1521-0103/353/2/380–391$25.00 http://dx.doi.org/10.1124/jpet.114.221804THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS J Pharmacol Exp Ther 353:380–391, May 2015Copyright ª 2015 by The American Society for Pharmacology and Experimental Therapeutics

Evaluation of Rosuvastatin as an Organic Anion TransportingPolypeptide (OATP) Probe Substrate: In Vitro Transport and InVivo Disposition in Cynomolgus Monkeys s

Hong Shen, Hong Su, Tongtong Liu, Ming Yao, Gabe Mintier, Lun Li, R. Marcus Fancher,Ramaswamy Iyer, Punit Marathe, Yurong Lai, and A. David RodriguesPharmaceutical Candidate Optimization (H.Sh., H.Su., T.L., M.Y., R.M.F., R.I., P.M., Y.L., A.D.R.), and Genomic Technologies(G.M.), Bristol-Myers Squibb Research and Development, Princeton, New Jersey; and Department of Bioanalytical Service, WuXiAppTec Co., Ltd., Shanghai, People’s Republic of China (L.L.)

Received December 6, 2014; accepted February 18, 2015

ABSTRACTOrganic anion transporting polypeptides (OATPs) mediate he-patic drug uptake and serve as the loci of drug–drug interactions(DDIs). Consequently, there is a major need to develop animalmodels and refine in vitro–in vivo extrapolations. Therefore, the invivo disposition of a model OATP substrate, [3H]rosuvastatin(RSV), was studied in the cynomolgus monkey and reported forthe first time. After monkeys had received a 3-mg/kg oral dose,mass balance was achieved after bile duct cannulation (meantotal recovery of radioactivity of 103.6%). Forty-two percent of theRSV dose was recovered in urine and bile, and the eliminationpathways were similar to those reported for human subjects;61.7%, 39.0%, and 2.9% of the dose was recovered in the feces,bile, and urine, respectively. The high levels of unchanged RSV

recovered in urine and bile (26%of the dose) and the relatively lowlevels of metabolites observed indicated that RSV was eliminatedlargely by excretion. Also, for the first time, the in vitro inhibitorypotential of cyclosporin A (CsA) toward cynomolgus monkeyOATPs and sodium-taurocholate cotransporting polypeptide wasstudied in vitro (primary hepatocytes and transporter-transfectedcells). It is concluded that one can study the CsA-RSV DDI in thecynomolgusmonkey. For example, the in vitro IC50 valueswerewithin2-fold (monkey versus human), and the increase (versus vehiclecontrol) in the RSV AUC0–inf (6.3-fold) and Cmax (10.2-fold) withCsA (100 mg/kg) was similar to that reported for humans. Theresults further support the use of the cynomolgus monkey asa model to assess interactions involving OATP inhibition.

IntroductionRosuvastatin (RSV) is a 3-hydroxy-3-methylglutaryl-coenzyme

reductase inhibitor (i.e., statin) used in the treatment of pa-tients with hypercholesterolemia. Drug–drug interactions(DDIs) involving RSV that result in an increase in its systemicexposure also might result in unwanted side effects includingmyopathy and rhabdomyolysis. Such DDIs are clinically sig-nificant when one considers that known transporter inhibitorssuch as cyclosporin A (CsA) increase RSV area under the plasmadrug concentration–time curve (AUC) and maximum plasmaconcentration (Cmax) 7.1- and 10.6-fold, respectively (Simonsonet al., 2004). BecauseRSV is not significantlymetabolized acrossseveral species (Nezasa et al., 2002b; Martin et al., 2003b) and isselectively transported and distributed into the liver (Nezasaet al., 2002a,b; Martin et al., 2003a,b), the overall disposition

profile of oral RSV is thought to be highly dependent on drugtransporters such as organic anion transporting polypeptide(OATP), sodium-taurocholate cotransporting polypeptide (NTCP),and breast cancer resistance protein (BCRP) (Ho et al., 2006;Pasanen et al., 2007; Kitamura et al., 2008; Keskitalo et al.,2009; Bi et al., 2013). It has been established that hepaticclearance and renal clearance are the primary pathways forthe elimination of RSV, accounting for 72% and 28% of totalbody clearance, respectively (Martin et al., 2003a). Active trans-port processes are responsible for approximately 90% of totalhepatic uptake clearance of RSV. Indeed, RSV is a substrate ofthe hepatic OATP1B1 and OATP1B3 in vitro, the former con-tributing 77% and the latter 23% to sodium-independent activeuptake in human hepatocytes. The contribution of both OATPsrepresents 70% of total hepatic active uptake; the remainingactive uptake (30%) is attributed to NTCP (Ho et al., 2006;Kitamura et al., 2008; Bi et al., 2013). Consistent with thecontribution of individual transporters to the overall hepaticuptake of RSV, OATPs and NTCP comprise two of the mostabundant sinusoidal uptake transporters in human liver tissues,

This study is supported by Bristol-Myers Squibb Company.dx.doi.org/10.1124/jpet.114.221804.s This article has supplemental material available at jpet.aspetjournals.org.

ABBREVIATIONS: ADME, absorption, distribution, metabolism, and excretion; ATV, atorvastatin; AUC, area under the concentration-time curve;BDC, bile duct cannulated; CsA, cyclosporin A; DDI, drug–drug interaction; ESI, electrospray ionization; HBSS, Hanks’ buffered saline solution;HEK, human embryonic kidney; HPLC, high-performance liquid chromatography; LC-MS/MS, liquid chromatography–tandem mass spectrometry;NTCP, sodium-taurocholate cotransporting polypeptide; OATP, organic anion transporting polypeptide; PCR, polymerase chain reaction; QC,quality control; RIF, rifampin; RSV, rosuvastatin; TCA, taurocholic acid.

380

http://jpet.aspetjournals.org/content/suppl/2015/03/04/jpet.114.221804.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

with relative abundance of quantifiable hepatic transportersof 29% and 13%, respectively, when quantified transporterprotein expression in a human liver bank (n 5 55) by liquidchromatography–tandem mass spectrometry (LC-MS/MS)(Wang et al., 2015). Furthermore, for most transporters, theexpression in the liver tissues was comparable to that in thecryopreserved hepatocytes (Wang et al., 2015). In addition to invitro assessment, RSV has often been used for kinetic studiesin vivo in several species, including humans (Schneck et al.,2004; Simonson et al., 2004; Prueksaritanont et al., 2014),monkeys (Shen et al., 2013), mice (Salphati et al., 2014), rats(Wen and Xiong, 2011), and pigs (Bergman et al., 2009). Whencompared with rodents, however, the absorption, distribution,metabolism, and excretion (ADME) profile of RSV in nonhumanprimates has not been studied extensively (Nezasa et al., 2002b;Martin et al., 2003b).It has been reported that cynomolgus monkey OATPs

(cOATPs) share a high degree of amino acid sequence identityand functional similarity to their human counterparts (Shenet al., 2013; Takahashi et al., 2013). Concomitant with theseidentities and similarities, there are investigations employingthe cynomolgus monkey as an in vivo preclinical model toassess OATP DDIs (Shen et al., 2013; Takahashi et al., 2013).The DDIs have been investigated in other animal models;however, the sequences and transporting profiles obtained forxenobiotics in other animal species are frequently differentfrom those obtained in humans, reducing the utility of suchanimals (Shitara et al., 2003; Shirasaka et al., 2010; Li et al.,2013). In fact, there is only one member of the OATP1Bfamily, Oatp1b2, that is the closest ortholog of both humanOATP1B1 and OATP1B3 (hOATP1B1 and hOATP1B3) andlikely arises from a gene duplication after divergence of therodent species (Hagenbuch and Meier, 2004). The comparisonof predicted Oatps in the genomes of dogs, cows, and horseshas suggested that there is only a single Oatp in the 1Bfamily; in the cynomolgus monkey, a species much closerrelated to humans, both cOATP1B1 and cOATP1B3 orthologshave been cloned (Shen et al., 2013). These results indicate,for many drugs, that cynomolgus monkey may represent anacceptable in vitro and in vivo model to investigate OATPs.Several reports have described various rodent models that

can be used to support the investigation of OATPs in vivo. Forexample, it is possible to delete the murine Oatp1b2 gene (Luet al., 2008; Zaher et al., 2008; Chang et al., 2014), introducehuman OATP genes (van de Steeg et al., 2009), and developcombinations of OATP knockout and knock-in mice (Higginset al., 2014; Salphati et al., 2014). However, the interpretationof data obtained with such models can be difficult in somecases when compensatorymechanisms are suspected (Klaassenand Lu, 2008; Iusuf et al., 2014; Salphati et al., 2014). Inaddition, quantitative translation of data obtained with knock-out and/or knock-in animals can be challenging. Indeed, un-expected results have been observed in mice and rats comparedwith humans. For example, the blood and liver concentrationsof atorvastatin in Oatp1a/b knockout mice were similar to wild-type animals (Chang et al., 2014). The expression of humanOATP1B1 and OATP1B3 in humanized mice did not signifi-cantly alter the liver or plasma concentration ratios of RSVor pitavastatin compared with Oatp1a/1b knockout controls(Salphati et al., 2014). Therefore, there is a need for a betterpreclinical model to enable the prediction of OATP-mediateddrug disposition and DDIs.

