. The stimulatorswere found to share the basic backbone structure of the physiologicsteroids, which suggests a potential in vivo relevance of in vitrostimulation of MRP2 transport. We concluded that CP-I could be analternative in vitro probe substrate replacing E17bG for appreciatingMRP2 interactions while minimizing potential false-negative resultsfor MRP2 inhibition due to stimulatory effects.

Introduction

Multidrug resistance–associated protein 2 [MRP2 (also calledABCC2)], a member of the ATP-binding cassette (ABC) transporters,is expressed exclusively on the apical membrane of hepatocytes,enterocytes, and kidney proximal tubule cells. MRP2 functionaldeficiency caused by the ABCC2 gene mutation in humans is themolecular basis of conjugated hyperbilirubinemia, known as Dubin-Johnson syndrome. MRP2 also transports many structurally diversedrugs and their metabolites and plays a key role in drug disposition anddetoxification processes (Nies and Keppler, 2007). For example,impaired function of MRP2 could increase the systemic exposure ofpravastatin due to the increased absorption and the reduced biliary and/orurinary excretion (Niemi et al., 2006). As a result, inhibition of MRP2can cause accumulation of compounds and/or their metabolites in theliver to reach a toxic level (Isley, 2003; Tang, 2007). Although drug-

drug interactions caused by MRP2 inhibition are still not well-definedin the clinic, given the importance of MRP2 in drug disposition andelimination, it has become one of the emerging transporters of clinicalimportance as recognized by the International Transporter Consortium(ITC) (Zamek-Gliszczynski et al., 2012; Hillgren et al., 2013). ITC hasalso rationalized the importance of MRP2 in hepatotoxicity due to itsrole in regulating drug concentration in the liver (Yoshida et al., 2013).As aforementioned, understanding the interaction of a new chemical

entity with MRP2 becomes important from drug-drug interaction andtoxicity perspective. Currently, two endogenous compounds, namelyestradiol-17b-glucuronide (E17bG) and leukotriene C4 (LTC4), arecommonly used in membrane vesicle uptake assays to understandwhether a compound is an inhibitor of MRP2 (Brouwer et al., 2013).LTC4 displays a very short duration of uptake linearity (only up to2 minutes) in membrane vesicles (Heredi-Szabo et al., 2008), leading toa chance of variability during experiments. These limitations hinder theuse of LTC4 in characterizing MRP2 inhibition. As a result, E17bG isthe endogenous in vitro probe substrate of choice for MRP2 interaction.However, E17bG uptake in membrane vesicles suffers from largeinterlaboratory variability in reported Michaelis-Menten constant (Km)

1Current affiliation: Drug Metabolism, Gilead Sciences, Inc. 333 Lakeside Drive,Foster City, California. E-mail yurong.lai@gilead.com.

https://doi.org/10.1124/dmd.116.074740.s This article has supplemental material available at dmd.aspetjournals.org.

ABBREVIATIONS: ABC, ATP binding cassette; CDF, carboxy-dichlorofluorescein; CP-I, coproporphyrin I; DMSO, dimethylsulfoxide; E17bG,estradiol-17b-glucuronide; GCDCA, glycochenodeoxycholic acid; ITC, International Transporter Consortium; Km, Michaelis-Menten constant; LC-MS/MS, liquid chromatography-tandem mass spectrometry; LTC4, leukotriene C4; MRP2, multidrug-resistance associated protein 2; TDCA,taurodeoxycholic acid.

604

http://dmd.aspetjournals.org/content/suppl/2017/03/21/dmd.116.074740.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from

and Vmax values (Supplemental Table 1) (Borst et al., 2006). E17bGhas displayed homotropic cooperativity with MRP2-mediated up-take, yielding sigmoidal kinetics in membrane vesicles (SupplementalTable 1). To explain cooperative interactions, a two–binding site theoryhas been proposed for MRP2: E17bG binds to the substrate binding siteS that mediates the transport of it; and as E17bG concentration increasesit binds to the modulator site M, modulating the affinity of the transportsite (Fig. 1). In addition, complex modulations (stimulation and/orinhibition) of MRP2-mediated E17bG transport have been well de-scribed in the literature (Bakos et al., 2000; Evers et al., 2000; Zelceret al., 2003a). The complex modulation suggests that MRP2-interactingcompounds could bind to S sites to inhibit, to M sites to stimulate, or toboth sites to stimulate and inhibit MRP2-mediated transport (also knownas “bell-shape” kinetics) (Zelcer et al., 2003a). This stimulatory effectis considered to be substrate dependent. For example, probenecid isreported to stimulate E17bG uptake but inhibits methotrexate uptake inMRP2 membrane vesicles (Zelcer et al., 2003b). In another study(Herédi-Szabó et al., 2009), benzbromarone, sulfasalazine, probenecid,and indomethacin are reported to stimulate both human and ratMRP2-mediated E17bG transport, albeit to different extents. AsMRP2-mediated E17bG transport displays sigmoidal kinetics, theself-cooperative effects form a unique interaction for each E17bGmodulator pair within the complex binding sites of MRP2 protein.Therefore, the stimulatory effects could potentially mask the MRP2inhibition of many modulating compounds and yield false-negativeMRP2 inhibition. Collectively, an alternative in vitro probe substratethat displays classic Michaelis-Menten kinetics is needed for betterunderstanding of MRP2 interactions with new chemical entities.Coproporphyrin I (CP-I) is one of coproporphyrin byproducts of

heme biosynthesis. In the liver, CP-I is taken up into hepatocytes byorganic anion–transporting polypeptide transporters and effluxed intobile likely by MRP2 (Benz-de Bretagne et al., 2011, 2014; Lai et al.,2016; Shen et al., 2016). As such, a higher proportion of CP-I is secretedinto the urine of subjects with Dubin-Johnson syndrome compared withnormal subjects. This information indicates a central role of MRP2 in

CP-I disposition. However, no in vitro corroboration is provided tosubstantiate these findings to date. In the present study, we aim to do thefollowing: 1) to provide a definitive in vitro proof of involvementof human MRP2 in transport of CP-I; 2) to characterize in depth thekinetics of CP-I transport in human MRP2 vesicles; and 3) to compareknown modulators of human MRP2 in the E17bG assay, with a specialemphasis on reported stimulators.

