Atorvastatin (ATV), a member of the class of 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors, also known as statins, is one of the most potent drugs for treatment of hypercholesterolemia (Nawrocki et al., 1995). ATV undergoes extensive metabolism primarily by cytochrome P450 3A (CYP3A) in the liver to form two active metabolites, ortho-hydroxy atorvastatin and para-hydroxy atorvastatin (Jacobsen et al., 2000). Black et al. (1999) demonstrated that ATV undergoes a similar metabolic pathway in rats, where both parent and metabolite compounds were exclusively excreted in the bile, indicating that hepatic metabolism and biliary excretion are the major routes of elimination.

Drug-drug interactions of ATV are often reported at the level of hepatic phase I or phase II enzymes (Kantola et al., 1998; Lennernas, 2003). Coadministration of ATV and an inhibitor for CYP3A or UDP-glucuronosyltransferase may increase ATV blood concentrations and the risk of rhabdomyolysis (Jacobson, 2004). However, interactions can also occur by inhibition of the relevant hepatic transporters that are located both upstream and downstream of hepatic enzymes. Our group has demonstrated previously, using digoxin and erythromycin as examples, that the ratio of intracellular drug concentrations of parent to metabolite may change when hepatic transporters (both uptake and efflux) are inhibited because the amount of drug available to the enzyme varies with transporter activity (Lam and Benet, 2004; Lau et al., 2004; Sun et al., 2004).

Two carrier-mediated processes govern the sequential transport across the hepatocytes: the sinusoidal (basolateral) uptake from portal/arterial circulation and the canalicular (apical) efflux into the bile. Canalicular efflux transporters P-glycoprotein (P-gp) and multidrug resistance associated protein 2 (MRP2/Mrp2; human/rodent) are the two major ATP-dependent efflux pumps for excretion of drugs into the bile (Keppler and Arias, 1997). At the basolateral membrane of hepatocytes, uptake transporters such as the organic anion transporting polypeptide (OATP/Oatp) are responsible for the influx of bulky amphipathic and hydrophilic compounds (Meier et al., 1997).

ATV has been shown to be a P-gp substrate in various transport studies (Wu et al., 2000; Hochman et al., 2004; Chen et al., 2005). ATV can inhibit MRP2-mediated efflux (Chen et al., 2005), but whether or not it is a substrate of MRP2 had not been confirmed, although other statins, such as pravastatin, have been shown to be substrates of MRP2 (Sasaki et al., 2002; Nezasa et al., 2003). Structurally, ATV consists of a lipophilic region and a more hydrophilic part. Its lipophilicity makes it a good substrate for CYP3A as well as for efflux transporters. In contrast, its hydrophilicity makes it a likely candidate for uptake transporters such as OATP/Oatp. In vitro studies have demonstrated that ATV and other statins, such as cerivastatin, pravastatin, and rosuvastatin, are transported by and are inhibitors of the liver-specific OATP1B1 (previously known as OATP-C or OATP2) (Hsiang et al., 1999; Tokui et al., 1999; Shitara et al., 2003; Simonson et al., 2004; Chen et al., 2005; Kameyama et al., 2005).

As of today, there is no published information on the influence of rat Oatps on the hepatic uptake of ATV and its two hydroxy metabolites. To estimate the contribution of Oatps on the hepatic disposition and metabolism of ATV, rifampicin (RIF) was chosen as a model inhibitor for Oatps since it has been shown to effectively inhibit the OATP/Oatp-mediated uptake of various anionic compounds, including taurocholate, estrone sulfate, as well as the more neutral digoxin, with relatively low K_i values (1-2 μM) (Fattinger et al., 2000; Shitara et al., 2002; Vavricka et al., 2002; Tirona et al., 2003). On the basis of in vitro studies using MRP2-overexpressing cells, Cui et al. (2001) showed that RIF can also inhibit the MRP2-mediated efflux of bromosulfophthalein and leukotriene C₄ at high concentrations. Moreover, RIF has also demonstrated an inhibitory effect on P-gp radiolabeling by the photoactivatable P-gp ligand azidopine (Fardel et al., 1995). This is not surprising, because OATP and MRP2 and/or P-gp often work in a concerted manner for drug elimination and share similar substrate and inhibitor specificity, even though the affinity and inhibition potency for uptake and efflux transporters might differ.

