Abstract
The mechanism by which drugs inhibit organic anion transporter 1 (OAT1) was examined. OAT1 was stably expressed in Chinese hamster ovary (CHO) cells, and para-aminohippurate (PAH) and 6-carboxyfluorescein were the substrates. Most compounds (10 of 14) inhibited competitively, increasing the Michaelis constant (Km) without affecting the maximal transport rate (Jmax). Others were mixed-type (lowering Jmax and increasing Km) or noncompetitive (lowering Jmax only) inhibitors. The interaction of a noncompetitive inhibitor (telmisartan) with OAT1 was examined further. Binding of telmisartan to OAT1 was observed, but translocation was not. Telmisartan did not alter the plasma membrane expression of OAT1, indicating that it lowers Jmax by reducing the turnover number. PAH transport after telmisartan treatment and its washout recovered faster in the presence of 10% fetal bovine serum in the washout buffer, indicating that binding of telmisartan to OAT1 and its inhibitory effect are reversible. Together, these data suggest that telmisartan binds reversibly to a site distinct from substrate and stabilizes the transporter in a conformation unfavorable for translocation. In the absence of an exchangeable extracellular substrate, PAH efflux from CHO-OAT1 cells was relatively rapid. Telmisartan slowed PAH efflux, suggesting that some transporter-mediated efflux occurs independent of exchange. Although drug-drug interaction predictions at OAT1 assume competitive inhibition, these data show that OAT1 can be inhibited by other mechanisms, which could influence the accuracy of drug-drug interaction predictions at the transporter. Telmisartan was useful for examining how a noncompetitive inhibitor can alter OAT1 transport activity and for uncovering a transport mode independent of exchange.
Introduction
The kidney is important for the urinary elimination of a variety of organic anions of physiologic and pharmacologic importance (Rizwan and Burckhardt, 2007). Renal excretion is accomplished by glomerular filtration, tubular secretion, and tubular reabsorption. For many hydrophilic organic anions, transporter-mediated tubular secretion is a major determinant of their overall urinary elimination. The organic anion transporter 1 (OAT1; SLC22A6) contributes to organic anion uptake across the peritubular membrane of proximal tubule cells, and a variety of transporters are implicated in apical efflux into the glomerular filtrate (Pelis and Wright, 2011). In a retrospective study that examined the elimination pathways of 200 of the top-prescribed therapeutic drugs, 32% are predominately eliminated in the urine, with 92% of these actively secreted by renal tubules (Morrissey et al., 2013). Of the drugs actively secreted, 28% are substrates of OAT1, highlighting the potential importance of OAT1 to pharmacokinetics (Morrissey et al., 2013).
OAT1 interacts with a wide variety of structurally diverse organic anions, making it a potential site of pharmacokinetic drug-drug interactions (Vanwert et al., 2010; Burckhardt, 2012). Given OAT1’s potential involvement in drug-drug interactions, the U.S. Food and Drug Administration and the European Medicines Agency recommend that pharmaceutical companies investigate whether investigational drugs inhibit OAT1 (CDER, 2012; CHMP, 2012). The studies are typically done in vitro using cell lines, such as human embryonic kidney or Chinese hamster ovary (CHO) cells expressing OAT1 (Giacomini et al., 2010). Depending on the potency with which investigational drugs inhibit OAT1 and the anticipated maximal therapeutic unbound plasma concentration of the investigational drug, a clinical drug-drug interaction study may be warranted. Thus, it is important to accurately assess the potency with which drugs inhibit OAT1, but to also investigate the mechanism of inhibition, as it could influence drug-drug interaction potential. Although competitive inhibition has been well described in the literature and is often presumed, it is not clear whether OAT1 is inhibited by other mechanisms.
We previously examined the effect of a physiologic plasma concentration of α-ketoglutarate on the interaction of ligands with OAT1 expressed in CHO cells (Ingraham et al., 2014). For some compounds, their IC50 value against OAT1 increased when α-ketoglutarate was present, but this was not the case for all compounds. From these data we speculated that for inhibitors where IC50 values shifted when α-ketoglutarate was present, they act in a competitive manner to reduce transport activity, whereas for inhibitors where a shift was not observed, they are noncompetitive inhibitors. Given our previous results, and recent data showing that a related transporter, organic cation transporter 2, is inhibited by multiple mechanisms (Harper and Wright, 2013), we hypothesized that OAT1 can be inhibited by multiple mechanisms as well.