Materials and Methods

Chemicals and Reagents. Metformin HCl was purchased from RT Corpo-ration (Laramie, WY), rosuvastatin calcium was purchased from AngeneInternational Limited (Hong Kong). Human MRP2-expressing inside-out mem-brane vesicles (protein concentration, 5 mg/ml) derived from Sf9 insect cells werepurchased from GenoMembrane, Co., Ltd (Yokohama, Japan). Assay incubationplates (96 well, ultralow attachment, polystyrene, flat bottom, clear; Costar) werepurchased fromCorning (Corning, NY). Assay plates (96well, black polystyrene)were used for fluorescence measurement. All other chemicals were purchasedfrom Sigma-Aldrich (St. Louis, MO).

Vesicular Transport Assay. MRP2-mediated transport assay was performedusing inside-out membrane vesicles, according to the manufacturer protocol.Briefly, membrane vesicles were diluted to an appropriate concentration withincubation buffer containing 50 mM 4-morpholinepropanesulfonic acid-Tris (pH7.0), 70 mM KCl, and 7.5 mM MgCl2. Diluted membrane vesicles (20ml) weretransferred into individual wells of 96-well plates and coincubated with 0.5 ml oftest substrates (CP-I or E17bG) at 37�C for 3 minutes. Then, the reaction wasinitiated by adding prewarmed incubation buffer (29.5 ml) containing 4 mMATPor 4 mM AMP and 2 mM glutathione. After incubation for a designated time at37�C on rotary shaker (Innova 40; New Brunswick Scientific Co., Edison, NJ) at100 rpm, the reaction was stopped by the addition of 150 ml of cold wash buffercontaining 40 mM 4-morpholinepropanesulfonic acid-Tris (pH 7.0) and 70 mMKCl. The reaction mixture was then transferred into a prewet, 96-well filter plate,which was placed onto a filtration device (FiltrEX 96 Well Filter Plates; CorningTechnologies India Pvt Ltd, Mumbai, India) and filtered by a rapid filtrationtechnique (MultiScreenHTS Vacuum Manifold; EMDMillipore, Billerica, MA).To remove excess reaction mixture, wells were washed five times, each time with200 ml of ice-cold wash buffer. After the final washing step, the filter plate wasdried at room temperature for 1 hour. Next, as an extraction solvent, 100 ml of

Fig. 1. Scheme of the two–binding site model for MRP2.Substrates or modulating compounds bind to the substratebinding site S that mediates the transport of the substrate andcan also bind to the modulator site M that modulates theaffinity of the transport site. The proposed mechanism ofE17bG is that it first binds to the S site followed by the Msite at a higher concentration, resulting in sigmoidal kinetics.conc., concentration.

Coproporphyrin-I as Novel In Vitro Probe Substrate for MRP2 605

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from

0.5% SDS dissolved in Milli-Q water (EMD Millipore) was added to the filterplate wells, and the plate was kept on microplate shaker (VWR International,Radnor, PA) for 15 minutes at 230 rpm. Then the filter plate was centrifuged(Eppendorf Inc., Westbury, NY) for 2 minutes at 778g along with a receiver plateattached to its bottom to receive the filtrate. In case of E17bG, 100 ml ofacetonitrile was used as an extraction solvent. The filtrate was then used toquantify the CP-I or E17bG levels using either a fluorimeter or LC-MS/MS,respectively.

Inhibition of MRP2-Mediated CP-I or E17bG Uptake in MembraneVesicles. Various compounds were selected from the literature as modulatorsto evaluate their inhibitory and/or stimulatory effects on MRP2-mediatedE17bG or CP-I transport in membrane vesicles. All modulators were tested atdifferent concentrations. The final assay concentrations of CP-I and E17bGwere 5 and 50 mM, respectively. For E17bG as a substrate, modulators weretested at 20 and 200 mM to determine their inhibitory or stimulatory effects onMRP2-mediated transport. With CP-I as a substrate, modulators were typicallytested at 0.1, 1, 10, 50, 100, 250, 500, and 1000 mM to determine IC50 values.All compound stock solutions were prepared in dimethylsulfoxide (DMSO),were spiked directly into assay mixture containing MRP2 vesicle protein (forCP-I and E17bG are 50 and 25 mg, respectively), and were preincubated for3 minutes at 37�C. The solubility of inhibitors and the stability of incubation athigh concentrations were monitored during the experiments. Then the reactionwas initiated with the addition of substrate containing ATP or AMP. The rest ofthe assay was conducted as described in the above section Vesicular TransportAssay. Controls included within each experiment were with substrate alone(either CP-I or E17bG) in the presence of ATP or AMP. ATP-dependenttransport of CP-I or E17bG was measured in the presence of modulators andcompared with control data. The effect of DMSO on the CP-I transport wasalso evaluated. It was observed that the final DMSO concentrations up to 2%have negligible effect on CP-I transport (data not shown). In this study, theDMSO content used was not more than 1.15%.