In this study, we evaluated the influence of individual hepatic uptake and efflux transporters on the disposition of ATV and its metabolites and characterized the effects of RIF on the hepatic uptake and efflux, as well as metabolism of ATV, using cellular systems and rat microsomes. To directly observe how the dynamic interplay between transporters and enzymes contributes to alterations in ATV pharmacokinetics, the isolated perfused rat liver (IPRL) system was used.

Materials and Methods

Materials. Atorvastatin (ATV), para-hydroxy atorvastatin (4-OH ATV) (PD142542, BMS-241423-01), and ortho-hydroxy atorvastatin (2-OH ATV) (PD152873, BMS243887-01) were kindly supplied by Parke-Davis (Ann Arbor, MI) and Bristol-Myers Squibb Co. (Princeton, CT). Fluvastatin (Novartis, Cambridge, MA) and GG918 (GF120918; GlaxoSmithKline, Research Triangle Park, NC) were kind gifts from the manufacturers. Rifampicin, NADPH, and sodium butyrate were purchased from Sigma-Aldrich (St. Louis, MO). All other chemicals were of reagent grade and purchased from either Sigma-Aldrich or Fisher Scientific Co. (Pittsburgh, PA). Human OATP1B1 and rat Oatp1a1 (Oatp1), Oatp1a4 (Oatp2), and Oatp1b2 (Oatp4) cDNA plasmids were kindly provided by Professor Richard Kim (Vanderbilt University, Nashville, TN). The MRP-2-overexpressing cell line MDCKII-canalicular multiorganic anion transporter (MII-cMOAT) and the respective wild-type cell line MII were a generous gift from Professor Piet Borst and Dr. Raymond Evers (The Netherlands Cancer Institute, Amsterdam, The Netherlands). MDCK and MDCK-MDR1 (M-MDR1) cell lines were generously provided by Dr. Ira Pastan (National Cancer Institute, National Institutes of Health, Bethesda, MD). Human embryonic kidney cells (HEK293) and all cell culture media were obtained from the University of California Cell Culture Facility (San Francisco, CA). Six-well plates were obtained from Corning Life Sciences (Acton, MA). Transwell inserts and poly-d-lysine-coated 12-well plates were obtained from BD Biosciences (Bedford, MA). The Lipofectamine 2000 transfection system was purchased from Invitrogen (Carlsbad, CA).

Transient Transfection and Uptake Transport Assays. All OATP/Oatp expression plasmids were sequence-verified, and when they were expressed in cells, they were shown to be transport-competent toward prototypical substrates (estrone sulfate, digoxin, cholecystokinin) (data not shown). HEK293 cells were cultured in Eagle's minimal essential medium with Eagle's balanced salt solution and l-glutamine plus 10% heat-inactivated fetal bovine serum (FBS), nonessential amino acids, sodium pyruvate, streptomycin, and penicillin. Cells were seeded into poly-d-lysine-coated 12-well plates at a density of 0.5 × 10⁶ cells/well 1 day prior to transient transfection with OATP1B1, Oatp1a1, Oatp1a4, and Oatp1b2 plasmids or pEF/V5-His vector control (Invitrogen) using the Lipofectamine 2000 transfection system according to the manufacturer's directions. Culture medium was replaced 24 h before the uptake studies with the same medium containing 10 mM sodium butyrate to induce the expression of transporters. Before initiation of the uptake study, cells were washed once with phosphate-buffered saline (PBS) prewarmed at 37°C. The uptake study was initiated by adding 0.5 ml of Opti-MEM I buffer (Invitrogen) containing substrates and incubating at 37°C for 3 min. Preliminary experiments had shown that the uptake rate was linear over this time period (data not shown). For the inhibition studies, inhibitors and substrates were added simultaneously. At the designated times, buffer was removed to terminate the reaction and the cells were washed three times with ice-cold PBS. The homogenate was centrifuged for 5 min at 13,000g, and the resulting supernatant was analyzed by LC/MS-MS. The protein concentrations in cell experiments were determined by the method of Lowry et al. (1951) using bovine serum albumin as a standard.

