The objective of the present study was to determine the efflux transporters responsible for acid and lactone statin drug efflux using transporter knockdown Caco-2 cells. The bidirectional transport was determined in Caco-2 cell monolayers in which the expression of P-glycoprotein (P-gp), breast cancer resistance protein (BCRP), or multidrug resistance associated protein 2 (MRP2) was knocked down by transduction with lentivirus containing human transporter-targeted small hairpin RNAs (shRNAs). Cells transduced with lentivirus containing nontargeted shRNA served as the vector control. Atorvastatin, lovastatin, and rosuvastatin displayed extremely low apical-to-basolateral (A-to-B) transport, which made the Papp,A-B values too unreliable to calculate the efflux ratio. Thus, transport comparisons were performed using the B-to-A permeability (Papp,B-A) values. Presented in the order of vector control, P-gp, BCRP, and MRP2 knockdown Caco-2 cells, the Papp,B-A values (×10−6, cm/s) were 28.1 ± 1.3, 8.6 ± 2.9, 20.3 ± 1.8, and 21.5 ± 1.6 for atorvastatin; 96.1 ± 7.1, 25.3 ± 3.5, 57.3 ± 9.8, and 48.2 ± 2.3 for fluvastatin; and 14.1 ± 1.9, 4.6 ± 1.7, 5.8 ± 0.7, and 6.6 ± 1.8 for rosuvastatin, respectively. Lovastatin and simvastatin showed no efflux in the vector control or knockdown cell monolayers in either lactone or acid forms. Results indicate that atorvastatin, fluvastatin, and rosuvastatin were transported by P-gp, BCRP, and MRP2. On the other hand, neither the lactone nor the resulting acid of lovastatin and simvastatin was transported by P-gp, BCRP, or MRP2. The current study demonstrated that the transporter knockdown Caco-2 cells are useful tools for studying drug-transporter interactions and should help eliminate some of the ambiguity associated with the identification of drug-transporter interactions based on chemical inhibitors alone.
Statins lower low-density lipoprotein cholesterol levels by competitively inhibiting HMG-CoA reductase activity, the rate-limiting step in cholesterol synthesis. They are among the most widely used drugs worldwide for the treatment and prevention of ischemic heart disease. Most statins are administered in the orally active β-hydroxy acid form, the exceptions being lovastatin and simvastatin, which are given as inactive lactone prodrugs. After oral administration, the fractional absorption of statin drugs ranges from 30 to >98%, and the bioavailability ranges from <5 to 60% (Neuvonen et al., 2006). Statin drugs are eliminated primarily through the hepatobiliary route, and renal excretion constitutes a minor route of elimination, except for pravastatin. Although generally well tolerated, statin drugs can cause side effects such as myopathy, whose symptoms range from benign myalgias to life-threatening rhabdomyolysis (Neuvonen et al., 2006). In addition, the side effects of statin drugs are frequently associated with drug-drug interactions resulting from their chronic administration to patients who are likely to be taking multiple drugs.
Previous statin transport studies have suggested that cellular uptake and excretion of these drugs is mediated by a variety of membrane transporters. Atorvastatin, lovastatin, and simvastatin have been reported to be effluxed by P-glycoprotein (P-gp) (Hochman et al., 2004; Chen et al., 2007; Kitamura et al., 2008), whereas the involvement of P-gp on the efflux of rosuvastatin is controversial (Huang et al., 2006; Kitamura et al., 2008). In contrast to ATP-binding cassette (ABC) efflux transporters, which differ in their affinity for and functional transport of various statins, the exchange transporters from the solute carrier family organic anion-transporting protein 1B1 were reported to mediate the cellular uptake of all statin drugs except fluvastatin (Kitamura et al., 2008; Ieiri et al., 2009).