Fluorimetry Analysis of CP-I. Fluorescence measurements were conductedwith a microplate reader (SpectraMaxM2e; Molecular Devices, Sunnyvale, CA),using 401 and 595 nm, respectively, as excitation and emission wavelengths forCP-1 ATP-dependent net transport were calculated by subtracting the valuesobtained in the presence of AMP from those obtained in the presence of ATP. Allthe experiments related to CP-I were conducted in dim light to minimizefluorescent bleaching.

LC-MS/MS Analysis of E17bG. LC-MS/MS analysis of E17bG wasperformed using a Waters Corporation (Milford, MA) UPLC system coupledwith a Triple Quad 5500 System (AB Sciex, Framingham, MA), fitted withElectrospray ionization source. Naphthyl glucuronide was used as internalstandard (IS). The analyte and IS were separated on a BEH C18-A 50 �2.1 mm column (Waters Corporation). An elution gradient with two solvents wasused: 1) water with 0.1% formic acid; and 2) acetonitrile with 0.1% formic acid.A linear gradient was performed as follows: 0.2 minute at 5% solvent B; in0.5 minute, solvent B was increased from 5% to 95% and remained constant at95% solvent B for 1.20 minutes. Then in 1.5 minutes solvent B was decreasedfrom 95% to 5%, and it remained constant for 0.5 minute. The flow rate was set at0.6 ml/min. The electrospray ionization source conditions were as follows:capillary voltage, �4500 V; drying gas temperature, 450�C; and nebulizer gaspressure, 50 psi (both nebulizer and drying gaswere high-purity nitrogen). E17bGand IS were monitored in negative ion mode with the transition of mass/chargeratios 447→ 112.7 and 547→ 112.7, respectively. Analyst version 1.6.2 was usedfor system control and data processing.

Data Analysis. All data were presented as the mean 6 SD. Fluorescenceintensities from three wells were used to generate the mean and SD. To determinethe kinetic parameters (Km and Vmax) and IC50 values, data were analyzed bynonlinear regression using GraphPad Prism software version 5.02 (GraphPadSoftware, San Diego, CA) with the following appropriate equations.

For CP-I, Km and Vmax were generated from the direct uptake transportmeasurements using the Michaelis-Menten equation:

V ¼ Vmax½s�Kmþ ½s�

For E17bG, the active transport followed sigmoidal fit. Therefore, to obtain bestfit, the Michaelis-Menten equation was modified as given below:

V ¼ Vmax½s�hKmh þ ½s�h

whereV is the velocity (pmol of substrate permilligram of protein perminute), [S]is the substrate concentration inmM, and h is the Hill coefficient characterizing thedegree of cooperativity.

Results

MRP2-Mediated CP-I and E17bG Transport in MembraneVesicles. The transport kinetics of E17bG or CP-I were first character-ized through concentration-dependent transport of E17bG or CP-I inmembrane vesicles overexpressing human MRP2 protein. Prior to thekinetics characterization, ATP-dependent activities and time linearity ofMRP2-mediated CP-I uptake were determined at a concentration of5 mM for different time points, including 5, 10, 15, 30, 45, and60 minutes, at 37�C in the presence of ATP or AMP. MRP2-mediatedCP-I transport in membrane vesicles was linear with time up to45 minutes (Supplemental Fig. 1). CP-I transport was negligible in thepresence of AMP (control) and 35-foldmore in the presence of ATP thancontrol. As such, all inhibition studies were optimized to a 30-minutetime point of incubation. For E17bG, as described previously (Zhanget al., 2016), a 15-minute incubation was chosen as the optimal timepoint of incubations.As depicted in Fig. 2A, the concentration-dependent MRP2-mediated

E17bG uptake in membrane vesicles appears to follow sigmoidalkinetics with cooperative interaction, which suggests the existence ofmultiple binding sites. The apparent Km and Vmax values (mean6 S.D.)for E17bG were 170 6 17 mM and 1447 6 137 pmol/mg protein/min,respectively. The Hill coefficient for E17bGwas 2.056 0.1. In contrast,MRP2-mediated CP-I uptake in membrane vesicles followed a hyper-bolic relationship, as the rate of uptake increased in a linear fashionat low concentrations and was saturated at high concentrations.The kinetics curve could be adequately modeled by the standardMichaelis-Menten equation (Fig. 2B), which suggests the involvementof single binding site in the transport of CP-I. The apparentKm and Vmax

values (mean6 S.D.) for CP-I were 7.76 0.7mMand 486 11 pmol/mgprotein/min, respectively.

Effects of Modulators on MRP2-Mediated E17bG Transport inMembrane Vesicles. To investigate the compound-dependent stimula-tion, we tested total of 97 compounds selected from the literature toevaluate their effect on E17bG transport by MRP2 in membranevesicles. All compounds for E17bG transport were tested at 20 and200 mM. Of those compounds tested, 51 compounds displayedstimulatory effects (.10% of control) with at least one concentration(Fig. 3; Supplemental Table 2). Stimulatory effects appeared to berelated to the concentration of modulators. Mixed effects were alsoobserved, as a subset of compounds showed stimulatory effect at lowerconcentration and inhibitory effect at higher concentration or vice versafor other subset molecules (Fig. 3).