Preparation of Cell Culture Monolayers and Transport Studies. MII-cMOAT and MII cells were cultured at 37°C and humidified 5% CO₂ atmosphere in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated FBS, 100 U/ml penicillin, and 100 U/ml streptomycin. M-MDR1 cells were cultured in the same medium containing 80 mg/ml colchicine for selected growth of transfected cells (Pastan et al., 1988). Cells grown to confluence in culture flasks were harvested and seeded into Transwell inserts in six-well plates at a density of ∼10⁶ cells/insert. Studies were conducted 5 to 6 days postseeding for the two cell lines. Medium was changed once every 2 days and 24 h before the experiment.

The transport experiments were adapted with modifications from Flanagan et al. (2002) and Cummins et al. (2004). In brief, cell monolayers were preincubated in transport buffer (Hanks' balanced salt solution containing 25 mM HEPES and 1% FBS, pH 7.4) for 20 min at 37°C. Transepithelial electrical resistance was measured in each well using a Millicell ERS voltohmmeter (Millipore Corporation, Bedford, MA) to assess the integrity of monolayers. The average transepithelial electrical resistance values obtained from MII, MII-cMOAT, and M-MDR1 cells were 170 ± 15 Ω · cm² (n = 12), 160 ± 10 Ω · cm² (n = 12), and 1780 ± 20 Ω · cm² (n = 12), respectively. For measuring drug secretion (B)asolateral → (A)pical, 2.5 ml of transport buffer containing ATV (5 μM) was put into the B side and 1.5 ml of buffer was put into the A side. At the selected times (1, 2, and 3 h), 150-μl samples were taken from the A side and replaced with fresh buffer. For measuring drug absorption (A → B), the drug solution was put into the A side and samples were taken from the B side. For inhibition studies, the inhibitor RIF (50 μM) or GG918 (0.5 μM) was put into both the A and B sides. During the studies, the cells were incubated in a shaking incubator (Boekel Scientific, Feasterville, PA). After the last time point (3 h), the apical solutions were removed by suction and each filter was dipped twice in ice-cold PBS. Intracellular measurements of ATV were obtained by solubilizing the cells on each culture insert with 0.4 ml of ice-cold MeOH/H₂O [7/3 (v/v)] and sonicating for 10 min. The homogenate was centrifuged for 5 min at 13,000g, and the resulting supernatant was analyzed by LC/MS-MS.

Metabolism by Rat Liver Microsomes. Rat liver microsomes were isolated and incubated as described previously by our laboratory (Jacobsen et al., 2000). In brief, microsomal proteins (0.75 mg/ml), 0.1 M phosphate buffer, pH 7.4, and ATV (in methanol, final concentration 1 μM) were preincubated for 5 min. NADPH (1 mM) was added to start the reaction at 37°C. To assess the effects of RIF on ATV metabolism, RIF (in dimethyl sulfoxide, final concentration 10-250 μM) was added to the liver microsomal preparations. The reaction was stopped by protein precipitation by the addition of an equal volume of ice-cold acetonitrile containing the internal standard fluvastatin (0.5 μM). The supernatants were stored at -80°C for LC/MS-MS measurement.