The purpose of this study was to identify the role of the most important intestinal ABC efflux transporters, P-gp, breast cancer resistance protein (BCRP), or multidrug resistance protein 2 (MRP2), in the intestinal efflux of statin drugs through the use of knockdown (KD) Caco-2 cell clones in which the expression of P-gp, BCRP, or MRP2 was silenced with lentiviral vectors containing shRNA targeted to unique mRNA sequences in each efflux transporter (Zhang et al., 2009). Human intestinal cell lines such as Caco-2 are widely used to study drug absorption and efflux pathways in vitro (Hidalgo, 2001; Elsby et al., 2008). By evaluating the bidirectional permeability of acid or lactone statins in the KD cells, we determined the role of various transporters on statin transport. In addition, the system permitted observation of a general distinction between acid and lactone statins in terms of their interaction with P-gp. Rosuvastatin was determined to be a substrate of various efflux transporters, and additional experiments with chemical inhibitors provided complementary evidence for the involvement of P-gp, BCRP, and MRP2 on rosuvastatin transport. We have shown that KD cell lines derived from Caco-2 cells constitute a valuable system for evaluating statin transport pathways. This system avoids the need to rely only on chemical inhibitors, which lack selectivity (Watanabe et al., 2005; Wang et al., 2008), to identify transporter-mediated drug-drug interactions. It also allows investigation of efflux mechanisms that may be extremely difficult to unravel using either noncell-based models such as membrane vesicles or animal cell lines transfected with human transporters; compared with the Caco-2 model, these systems are less representative of the physiological factors found in human intestine (Polli et al., 2001; Elsby et al., 2008; Kitamura et al., 2008).
Materials and Methods
Atorvastatin calcium, fluvastatin sodium, lovastatin, rosuvastatin calcium, and simvastatin were obtained from Toronto Research Chemicals, Inc. (North York, ON, Canada). Digoxin, verapamil, bromosulfophthalein (BSP), estrone-3-sulfate (E3S), (E)-3-[[[3-[2-(7-chloro-2-quinolinyl)ethenyl]phenyl]-[[3-dimethylamino)-3-oxopropyl]thio]methyl]thio]-propanoic acid (MK571) sodium salt, fumitremorgin C (FTC), propranolol HCl, and atenolol were purchased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum was obtained from Omega (Tarzana, CA). Penicillin-streptomycin, nonessential amino acids, l-glutamine, trypsin, and puromycin were obtained from Cellgro (Manassas, VA). Lucifer yellow, Hanks' balanced salt solution (HBSS), Dulbecco's modified Eagle's medium, Dulbecco's phosphate-buffered saline, and HEPES were obtained from Invitrogen (Carlsbad, CA).
Transduction of Caco-2 Cells with shRNAs.
Parental Caco-2 cells were obtained from American Type Culture Collection (Manassas, VA; ATCC accession number CRL-2102). The cells were maintained in Dulbecco's modified Eagle's medium with 4.5 g/l glucose, supplemented with 10% fetal bovine serum, 1% nonessential amino acids, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 and 95% air. Lentivirus plasmid vectors containing shRNA inserts targeting human P-gp (GenBank accession number NM_000927), BCRP (GenBank accession number NM_004827), and MRP2 (GenBank accession number NM_000392) genes were obtained from Sigma-Aldrich. The transduction procedure was described previously (Zhang et al., 2009). In brief, the cells were plated in 96-well tissue culture plates at a density of 5 × 106 cells/well and transduced with 1 μg of P-gp-, BCRP-, or MRP2-targetting shRNA viral vector. Transduced cells were selected in culture medium containing 10 μg/ml puromycin. As a transduction control, parental Caco-2 cells were also transduced with a lentivirus plasmid vector containing shRNA that does not match any known human genes (Sigma-Aldrich), and the transduction procedures were identical to those used with P-gp, BCRP, and MRP2 shRNA vectors. Transfected cells were maintained under selection pressure of 10 μg/ml puromycin for at least five passages. After down-regulation of the efflux transporter of interest was verified by real-time polymerase chain reaction (PCR), the cells were seeded and grown on a Transwell plate for transport experiment.