Effects of Modulators on MRP2-Mediated CP-I Uptake inMembrane Vesicles. We further tested 47 compounds that show eitherstimulatory effects in the inhibition of MRP2-mediated E17bG uptakeor bile acids and statins that share the cholestane structure and arepotential stimulators to assess their effects on ATP-dependent CP-Itransport by MRP2. The compounds selected were reported in theliterature for their inhibitory and/or stimulatory effects on E17bGtransport by MRP2. For example, stimulatory effects of 32 of 47 com-pounds selected on MRP2-meidated E17bG transport were reported inliterature (Pedersen et al., 2008; Morgan et al., 2013). In addition, a fewrepresentative statins, bile acids, and steroidal compounds were alsoincluded to test their effects on CP-I uptake transport by MRP2.

606 Gilibili et al.

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from

Of the compounds assessed, 30 compounds were determined to beinhibitors forMRP2-mediated CP-I transport (Table 1). Of 30 inhibitors,7 compounds were obtained with an IC50 of ,100 mM, ranging from11 to 84 mM as benzbromarone . bromosulfophthalein . MK-571 .troglitazone . rifampicin . atorvastatin . losartan potassium. Tencompounds, including nodolol, alpha-bilirubin, metformin, acetamino-phen, taurocholic acid, and cholic acid, displayed no effect on ATP-dependent CP-I uptake transport by MRP2. Eight compounds (17%)were identified as stimulators, comparedwith 32 of 47 compounds whenE17bG was used as the probe substrate. The net changes in stimulationwith these compounds varied from 34% to 181% (Table 1). A bell-shaped inhibition curve of MRP2-mediated CP-I transport was observedwith E17bG, with stimulatory effects appearing at low concentrations(0.1–100 mM) and inhibitory effects appearing at high concentra-tions .100 mM, with an IC50 value of 187 mM.Structure Similarity of Stimulators of MRP2-Mediated CP-I

Uptake in Membrane Vesicles. The chemical structures of thestimulators of MRP2-mediated CP-I uptake in membrane vesicles aredepicted in Fig. 5. These stimulators, except for mitoxantrone and

pyrimethamine, share the general pattern of the 17–carbon ringbackbone, which is the same structure used by other biologicallyimportant steroid molecules including steroid hormones, bile acids, andvitamins (Fig. 5). Mitoxantrone has two identical side chains containingboth amino and hydroxyl moieties that were rich in oxygen- andnitrogen-containing planar tricyclic anthraquinone rings (9.08 pKa).Pyrimethamine (7.34 pKa) belongs to phenylpyrimidines that contain abenzene ring linked to a pyrimidine ring through a C-C or C-N bond(Fig. 5).

Discussion

ABC transporter MRP2 transports its substrates from the inside to theoutside of cells. As a result, identifying a substrate in an intact cellsystem overexpressing MRP2 protein is difficult to accomplish throughthe monitoring of the direct transport of the substrate. Alternatively,substrate transport can be determined using inside-out–oriented mem-brane vesicles demonstrating ATP-dependent transport of a substrateinto the vesicles. For inhibition studies, probe substrates should ideally

Fig. 3. Graphical representation of effect of various modulators on MRP2-mediated E17bG transport. MRP2-mediated E17bG transport was measured in inside-outmembrane vesicles, derived from MRP2-expressing Sf9 insect cells, after incubation with E17bG at 50 mM with and without modulators at 20 and 200 mM. A total of96 modulators were tested and, among those, 51 modulators (53%) stimulated MRP2-mediated E17bG transport by $10%. Sixteen modulators (16%) significantly inhibitedE17bG transport at least either one or both of the concentrations tested. The data were presented as the mean from a single experiment (CV ,20%).

Fig. 2. Concentration-dependent uptake transport of E17bG (A) and CP-I (B) in MRP2-expressing Sf9 membrane vesicles. MRP2 membrane vesicles were incubated at37�C for 15 minutes (E17bG) or 30 minutes (CP-I). ATP-dependent uptake is measured by subtracting uptake in the presence of AMP from that in the presence of ATP. ForMRP2-mediated E17bG uptake, Km and Vmax values were 170 6 17 mM and 1447 6 137 pmol/mg protein/min, respectively. The Hill plot was shown as insert graph, andthe Hill coefficient is 2.046 6 0.1. For MRP2-mediated CP-I uptake, Km and Vmax values were 7.7 6 0.7 mM and 48 6 11 pmol/mg protein/min, respectively. The Hillcoefficient is 1 for the best fit. The data are presented as the mean and SD of a single experiment performed in triplicate.

Coproporphyrin-I as Novel In Vitro Probe Substrate for MRP2 607

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from

be potential victim drugs used in the clinic. However, technicaldifficulties, such as extensive substrate diffusion into membrane vesiclesand the lack of an optimal dynamic window of ATP-dependent uptake,make most xenobiotic substrates unsuitable in the vesicle assay. Currently,LTC4 and E17bG are recommended by the ITC as probe substrates forassessing MRP2 inhibition (Brouwer et al., 2013).In the present study, we first tested 97 compounds in MRP2-mediated