Surgery and Perfusion of Isolated Rat Livers. Male Sprague-Dawley rats (300-400 g; Bantin and Kingman, San Leandro, CA) were anesthetized with ketamine/xylazine (80 mg; 12 mg/ml) before surgery. The hepatic portal vein and superior vena cava were cannulated after the approval of protocols by the Committee on Animal Research, University of California (San Francisco, CA). The livers were isolated for perfusion ex situ using standard techniques as described previously by our laboratory (Wu and Benet, 2003; Lau et al., 2004). In brief, recirculating perfusion was performed at 37°C from a reservoir containing 110 ml of perfusate composed of Krebs-Henseleit buffer, pH 7.4, supplemented with sodium taurocholate (220 nmol/min), 1% bovine serum albumin, and glucose (10 mM), through the liver via a catheter inserted in the portal vein. The perfusate in the reservoir was oxygenated directly using carbogen (95% O₂/5% CO₂) and stirred continuously. Measures of liver viability included oxygen consumption, portal vein pressure (20-30 mm Hg), pH 7.35-7.45, and metabolic capability. Livers were allowed to stabilize for 20 min. RIF was added to the reservoir 5 min before ATV addition. Perfusate samples (0.5 ml) were collected immediately (0 min) and at 3, 5, 10, 15, 20, 30, 45, and 60 min after the addition of ATV, and accumulative bile samples were collected for up to 60 min. At the end of experiment, the liver was removed, blotted dry, and weighed. An aliquot of liver was homogenized with ice-cold PBS in a 1:2 ratio and maintained frozen at -80°C before analysis. To examine the effects of RIF on ATV hepatic disposition, 20 rats were divided equally into four groups, and each group was perfused with a bolus dose of ATV to yield an initial perfusate concentration of ∼1 μM. Whereas one group served as the control, the other three groups served as treatment groups for inhibition studies, with three different final perfusate concentrations of RIF (5, 10, and 50 μM). Samples were prepared by liquid-liquid extraction as described previously (Lau et al., 2004). In brief, 3 ml of methyl tertiary butyl ether with internal standard fluvastatin (0.5 μM) was added to each liver, bile, and perfusate sample. After centrifugation, separation of the organic layer in a methanol dry ice bath, and evaporation of the organic layer under nitrogen gas, the dried solutes were reconstituted with methanol for analysis by LC/MS-MS.

LC/MS-MS Measurement of ATV and Metabolites. A Micro-mass Quattro Ultima instrument (Waters, Milford, MA) with electrospray-positive ionization was used. The multiple reaction monitor was set at 559.6-440.8 m/z for ATV, 575.2-440.5 m/z for 2-OH ATV and 4-OH ATV, and 412.3-265.9 m/z for the internal standard fluvastatin. The cone voltage and collision energy were set at 30 V and 20 eV, respectively. The analytical Agilent XDB C18 column (4.6 × 50 mm, 5-μm particle size; Agilent Technologies, Palo Alto, CA) was used. The mobile phase consisted of 46% acetonitrile containing 0.05% acetic acid and 5 mM ammonium acetate. Five-microliter aliquots were injected, and the flow rate was set at 1.0 ml/min with a one-fourth split into the mass system.

HPLC Measurement of RIF. The HPLC method for analyzing RIF was described previously (Lau et al., 2004). In brief, liver samples were analyzed on a HP1100 HPLC system (Hewlett Packard, Palo Alto, CA). RIF was resolved on a Microsorb-MV C₈ analytical column (4.6 × 250 mm, 4-μm particle size; Varian, Inc., Walnut Creek, CA). The mobile phase was composed of 0.05 M potassium dihydrogen phosphate/acetonitrile [55/45(v/v)] at a flow rate of 1 ml/min. The UV absorbance was monitored at a wavelength of 340 nm.

Data Analysis. The kinetics parameters for the uptake of ATV in the transporter-expressing HEK293 cells were obtained by using the following equation,

where v₀ is initial uptake rate, S is the substrate concentration, K_m is the Michaelis constant, V_max is the maximal uptake rate, and P_dif is the nonsaturable uptake clearance. The above equation was fitted to the data, using a nonlinear least-squares method by SigmaPlot (Version 5.0; SPSS Inc., Chicago, IL). To obtain the IC₅₀ value of RIF inhibition on ATV uptake, the data were fitted using WinNonlin (Scientific Consulting, Inc., Apex, NC). For transport studies, the permeability (P_app) values were calculated as follows where the rate of transport was measured from the flux of drug across the cells.

For the IPRL studies, values for area under the concentration time curve (AUC) were calculated using the linear trapezoidal method. Biliary clearance (CL_b) was the quotient of the cumulative biliary amount of ATV and the AUC of ATV. Student's t test was used to analyze differences between two groups. Analysis of variance was used to analyze differences among more than two groups, and the significance of difference between two means in these groups was evaluated using Tukey's post hoc test. The p value for statistical significance was set at <0.05.