We studied the ADME of RSV after administration of oral[3H]RSV (3 mg/kg) to bile duct–annulated (BDC) cynomolgusmonkeys. Because RSV has been shown to be a substrate ofhuman OATPs and NTCP (Bi et al., 2013), we investigatedCsA as an inhibitor of both monkey OATP- and NTCP-mediated RSV uptake in vitro. Finally, the CsA-RSV DDI wasstudied in vivo in cynomolgus monkeys. Our results indicatethat there is a consistency in the elimination pathway of oralRSV between humans and cynomolgus monkeys, and that themonkey can be used as a model to assess inhibition of OATPboth in vitro and in vivo.

Materials and MethodsChemicals, Hepatocytes, and Cynomolgus Monkeys. All

chemicals and solvents of reagent or high-performance liquidchromatography (HPLC) grade were purchased from Sigma-Aldrich(St. Louis, MO) unless otherwise stated. Nonradiolabeled RSV waspurchased from Toronto Research Chemicals (Toronto, ON, Canada),and CsA oral solution (Neoral, 100 mg/ml) was purchased fromNovartis Pharmaceuticals (East Hanover, NJ). [3H]Taurocholic acid(TCA) (5.0 Ci/mmol) was purchased from PerkinElmer Life andAnalytical Sciences (Waltham, MA). [3H]RSV calcium (10.0 mCi/mmol) and [3H]atorvastatin (ATV) sodium (20.0 mCi/mmol) wereobtained from American Radiolabeled Chemicals (St. Louis, MO). Theradiochemical purity of all compounds was determined to be greaterthan 98.2% by HPLC. Cell culture media and reagents werepurchased from Invitrogen (Carlsbad, CA) or Mediatech (Manassas,VA). Cryopreserved male human and cynomolgus hepatocytes werepurchased from Bioreclamation IVT (Baltimore, MD). The 24-wellpoly-D-lysine–coated plates were purchased from BD Biosciences (SanJose, CA). Intact and BDC male cynomolgus monkeys were procuredfrom Charles River Laboratories (Wilmington, MA).

BDC Cynomolgus Monkey Study. All animal studies wereperformed under the standards recommended by the Guide for theCare and Use of Laboratory Animals (NRC, 1996) and were approvedby BMS Institutional Animal Care and Use Committee. The excretionof radioactivity into bile, feces, and urine and metabolism of RSV wasinvestigated in BDC male cynomolgus monkeys (Charles RiverLaboratories) after administration of [3H]RSV. BDC monkeys (N 53, weighing approximately 5.5 to 6.5 kg) were individually housed inmetabolism cages and were freely mobile during the entire study withthe exception of brief manual restraint for oral dosing. Each animalreceived a single oral dose of [3H]RSV administered by gavage ata dose level of 3 mg/kg (approximately 15 mCi/kg). Animals werefasted overnight before dosing. Approximately 4 hours after dosing,animals were fed Certified Primate Diet 5048 (PMI NutritionInternational, Shoreview, MN). Bile was collected before dosingand from 0 to 8, 8 to 24, and 24 to 72 hours after dosing intocontainers that were surrounded by dry ice. Urine and feces werecollected before dosing and over 24-hour intervals through 168 hoursafter dosing. Blood samples were collected into tubes containingK2EDTA from the femoral artery before dosing and at 1, 2, 6, 12, 24,and 48 hours after dosing. The blood samples were then centrifugedto obtain plasma, and the plasma samples were frozen at –70°C untilanalysis.

Determination of Radioactivity in Biologic Matrices fromBDC Cynomolgus Monkeys. Plasma, urine, bile, and fecal sampleswere analyzed for radioactivity concentration using liquid scintilla-tion counting. Portions of plasma (50–100 ml), urine (50 ml), and bile(10 ml) were mixed with 5 ml of Ecolite scintillation cocktail(PerkinElmer Life and Analytical Sciences) into polystyrene tubes.For fecal samples, two portions (approximately 0.2 g each) of fecalhomogenate were weighed individually in a scintillation vial, mixedwith 1 ml of Soluene-350, and shaken slowly at room temperatureovernight. The solubilized homogenate mixtures were bleached with

Use of Cynomolgus Monkey to Assess DDIs Involving OATPs 381

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

1 ml of 20% hydrogen peroxide, and then neutralized with 0.1 ml ofa solution containing saturated sodium pyruvate in methanol, glacialacetic acid, and methanol (4:3:1, by volume). After addition of Ecolitecocktail, the samples were mixed and stored under refrigeratedconditions in the dark overnight. Radioactivity was determined byLSC6000 or LS6500 liquid scintillation counter (Beckman Coulter,Fullerton, CA) for 5 minutes.

Metabolite Profiling of Cynomolgus Monkey Bile andUrine. Pooled bile and urine samples from BDC monkeys wereprepared by combining a constant percentage of bile and urine volumeacross animals (3% and 2%, respectively). The pooled samples werethen centrifuged at 14,000g for 5 minutes. A portion of supernatant(25–50 ml) was injected into the HPLC for biotransformation profilingand mass spectral analysis.

The HPLC analysis was conducted on an Agilent 1290 Infinity LCsystem (Agilent Technologies, Santa Clara, CA) interfaced toa linear ion trap mass spectrometer (Thermo Fisher Scientific,Waltham, MA). Separation was achieved on a Agilent Zorbax SB-C18 column (4.6 mm � 250 mm, 5 mm) using a mobile phaseconsisting of 0.1% formic acid and 0.1% acetonitrile in water(solvent A) and 0.1% formic acid in acetonitrile (solvent B) ata constant flow rate of 0.8 ml/min. The gradient was as follows: 0–5minutes, 20% B; 40 minutes, 25% B; 47 minutes, 50% B; 50 minutes,90% B; and 60 minutes, 20% B.

The HPLC eluate was split via a Gilson Model FC 204 fractioncollector (Gilson, Middleton, WI), with which 25% of the eluate wasdirected into the linear ion trap mass spectrometer for profiling. Theremaining 75% of the eluate was collected into ScintiPlate 96-wellplates at 0.2-minute intervals per well. Ecolite scintillation cocktail(200 ml) was then added to each well, and the radioactivity wascounted for 20 minutes per well with a PerkinElmer 1450 MicroBetaWallac TRILUX Liquid Scintillation and Luminescence Counter(PerkinElmer Life and Analytical Sciences).

Radiochromatographic profiles were prepared by plotting thenet counts per minute values obtained from the counter versustime after injection. Mass spectral analyses were performed usingelectrospray ionization (ESI) in the positive ion mode. The cap-illary temperature was 300°C, and the ESI voltage was maintainedat 4.5 kV for all analyses. The collision energy was 15.0% forLC-MS/MS analysis. Other instrument parameters were adjustedto give maximum sensitivity or fragmentation of drug-relatedcomponents.

Generation of Stably Expressed Cynomolgus Monkey NTCPin HEK-293 Cells. Cloning and stable transfection of humanembryonic kidney 293 (HEK-293) cells with cynomolgus monkeyNTCP (cNTCP) were performed as described elsewhere (Shen et al.,2013). In brief, cNTCP was cloned out of cDNA synthesized fromMauritian cynomolgus monkey liver total RNA. For polymerase chainreaction (PCR) on double-stranded cDNA from monkey liver, thefollowing degenerate oligonucleotides, derived from the human andrhesus monkey NTCP gene sequences and predicted cNTCP tran-script sequence, were used: 59-CTT CCA CTG CCT CAC AGG AGG-39(forward primer, corresponding to the nucleotide positions 114–134 ofhumanNTCP cDNANM_003049.3), and 59-AAGGGCTAGGCTGTGCAA GG-39 (reverse primer, reverse complementary to positions1170–1189).

The PCR products were cloned into pJet1.2 (Fermentas, Vilnius,Lithuania), and several clones for each cDNA were sequenced.Sequences were deposited to GenBank: cNTCP (KP453714). Thecoding sequence of the cDNA was subcloned into the Gateway entryvector pDONR221 (Invitrogen) using standard methods. The Gate-way entry clones were recombined into a Gateway adapted version ofthe expression vector pcDNA5/FRT/TO (Invitrogen) using LR ClonaseII (Life Technologies, Carlsbad, CA) according to the manufacturer’sprotocol. Expression constructs were analyzed by agarose gelelectrophoresis, and the sequence was confirmed.

Flp-In HEK-293 cells were cultured in Dulbecco’s modified Eagle’smedium supplemented with 10% fetal bovine serum, 0.1 mMnonessential

amino acids, and 2 mM L-glutamine on a 6-well plate as describedelsewhere (Shen et al., 2013). After washing with serum-free medium,the culture well was incubated at 37°C for 4 hours with 1 ml of serum-free medium that contained 10 ml Lipofectamine 2000 (Invitrogen)and 2.5 mg plasmid DNA. The cells stably transfected with pcDNA5/FRT/TO/cNTCP were then selected using hygromycin B (200 mg/ml)according to the protocol of the vendor (Invitrogen). Expression ofcNTCP was verified by reverse-transcription PCR and functionalcharacterization.