Cells were seeded at 60,000 cells/cm2 onto collagen-coated, microporous, polycarbonate membranes in 12-well Costar Transwell plates (1.13-cm2 insert area, 0.4-μm pore size; Corning Life Sciences, Lowell, MA). The culture medium was changed 24 h after seeding to remove cell debris and dead cells; afterward, the medium was changed every other day. The cells were maintained at 37°C in a humidified atmosphere of 5% CO2 in air for 3 weeks to form confluent monolayers. The transport assay buffer was Hanks' balanced salt solution containing 10 mM HEPES and 15 mM glucose at pH 7.4 (HBSSg buffer). Batches of cell monolayers were certified by measuring the transepithelial electrical resistance (TEER) values of the monolayers and the apparent permeability coefficient (Papp) of control compounds: Lucifer yellow (low permeability marker), atenolol (low permeability marker), propranolol (high permeability marker), and the efflux ratios of digoxin, E3S, and BSP (P-gp, BCRP, and MRP2 probe substrates, respectively). Lucifer yellow was assayed at 0.5 mM, and the other probes (atenolol, propranolol, digoxin, E3S, and BSP) were assayed at 10 μM. The acceptance criteria for batches of cell monolayers usable in transport experiments were as follows: TEER, >450 Ω×cm2; Lucifer yellow, Papp < 0.4 × 10−6 cm/s; atenolol, Papp < 0.5 × 10−6 cm/s; propranolol, Papp between 15 × 10−6 and 25 × 10−6 cm/s. Before the transport experiment, TEER values were measured for all cell monolayers, and only monolayers with TEER values that met the acceptance criteria were used in the study.
Efflux Transporter Gene Expression by Real-Time PCR.
Total RNA was isolated from mature cell monolayers using the RNeasy Mini kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's protocol. Synthesis of cDNA was performed from 1 μg of total RNA using the QuantiTect RT kit (QIAGEN GmbH) for reverse transcription-PCR with random hexamer primers according to the manufacturer's protocol. Real-time quantitative PCR (qPCR) was performed using the LightCycler 480 system (Roche Diagnostics, Mannheim, Germany). Primer and probes were designed using the Universal Probe Library (Roche, Basel, Switzerland). Each PCR contained 10 ng of cDNA, LightCycler 480 probes master kit with 0.1 μM probe, and 0.5 μM primers. The PCR was run at 95°C for 10 min, followed by 45 amplification cycles of 95°C for 10 s, 60°C for 15 s, and 72°C for 1 s. Quantification of relative gene expression was performed using the ΔΔCT method. The housekeeping gene β-actin was used for normalization. Fold changes were calculated using vector control (VC) Caco-2 cells as the calibrator.
Western Blot Analysis for Efflux Transporter Proteins.
Cells were lysed with ice-cold radioimmunoprecipitation assay lysis buffer (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), and proteins were isolated following the manufacturer's protocol. Proteins (45 μg per lane) were separated on 4 to 20% SDS-polyacrylamide gel, transferred to a polyvinylidene difluoride membrane, and blocked with 0.2% I-block solution in Tris-buffered saline containing 0.05% Tween 20. Blocked membranes were probed sequentially with antibodies against the proteins of interest: P-gp (C-219; Abcam Inc., Cambridge, MA), BCRP (BXP-21), MRP2, and β-actin (Sigma-Aldrich). After washing, the membranes were reacted with the secondary antibodies. Protein-antibody complexes were visualized using SuperSignal West Femto Chemiluminescent Substrate (Pierce Chemical, Rockford, IL).
Statin Transport Assays.