Results

Uptake Studies in HEK293 Cells with Transient Expression of Uptake Transporters. The uptake of ATV into various rat Oatps (Oatp1a1, Oatp1a4, and Oatp1b2) and human OATP1B1-transfected cells and vector-transfected cells is shown in Fig. 1. The uptake of ATV was significantly higher in Oatp1a4-, Oatp1b2-, and OATP1B1-transfected cells relative to that into vector-transfected cells (Fig. 1). Oatp1a4 also showed a lower apparent transport efficiency compared with Oatp1b2 using prototypical substrates (data not shown). The concentration-dependent uptake of ATV was examined in Oatp1a4- and Oatp1b2-transfected cells (Fig. 2, A and B). Both a saturable and a nonsaturable component were observed as demonstrated by the Eadie-Hofstee plots (Fig. 2, A and B, insets). The obtained K_m, V_max, and P_dif values for the uptake of ATV in Oatp1a4- transfected cells were 22.2 ± 11.9 μM, 106.0 ± 49.7 pmol/min/mg protein, and 0.80 ± 0.59 μl/min/mg protein, respectively. The saturable component estimated by V_max/K_m accounts for approximately 86% of the total uptake. The corresponding values for the uptake of ATV in Oatp1b2-transfected cells were 7.12 ± 3.10 μM for K_m, 37.1 ± 11.1 pmol/min/mg protein for V_max, and 1.50 ± 0.15 μl/min/mg protein for P_dif. Saturable uptake accounts for approximately 79% of the total uptake. Oatp1a4- and Oatp1b2-mediated uptake of ATV was inhibited by RIF in a concentration-dependent manner (Fig. 3). The K_i value was estimated by IC₅₀/[1 + S/K_m]. The estimated IC₅₀ and K_i values for RIF inhibition on Oatp1a4-mediated uptake of ATV were 3.05 ± 1.44 and 2.88 ± 1.33 μM, respectively. The corresponding values for RIF inhibition on Oatp1b2-mediated uptake of ATV were 0.95 ± 0.25 and 0.79 ± 0.13 μM. In addition, Oatp1b2 was capable of transporting 2-OH ATV and 4-OH ATV (Fig. 4). In the presence of excess RIF (50 μM), the uptake of these metabolites was reduced to almost the same level as that in vector-transfected cells (Fig. 4). ATV (1 μM) uptake was inhibited by both 2-OH ATV and 4-OH ATV at concentrations of 1 and 10 μM (Fig. 5).

Transport Studies of ATV in MII, MII-cMOAT, and M-MDR1 Cell Systems. A significantly (p < 0.01) higher B → A transport of ATV across MII-cMOAT cells (P_app = 2.65 × 10^-6 cm s^-1) than across MII cells (P_app = 1.52 × 10^-6 cm s^-1) was observed resulting in a net flux ratio (B → A/A → B) ratio of 3.03 ± 0.43 for MII-cMOAT cells (Table 1). In MII cells, the net flux ratio was also higher than 1. This could be explained by the presence of endogenous canine P-gp as well as MRP2 present in the cells, as previously shown in our laboratory (Flanagan et al., 2002) by Western blot, which would be expected to generate partial transepithelial transport of ATV. With the addition of RIF (50 μM), the B → A permeability of ATV across MII-cMOAT was significantly reduced by ∼50%, resulting in a 3-fold decrease in the net flux ratio (Table 1). The intracellular ATV accumulation also significantly increased (p < 0.05) compared with the MII-cMOAT control (Table 1). In MII cells, the B to A transport of ATV was also decreased by RIF because of the presence of modest expression levels of endogenous MRP2 (Flanagan et al., 2002). RIF demonstrated no inhibitory effect on P-gp at 50 μM using P-gp-overexpressing M-MDR1 cells (Fig. 6A); however, GG918, a relatively specific inhibitor for P-gp (Cummins et al., 2004), at 0.5 μM completely eliminated the bidirectional difference in transport (Fig. 6B).

Metabolic Studies of ATV by Rat Microsomes. The inhibitory effects of RIF (0-250 μM) on the metabolism of ATV were examined (Table 2). RIF did not alter the metabolic formation of 2-OH ATV up to a concentration of 25 μM and reduced it to 57% of the control value at 250 μM. RIF did not significantly affect the formation of 4-OH ATV at concentrations lower than 100 μM.