Uptake Studies Using Transporter-Expressing HEK-293Cells. The HEK-293 cells individually expressing human OATP1B1(hOATP1B1), hOATP1B3, cynomolgusmonkeyOATP1B1 (cOATP1B1),cOATP1B3, or cNTCP were prepared and used as described previouslyelsewhere (Shen et al., 2013). In brief, all transporter- and vector-transfected cells were cultured at 37°C in an atmosphere of 95% air and5%CO2 and subcultured once a week. Cells were seeded in 24-well poly-D-lysine–coated plates at a density of 5 � 105 cells per well, and wereready for experiment after 48 to 72 hours. Before uptake experiments,cells were washed twice with 1.5 ml of prewarmed Hanks’ bufferedsaline solution (HBSS). The cells were then preincubated with uptakebuffer (HBSS with 10 mM HEPES, pH 7.4) containing CsA only(0.023–16.7 mM) at 37°C for 15 minutes. Subsequently, the cells wereincubated with the test solution containing CsA and the substrate RSV(0.1mM)at 37°C for 2minutes. Over this time period, linearity of uptakewas confirmed (Shen et al., 2013). Cells were washed 3 times with 1 mlof ice-cold HBSS buffer and lysed with 0.3 ml of 0.1% Triton X-100. Theintracellular accumulation of RSV was determined using LC-MS/MS asdescribed later.

Assessment of uptake of [3H]TCA, [3H]RSV, and [3H]ATV intoHEK-293 cells expressing cNTCP involved a 5-minute incubation.The cells were lysed with 0.3 ml of 0.1% Triton X-100, and theradioactivity was determined by liquid scintillation counting. Accu-mulation was normalized to the protein content of the HEK-293 cellsin each well, as measured using a bicinchoninic acid assay (BCAProtein Assay Kit; Pierce Chemical, Rockford, IL).

LC-MS/MS Analysis of RSV in Cell Lysates from In VitroTransporter Studies. Cells were solubilized and lysed in 300 ml of0.1% Triton X-100 in water. After 60 minutes incubation at roomtemperature, the samples were transferred by a Tecan liquid handler(Tecan Group, Männedorf, Switzerland) from a 24-well plate to thefirst 96-well filter plate for LC-MS/MS analysis and the second 96-well plate for protein measurement. Protein concentrations in the celllysates (20 ml) were measured using the BCA protein assay kit (PierceChemical). The lysed samples (150 ml) were mixed with 200 ml ofacetonitrile containing d6-rosuvastatin (internal standard) in a 96-well hydrophilic filter plate (Millipore, Billerica, MA), stacked witha deep 96-well receiver plate, and centrifuged. The filtrates were driedunder stream of nitrogen to half of the initial volume in the receiverplate, vortexed, and analyzed by LC-MS/MS. Cellular uptake wasnormalized to the protein content.

The liquid chromatography system was the Shimadzu SCL 30ADNexera, composed of two LC-30AD pumps and a Shimadzu SIL-30ACautosampler (Shimadzu Scientific Instruments, Columbia, MD). RSVand d6-rosuvastatin were separated under gradient elution ona Waters Acquity HSS T3 (50 mm � 2.1 mm internal diameter, 1.8mm particle size) (Waters Corporation, Milford, MA). The column wasmaintained at room temperature. The mobile phase was a mixture of0.1% formic acid (A) and acetonitrile (B) at a flow rate of 0.60 ml/min.The gradient programwas set as follows: from 0 to 0.01 minutes, keepB for 20%; from 0.01 to 2.0 minutes, increase B linearly from 20% to60%; from 2.0 to 2.1 minutes, increase B linearly from 60% to 95%;maintain 95%B for 0.4 minutes; and from 2.5 to 2.6 minutes, decreaseB from 95% to 20%. We then equilibrated the column for 0.9 minutesfor the next injection. The retention time for rosuvastatin andd6-rosuvastatin was 1.86 minutes.

An AB Sciex Qtrap 6500 system (Applied Biosystems/MDSAnalytical Technologies, Foster City, CA) equipped with ESI sourcewas used for mass spectrometric detection. Analyst version 1.62 (AB

382 Shen et al.

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

Sciex, Framingham, MA) was used as the data acquisition software.The ESI source was used in the positive ion mode. The LC-MS/MSdetector was operated at unit resolution in the multiple reactionmonitoring mode using the transitions of the protonated forms of RSVat m/z 482.1 . 258.07 and d6-rosuvastatin at m/z 488.16 . 263.9.Optimized parameters were as follows: curtain gas, gas 1 and gas 2(nitrogen) 60 and 75 units, respectively; dwell time 100 milliseconds;source temperature 400°C; and IonSpray voltage 5000 V. Declusteringpotential and collision energy were 116 V and 45 eV for both RSV andd6-rosuvastatin.

In addition, a divert valve was used to minimize the Triton X-100introduced into the MS source. At 0 to 0.8 minutes, the ultraper-formance liquid chromatography eluent was diverted to waste thenswitched to MS source after 1 minute. After 3.0 minutes, the eluentwas diverted back to waste again. The range of standard curve wasbetween 0.02 to 10.0 nM for RSV.

Uptake Studies in Cryopreserved Hepatocytes. Cryopreservedhuman (Lots OJE, GST, LIO) and cynomolgusmonkey hepatocytes (poolof four males; Lot VCE) were thawed according to the manufacturer’sinstructions and resuspended in InVitroGRO Krebs-Henseleit buffer(Bioreclamation IVT). Human hepatocytes from three donors were thenpooled together after thawing.

Cell number and viability were assessed using trypan blue ex-clusion. Cells with viability greater than 80% at the time of uptakeassay were selected to reduce the background signal, and resuspendedat a concentration of 2 � 106 viable cells/ml in InVitroGRO Krebs-Henseleit buffer. Centrifuge tubes (0.4 ml) were prepared by fillingwith 50 ml of 3N KOH solution and layering 100 ml of filtration oil ontop of the KOH. The filtration oil was prepared as a mixture of 82parts silicone oil to 18 parts mineral oil, resulting in an oil mixturewith a density of 1.015 g/ml. Cell suspensions (100 ml) containingappropriate concentrations of CsA were added to glass tubes andpreincubated in a water bath at 37°C for 5 minutes.

Uptake studies were initiated by an addition of 100 ml ofprewarmed buffer containing [3H]RSV (0.2 mM). Aliquots (80 ml) ofthe mixture were removed at 15 and 90 seconds, and then transferredto centrifuge tubes and layered carefully on top of an oil layer. Thetubes were centrifuged immediately at 13,000g for 15 seconds. Aftercentrifugation of hepatocytes through the oil layer, the cell suspensionbuffer containing RSV was left in the supernatant.

The centrifuge tubes were incubated for at least 2 hours at anambient temperature, then were frozen in a 280°C freezer or on dryice just before cutting. The tubes were cut in the middle of the oillayer, allowing the bottom section to drop into a 20-ml scintillationvial. The cell pellets were resuspended in 120 ml 2 N HCl to neutralizethe solution and read in a scintillation counter after addition of thescintillation cocktail.

In Vivo DDI Study. The in vivo DDI pharmacokinetic studieswere conducted by WuXi AppTec Corporation (Suzhou, P. R. China)using three young adult male cynomolgus monkeys weighing between3.4 and 4.2 kg. The animals were housed in a temperature- andhumidity-controlled room with a 12-hour light/dark cycle. Animalswere fed approximately 120 g of Certified Monkey Diet daily (BeijingVital Keao Feed Corporation, Beijing, P. R. China). Themonkeys werefasted for 12 hours before dose administration, and water was madeavailable ad libitum.

In the first period, the three monkeys received an oral dose of RSVat 3 mg/kg of body weight dissolved in sterile water by oral gavagefollowed by a water rinse. After a 1-week washout period, the sameanimals were administered an equivalent oral dose of RSV 1 hourafter CsA administration (100 mg/kg p.o.; Neroral cyclosporine oralsolution) in the second period. There was no substantial weightfluctuation of the animals between the times of RSV alone andcoadministration.

Approximately 1 ml of blood was collected in tubes containingpotassium (K2) EDTA by venipuncture via the cephalic or saphenousveins at 0.25, 0.5, 0.45, 1, 2, 3, 5, 7, 24, and 48 hours after RSVadministration. In period 2 (i.e., coadministration treatment), 300 ml

of whole blood was aliquoted and stored at #215°C for CsA analysis.The remaining blood volume was centrifuged (3000g for 10 minutes at2 to 8°C) within 1 hour of collection, transferred into polypropylenemicrocentrifuge tubes, and stored frozen at270°C. The animals wereplaced in metabolic cages, and urine was collected at 0–7, 7–24, and24–48 hours after dosing into containers cooled by wet ice. The totalvolume of urine samples for different periods wasmeasured at the endof collection and recorded. Aliquots (5 ml) of urine samples collectedfrom each period were frozen on dry ice and kept at 270°C untilanalysis.

LC-MS/MS Analysis of RSV in Plasma and Urine. Stock solu-tion (1 mg/ml) of RSV was prepared in acetonitrile/0.1 M ammoniumacetate pH 4.0 buffer (80:20 v/v). A stock solution (1 mg/ml) ofrosuvastatin-5S-lactone, a metabolite of RSV, was prepared inacetonitrile/0.1M ammonium acetate pH 4.0 buffer (50:50 v/v), anda stock solution (0.2mg/ml) of ATV (internal standard) was prepared in50% acetonitrile. For plasma assay, the following RSV calibrationstandards were prepared in monkey plasma: 0.1, 0.2, 1, 10, 50, 100,200, and 250 ng/ml. Quality control (QC) samples were also prepared inmonkey plasma at 0.3, 5, 90, and 190 ng/ml. For urine assay, the RSVcalibration standards were prepared in monkey plasma at 0.3, 0.6, 3,30, 150, 300, 600, and 750 ng/ml. QC samples were prepared inmonkeyurine with 0.75 mg/ml Triton X-100 at 4.5, 75, 1350, and 2850 ng/ml,and diluted with plasma (1:4, urine/plasma v/v), and quantified byplasma calibration curve.