E17bG transport at two concentrations, 20 and 200 mM (Figs. 3 and 4).Of these compounds, 51 (53%) were found to be the stimulators ofMRP2-mediated E17bG transport in at least one of the two concentra-tions tested. Our data are in good agreement with results reportedpreviously, where inmost instances, marketed drugs have been shown tostimulate the MRP2-mediated E17bG transport (Pedersen et al., 2008;Morgan et al., 2013). For some compounds (dacarbazine and sulfasa-lazine), a strong stimulation was observed at a lower concentration(20 mM), whereas E17bG transport was inhibited at a high concentra-tion (200 mM), demonstrating a bell-shaped curve of modulation. Thestimulatory effects of compounds on MRP2-mediated E17bG transportin membrane vesicles limits the use of E17bG, as concerns of false-negative inhibition remain for many modulating compounds. Theconcern for false-negative inhibition becomes important in explaininghyperbilirubinemia caused by MRP2 inhibition. A compound can causehyperbilirubinemia by causing direct hepatotoxicity or by inhibitingMRP2-mediated bilirubin transport. Hyperbilirubinemia mediated byMRP2 inhibition, may not be fatal for the progression of the compounddown clinical development; however, hepatotoxicity signal can prevent

a compound from further progress. Therefore, providing a clini-cally relevant MRP2 probe that is devoid of the risk of identifying acompound as having false-negative inhibitor, becomes very importantfor further use a compound. In this relation, CP-I as a probe substrateidentified 64% of the 47 compounds as inhibitors (Fig. 4). This will helpto differentiate hyperbilirubinemia as a consequence ofMRP2 inhibitionfrom hyperbilirubinemia via hepatotoxicity. In contrast, using E17bG asa probe substrate, it is difficult to classify compounds as inhibitors,because compounds vary significantly in their behavior depending onconcentration (bell-shaped curves). Therefore, in our study, we sepa-rated the compounds into stimulators and inhibitors and/no effects basedon their behavior using E17bG as the probe substrate (Fig. 4).Although the urinary concentration ratio of CP-I to that of sum of CP-I

and CP-III has been used as a surrogate marker to assess MRP2 activityclinically (Benz-de Bretagne et al., 2011), in vitro evidence of thetransport of CP-I by MRP2 was lacking. Herein, we have shown for thefirst time that CP-I is taken up into membrane vesicles overexpressingMRP2 and that the ATP-dependent uptake was about 30-fold higherthan the control, which provided an excellent dynamic window to assessMRP2 inhibitory effects. In addition, the transport of CP-I byMRP2waslinear up to 45 minutes (Supplemental Fig. 1), which provides a timerange long enough to decrease variations, comparedwith LTC4 (Heredi-Szabo et al., 2008). Low Km values of LTC4 (695 nM) results in rapiddeviation from linearity of uptake, therefore, making the experimentsvariation prone. In addition, because CP-I is fluorescent, it is easilymeasurable compared with LTC4, which requires radioactivity analysis.

TABLE 1

List of compounds tested for their modulatory effect against CP-I as substrate, and their classification into type of modulator: compounds with no modulation are listed first,followed by inhibitors in order of increasing IC50, followed by stimulators

Compound Type of modulatorIC50

(mean 6 SD)Comments Compound

Type ofmodulator

IC50

(mean 6 SD)Comments

mM mM

Acetaminophen None ND No effect E17bG Inhibitor/stimulator 188 6 32 Moderate inhibitor; maximumnet stimulation 39% at 1 mM

Etoposide None ND No effect Indomethacin Inhibitor 214 6 14 Weak inhibitorNodolol None ND No effect TDCA Inhibitor 228 6 22 Weak inhibitorTolbutamide None ND No effect Lovastatin Inhibitor 237 6 9 Weak inhibitora-Bilirubin None ND No effect DCA Inhibitor 252 6 2 Weak inhibitorCA None ND No effect Glyburide Inhibitor 280 6 79 Weak inhibitorGCA None ND No effect Glimpride Inhibitor 305 6 43 Weak inhibitorTCA None ND No effect GDCA Inhibitor 325 6 20 Weak inhibitorPravastatin None ND No effect GCDCA Inhibitor 410 6 40 Weak inhibitorMetformin None ND No effect Rosuvastatin Inhibitor 412 6 183 Weak inhibitorBenzbromarone Inhibitor 11 6 0.5 Strong inhibitor Irinotecan Inhibitor 532 6 171 Weak inhibitorBromosulfophthalein Inhibitor 26 6 2 Strong inhibitor Bumetanide Inhibitor 541 6 172 Weak inhibitorMK-571 Inhibitor 39 6 3 Strong inhibitor Gemfibrozil Inhibitor 648 6 71 Weak inhibitorTroglitazone Inhibitor 53 6 3 Strong inhibitor Phenylbutazone Inhibitor 777 6 47 Weak inhibitorRifamycin SV Inhibitor 59 6 1 Strong inhibitor Nimodipine Inhibitor 793 6 35 Weak inhibitorAtorvastatin Inhibitor 60 6 3 Strong inhibitor Probenecid Inhibitor 1121 6 67 Weak inhibitorLosartan potassium Inhibitor 74 6 3 Strong inhibitor Budesonide Stimulator ND Maximum net stimulation

82% at 250 mMRifampicin Inhibitor 83 6 8 Strong inhibitor DHA3S Stimulator ND Maximum net stimulation

62% at 100 mMCDF Inhibitor 130 6 20 Moderate inhibitor Mitoxantrone Stimulator ND Maximum net stimulation

180% at 100 mMSimvastatin Inhibitor 132 6 24 Moderate inhibitor Progesterone Stimulator ND Maximum net stimulation

95% at 50 mMTerfenadine Inhibitor 150 6 2 Moderate inhibitor Testosterone Stimulator ND Maximum net stimulation

79% at 100 mMSulfasalazine Inhibitor 161 6 45 Moderate inhibitor E3S Stimulator ND Maximum net stimulation

34% at 100 mMDiclofenac Inhibitor 168 6 14 Moderate inhibitorCDCA Inhibitor 175 6 21 Moderate inhibitor Pyrimethamine Stimulator ND Maximum net stimulation

49% at 1 mM

CA, cholic acid; CDCA, chenodeoxycholic acid; CDF, carboxydichlorofluorescein; DCA, deoxycholic acid; E3S, estrone 3-sulfate; GCA, glycocholic acid; moderate inhibitor, IC50 = 100–200 mM;ND, not determined; strong inhibitors, IC50 ,100 mM; TCA, taurocholic acid; weak inhibitor, IC50 .200 mM.