Our study examined the mechanism by which drugs in diverse classes inhibit OAT1. We examined 12 drugs in seven different classes (uricosuric, antiviral, nonsteroidal anti-inflammatory, loop diuretic, angiotensin II receptor antagonist, proton pump inhibitor, and statin); morin, a flavonoid; and estrone-3-sulfate, a metabolite of the endogenous hormone estrone. These compounds were chosen because they inhibit OAT1, they are structurally diverse, and a previous study suggested that some may inhibit OAT1 via mechanisms other than simple competition (Ingraham et al., 2014). Consistent with our hypothesis, we observed other types of inhibition than just competition. The interaction of telmisartan (a noncompetitive inhibitor) with OAT1 was further examined as it was useful for examining how a noncompetitive inhibitor can alter OAT1 transport activity and for uncovering a transport mode independent of exchange.
Materials and Methods
Reagents and Chemicals.
We obtained [3H]-para-aminohippurate (60 Ci/mmol) and [3H]-telmisartan (15 Ci/mmol) from American Radiochemicals (St. Louis, MO). The F12 Kaighn’s modification medium, fetal bovine serum (certified, U.S. origin), 1% penicillin-streptomycin solution, zeocin, hygromycin B, 4%–12% Tris-glycine gels, and 6-carboxyfluorescein were purchased from Life Technologies (Burlington, ON, Canada). The bicinchoninic acid protein assay kit, Sulfo-NHS-SS-biotin reagent, Streptavidin UltraLink resin, and the SuperSignal West Pico chemiluminescent substrate were from Thermo Scientific (Rockford, IL). Olmesartan was obtained from Tocris Biosciences (Bristol, United Kingdom). All other chemicals were of the highest purity possible and were obtained from Sigma-Aldrich (St. Louis, MO). Waymouth’s buffer (WB) used for transport experiments, contained (in mM): 135 NaCl, 28 d-glucose, 5 KCl, 1.2 MgCl2, 2.5 CaCl2, 0.8 MgSO4, and 13 HEPES-NaOH, pH 7.4.
Cell Culture.
Cloning and stable expression of the human ortholog of OAT1 in CHO Flp-In cells (CHO-OAT1 cells) was described elsewhere (Ingraham et al., 2014). The CHO cells were grown in complete medium containing F12 Kaighn’s modification medium, 10% fetal bovine serum, and 1% penicillin-streptomycin. The medium used for culturing the CHO-OAT1 cells was supplemented with hygromycin B (200 μg/ml, final concentration), whereas the medium used for culturing the parental CHO cells (CHO parental cells) was supplemented with zeocin (100 μg/ml, final concentration). Cells were grown at 37°C in a humidified atmosphere of 5% CO2/95% air.
Cell Surface Biotinylation and Western Blot Analysis.
Cell surface biotinylation was performed to determine whether telmisartan exposure results in loss of OAT1 at the cell surface. CHO-OAT1 cells were treated without or with telmisartan (5 μM) diluted in WB for 10 seconds, and the telmisartan was removed by rinsing the cells rapidly 3 times (0.5 ml each) with room temperature WB. The cells were then placed in room temperature WB for 2, 5, 20, 40, or 60 minutes (recovery periods), and they were rinsed once with WB before performing the cell surface biotinylation. All solutions were kept ice-cold for cell surface biotinylation, and long incubations were conducted on ice with gentle shaking. One well of a 12-well plate containing CHO-OAT1 cells grown to confluence was used for each biotinylation reaction. Cell surface biotinylation was performed using Sulfo-NHS-SS-biotin (0.5 mg/ml) using procedures identical to those described by Astorga et al. (2011). The biotinylated proteins were separated on 4%–12% Tris-glycine gels. The protocol for detection of OAT1 protein by immunoblotting is essentially the same as described by Astorga et al. (2011).
Cellular Accumulation Studies with Radiolabeled Compounds.
All transport experiments using [3H]para-aminohippurate ([3H]PAH) and [3H]telmisartan were conducted using cells grown to confluence in 24-well flat bottom plates. The transport solution consisted of WB containing the radiolabeled compound and, in some cases, an inhibitor and/or unlabeled compound. Before the uptake period, the medium was aspirated, and the cells were rinsed once with room temperature WB. In some cases, the cells were permeabilized for 5 minutes in phosphate-buffered saline containing 0.1% saponin, followed by two rapid rinses with WB before measuring the cellular accumulation of radiolabeled compounds. After the uptake period, the cells were rinsed 3 times with ice-cold WB. The cells were lysed with 0.5 N NaOH/1% SDS (0.4 ml) for ∼30 minutes on an orbital shaker after which the NaOH was neutralized with 1 N HCl (0.2 ml). The cell lysates (0.5 ml) were transferred to scintillation vials, and liquid scintillation cocktail (CytoScint ES; MP Biomedicals, Santa Ana, CA) was added (7 ml). Cellular radioactivity content was determined with a Beckman LS6500 liquid scintillation counter (Beckman Coulter, Brea, CA). Protein content in the wells was determined using the bicinchoninic acid method.