Five statin drugs (atorvastatin, fluvastatin, rosuvastatin, lovastatin, and simvastatin) were selected for this study because they reflect a wide range of physicochemical properties (i.e., lipophilicity), chemical structures (acid and lactone), and interactions with transporters and/or metabolizing enzymes (Neuvonen et al., 2006). The transport assay included apical-to-basolateral (A-to-B) and basolateral-to-apical (B-to-A) transport rate determinations for each statin in each cell line used in this study. The statin compound was prepared at 10 μM in HBSSg buffer containing the high and low permeability reference standards (10 μM atenolol and 10 μM propranolol, respectively). For A-to-B transport, 0.55 ml of compound-containing solution was dosed to the apical (donor) chamber, and 1.5 ml of blank HBSSg buffer was applied to the basolateral (receiver) chamber. For B-to-A transport, 1.55 ml of compound-containing solution was dosed to the basolateral (donor) chamber, and 0.5 ml of blank HBSSg was added to the apical (receiver) chamber. At 30, 60, 90, and 120 min, 0.2-ml samples were withdrawn from the receiver chambers. The sample volume removed from the receiver chambers was replaced with an equal volume of prewarmed blank HBSSg buffer. At 0 and 120 min, 0.05-ml samples were collected from the donor chambers. Upon collection, the samples were immediately combined with 0.2 ml of 100% acetonitrile and stored at 4°C until sample analysis. For rosuvastatin kinetic studies, B-to-A transport rates of rosuvastatin across Caco-2 cell monolayers were measured as a function of rosuvastatin concentration at 37 and 4°C. For rosuvastatin inhibition studies, Caco-2 cell monolayers were preincubated with an inhibitor in both apical and basolateral chambers for 30 min, and bidirectional transport of rosuvastatin was conducted in the presence of the inhibitor in both chambers.
Concentrations of statin drugs were analyzed using reverse-phase liquid chromatography with triple-quadrupole tandem mass spectrometry methods. Samples were diluted with acetonitrile/HBSS (pH7.4, 50:50 v/v) as follows: receiver samples were diluted 4-fold; donor samples were diluted 20-fold. The high-performance liquid chromatography equipment consisted of a Perkin-Elmer series 200 autosampler and a Perkin-Elmer series 200 micro pump (PerkinElmer Life and Analytical Sciences, Waltham, MA). Chromatography was performed at ambient temperature using a 50 × 2.0 mm i.d., 4 μm Synergi Polar-RP analytical column (Phenomenex, Torrance, CA). The mobile phase consisted of 0.2% formic acid (aqueous phase) and 90:10 ACN/MeOH (organic phase). Typical gradients started at 10% aqueous phase, changed linearly to 95% organic phase over 3.0 min, held at 95% organic phase for 1.0 min, returned back to 10% aqueous phase in 0.2 min, and re-equilibrated at 10% aqueous phase for 0.8 min. The flow rate was 300 μl/min, and the injection volume was 5 μl. Mass spectrometry analyses were performed on a Sciex API 4000 triple-quadrupole mass spectrometer equipped with a TurboIonSpray interface in the positive electrospray ionization mode (Applied Biosystems/MDS, San Francisco, CA). The multiple reaction monitoring transitions and instrument settings were optimized for each compound. Equipment operation, data acquisition, and data integration were performed using Analyst version 1.4.2. software (Applied Biosystems, Foster City, CA).
Apparent permeability coefficients (Papp) were calculated using the following equation: where VR is the volume in the receiver chamber (milliliters), A is the filter surface area (1.12 cm2), C0 is the initial concentration (micromoles) in the donor chamber, and dC/dt is the maximal (or initial) slope of the concentration (micromoles) versus time (seconds) curve. The slope was calculated from the interval yielding the most linear rate of transfer spanning at least two sampling intervals.
Kinetic analysis of rosuvastatin was performed using GraphPad Prism 5.02 for Windows (GraphPad Software Inc., San Diego, CA). The active portion of B-to-A rosuvastatin transport across Caco-2 cell monolayers was isolated by subtracting the rate at 4°C from the corresponding rate at 37°C, and the differential transport rates were fit to the Michaelis-Menten equation: where V is the B-to-A rosuvastatin flux (pmol/cm2 · s) at a given concentration, C, of rosuvastatin; Vmax is the maximal B-to-A flux; and Km is the concentration of rosuvastatin at half the maximal B-to-A rate.
The significance of differences among multiple groups was determined by one-way ANOVA followed by post hoc Tukey tests. Differences were considered statistically significant when p < 0.05.
Suppression of Efflux Transporters by shRNA in Caco-2 Cells.