IPRL Studies of ATV.Figure 7A illustrates the disposition of ATV in perfusate over time in the absence (control) and presence of various concentrations of RIF (5, 10, and 50 μM). In all RIF-treated groups, ATV concentrations at all of the sampling times were elevated (Fig. 7A). The AUC of ATV over the 1-h sampling period was significantly increased by 56, 108, and 151% for 5, 10, and 50 μM RIF-treated groups, respectively, as shown in Table 3. The concentrations of 2-OH ATV and 4-OH ATV were detectable 15 min after the start of perfusion and increased steadily up to 60 min (Fig. 7, B and C), indicating that hepatic enzymatic activity of CYP3A was functional throughout the perfusion period. The observed increase in metabolites perfusate concentrations without a concomitant decrease in ATV perfusate concentrations after the rapid initial distribution phase might be due to significant evaporation from perfusate or cis-inhibition of ATV uptake by the metabolites accumulated in the perfusate. Both metabolites could inhibit ATV uptake at the same or 10-fold higher molar concentration of ATV in cellular uptake studies (Fig. 5).

The AUC of both hydroxy metabolites were increased by ∼20% for 5 μM RIF, ∼80% for 10 μM RIF (p < 0.01), and ∼100% by 50 μM RIF (p < 0.001) treated groups (Table 3). The ratio of the sum of the metabolite AUC to parent AUC (OH ATV AUC/ATV AUC) was comparable for all RIF-treated groups and control group.

The amounts of ATV and metabolites recovered in various matrices from IPRL studies in the absence and presence of RIF are presented in Table 4. The amounts of RIF recovered in the liver homogenate increased linearly with the RIF dose administered (y = 2.49x - 3.71; r² = 0.998). The majority of the dose recovered was present in the liver homogenate as ATV at the end of the experiment (1 h). ATV amount was significantly decreased in 5 μM RIF-treated livers and slightly decreased in 10 μM RIF-treated livers. In 50 μM RIF-treated livers, however, there was an increase in ATV amount, although not significant. Liver retention of metabolites was reduced, with a significant decrease of ∼50% for total metabolites (sum of 2-OH ATV and 4-OH ATV) in all RIF-treated groups. The ratios of total metabolites to parent retention in liver were calculated for all groups (Table 3). There was a gradual decrease of the ratios with increasing RIF concentration (19, 35, and 60% decrease for 5, 10, and 50 μM RIF, respectively). However, only the 50 μM RIF-treated group exhibited a significant decrease.

RIF exhibited a concentration-dependent inhibitory effect on the biliary excretion of ATV and metabolites, in which the cumulative amounts were significantly reduced in the 10 and 50 μM RIF-treated groups (Table 4; Fig. 8, A and B). The biliary clearance of ATV significantly decreased 51, 84, and 95% by 5, 10, and 50 μM RIF, respectively. The ratios of amounts of ATV in bile/liver were statistically unchanged for 5 and 10 μM RIF-treated groups but significantly decreased 88% by 50 μM RIF (Table 3). The bile/liver ratio for total OH ATV was not significantly changed between groups. Mass balance calculations (ATV, 2-OH ATV, and 4-OH ATV in perfusate, liver, and bile) show no statistical significant differences among the four studies (79.3 ± 8.4% for control, 70.0 ± 6.5% with 5 μM RIF, 79.4 ± 8.2% with 10 μM RIF, and 93.1 ± 13.9% with 50 μM RIF). However, for the 5 and 10 μM RIF treatment groups where we expect no effect on the enzymatic process, the increase in perfusate drug and metabolite amounts seems to be balanced by the decreased amounts in the liver so that the total percentage of metabolites (perfusate, liver plus bile) is essentially the same as that for control (28.7% for control, 26.4% for 5 μM RIF, and 30.8% for 10 μM RIF) (Table 4). This unexpected finding differs from our previous digoxin reports (Lam and Benet, 2004; Lau et al., 2004).

Discussion

It is now well recognized that the hepatic drug elimination process is mainly regulated by three distinct groups of proteins: uptake transporters, metabolizing enzymes, and efflux transporters arranged in a sequential manner. The goal of the present study was to investigate the complex interplay between these groups of proteins by applying the IPRL as a proof of concept model complemented by cellular studies to examine individually the relevance of the potential players in the disposition of ATV. In all of the cases, the ability of RIF to inhibit transporters and enzymes was tested.