608 Gilibili et al.

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from

In addition to LTC4, carboxy-dichlorofluorescein (CDF) is also recom-mended as a probe substrate for MRP2 by the ITC (Brouwer et al.,2013). CDF follows the Michaelis-Menten kinetics and is a fluorescent

substrate; however, it is not endogenous. CDF is administered in cellsystem as di-acetate prodrug, because CDF is a poor permeablemolecule. CDF transport is also reported to be stimulated by different

Fig. 4. Graphical representation of modulators of E17bG and CP-I transport. Pie chart of the segregation of compounds based on their transport properties with either 50 mME17bG (A) or 5 mM CP-I (B) in MRP2-transfected Sf9 membrane vesicles. For E17bG transport, compounds were classified into two classes: stimulators (showing .10%stimulation of E17bG transport), and no effect/inhibitors. For CP-I transport, compounds were classified into three classes: inhibitors, stimulators, and compounds with no effect.

Fig. 5. Chemical structure of stimulators for MRP2-mediated CP-I transport.

Coproporphyrin-I as Novel In Vitro Probe Substrate for MRP2 609

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from

compounds such as verapamil, budesonide, and thioridazine (Muni�cet al., 2011; Kidron et al., 2012). However, it is challenging to furtherunderstand the significance of the stimulation of MRP2-mediated CDFtransport in vivo, because of the permeability and prodrug factor.Unlike E17bG, CP-I uptake in MRP2-expressed membrane vesicles

followedMichaelis-Menten kinetics (Fig. 2B), which indicates that CP-Idoes not have affinity for the modulator site for its own transport. Mostimportantly, although 32 of the 47 compounds chosen to assess CP-Itransport are identified as stimulators in E17bG transport (Fig. 3), eitherin the literature (Pedersen et al., 2008;Morgan et al., 2013) or in our data;only 8 compounds were found to stimulate CP-I transport. Therefore,with CP-I as the probe substrate the percentage of compounds showingin vitro stimulation was largely reduced. Although stimulators of E17bGtransport varied widely in their structural features, stimulators of CP-Itransport displayed common structural feature of 17–carbon ring steroidmoiety (Fig. 5). The other two compounds, mitoxantrone and pyrimeth-amine, are strongly basic. Mitoxantrone forms a head-to-tail dimer andbinds at two opposite grooves of the G-quadruplex. These structuralfeatures are being investigated with in vivo studies to understand thephysiologic relevance of the stimulators with structure similarities.E17bG kinetics is usually explained by distinct modulator and

transport sites (Zelcer et al., 2003a; Borst et al., 2006). The substantialstimulatory effect of the binding of E17bG to themodulator site changesthe transport kinetics to a sigmoidal nature. However, E17bG stimulatedthe transport of CP-I only to a low extent (,40% to 10 mM, whereasE17bG was an inhibitor of CP-I transport at higher concentrations).Therefore, an assumption can be made that the binding of E17bG to themodulator site has a minimum effect on transport of CP-I. This provesthat the stimulation is a probe substrate–dependent phenomenon, furthernecessitating the importance of identifying a clinically relevant probesubstrate.Among the compounds assessed, bile acids inhibitedMRP2-mediated

CP-I transport. Previous studies indicated GCDCA, taurochenodeox-ycholic acid, TDCA, and glycocholic acid as stimulators of MRP-2–mediated E17bG transport (Bodo et al., 2003). With CP-1 as substrate,chenodeoxycholic acid, deoxycholic acid, GCDCA, glycodeoxycholicacid, and TDCA were found to be inhibitors of MRP2. MRP2 is re-sponsible for bile acid–independent bile flow in the case of cholestasis.As a result, accumulated intrahepatic bile acids due to cholestasismight inhibit MRP2 function and result in hyperbilirubinemia,providing another mechanism of bilirubin increase apart from directhepatotoxicity.Although diclofenac, indomethacin, glyburide, losartan, and troglita-

zone are stimulators of E17bG transport (Fig. 3) (Morgan et al., 2013),these compounds appeared to be inhibitors in MRP2-mediated CP-Itransport (Table 1). Troglitazone was withdrawn from the market for thecases of liver failure. Cholestasis mediated by the bile salt export pump(ABCB11) inhibition and reactive metabolite formation from troglita-zone sulfate have been shown to be two major mechanisms of toxicity.Although it has been reported to downregulateMRP2 expression (Fosteret al., 2012), MRP2 functional inhibition by troglitazone can poten-tially contribute to the observed cholestatic injury with troglitazone. Inaddition, troglitazone sulfate has been reported (Funk et al., 2001) tobe amore potent inhibitor of canalicular transporter bile salt export pumpand hence its hepatic accumulation increases toxicity. Troglitazonesulfate is also reported to be a substrate of MRP2 (Kostrubsky et al.,2001). Troglitazone, by inhibiting MRP2, can contribute to the higheraccumulation of troglitazone sulfate and hence to toxicity. Higherintrahepatic levels of troglitazone sulfate corroborate our findings(Chojkier, 2005). Therefore, the interaction of troglitazone sulfate withMRP2 and its contribution to the accumulation of troglitazone and itssulfate conjugates remains to be further determined. Similarly, losartan