After suppression of the expression of the major efflux transporters (P-gp, BCRP, and MRP2) in Caco-2 cells, by means of transduction with lentivirus vectors containing shRNA targeting transporter mRNAs, the transduced cells with the highest transporter suppression were selected. To determine potential off-target effects of the transduction process on these transporters, we transduced Caco-2 cells with a viral vector containing shRNA that does not match any known human genes using the same transduction procedure. The suppression efficiency and selectivity were examined by measuring the mRNA expression of these efflux transporters using qPCR (Table 1). In VC cells, the expression of P-gp, BCRP, and MRP2 mRNA was between 70 and 100% of the levels seen in untransduced Caco-2 cells. In the KD cells, expression of targeted transporter genes was suppressed to 20% or less of the VC, whereas expression of nontargeted transporter genes remained 60 to 100% of the VC. The suppression of efflux transporter protein expression by transduction with the shRNA-containing lentivirus was confirmed by Western blot analysis (Fig. 1). The transporter shRNA lentivirus-transduced cells (Fig. 1, lanes B–D) produced much less corresponding transporter proteins than the VC-transduced cells (Fig. 1, lane A). Function of efflux transporters in the KD cells was assessed using prototypical substrates (Table 2). The efflux ratio of the P-gp substrate digoxin was 17.7 in VC cells and decreased to 7.4 (42% of VC) in P-gp KD cells without showing an appreciable change in BCRP KD and MRP2 KD cells. The BCRP substrate, E3S, exhibited a high efflux ratio (i.e., 48.5) in the VC cells, which was reduced to 10.2 (21% of VC cells) in the BCRP KD cells but remained roughly unchanged in P-gp KD and MRP2 KD cells. The efflux ratio of BSP, a MRP2 substrate, was reduced from 48.3 in VC cells to 2.6 in MRP2 KD cells.
Transport of Statin Drugs in Transporter KD Caco-2 Cells.
Atorvastatin, fluvastatin, and rosuvastatin are commercially available in the acid form, whereas lovastatin and simvastatin are as lactones. Samples were analyzed by reverse-phase liquid chromatography with triple-quadrupole tandem mass spectrometry for both the acid and lactone forms of each compound. After dosing, the statin acid drugs remained as acid form, with little conversion to the lactone form. On the other hand, the two lactone drugs showed substantial conversion to acid form during the course of the experiments. Atorvastatin acid was highly effluxed across the monolayers of VC cells. Its efflux was significantly diminished across monolayers of P-gp KD cells and less so across monolayers of BCRP and MRP2 KD cells (Fig. 2A). The results indicate that P-gp was the major efflux transporter responsible for atorvastatin acid efflux in this cell model, along with some contribution from both BCRP and MRP2. The decrease in atorvastatin acid efflux across transporter KD cell monolayers was mostly due to a decrease in the B-to-A transport as there was little change in the A-to-B transport compared with VC cell monolayers. Fluvastatin acid showed a similar pattern of permeability differences among the different cell lines (Fig. 2B). It was most highly effluxed across VC cell monolayers followed by across MRP2, BCRP, and P-gp KD cell monolayers. Fluvastatin acid, similarly to the other statin acids, appears to be primarily effluxed by P-gp, with BCRP and MRP2 playing minor roles in the efflux of fluvastatin acid. The decreased efflux across monolayers of P-gp KD cells was associated with a significant decrease in the B-to-A permeability, similar to the observations obtained with atorvastatin. Rosuvastatin acid was also highly effluxed in VC cells (Fig. 2C). The B-to-A permeability coefficients and the efflux ratios for rosuvastatin acid are diminished in all three KD cell lines, compared with those in the VC cells, suggesting all three of the targeted efflux transporters play a role in excreting rosuvastatin across the apical membrane of the intestinal cells.
After dosing the lactone form, more than half of lovastatin in the receiver compartments was converted to the acid, and the drug was not effluxed in any of the cell lines (Fig. 3). The results indicate that the efflux transporters play no significant role in lovastatin lactone transport. In the same experiment after lovastatin lactone dosing, formation of lovastatin acid was also monitored, and the permeability coefficients of formed lovastatin acid were high and comparable in both A-to-B and B-to-A directions (Fig. 3B). This high permeability would likely overwhelm any contribution of efflux or uptake transporters to the absorption of lovastatin acid in this model. Similar to lovastatin lactone, simvastatin lactone showed approximately 50% acid conversion in the receiver compartment and no apparent efflux in any of the cell clones (Fig. 4). Simvastatin acid showed high transport rates in both A-to-B and B-to-A directions across the monolayers of VC cells, and there was no significant difference between the A-to-B and B-to-A permeability coefficients. The KD cells did not show significant changes in the transport rates of simvastatin acid. Based on these findings, one can conclude that efflux transporters likely play a negligible role in limiting the absorption of highly permeable statin lactones, such as lovastatin or simvastatin.