Our uptake study results demonstrate that ATV is a substrate of OATP1B1, Oatp1a4, and Oatp1b2 (Fig. 1), the rat ortholog of OATP1B1 (Kakyo et al., 1999; Cattori et al., 2001). RIF is capable of inhibiting ATV uptake mediated by both Oatps in a concentration-dependent manner (Fig. 3). Recently, it has been demonstrated that RIF is a substrate of Oatp1b2 (Tirona et al., 2003), which is liver-specific and exhibits the highest amino acid sequence identity with OATP1B1 (Cattori et al., 2001). It is likely that RIF competitively reduces the sinusoidal uptake of ATV via this transporter as well as the more well studied Oatp1a4 (Fattinger et al., 2000; Shitara et al., 2002).

Similar to P-gp, MRP2 is an efflux transporter expressed at high levels in the canalicular domains of hepatocytes (Evers et al., 1998). Transport studies were used as a tool to evaluate the influence of MRP2 on drug export, because the cellular B → A transport mimics the orientation of apical efflux transporters in the liver. The results generated from the MII-cMOAT cells indicate that ATV is a cosubstrate of both P-gp (Fig. 6B) and MRP2 (Table 1), with a net flux ratio of 3 using the MII-cMOAT cells, a value that reflects the leaky nature of this cell line. Incubation with 50 μM RIF results in decreased P_app (B → A), a change in the flux ratio to approximate unity, along with a significant increase in intracellular ATV accumulation (Table 1). These results support the report of Cui et al. (2001) in that RIF is capable of inhibiting MRP2 as well as OATP.

The working concentration of ATV was carefully chosen at 1 μM in the IPRL study, which is well below the K_m values (23.8 and 19.8 μM for 2-OH ATV and 4-OH ATV formation, respectively) obtained from microsomal studies (data not shown) and uptake studies (Fig. 2, A and B), to prevent saturation of both enzyme and Oatp transporters. As expected, RIF significantly increased the AUC of ATV in the IPRL studies at all of the three tested concentrations (5, 10, and 50 μM).

The influence of Oatps on hepatic transport of the two ATV active hydroxy metabolites has not been reported previously. In the present IPRL studies, we observed an increase in AUC for both metabolites by RIF in a concentration-dependent manner. This phenomenon is consistent with the uptake study in which both metabolites are substrates of Oatp1b2 and that RIF significantly inhibits their uptake (Fig. 4). Our previous digoxin IPRL study (Lau et al., 2004) showed that inhibition of digoxin hepatic uptake led to an increase in parent drug AUC but a decrease in metabolite AUC; hence, a decrease in metabolite/parent AUC ratio was observed. We suspect that this decrease in ratio is a result of the metabolite not being a substrate, at least not as good a substrate for the uptake transporter as the parent drug. In the present study, the metabolite/parent AUC ratios remain statistically unchanged, indicating that the disposition of the two ATV metabolites were inhibited by RIF to a similar extent as the parent ATV.

Overlap in inhibitor selectivity for enzyme versus transporter makes it difficult to differentiate the roles of uptake transporter and enzyme in drug disposition, because inhibition of either can lead to an increase in drug exposure. Therefore, we examined the inhibitory effect of RIF on CYP-mediated metabolism of ATV in microsome. RIF seems to have only a moderate effect on metabolism, showing inhibition only at higher concentrations (≥50 μM) (Table 2). RIF liver concentrations reached values of 11.2 and 18.4 μM for the 5 and 10 μM RIF doses, respectively (Table 3), concentrations well below 50 μM. Therefore, less ATV and metabolites were detected in liver tissues for 5 and 10 μM RIF doses (Table 4), indicating that RIF has an inhibitory effect on uptake but no effect on metabolism at concentrations less than 50 μM in the liver. Although not statistically significant, the metabolites/parent liver ratios between the control and the 5 and 10 μM RIF treatment groups seem to decrease gradually with increasing RIF concentration (Table 3). Because the metabolites seem to be more susceptible than the parent to Oatp-mediated uptake, more metabolites would stay in the perfusate as opposed to the liver (Fig. 7, B and C); hence, the extent of decrease in liver retention for metabolites would be greater than that for parent drug.