and glyburide are drugs associated with clinical liability of liver toxicity.They are reported to be MRP2 inhibitors in our study for the first time.Although the relationship of MRP2 inhibition and liver toxicity remainsto be further explored, our data could shed new light on the mechanismof hepatotoxicity caused by troglitazone, losartan, and glyburide.To summarize, to better understand MRP2 inhibition and the in vivo

consequence, an alternative in vitro probe substrate is needed to min-imize the stimulatory effects and consequent false-negative occur-rences in membrane vesicle assays using E17bG as a substrate. Herein,we have for the first time shown CP-I to be an MRP2 in vitro substratefor appreciating MRP2 inhibition. The transport kinetics were obtainedin membrane vesicles transfected with MRP2. Only eight compoundswith structural similarity stimulated CP-I transport among the 47 com-pounds evaluated. With CP-1, previously classified stimulators in-cluding diclofenac, indomethacin, glyburide, losartan and troglitazone inMRP2-mediated E17bG vesicle uptake were redetermined to be potentMRP2 inhibitors. In addition, bile acids and atorvastatin, simvastatin,lovastatin, and rosuvastatin were also found to be inhibitors of MRP2-mediated CP-I transport. We conclude that CP-I is a better probesubstrate in membrane vesicle uptake assays for characterizing MRP2inhibition effects.

Acknowledgments

The authors thank Punit Marathe, William Humphreys, Devang Shah, MuraliSubramanian, andMike Sinz for help during various aspects of this work, such asdata analysis, experimental setup, and reviewing of the manuscript.

Authorship ContributionsParticipated in research design: Lai, Chatterjee and Gilibili.Conducted experiments: Gilibili, Chatterjee, Bagul, Mosure, Murali.Contributed new reagents or analytic tools: Gilibili, Chatterjee, Bagul,

Mosure and Lai.Performed data analysis: Lai, Chatterjee, Gilibili, Mosure, Bagul and

Mariappan.Wrote or contributed to the writing of the manuscript: Chatterjee, Lai,

Gilibili, Mosure, Bagul, Mandlekar and Mariappan.

References

Bakos E, Evers R, Sinkó E, Váradi A, Borst P, and Sarkadi B (2000) Interactions of the humanmultidrug resistance proteins MRP1 and MRP2 with organic anions. Mol Pharmacol 57:760–768.

Benz-de Bretagne I, Respaud R, Vourc’h P, Halimi JM, Caille A, Hulot JS, Andres CR, and LeGuellec C (2011) Urinary elimination of coproporphyrins is dependent on ABCC2 polymor-phisms and represents a potential biomarker of MRP2 activity in humans. J Biomed Biotechnol2011:498757.

Benz-de Bretagne I, Zahr N, Le Gouge A, Hulot JS, Houillier C, Hoang-Xuan K, Gyan E,Lissandre S, Choquet S, and Le Guellec C (2014) Urinary coproporphyrin I/(I + III) ratio as asurrogate for MRP2 or other transporter activities involved in methotrexate clearance. Br J ClinPharmacol 78:329–342.

Bodo A, Bakos E, Szeri F, Varadi A, and Sarkadi B (2003) Differential modulation of the humanliver conjugate transporters MRP2 and MRP3 by bile acids and organic anions. J Biol Chem 278:23529–23537.

Borst P, Zelcer N, and van de Wetering K (2006) MRP2 and 3 in health and disease. Cancer Lett234:51–61.

Brouwer KL, Keppler D, Hoffmaster KA, Bow DA, Cheng Y, Lai Y, Palm JE, Stieger B,and Evers R; International Transporter Consortium (2013) In vitro methods to support transporterevaluation in drug discovery and development. Clin Pharmacol Ther 94:95–112.

Chojkier M (2005) Troglitazone and liver injury: in search of answers. Hepatology 41:237–246.Evers R, Kool M, Smith AJ, van Deemter L, de Haas M, and Borst P (2000) Inhibitory effect of thereversal agents V-104, GF120918 and Pluronic L61 on MDR1 Pgp-, MRP1- and MRP2-mediated transport. Br J Cancer 83:366–374.

Foster JR, Jacobsen M, Kenna G, Schulz-Utermoehl T, Morikawa Y, Salmu J, and Wilson ID(2012) Differential effect of troglitazone on the human bile acid transporters, MRP2 and BSEP,in the PXB hepatic chimeric mouse. Toxicol Pathol 40:1106–1116.

Funk C, Pantze M, Jehle L, Ponelle C, Scheuermann G, Lazendic M, and Gasser R (2001)Troglitazone-induced intrahepatic cholestasis by an interference with the hepatobiliary export ofbile acids in male and female rats. Correlation with the gender difference in troglitazone sulfateformation and the inhibition of the canalicular bile salt export pump (Bsep) by troglitazone andtroglitazone sulfate. Toxicology 167:83–98.

Herédi-Szabó K, Glavinas H, Kis E, Méhn D, Báthori G, Veres Z, Kóbori L, von Richter O,Jemnitz K, and Krajcsi P (2009) Multidrug resistance protein 2-mediated estradiol-17beta-D-glucuronide transport potentiation: in vitro-in vivo correlation and species specificity. DrugMetab Dispos 37:794–801.