Plasma sample extraction for RSV was conducted in 96-well platesusing protein precipitation with acetonitrile containing 0.5% formicacid. In brief, 50 ml of calibration standards, QC samples, and studysamples were first mixed with 50 ml of chilled (ice/water bath) 0.1 Mammonium acetate (pH 4.5) buffer followed by 20 ml of chilled 100 ng/mlatorvastatin solution in acetonitrile/0.1 M ammonium acetate (pH 4.0)buffer (75:25 v/v). After addition of 300 ml of chilled 0.5% formic acidin acetonitrile, vortex-mixing, and centrifugation at 4°C, an aliquotof 300 ml of each sample was transferred to a new 96-well plate,evaporated to dryness under N2 at room temperature, and recon-stituted with 200 ml of chilled acetonitrile/0.1 M ammonium acetate(pH 4.0) buffer (50:50 v/v). A 10-ml aliquot was injected for analysisby LC-MS/MS.

Urine sample extraction for RSV was conducted in 96-well platesusing protein precipitation with acetonitrile containing 0.5% formicacid. In brief, 50 ml of calibration standards, diluted QC samples, anddiluted study samples were first mixed with 50 ml of chilled (ice/waterbath) 0.1 M ammonium acetate (pH 4.5) buffer followed by 20 ml ofchilled 100 ng/ml atorvastatin solution in acetonitrile/0.1 M ammo-nium acetate (pH 4.0) buffer (75:25 v/v). After addition of 300 mlchilled 0.5% formic acid in acetonitrile, vortex-mixing, and centrifu-gation at 4°C, an aliquot of 200 ml of each sample was transferred toa new 96-well plate. Chilled acetonitrile/0.1M ammonium acetate (pH4.0) buffer (200 ml) (50:50 v/v) was added and a 10-ml aliquot wasinjected for analysis by LC-MS/MS.

The HPLC system consisted of an Agilent 1200 pump, Agilent1200 column oven (Agilent Technologies), and a CTCPAL autosampler(CTC Analytics AG, Zwingen, Switzerland) maintained at 4°C duringanalysis. The analytic column used was a Shiseido CapCellPak MGC18 (2.0� 50mm, 5.0mm) (Shiseido, Tokyo, Japan) andwasmaintainedat 40°C. The mobile phase consisted of ammonium acetate pH 4.0 buffer(eluent A) and 100% acetonitrile (eluent B). The following gradientelution was used: start and maintain at 30% B from 0 to 0.5 minutes;ramp from 30 to 70% B from 0.5 to 1.5 minutes; hold at 70% B from1.5 to 2.5 minutes; ramp to 30% B from 2.5 to 2.51 minutes; holdat 30% B until 3.7 minutes before the next injection. The flow ratewas 0.4 ml/min. The retention times of RSV and ATV were 2.15 and2.50 minutes, respectively.

The HPLC was interfaced to a Sciex API 5000 mass spectrometer(AB Sciex). The following positive ESI source/gas conditions wereused: curtain gas at 25, ion spray voltage at 5500 V, temperatureat 550°C, ion source gas 1 at 40, and ion source gas 2 at 50. Thecompound-dependent parameters used for RSV and ATV were,

Use of Cynomolgus Monkey to Assess DDIs Involving OATPs 383

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

respectively, as follows: declustering potential of 80,80; collisionenergy of 48,30; collision cell exit potential of 14,14. The multiplereaction monitoring transitions used were as follows:m/z 482.2→m/z258.1 for RSV, and m/z 559.3 → m/z 440.2 for ATV.

To avoid the conversion of rosuvastatin-5S-lactone metabolite toRSV, plasma samples stored were thawed at 4°C (ice/water), and thealiquots for analysis were mixed with a pH 4.5 buffer. For the samereason, the acetonitrile used for protein precipitation and the

reconstitution solution were also acidified. To avoid assay bias dueto potential conversion of the metabolite in the mass spectrometer,the chromatographic conditions were optimized to achieve separationof the lactone metabolite from rosuvastatin. In addition, a special QCsample containing only rosuvastatin-5S-lactone (lactone-only QC)was included in every sample analysis run to gauge any conversion ofthe metabolite to RSV. The performance of this QC showed that theconversion was minimal (less than 5%) ensuring no contribution tothe measured RSV concentrations.

LC-MS/MS Analysis of CsA in Blood. Stock solutions of CsA(0.2 mg/ml) and of Cyclosporine A-13C2,d4 (1 mg/ml, internal standard)were prepared in methanol/acetone: (50:50 v/v). The concentrations ofCsA calibration standards prepared in monkey whole blood were: 1,2.5, 10, 25, 100, 400, and 1500 ng/ml. The QC samples were preparedinmonkey whole blood at concentrations of 3, 40, 400, and 1200 ng/ml.

Whole blood sample extraction for CsA was conducted in poly-propylene tubes using protein precipitation with methanol/ acetoni-trile (50:50 v/v). In brief, 450 ml of methanol/ acetonitrile (50:50 v/v)was added into each polypropylene tube,mixedwith 50ml of CsA-13C2,d4

(internal standard) in methanol/ acetonitrile (50:50 v/v) followed by50 ml sample. After vortex-mixing and centrifugation at room tem-perature, an aliquot of 200 ml supernatant of each sample wastransferred to a 96-well plate, 200 ml of methanol/ acetonitrile (50:50v/v) was added and mixed. A 20-ml aliquot was injected for analysisby LC-MS/MS.

The HPLC system consisted of Agilent 1200 HPLC pump, Agilent1200 column oven and a CTC PAL autosampler maintained at 20°C.The analytic column used was a Venusil XBP-C8, 5 mm (2.1 mm

TABLE 1Cumulative excretion of radioactivity in the bile, urine, and feces of bileduct–cannulated cynomolgus monkeys (n = 3) after a single oraladministration of [3H]RSVData are reported as mean 6 S.D., n = 3 different animals.

Collection IntervalExcretion of Radioactivity (% of Dose)

Bile Urine Feces

h

0–8 20.8 6 7.3 0.9 6 0.3 NSa

8–24 13.5 6 4.4 0.7 6 0.2 30.1 6 5.524–72 4.7 6 2.9 0.8 6 0.5 20.5 6 17.072–168 NSb 0.6 6 0.3 11.2 6 12.40–168 39.0 6 3.0 2.9 6 1.3 61.7 6 25.1Total recovery 103.6 6 27.5

NS, no sample.aOnly 0- to 24-hour fecal samples were collected and analyzed.bOnly 0- to 72-hour bile samples were collected and analyzed.

Fig. 1. Representative radiochromatogram of pooled bile (A) and urine (B) from samples taken 72 and 168 hours after administration of a single oraldose of [3H]RSV at 3 mg/kg to bile duct–cannulated cynomolgus monkeys. Samples were profiled with HPLC. The retention time for RSV wasapproximately 32 minutes. The percentage (of the radioactivity collected in bile or urine samples) as parent RSV is indicated.

384 Shen et al.

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

I.D. � 50 mm) (Bonna-Agela Technologies, Wilmington, DE) and wasmaintained at 80°C. The mobile phase consisted of 0.2% acetic acid inwater (eluent A) and 0.2% acetic acid in acetonitrile (eluent B). Thefollowing gradient elution was used: start at 40% B, ramp from 40%to 100% B from 0.0 to 1.2 minutes; hold at 100% B from 1.2 to 2.0minutes; ramp to 40% B from 2.0 to 2.1 minutes; hold at 40% B until2.8 minutes before the next injection. The flow rate was 0.6 ml/min.Retention times of CsA and CsA-13C2,d4 were 1.54 and 1.54 minutes,respectively.

The HPLC was interfaced to a Sciex API 5000 mass spectrometer(AB Sciex). The following positive ESI source/gas conditions were used:curtain gas at 20; ion spray voltage at 5500 V; temperature at 500°C;ion source gas 1 at 60; ion source gas 2 at 20. The compound-dependentparameters used for CsA and CsA-13C2,d4 were, respectively, as follows:declustering potential of 140,140; collision energy of 25,25; collision cellexit potential of 10,10. The multiple reaction monitoring transitionsused were as follows: m/z 1219.45 → m/z 1202.70 for CsA, and m/z1225.45 → m/z 1208.70 for CsA-13C2,d4.

Data Analysis. Within the in vitro transporter studies, each assaywas conducted in triplicate. The active transport of RSV by hOATPand cOATP was calculated after subtracting the uptake in mock-transfected cells from the total uptake in the transporter-expressingcells. IC50 values, the concentration of CsA required for 50% inhibitionof transport of RSV, were calculated with WinNonlin (Pharsight,Mountain View, CA) using following equation:

V5V0 2

�Imax •Cg

IC50g 1Cg

�

where V is the rate of RSV transport measured at given CsAconcentration, g is the slope factor, C is the CsA concentration, V0 isthe rate of RSV transport measured in the absence of CsA, and Imax isthe maximum inhibitory effect.