A greater and significant decrease in the metabolite/parent liver ratio was observed for the 50 μM RIF treatment group (Table 3). Here, RIF liver concentration reached a value of 120 μM, a concentration high enough to inhibit the formation of both metabolites (Table 2). However, the increase in ATV retention in liver cannot be attributed to enzymatic inhibition only. Note that the bile/liver amount ratio for ATV was relatively constant across control and the 5 and 10 μM RIF treatment groups (Table 3), suggesting that the change of parent drug and metabolites in bile is reflective of what is observed in the liver. However, the ratio was significantly decreased in the 50 μM RIF treatment group (Table 3), suggesting that apical transporters mediating the efflux of ATV might be inhibited at this concentration. Determining whether the metabolites are substrates of P-gp and Mrp2 is the subject of continuing studies in our laboratory. However, it is likely that increased ATV retention in liver is partially caused by Mrp2 inhibition based on our present transport study results (Table 1). We also tested the potential for 50 μM RIF to inhibit P-gp, using P-gp overexpressing cells. No inhibition of transport activity could be detected (Fig. 6A). Obviously, to unambiguously characterize the pharmacokinetics of ATV, the contributions of other transporters such as breast cancer resistance protein (Bcrp) and bile salt export pump (Bsep) at the canalicular border and Mrp3 and sodium/taurocholate-cotransporting polypeptide (Ntcp) on the basolateral side should be further examined.

It has been suggested that the concerted action of uptake and efflux transporters govern the movement of drugs across hepatocytes and that combined inhibition of both types of transporters may account for the observed drug interactions in vivo (Cvetkovic et al., 1999; Cui et al., 2001; Sasaki et al., 2004). However, it is essential to recognize that uptake and efflux transporters may exert opposing effects on liver drug concentrations of both parent drug and metabolites. Recent studies from our laboratory have shown that both decreased and increased metabolism of digoxin were observed when Oatp1a4 and P-gp were inhibited, respectively (Lam and Benet, 2004; Lau et al., 2004). When efflux is blocked, more drug is available to be metabolized by the enzyme, whereas less drug can access the enzyme when uptake is inhibited. In the present study, inhibition of Oatp-mediated uptake also reduced the partition of ATV in the liver, as reflected by the decrease in parent drug and metabolites detected in liver for the 5 and 10 μM RIF groups. However, the extent of metabolism remained unaltered as demonstrated by the similar OH ATV recovered in the entire system between control and 5 and 10 μM RIF treatment groups (Table 4), which we cannot explain at the present time. Although ATV liver retention was increased due to a combination of inhibition of enzyme and efflux transporter at 50 μM RIF, the perfusate AUC changes still suggest that uptake inhibition was in effect.

Inhibition of ATV hepatic uptake has implications beyond pharmacokinetics, because the liver is the site of action for ATV, as well as its two active metabolites. Considering that significant hydroxy metabolites are formed before reaching the liver due to intestinal CYP3A-mediated metabolism (Lennernas, 2003), a decrease in 3-hydroxy-3-methylglutaryl-CoA reductase inhibitory effect will probably occur under Oatp inhibition because hydroxy metabolites are substrates for Oatp and are equipotent as the parent drug.

A recent study examining the relationship between an OATP1B1 variant and ATV and pravastatin-induced rhabdomyolysis found that the OATP-C^*15 mutant allele is significantly higher in patients who experienced statin-induced myopathy, suggesting the significance of OATP1B1 down-regulation to statin exposure (Morimoto et al., 2004). It has also been previously demonstrated that OATP/Oatp could be the major determinant in causing clinical drug-drug interactions between cerivastatin and cyclosporine (Shitara et al., 2003) and between bosentan and cyclosporine (Treiber et al., 2004). Our study further confirms the dominance of Oatp in mediating the interactions between ATV and the Oatp inhibitor RIF. RIF is a much weaker inhibitor of CYP3A compared with cyclosporine and therefore a better compound for differentiating uptake versus metabolism. Based on the results of the present study, we conclude that inhibition of hepatic uptake might be one of the major mechanisms for drug-drug interaction, particularly for metabolized compounds such as ATV. Future studies in our laboratory are examining the ATV-RIF interaction in whole animal and clinical studies.