Involvement of drug efflux transporters in the transport of statin acids was further assessed using chemical inhibitors to compare the results using transporter KD cell lines or chemical inhibitors to dissect absorption and excretion pathways. Rosuvastatin acid showed the highest efflux among the statin acids tested in the study; therefore, this drug was chosen for the inhibition experiments. Before inhibition experiments, the kinetics of rosuvastatin B-to-A transport across Caco-2 monolayers was determined as a function of rosuvastatin concentration from 2 to 200 μM at 37 and 4°C. Rosuvastatin B-to-A transport across Caco-2 monolayers was saturable and temperature dependent (Fig. 5). Fitting differential transport rates between 37 and 4°C to the Michaelis-Menten equation gave an apparent Km of 98 ± 33 μM and a Vmax of 137 ± 21 pmol/cm2 · min. The concentration and temperature dependence confirmed the involvement of a carrier-mediated efflux system(s) in rosuvastatin transport across Caco-2 monolayers. Rosuvastatin B-to-A transport was inhibited by various compounds (Fig. 6). Cyclosporine A (CsA), a broad inhibitor of efflux transporters; FTC, a selective inhibitor of BCRP; elacridar (GF120918), an inhibitor of both P-gp and BCRP; MK571, an inhibitor of MRPs, including MRP2; and verapamil, a selective inhibitor of P-gp, all decrease rosuvastatin efflux to some extent, suggesting the involvement of multiple efflux transporters in the efflux of this compound.
The ABC efflux transporters P-gp, BCRP, and MRP2, located in the apical membrane of enterocytes and in the canalicular membrane of hepatocytes, play important roles in the absorption of xenobiotics and facilitate the hepatobiliary elimination of xenobiotics and/or their metabolites. In this study, we used Caco-2 cells in which efflux transporters had been knocked down using shRNA (Zhang et al., 2009), to investigate the role of efflux transporters in statin transport. Given the gene sequence specificity of RNA interference (RNAi) silencing (Montgomery and Fire, 1998), shRNA-induced transporter knockdown is expected to avoid the cross-inhibition associated with chemical inhibitors (Watanabe et al., 2005; Wang et al., 2008). Despite the original expectation that RNAi-mediated gene silencing was specific, subsequent studies found that RNAi-mediated silencing may be associated with down-regulation of nontarget genes, so called “off-target effects” (Jackson et al., 2003). Furthermore, in certain animal models, silencing or down-regulation of a transporter gene may lead to compensatory up-regulation of other genes (Chen et al., 2005b). To determine the knockdown efficiency and nonspecific effects of RNAi silencing in the current study, we measured the expression levels of the efflux transporters at both the mRNA and protein levels by qPCR and Western blot analysis. Knockdown of the transporter of interest was confirmed by both qPCR (Table 1) and Western blot results (Fig. 1). The qPCR analyses further revealed that up-regulation of nontarget efflux transporters was virtually absent; instead, a slight down-regulation of certain nontarget transporters was observed. Because down-regulation of off-target mRNA transcripts can hardly be avoided by the in silico design of siRNA sequences (Echeverri and Perrimon, 2006; Tschuch et al., 2008), additional methods were proposed to reduce the off-target effects. One approach is to use “nonspecific” or “scrambled” siRNAs that have no physiological target as negative controls (Echeverri and Perrimon, 2006). Another approach is to target a single mRNA of interest with a pool of several different endonuclease-prepared siRNA molecules (Kittler et al., 2005). Furthermore, the viral-based shRNA approach using lentiviruses or retroviruses is another option for reducing off-target effects, because viral delivery can be nearly 100% efficient in almost all cell types and the level of stable shRNA expression is comparatively modest, which causes minimal off-target effects with sustained silencing of the target gene (Tschuch et al., 2008). Therefore, in this current study we used lentivirus-based shRNA gene silencing to generate the transporter KD cell lines and used the cells transduced with nonspecific shRNA vector as a negative control in the transport experiments.