For the inhibition studies employing hepatocytes, the uptake rateof RSV (Vðx2 yÞ pmol/min/106 cells), determined from time x to time y(where y 5 0.25 minutes when x 5 1.5 minutes; or y 5 15 secondswhen x 5 90 seconds), was calculated using following equation:

Vðx2 yÞ 5Vx 2Vy

ðx2 yÞ

The IC50 values were then estimated by fitting V, the hepatic uptakeat different CsA concentrations, to this equation using WinNonlin.

For the pharmacokinetic analysis, noncompartmental analysis ofRSV and CsA concentration–time data was performed, and thefollowing standard parameters were estimated using Kinetica(Thermo Fisher Scientific, Waltham, MA):

CL=F5Dose

AUCð02 infÞ

Vd=F5Dose •MRTAUCð02 infÞ

CLR 5Xe; 02 48 h

AUCð0248 hÞ

where CL/F and Vd/F are apparent clearance and volume of dis-tribution after oral administration, MRT is mean residence time, CLR

is renal clearance, and Xe, 0–48 h is the amount recovered in urine over48 hours. The AUC from time 0 to 48 hours (AUC0–48 h) was calculatedusing the log-linear-trapezoidal method. The AUC from time 0 toinfinity (AUC0–inf) was calculated as the sum of AUC0–48 h and Ct/l,where l is the apparent terminal rate constant and Ct is the pre-dicted concentration at last time point. For each animal, l was cal-culated by regression of the terminal log-linear portion of the plasmaconcentration-time profile, and the apparent terminal elimination

Fig. 2. Effect of CsA on the uptake of RSV (0.1 mM) in hOATP1B1 (A), hOATP1B3 (B), cOATP1B1 (C), and cOATP1B3 (D). The transporter-mediateduptake was obtained by subtracting the transport rate during the first 2 minutes of incubation at 37°C in vector control cells from that in transporter-transfected cells. The data are expressed as a percentage of the initial uptake rate in the presence or absence of CsA. IC50 values were obtained by fittingthe uptake data to the equation (seeMaterials and Methods) and are summarized in Table 2. Values shown are mean6 S.D. for experiments performedin triplicate.

Use of Cynomolgus Monkey to Assess DDIs Involving OATPs 385

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

half-life (t1/2) was calculated as the quotient of the natural log of 2 (ln[2]) and l. The geometric mean ratio and its 90% confidence intervalwere calculated by calculating the arithmetic difference of the logmean and its associated 90% confidence intervals and thenexponentiating these results back to the original (geometric meanratio) scale. The t-distribution was assumed in the calculation of all.

ResultsExcretion of Radioactivity in Bile, Urine, and Feces

of Cynomolgus Monkeys. Recovery of radioactivity in bile,urine and feces was complete (103.6% 6 27.5%) by 168 hoursafter administration of a single oral dose of [3H]RSV (3 mg/kg)to male BDC cynomolgus monkeys (Table 1). Fecal excretionwas major and accounted for 61.7% 6 25.1% of the dose(Table 1). The balance was recovered in bile and in urine with39.0% 6 3.0% and 2.9% 6 1.3% of the radioactive dose,respectively. Approximately 90% of the biliary excretion wascomplete between 0 and 24 hours with evidence of slowerexcretion beyond this time (0–24 hours versus 24–72 hours:34.3% versus 4.7%). A majority of fecal excretion occurred by72 hours. Although the urinary excretion was small comparedwith biliary excretion, it continued up to the final collectionperiod (0–24 hours versus 24–72 hours versus 72–168 hours:1.6% versus 0.8% versus 0.6%), indicating a slower rate ofexcretion for a fraction of the dose (Table 1).The radioactivity in the pooled bile and urine samples was

analyzed by radio-HPLC. In the pooled bile collected from 0 to72 hours, unchanged RSV was the predominant component,accounting for, on average, 62.8% of the sample radioactivityand 24.5% of the dose (Fig. 1A). In addition to RSV, severalmetabolites were detected which accounted for 26% of thesample radioactivity (Supplemental Figs. 1 and 2).A similar chromatogram was obtained from the pooled urine

samples, as unchanged RSV was the main radioactive com-pound accounting for 45.7% of the sample radioactivity (Fig. 1B).Metabolites accounted for 37.2% of the sample radioactivity,and additional trace peaks were detected compared with biliarychromatogram (Supplemental Figs. 1 and 3).CsA as an Inhibitor of RSV UptakeMediated by cOATP

and cNTCP. CsAwas evaluated as an inhibitor of RSV (0.1mM)uptake after incubation with HEK-293 cells containing individ-ually expressed human and monkey OATP1B1 and OATP1B3and monkey NTCP. In these studies, cells were preincubatedwith CsA (0.02–16.7 mM) at 37°C for 15 minutes before theaddition of RSV. CsA inhibited uptake for hOATP1B1,hOATP1B3, cOATP1B1, cOATP1B3, and cNTCP with anestimated IC50 of 0.21 6 0.10 mM, 0.13 6 0.06 mM, 0.28 60.11 mM, 0.25 6 0.09 mM, and 3.9 6 2.0 mM, respectively

(Figs. 2 and 3B; Table 2). In addition, the expression ofcNTCP was confirmed at a functional level by measuring theuptake of human NTCP model substrates (TCA, RSV, andATV) into the HEK-293 cells transfected with cNTCPcompared with the mock cells (Fig. 3A).Parallel experiments were also conducted to evaluate the

inhibitory effects of CsA on the uptake of 0.1 mM [3H]RSV in

Fig. 3. (A) Uptake of [3H]TCA, [3H]RSV, and [3H]ATV in mock- andcNTCP‐HEK cells was determined after 5 minutes of incubation. Eachvalue represents the mean 6 S.D. (n = 3). ***P , 0.001, uptake wassignificantly different from mock HEK cells. (B) Effect of CsA on cNTCP-mediated uptake of RSV (0.1 mM) was determined by subtracting thetransport rate during the first 2 minutes of incubation at 37°C in vectorcontrol cells from that in transporter-transfected cells in the presence ofvarious concentrations of CsA. The data are expressed as a percentage ofthe initial uptake rate in the presence or absence of CsA. IC50 values wereobtained by fitting the uptake data to either the equation (see Materialsand Methods) and are summarized in Table 2. Values shown are mean 6S.D. for experiments performed in triplicate.

TABLE 2In vitro evaluation of CsA as an inhibitor of RSV uptake into hepatocytes and HEK-293 cells expressingindividual transporterData are represented as mean 6 S.D. (n = 3 determinations) (see Materials and Methods).

CellsIC50

a

Hepatocytes OATP1B1 OATP1B3 NTCP

mM

Human 0.30 6 0.08 0.21 6 0.10 0.13 6 0.06 2.1 6 0.9b

Cynomolgus monkey 0.29 6 0.11 0.28 6 0.11 0.25 6 0.09 3.9 6 2.0

aIC50 values were generated at a low concentration of RSV (0.1 mM).bIC50 was average of the values generated from 55 different test occasions using [3H]TCA (1 mM) as probe substrate.

386 Shen et al.

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

human and cynomolgus monkey hepatocytes (Fig. 4). Consis-tent with being a potent inhibitor of human and monkeyOATPs using transporter-overexpressing cells, CsA demon-strated significant concentration-dependent inhibition of RSVuptake in human and monkey hepatocytes (maximal ∼80%inhibition) with IC50 values of 0.306 0.08 mMand 0.296 0.11mM, respectively (Table 2). The IC50 values generated withhepatocyte suspensions were closer to the transfected OATP-derived values, which implies that RSV (0.1 mM) uptake islargely dominated by OATPs.Impact of CsA onRSVPharmacokinetics in Cynomolgus

Monkeys. To assess the inhibitory effect of CsA on pharma-cokinetics of RSV in vivo, a DDI study was conducted in malecynomolgus monkeys. The plasma concentrations of RSVwere increased significantly when RSV was coadministeredwith CsA (Fig. 5A), and the pharmacokinetic results are sum-marized in Table 3. CsA increased the RSV AUC0–inf by6.3-fold and Cmax by 10.2-fold relative to RSV administeredalone. The mean CL/F and Vd/F for RSV alone are 784.6 6227.3 ml/min per kg and 594.0 6 216.7 l/kg, respectively. Themean CL/F and Vd/F for RSV when coadministered with CsAwere 129.8 6 54.8 ml/min per kg and 48.7 6 29.1 l/kg, re-spectively. The t1/2 of RSV decreased with coadministration ofCsA (9.1 6 3.1 hours versus 4.3 6 1.3 hours; Table 3).

Urinary excretion of RSVwas also increased in the presenceof CsA (Fig. 6; Table 4). CsA increased the total amount ofRSV excreted in urine Xe, 0–48 h by approximately 6-fold (214.16 81.3 nmol versus 1234.8 6 506.3 nmol). However, the renalclearance (CLR) of RSV did not change significantly withcoadministered CsA (Table 4).At an oral dose of 100 mg/kg Neoral oral solution, CsA blood

levels reached a peak of 1.1 6 0.3 mM, with a tmax of 4.3 6 0.4hours (Fig. 5B). The AUC over a 49-hour period was 10.06 3.0mM×h, with an average concentration of 0.20 mM. The averageconcentration is comparable to the IC50 values of CsA obtainedfrom monkey hepatocytes and cOATP-expressing HEK-293cells. The CsA systemic exposures obtained in this study werecomparable to those in patients at a therapeutic dose (Novartis,2005) and in monkeys at 50 and 100 mg/kg (Schuurman et al.,2001).