610 Gilibili et al.

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from

Heredi-Szabo K, Kis E, Molnar E, Gyorfi A, and Krajcsi P (2008) Characterization of 5(6)-carboxy-2,‘7’-dichlorofluorescein transport by MRP2 and utilization of this substrate as afluorescent surrogate for LTC4. J Biomol Screen 13:295–301.

Hillgren KM, Keppler D, Zur AA, Giacomini KM, Stieger B, Cass CE, and Zhang L; InternationalTransporter Consortium (2013) Emerging transporters of clinical importance: an update from theInternational Transporter Consortium. Clin Pharmacol Ther 94:52–63.

Isley WL (2003) Hepatotoxicity of thiazolidinediones. Expert Opin Drug Saf 2:581–586.Kidron H, Wissel G, Manevski N, Häkli M, Ketola RA, Finel M, Yliperttula M, Xhaard H,and Urtti A (2012) Impact of probe compound in MRP2 vesicular transport assays. Eur J PharmSci 46:100–105.

Kostrubsky VE, Vore M, Kindt E, Burliegh J, Rogers K, Peter G, Altrogge D, and Sinz MW (2001)The effect of troglitazone biliary excretion on metabolite distribution and cholestasis intransporter-deficient rats. Drug Metab Dispos 29:1561–1566.

Lai Y, Mandlekar S, Shen H, Holenarsipur VK, Langish R, Rajanna P, Murugesan S, Gaud N,Selvam S, Date O, et al. (2016) Coproporphyrins in plasma and urine can be appropriate clinicalbiomarkers to recapitulate drug-drug interactions mediated by organic anion transporting poly-peptide inhibition. J Pharmacol Exp Ther 358:397–404.

Morgan RE, van Staden CJ, Chen Y, Kalyanaraman N, Kalanzi J, Dunn, 2ndRT, Afshari CA,and Hamadeh HK (2013) A multifactorial approach to hepatobiliary transporter assessmentenables improved therapeutic compound development. Toxicol Sci 136:216–241.

Muni�c V, Hlevnjak M, and Erakovi�c Haber V (2011) Characterization of rhodamine-123, calceinand 5(6)-carboxy-29,79-dichlorofluorescein (CDCF) export via MRP2 (ABCC2) in MES-SA andA549 cells. Eur J Pharm Sci 43:359–369.

Niemi M, Arnold KA, Backman JT, Pasanen MK, Gödtel-Armbrust U, Wojnowski L, Zanger UM,Neuvonen PJ, Eichelbaum M, Kivistö KT, et al. (2006) Association of genetic polymorphism inABCC2 with hepatic multidrug resistance-associated protein 2 expression and pravastatinpharmacokinetics. Pharmacogenet Genomics 16:801–808.

Nies AT and Keppler D (2007) The apical conjugate efflux pump ABCC2 (MRP2). Pflugers Arch453:643–659.

Pedersen JM, Matsson P, Bergström CA, Norinder U, Hoogstraate J, and Artursson P (2008)Prediction and identification of drug interactions with the human ATP-binding cassette trans-porter multidrug-resistance associated protein 2 (MRP2; ABCC2). J Med Chem 51:3275–3287.

Shen H, Dai J, Liu T, Cheng Y, Chen W, Freeden C, Zhang Y, Humphreys WG, Marathe P,and Lai Y (2016) Coproporphyrins I and III as functional markers of OATP1B activity:in vitro and in vivo evaluation in preclinical species. J Pharmacol Exp Ther 357:382–393.

Tang W (2007) Drug metabolite profiling and elucidation of drug-induced hepatotoxicity. ExpertOpin Drug Metab Toxicol 3:407–420.

Yoshida K, Maeda K, and Sugiyama Y (2013) Hepatic and intestinal drug transporters: predictionof pharmacokinetic effects caused by drug-drug interactions and genetic polymorphisms. AnnuRev Pharmacol Toxicol 53:581–612.

Zamek-Gliszczynski MJ, Hoffmaster KA, Tweedie DJ, Giacomini KM, and Hillgren KM (2012)Highlights from the international transporter consortium second workshop. Clin Pharmacol Ther92:553–556.

Zelcer N, Huisman MT, Reid G, Wielinga P, Breedveld P, Kuil A, Knipscheer P, Schellens JH,Schinkel AH, and Borst P (2003a) Evidence for two interacting ligand binding sites inhuman multidrug resistance protein 2 (ATP binding cassette C2). J Biol Chem 278:23538–23544.

Zelcer N, Saeki T, Bot I, Kuil A, and Borst P (2003b) Transport of bile acids in multidrug-resistance-protein 3-overexpressing cells co-transfected with the ileal Na+-dependent bile-acidtransporter. Biochem J 369:23–30.

Zhang Y, Han YH, Putluru SP, Matta MK, Kole P, Mandlekar S, Furlong MT, Liu T, Iyer RA,Marathe P, et al. (2016) Diclofenac and its acyl glucuronide: determination of in vivo exposure inhuman subjects and characterization as human drug transporter substrates in vitro. Drug MetabDispos 44:320–328.

Address correspondence to: Sagnik Chatterjee, Pharmaceutical CandidateOptimization, Biocon Bristol-Myers Squibb R&D Center (BBRC), Syngene In-ternational Ltd., Bangalore 560100, India. E-mail: sagnik.chatterjee@syngeneintl.com

Coproporphyrin-I as Novel In Vitro Probe Substrate for MRP2 611

at ASPE

T Journals on D

ecember 11, 2020

dmd.aspetjournals.org

Dow

nloaded from