The large decrease in B-to-A transport of atorvastatin in P-gp KD Caco-2 cells (Fig. 1) indicates that P-gp is an important mediator of atorvastatin efflux, an observation consistent with the conclusions of two previous studies in Caco-2 cells (Wu et al., 2000; Hochman et al., 2004). The smaller reduction in BCRP and MRP2 KD Caco-2 cells suggests that these transporters play lesser roles in atorvastatin transport. Despite in vitro evidence showing the interaction of atorvastatin with P-gp (Hochman et al., 2004; Chen et al., 2005a), the in vivo role of P-gp in statin drug interactions remains unclear (Holtzman et al., 2006). This is partly because atorvastatin is metabolized by CYP3A4, and drugs that interact with atorvastatin are frequently substrates/inhibitors of both P-gp and CYP3A4. Despite these limitations, numerous pharmacokinetic drug interactions associated with atorvastatin have been reported (Boyd et al., 2000; Mazzu et al., 2000; Asberg et al., 2001). The affinity of atorvastatin for all three intestinal luminal efflux transporters could hamper its intestinal permeation, which may explain its moderate (30%) absorption in humans despite its relatively high lipophilicity (Neuvonen et al., 2006).
The two other acid statins, fluvastatin acid and rosuvastatin acid, were found to be substrates for P-gp, BCRP, and MRP2 in the transporter KD Caco-2 cell system (Fig. 2, B and C). These results agree with a study in which MRPs were reported to mediate the efflux of fluvastatin in Caco-2 cells (Lindahl et al., 2004). As is the case with atorvastatin, fluvastatin and rosuvastatin were identified as P-gp substrates in vitro, but evidence of their involvement in P-gp-mediated drug interactions in vivo is scarce. In fact, coadministration of these statins with the P-gp substrate digoxin did not cause clinically significant adverse reactions to either statin drugs or digoxin (Garnett et al., 1994; Martin et al., 2002). The apparent absence of P-gp-based drug interactions might be due to the involvement of multiple efflux transporters as for atorvastatin. The report that BCRP polymorphism alters the area under the curve and Cmax but not the elimination half-life (t½;) suggests that the effect of BCRP on fluvastatin and rosuvastatin pharmacokinetics occurs primarily in the intestinal mucosa during the absorption of the acid forms of the drugs (Zhang et al., 2006; Keskitalo et al., 2009a,b).
Another similarity shared by atorvastatin, fluvastatin, and rosuvastatin, the three acid statins, is the little acid-to-lactone conversion during transport. In light of a study showing that after oral administration of atorvastatin acid to healthy volunteers, atorvastatin lactone was found in plasma (Kantola et al., 1998), our results suggest that acid statins exist predominantly in the native acid form in the gastrointestinal tract, with their lactonization occurring after absorption, most likely in the liver through the CoASH-dependent and glucuronidation pathways (Prueksaritanont et al., 2002). In contrast to the acid statins, the lactone statins (lovastatin and simvastatin) showed substantial lactone-to-acid conversion during transport as nearly equal amounts of the lactone and acid forms appeared in the receiver chambers (Figs. 3 and 4). Hydrolysis of statin lactones to statin acids in Caco-2 assays seemed to involve both spontaneous chemical conversion and enzymatic processes. Because lipophilic compounds often can permeate cell membranes and show low efflux in cell monolayers, the relatively high lipophilicity of lovastatin and simvastatin (Neuvonen et al., 2006) suggests that efflux transporter activity is unlikely to be a significant barrier for these compounds, consistent with a suggestion made in a previous study (Polli et al., 2001). Our finding that lactone statins were either nonsubstrates or weak substrates for P-gp is in agreement with other in vitro studies (Hochman et al., 2004; Chen et al., 2005a).