DiscussionOATP1B-mediated DDIs are a major concern in drug de-

velopment and clinical practice (International TransporterConsortium et al., 2010). OATP inhibition likely is not onlydose dependent but also time dependent, so the extent andduration of inhibition is dynamic. As a result, a mechanisticstatic model will often overpredict a clinical DDI involvingOATP inhibition, especially when assuming that the inhibi-tor concentration is represented by the maximum unbound

Fig. 4. Effect of CsA on the uptake of [3H]RSV (0.1 mM) by human (A) andcynomolgus monkey hepatocytes (B). The data are expressed as a per-centage of the initial uptake rate measured at 15 and 90 seconds at 37°C inthe presence or absence of CsA. IC50 values were obtained by fitting theuptake data to the equation (see Materials and Methods) and are sum-marized in Table 2. Values shown are mean 6 S.D. for experiments per-formed in triplicate.

Fig. 5. Mean plasma concentrations of RSV (A) and whole blood con-centrations of CsA (B) after a single oral dose of RSV (3 mg/kg) alone (s) orwith CsA (100 mg/kg p.o.) (d or m) in cynomolgus monkeys. Inset inA depicts same data on semilogarithmic scale. CsA was dosed 1 hour aheadof RSV. Data are expressed as mean 6 S.D. (n = 3 animals).

Use of Cynomolgus Monkey to Assess DDIs Involving OATPs 387

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

concentration in the portal vein. The cynomolgus monkeymodel can be used to examine in vivo DDI, thus bridging thein vitro inhibitory potential to the extent of in vivo inhibition.However, the assumption is that the pharmacokinetics of probesubstrate in monkeys is similar to that of human subjects. Inrecent studies, themonkeymodel has been successfully appliedby investigators to quantitatively predict transport-basedDDIs (Shen et al., 2013; Takahashi et al., 2013). The objectiveof our current study was to evaluate absorption, metabolism,and excretion of RSV in cynomolgusmonkeys (BDC study), andin vitro and in vivo inhibitory effects of CsA on the transport ofRSV, a probe substrate most commonly used in clinical DDIstudies.Given that the prediction ofDDI ofmanyOATP1B substrates,

including RSV, is complex, requiring input parameters of frac-tion eliminated via each pathway, we evaluated the dispositionprofile of RSV in BDC monkeys using [3H]RSV. Recovery ofradioactivity was complete after 168 hours after dosing. Ex-cretory profiles of radioactivity in the urine, bile, and fecessuggested that fecal excretion followed by biliary eliminationwas the major route of elimination of drug-derived radioactivity(61.7% and 39.0% of dose, respectively; Table 1). Metabolite

profiles in urine and bile were qualitatively similar. UnchangedRSV was detected as the major component in both pooled bileand urine from monkeys (62.8% and 45.7% of the radioactivitywere collected in bile and urine, respectively; Fig. 1), indicatingthat RSV did not undergo extensive metabolism before excre-tion in monkeys. The radioactivity in plasma was too low toexamine the metabolite profile. Unfortunately, the metabolicprofile of fecal samples was not examined.If we assume no significant intestinal secretion and gastro-

intestinal metabolism, the total unchanged RSV recovered inbile and feces would be 86.2% of the dose administered in BDCmonkeys. This is in agreement with what is observed in intactcynomolgus monkeys and humans (75.0% and 76.8%, respec-tively; Table 5) (Martin et al., 2003b; http://www.accessdata.fda.gov/drugsatfda_docs/nda/2003/21-366_Crestor_Pharmr_P2.pdf).The extent of intestinal secretion of RSV is known to be low indogs, with 3.3% of the radioactive dose found in the fecescollected from 0 to 72 hours after intravenous administration of5 mg/kg [14C]RSV to BDC dogs (http://www.accessdata.fda.gov/drugsatfda_docs/nda/2003/21-366_Crestor_Pharmr_P2.pdf).The percentage of RSV absorbed in the BDC monkeys, es-timated by adding the total radioactivity in bile and urine, wasat least 41.9%, suggesting that the compound is reasonablyabsorbed in monkeys. This is comparable to the oral absorptionfraction estimate of approximately 50% in humans (Table 5)(Martin et al., 2003a; http://www.accessdata.fda.gov/drugsatfda_docs/nda/2003/21-366_Crestor_Pharmr_P2.pdf). Putting thesefindings together, the similarities observed in absorption,metabolism, and excretion properties between cynomolgusmonkeys and humans suggest that the monkey is a suitablesurrogate animal model for further preclinical pharmacokineticstudies for RSV.Previously we reported that the cynomolgus monkey is an

effective model to assess investigational drugs for OATPinteraction in humans. Moreover, we provided evidence thatRSV-rifampicin (RIF) is an OATP1B probe substrate/referenceinhibitor combination applicable in several assay systems invitro and in vivo (Shen et al., 2013). In the present study, weextended the application of the cynomolgus monkey model byvalidation of RSV-CsA as OATP1B probe substrate/referenceinhibitor combination becauseCsA is commonly used as a clinicalOATP1B inhibitor.The potential hepatic transporter-mediated DDI between

RSV with CsA was first investigated in monkey and humanOATP1B and NTCP recombinant systems. CsA inhibited RSVuptake mediated by hOATP1B1 and cOATP1B1, with IC50

values of 0.216 0.10 and 0.286 0.11 mM, respectively (Fig. 2,A and C; Table 2). CsA also inhibited RSV uptake mediated byhOATP1B3 and cOATP1B3, in a similarly potent manner,with IC50 values of 0.13 6 0.06 and 0.25 6 0.09 mM, respec-tively (Fig. 2, B and D; Table 2). Moreover, CsA decreased theRSV uptake mediated by cNTCP in a concentration-dependentmanner, with an IC50 value of 3.9 6 2.0 mM, which is com-parable to that of humanNTCPgenerated from 55 different testoccasions at Bristol-Myers Squibb (Fig. 3B; Table 2).Furthermore, the potential hepatic transporter-mediated

DDI between RSV with CsA was investigated in monkey andhuman hepatocyte inhibition assays. CsA significantly re-duced uptake of RSV in both human and monkey hepatocytesin a concentration-dependent manner, with IC50 values of0.306 0.08 and 0.296 0.11 mM, respectively (Fig. 4; Table 2).These data are in agreement with the recombinant data, which

TABLE 3Pharmacokinetics of RSV and CsA after a single oral dose of 3 mg/kg RSVwith and without 100 mg/kg oral CsA in cynomolgus monkeysData reported as mean 6 S.D., n = 3 different animals.

Parameter RSV Alone CsA + RSVa GMR (90% CI)

RSVCmax (nM) 14.3 6 5.2 145.5 6 53.1 10.2 (3.3–31.9)tmax (h) 2.7 6 0.6 2.7 6 0.6t1/2 (h) 9.1 6 3.1 4.3 6 1.3AUC0–inf (nM×h) 133.9 6 34.5 897.5 6 462.0 6.3 (3.1–12.9)

CsAa

Cmax (mM) 1.1 6 0.3tmax (h) 4.3 6 1.5t1/2 (h) 5.9 6 0.4AUC0–49 (mM×h) 10.0 6 3.0C25 h (mM) 0.05 6 0.01

AUC0–inf area under the concentration-time curve from time 0 to infinity; C25 hconcentration 25 hours after the dose; CI, confidence interval; GMR geometric meanratio; t1/2 terminal elimination half-life; tmax, time of Cmax.

aCsA was dosed 1 hour before RSV (see Materials and Methods).

Fig. 6. Effects of CsA on urinary exertion of RSV after single oral dose ofRSV (3 mg/kg) alone or with CsA (100 mg/kg p.o.) in cynomolgus monkeys.The amount of RSV excreted in urine in the monkeys was determined forthe RSV alone (open bar) and CsA-treated (closed bar) groups. CsA wasdosed 1 hour ahead of RSV. Data are expressed as mean 6 S.D. (n = 3animals).