Last, we used rosuvastatin as a model compound to compare results from the transporter KD cells with results from chemical inhibitors in identifying the involvement of efflux transporters (Fig. 6). Inhibition of apical efflux transporters normally results in a decrease in B-to-A transport accompanied by an increase in A-to-B transport of a substrate (Polli et al., 2001). CsA, a broad inhibitor of ABC efflux transporters, partially inhibited the B-to-A transport of rosuvastatin in Caco-2 cells, indicating the involvement of these transporters. Inhibitions by FTC and GF120918 are consistent with rosuvastatin being a substrate of BCRP. A slight inhibition by verapamil suggests that P-gp plays only a minor role in rosuvastatin transport, an observation that contrasts with results from transporter KD cells and with findings from a previous study (Kitamura et al., 2008). The finding that the MRP inhibitor MK571 completely abolished rosuvastatin efflux was unexpected because it seems to suggest that rosuvastatin was solely transported by MRP2 in Caco-2 cells, a possibility that conflicts with evidence of BCRP involvement found in this and other studies (Zhang et al., 2006; Kitamura et al., 2008; Keskitalo et al., 2009b). An alternative explanation is that MK571 cross-inhibits other efflux transporters as reported previously (Wang et al., 2008). Together, these results highlight the difficulties associated with relying only on chemical inhibitors to identify drug transporter interactions.
In conclusion, the transporter KD Caco-2 system permitted us to identify the involvement of efflux transporters in statin transport. The statins in active β-hydroxy acid form were transported, to some extent, by P-gp, BCRP, and MRP2 in Caco-2 cells. The relatively high permeability of the lactone statins makes it difficult to detect directional transport in monolayer transport experiments. Our results showed that lactone statins are significantly hydrolyzed to the acid form in Caco-2 and KD cells. In vivo, the cooperative roles of intestinal transporters (of the acid form) and hydrolysis of the lactone form influence the pharmacokinetics of statin drugs (Keskitalo et al., 2008, 2009a). The current study demonstrates that the transporter KD Caco-2 system is a straightforward approach to assess drug-transporter interactions. Its human origin makes results relevant to human studies and avoids the species differences in statin interactions with efflux transporters that have been observed (Kitamura et al., 2008). In addition, Caco-2 cells express a basolateral uptake system that is essential for statins to reach apical efflux transporters. As a result of the absence of this basolateral uptake transporter in MDCK cells, statin efflux could not be detected in nontransfected or single-transfected MDCK cells (Kitamura et al., 2008). Overall, the transporter KD Caco-2 cells provide a valuable tool to study drug interactions with human transporters in vitro (Watanabe et al., 2005).
Participated in research design: Li, Hidalgo, Volpe, Owen, Bode, and Zhang.
Conducted experiments: Volpe and Wang.
Performed data analysis: Li and Volpe.
Wrote or contributed to the writing of the manuscript: Li, Hidalgo, Volpe, and Owen.
We thank Samantha Allen, Erica A. Weiskircher, James L. Merdink, Grainne McMahon Tobin, and Yuehua Huang for expert technical assistance.
This work was supported in part by a research grant from the Food and Drug Administration [Grant 1R43FD003482-01].
The findings and conclusions in this article have not been formally disseminated by the Food and Drug Administration and should not be construed to represent any Agency determination or policy.
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
- ATP-binding cassette
- breast cancer resistance protein
- multidrug resistance-associated protein 2
- vector control
- apparent permeability
- (E)-3-[[[3-[2-(7-chloro-2-quinolinyl)ethenyl]phenyl]-[[3-dimethylamino)-3-oxopropyl]thio]methyl]thio]-propanoic acid
- fumitremorgin C
- Hanks' balanced salt solution
- transepithelial electrical resistance
- polymerase chain reaction
- quantitative polymerase chain reaction
- small hairpin RNA
- small interfering RNA
- cyclosporine A
- RNA interference.
- Received January 7, 2011.
- Accepted March 28, 2011.
- U.S. Government work not protected by U.S. copyright