388 Shen et al.

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

showed that CsA inhibited the hOATP1B- and cOATP1B-mediated transport of RSV, with IC50 ranging from 0.13 to0.28 mM.No attempt was made to study OATP2B1 inhibition by CsA

because previous studies suggested that OATP2B1 is less likelyto play a significant role in RSV disposition in both species.Moreover, CsAwas aweak inhibitor of bothmonkey and humanOATP2B1 (Shen et al., 2013). Similarly, Prueksaritanont et al.(2014) have shown via RIF inhibition andDDI data that humanOATP2B1 contributed minimally to the hepatic uptake of RSV.Using in vitro transporter inhibition studies as screening

tools to evaluate the potential for DDI in vivo is based on theassumption that the victim drugs analyzed share the sametransport kinetics between cynomolgus monkey and humanOATP1B. Previous concentration-dependent transport studiesusing stably transfected HEK-293 cells, expressing individualmonkey and human OATP1B1, OAPT1B3, and OATP2B1, in-dicated that the RSV apparent Michaelis–Menten constant Km

(9.6–15.3 mM) is comparable across the three transporters(Shen et al., 2013). The transport kinetics of RSV in hepatocyteshas been examined also. In this instance, the kinetics of RSVuptake were best described by a mixed model consisting of botha single saturable process and a passive component, yieldingKm

of 6.7 and 10.3 mM for monkey and human, respectively, whichagreed well with the Km values obtained from the recombinantsystems. These studies suggest no species difference in thetransport kinetics of RSV in monkey and human OATP1B1-,OATP1B3-, and OATP2B1-overexpressing cells and hepato-cytes (Shen et al., 2013).Coadministration of CsA with RSV markedly increased the

AUC0–inf and Cmax of RSV in the cynomolgus monkey by 6.3-fold and 10.2-fold, respectively (Table 3). After oral adminis-tration of 100 mg/kg CsA, blood concentrations of CsA in therange of 1.1 to 0.05 mM were achieved for the first 25 hours(the first 24 hours after RSV administration) (Fig. 5B). The

Cmax of CsA after a single oral dose in monkeys was com-parable to that at steady state after multiple dosing in pa-tients (1.1 versus 1.0 mM) (Novartis, 2005).Interestingly, the impact of CsA on RSV in monkeys

appeared to be identical to that in heart transplant patientstaking CsA (7.1- and 10.6-fold for AUC and Cmax, respectively)(Simonson et al., 2004). This finding, although not unantic-ipated, suggests the cynomolgus monkey is an appropriatemodel for the assessment of OATP-mediated DDIs in a non-clinical setting. It is not clear, however, whether this can beextended to other substrates of these transporters. First,although cynomolgus monkey OATPs share a high degree ofamino acid sequence identity and functional similarity totheir human counterparts, subtle amino acid differences cangreatly impact substrate specificity. For example, DeGorteret al. (2012) reported that site-directed mutagenesis of threeamino acid residues in OATP1B1 transmembrane domains 1and 10, and extracellular loop 6 to the corresponding residuesin OATP1B3 resulted in a gain of CCK-8 transport byOATP1B1, which is a high-affinity substrate for OATP1B3but not OATP1B1. Second, this in vitro–in vivo extrapolationapproach only makes a reasonable prediction if the relativecontribution of OATP-mediated uptake clearance to the totalbody clearance and ADME profiles are well understood inboth species. The BDCmonkey study indicated that RSV is anappropriate in vivo probe for OATP-mediated DDI study inmonkeys. These conclusions likely extend to pitavastatin.Takahashi et al. (2013) reported that the magnitude ofhepatic OATP DDI was comparable between the monkeystudy and the clinical study by using pitavastatin as a sub-strate. They concluded that pharmacokinetic studies usingpitavastatin as a probe in combination with drug candidatesin cynomolgus monkeys are useful to support the assess-ment of potential clinical DDIs involving hepatic uptaketransporters.

TABLE 4Urinary excretion of RSV after a single oral dose of 3 mg/kg RSV with and without 100 mg/kg oral CsA in cynomolgusmonkeysData reported as mean 6 S.D., n = 3 different animals.

ParameterAnimal 1 Animal 2 Animal 3 Average

RSV Alone CsA + RSV RSV Alone CsA + RSV RSV Alone CsA + RSV RSV Alone CsA + RSV

Body weight (kg) 3.4 3.4 3.9 3.7 4.1 4.2 3.8 6 0.4 3.8 6 0.4Xe, 0–48 h (nmol) 120.6 925.1 253.9 1819.0 267.9 960.2a 214.1 6 81.3 1234.8 6 506.3Xe, 0–48 h (% dose) 0.6 4.5 1.1 8.1 1.1 3.8a 0.92 6 0.29 5.5 6 2.3AUC0–inf (nM×h) 95.8 706.3 162.9 1424.4 143.0 561.8 133.9 6 34.5 897.5 6 462.0CLR (ml/min/kg) 6.1 6.3 6.7 5.7 7.6 6.8a 6.8 6 0.7 6.3 6 0.6

CLR renal clearance; Xe, 0–48 h, amount or percentage of RSV recovered in urine over 48 hours.aThe urine sample collected between 0 and 7 hours from animal 3 was contaminated, the amount of RSV excreted into urine during the interval

was not included into the cumulative excretion.

TABLE 5Percentage of dose recovery in monkeys and humans after oral administration of radiolabeled RSV

Species Oral Dose% Dose

Urine (Parent) Feces (Parent) Bile (Parent) Total Recovery

BDC monkey 3 mg/kg 2.9 (1.3) 61.7 (ND) 39 (24.5) 103.6Intact monkeya 10 mg/kg 5.0 (1.0) 90.7 (75.0) 95.7Humanb 20 mg 10.6 (4.9) 90.0 (76.8) 100.6

ND, not determined.ahttp://www.accessdata.fda.gov/drugsatfda_docs/nda/2003/21-366_Crestor_Pharmr_P2.pdf.bMartin et al., 2003a.

Use of Cynomolgus Monkey to Assess DDIs Involving OATPs 389

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

RSV has been shown in vitro to be a substrate of trans-porters other than OATP1B1, OATP1B3, and OATP2B1. Forexample, it has been reported that human NTCP plays animportant role in hepatic uptake of RSV in human hepato-cytes (Ho et al., 2006; Bi et al., 2013), although the relevanceof this transporter in vivo has yet to be confirmed. In addition,BCRP and OAT3 (organic anion transporter 3) may play a rolein RSV intestinal absorption and renal elimination, respec-tively (Yoshida et al., 2012). In the case of the former, patientsexpressing the ABCG2 variant 421C.A, a single-nucleotidepolymorphism (SNP) associated with reduced efflux activityin vitro, showed a 140% increase in RSV exposure (Keskitaloet al., 2009). This implies that inhibition of intestinal BCRPcan also bring about increased RSV exposure. Therefore, theimpact on RSV pharmacokinetics will likely be determined byinhibition of OATPs, NTCP, BCRP, and OAT3. Considerationof OAT3 has been ruled out in this instance, because CsA didnot impact the renal clearance of RSV (Table 4).In our present study, we report for the first time that CsA is

a potent inhibitor of cOATPs and cNTCP (IC50 values of ∼0.3mM and 3.9 mM, respectively; Table 2). In addition, CsA in-hibits these monkey hepatic transporters to a similar extentcompared with the human counterparts. Consistent withhuman data also is the marked effect of CsA on RSV exposurewhen compared with RIF (6.3- to 7.1-fold versus 2.9- to 4.4-fold increase) (Simonson et al., 2004; Shen et al., 2013). Suchresults are consistent with the fact that DDI studies with CsAhave been accepted by regulatory agencies as the worst-casescenario for substrates of transporters. Although RIF is alsoa potent inhibitor of OATPs (IC50: 0.42–1.69mM), it is a weakerinhibitor of NTCP (IC50 of 277 mM versus 3.9 mM) and BCRP(IC50 of 14 mMversus ∼7 mM) (Nezasa et al., 2002b; Shen et al.,2013; Prueksaritanont et al., 2014), so it is hypothesized thatthe extent of inhibition of liver basolateral NTCP and OATP,and intestinal apical BCRP, is greater for CsA (versus RIF) inmonkeys and humans. In addition, the decreased hepaticuptake could impact the extent of metabolism of RSV.In summary, an in vitro–in vivo assay system previously

used to evaluate the interaction between RIF with RSV incynomolgus monkeys was extended to include CsA. Here weprovided further evidence that the disposition of radiolabeledRSV in cynomolgus monkeys is comparable to that in humansafter an oral dose. In addition, the magnitude of the DDIbetween CsA and RSV is similar to that reported clinically.Although additional transporter data are needed for monkeyBCRP, the results described herein do suggest that RSV canbe a useful substrate for probing OATP-mediated pharmaco-kinetics and DDIs in humans and monkeys.

Acknowledgments

The authors thank Dr. Lisa Elkin for providing IC50 values of CsAagainst hNTCP, which are generated from 55 different test occasionsat Bristol-Myers Squibb.

Authorship Contributions

Participated in research design: Shen, Su, Mintier, Iyer, Marathe,Lai, Rodrigues.

Conducted experiments: Shen, Su, Liu, Yao, Mintier, Li.Contributed new reagents or analytic tools: Shen, Liu, Mintier,

Fancher.Performed data analysis: Shen, Su, Liu, Mintier, Lai, Rodrigues.Wrote or contributed to the writing of the manuscript: Shen,

Marathe, Lai, Rodrigues.

References

Bergman E, Lundahl A, Fridblom P, Hedeland M, Bondesson U, Knutson L,and Lennernäs H (2009) Enterohepatic disposition of rosuvastatin in pigs and theimpact of concomitant dosing with cyclosporine and gemfibrozil. Drug Metab Dis-pos 37:2349–2358.

Bi YA, Qiu X, Rotter CJ, Kimoto E, Piotrowski M, Varma MV, Ei-Kattan AF, and LaiY (2013) Quantitative assessment of the contribution of sodium-dependenttaurocholate co-transporting polypeptide (NTCP) to the hepatic uptake ofrosuvastatin, pitavastatin and fluvastatin. Biopharm Drug Dispos 34:452–461.

Chang JH, Ly J, Plise E, Zhang X, Messick K, Wright M, and Cheong J (2014)Differential effects of Rifampin and Ketoconazole on the blood and liver concen-tration of atorvastatin in wild-type and Cyp3a and Oatp1a/b knockout mice. DrugMetab Dispos 42:1067–1073.

DeGorter MK, Ho RH, Leake BF, Tirona RG, and Kim RB (2012) Interaction of threeregiospecific amino acid residues is required for OATP1B1 gain of OATP1B3substrate specificity. Mol Pharm 9:986–995.

Hagenbuch B and Meier PJ (2004) Organic anion transporting polypeptides of theOATP/ SLC21 family: phylogenetic classification as OATP/ SLCO superfamily, newnomenclature and molecular/functional properties. Pflugers Arch 447:653–665.

Higgins JW, Bao JQ, Ke AB, Manro JR, Fallon JK, Smith PC, and Zamek-Gliszczynski MJ (2014) Utility of Oatp1a/1b-knockout and OATP1B1/3-humanizedmice in the study of OATP-mediated pharmacokinetics and tissue distribution: casestudies with pravastatin, atorvastatin, simvastatin, and carboxydichlorofluorescein.Drug Metab Dispos 42:182–192.

Ho RH, Tirona RG, Leake BF, Glaeser H, Lee W, Lemke CJ, Wang Y, and Kim RB(2006) Drug and bile acid transporters in rosuvastatin hepatic uptake: function,expression, and pharmacogenetics. Gastroenterology 130:1793–1806.

International Transporter Consortium, Giacomini KM, Huang SM, Tweedie DJ,Benet LZ, Brouwer KL, Chu X, Dahlin A, Evers R, Fischer V, Hillgren KM, et al.(2010) Membrane transporters in drug development. Nat Rev Drug Discov 9:215–236.

Iusuf D, Ludwig M, Elbatsh A, van Esch A, van de Steeg E, Wagenaar E, van derValk M, Lin F, van Tellingen O, and Schinkel AH (2014) OATP1A/1B transportersaffect irinotecan and SN-38 pharmacokinetics and carboxylesterase expression inknockout and humanized transgenic mice. Mol Cancer Ther 13:492–503.

Keskitalo JE, Zolk O, Fromm MF, Kurkinen KJ, Neuvonen PJ, and Niemi M (2009)ABCG2 polymorphism markedly affects the pharmacokinetics of atorvastatin androsuvastatin. Clin Pharmacol Ther 86:197–203.

Kitamura S, Maeda K, Wang Y, and Sugiyama Y (2008) Involvement of multipletransporters in the hepatobiliary transport of rosuvastatin. Drug Metab Dispos 36:2014–2023.

Klaassen CD and Lu H (2008) Xenobiotic transporters: ascribing function from geneknockout and mutation studies. Toxicol Sci 101:186–196.

Li L, Nouraldeen A, and Wilson AG (2013) Evaluation of transporter-mediated he-patic uptake in a non-radioactive high-throughput assay: a study of kinetics, spe-cies difference and plasma protein effect. Xenobiotica 43:253–262.

Lu H, Choudhuri S, Ogura K, Csanaky IL, Lei X, Cheng X, Song PZ, and Klaassen CD(2008) Characterization of organic anion transporting polypeptide 1b2-null mice:essential role in hepatic uptake/toxicity of phalloidin and microcystin-LR. ToxicolSci 103:35–45.

Martin PD, Warwick MJ, Dane AL, Brindley C, and Short T (2003a) Absolute oralbioavailability of rosuvastatin in healthy white adult male volunteers. Clin Ther25:2553–2563.

Martin PD, Warwick MJ, Dane AL, Hill SJ, Giles PB, Phillips PJ, and Lenz E (2003b)Metabolism, excretion, and pharmacokinetics of rosuvastatin in healthy adult malevolunteers. Clin Ther 25:2822–2835.

National Research Council (NRC) (1996) Guide for the Care and Use of LaboratoryAnimals. 7th ed. National Academies Press, Washington, DC.

Nezasa K, Higaki K, Matsumura T, Inazawa K, Hasegawa H, Nakano M, and KoikeM (2002a) Liver-specific distribution of rosuvastatin in rats: comparison withpravastatin and simvastatin. Drug Metab Dispos 30:1158–1163.

Nezasa K, Takao A, Kimura K, Takaichi M, Inazawa K, and Koike M (2002b) Phar-macokinetics and disposition of rosuvastatin, a new 3-hydroxy-3-methylglutarylcoenzyme A reductase inhibitor, in rat. Xenobiotica 32:715–727.

Novartis (2005) Prescribing information on Neoral soft gelatin capsules and Neoralcyclosporine oral solution, in Physicians’ Desk Reference, vol 59, pp 2346–2347,Medical Economics Co., Montvale, NJ.

Pasanen MK, Fredrikson H, Neuvonen PJ, and Niemi M (2007) Different effects ofSLCO1B1 polymorphism on the pharmacokinetics of atorvastatin and rosuvastatin.Clin Pharmacol Ther 82:726–733.

Prueksaritanont T, Chu X, Evers R, Klopfer SO, Caro L, Kothare PA, Dempsey C,Rasmussen S, Houle R, Chan G, et al. (2014) Pitavastatin is a more sensitive andselective organic anion-transporting polypeptide 1B clinical probe than rosuvastatin.Br J Clin Pharmacol 78:587–598.

Salphati L, Chu X, Chen L, Prasad B, Dallas S, Evers R, Mamaril-Fishman D, GeierEG, Kehler J, Kunta J, et al. (2014) Evaluation of organic anion transportingpolypeptide 1B1 and 1B3 humanized mice as a translational model to study thepharmacokinetics of statins. Drug Metab Dispos 42:1301–1313.

Schneck DW, Birmingham BK, Zalikowski JA, Mitchell PD, Wang Y, Martin PD,Lasseter KC, Brown CD, Windass AS, and Raza A (2004) The effect of gemfibrozilon the pharmacokinetics of rosuvastatin. Clin Pharmacol Ther 75:455–463.

Schuurman HJ, Slingerland W, Mennninger K, Ossevoort M, Hengy JC, Dorobek B,Vonderscher J, Ringers J, Odeh M, and Jonker M (2001) Pharmacokinetics ofcyclosporine in monkeys after oral and intramuscular administration: relation toefficacy in kidney allografting. Transpl Int 14:320–328.

Shen H, Yang Z, Mintier G, Han YH, Chen C, Balimane P, Jemal M, Zhao W, ZhangR, Kallipatti S, et al. (2013) Cynomolgus monkey as a potential model to assessdrug interactions involving hepatic organic anion transporting polypeptides: invitro, in vivo, and in vitro-to-in vivo extrapolation. J Pharmacol Exp Ther 344:673–685.

390 Shen et al.

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from

Shirasaka Y, Kuraoka E, Spahn-Langguth H, Nakanishi T, Langguth P, and Tamai I(2010) Species difference in the effect of grapefruit juice on intestinal ab-sorption of talinolol between human and rat. J Pharmacol Exp Ther 332:181–189.

Shitara Y, Li AP, Kato Y, Lu C, Ito K, Itoh T, and Sugiyama Y (2003) Function ofuptake transporters for taurocholate and estradiol 17b-D-glucuronide in cryopreservedhuman hepatocytes. Drug Metab Pharmacokinet 18:33–41.

Simonson SG, Raza A, Martin PD, Mitchell PD, Jarcho JA, Brown CD, Windass AS,and Schneck DW (2004) Rosuvastatin pharmacokinetics in heart transplantrecipients administered an antirejection regimen including cyclosporine. ClinPharmacol Ther 76:167–177.

Takahashi T, Ohtsuka T, Yoshikawa T, Tatekawa I, Uno Y, Utoh M, Yamazaki H,and Kume T (2013) Pitavastatin as an in vivo probe for studying hepatic organicanion transporting polypeptide-mediated drug-drug interactions in cynomolgusmonkeys. Drug Metab Dispos 41:1875–1882.

van de Steeg E, van der Kruijssen CM, Wagenaar E, Burggraaff JE, MesmanE, Kenworthy KE, and Schinkel AH (2009) Methotrexate pharmacoki-netics in transgenic mice with liver-specific expression of human organicanion-transporting polypeptide 1B1 (SLCO1B1). Drug Metab Dispos 37:277–281.

Wang L, Prasad B, Salphati L, Chu X, Gupta A, Hop CE, Evers R, and Unadkat JD(2015) Interspecies variability in expression of hepatobiliary transporters acrosshuman, dog, monkey, and rat as determined by quantitative proteomics. DrugMetab Dispos 43:367–374.

Wen JH and Xiong YQ (2011) The effect of herbal medicine danshensu and ursolicacid on pharmacokinetics of rosuvastatin in rats. Eur J Drug Metab Pharmacokinet36:205–211.

Yoshida K, Maeda K, and Sugiyama Y (2012) Transporter-mediated drug—druginteractions involving OATP substrates: predictions based on in vitro inhibitionstudies. Clin Pharmacol Ther 91:1053–1064.

Zaher H, Meyer zu Schwabedissen HE, Tirona RG, Cox ML, Obert LA, Agrawal N,Palandra J, Stock JL, Kim RB, and Ware JA (2008) Targeted disruption of murine organicanion-transporting polypeptide 1b2 (Oatp1b2/Slco1b2) significantly alters disposition ofprototypical drug substrates pravastatin and rifampin. Mol Pharmacol 74:320–329.

Address correspondence to: Hong Shen, F1.3802, Route 206 and ProvinceLine Road, Bristol-Myers Squibb Company, Princeton, NJ 08543-4000. E-mail:hong.shen1@bms.com

Use of Cynomolgus Monkey to Assess DDIs Involving OATPs 391

at ASPE

T Journals on A

ugust 20, 2020jpet.aspetjournals.org

Dow

nloaded from