Cellular Accumulation Studies with 6-Caboxyfluorescein.
Cells grown to confluence in 96-well black-walled flat-bottom plates were used for transport experiments with 6-carboxyfluorescein. The transport solution consisted of WB containing 6-carboxyfluorescein and, in some cases, inhibitor. Briefly, cells were rinsed once with WB after which the uptake of 6-carboxyfluorescein was determined. After the uptake period the cells were rinsed 3 times with ice-cold WB. After the last rinse, the WB was aspirated, and 6-carboxyfluorescein in the cells was determined at excitation and emission wavelengths of 485 and 525 nm, respectively, using a Cytation 3 multimodal plate reader (BioTek, Winooski, VT). We determined the 6-carboxyfluorescein concentration from a standard curve, and it was expressed relative to the surface area of the bottom of an individual well.
Data Analysis.
Data are reported as mean ± S.E.M. All experiments were performed in triplicate using cells of the same passage, with the number of observations based on the number of experiments performed using cells of a different passage. Comparison of sample means was performed using a two-tailed unpaired Student’s t test or one-way analysis of variance followed by the Newman-Keuls post hoc test. The IC50 values were determined by nonlinear regression, as described previously elsewhere (Ingraham et al., 2014). The maximal transport rate (Jmax) and Michaelis constant (Km) of PAH and 6-carboxyfluorescein transport were determined by nonlinear regression analysis using the Michaelis-Menten equation as described previously elsewhere (Ingraham et al., 2014). All graphing, nonlinear regression analyses, and statistical analyses were performed with GraphPad Prism (version 5.04; GraphPad Software, San Diego, CA).
Results
The inhibitory effect of ibuprofen, telmisartan, and omeprazole on PAH and 6-carboxyfluorescein transport is shown in Fig. 1; these compounds were chosen as examples because of their different inhibition mechanisms (see below). We chose to examine two different substrates to determine the potential for inhibition mechanism to be substrate dependent. In the examples in Fig. 1, ibuprofen, telmisartan, and omeprazole inhibited [3H]PAH uptake by CHO-OAT1 cells with IC50 values of 2.7 μM, 0.36 μM, and 10.4 μM, respectively, and the drugs were near equipotent inhibitors of 6-carboxyfluorescein transport as PAH transport (Fig. 1).
Representative experiments (n = 1) showing the inhibitory effect of increasing concentrations of ibuprofen, telmisartan, or omeprazole on either [3H]PAH (A, C, and E) or 6-carboxyfluorescein (B, D, and F) uptake by CHO-OAT1 cells. The concentrations of [3H]PAH and 6-carboxyfluorescein in the transport solution (room temperature) were 20 nM and 5 μM, respectively. The [3H]PAH and 6-carboxyfluorescein uptake experiments were conducted for 1 and 5 minutes, respectively. These were determined to be initial rate time points in preliminary experiments (data not shown).
Under control conditions, and when the individual inhibitors were present at a fixed concentration, PAH uptake (Fig. 2, A, C, and E) and 6-carboxyfluorescein uptake (Fig. 2, B, D, and F) by CHO-OAT1 cells were a saturable process. Ibuprofen caused a significant increase in the Km value without affecting the Jmax value for both PAH transport (Fig. 2A and Table 1) and 6-carboxyfluorescein transport (Fig. 2B and Table 2), an effect consistent with competitive inhibition. The Jmax value for both PAH transport and 6-carboxyfluorescein transport was significantly reduced by telmisartan, but the Km value was unchanged, indicating that telmisartan is a noncompetitive inhibitor of OAT1 (Fig. 2, C and D and Table 1 and Table 2). Omeprazole showed mixed inhibition of both PAH transport (Fig. 2E and Table 1) and 6-carboxyfluorescein transport (Fig. 2F and Table 2), causing changes in both Km (an increase) and Jmax (a decrease) values.
Michaelis-Menten curves showing the kinetics of PAH (A, C, and E) or 6-carboxyfluorescein (B, D, and F) uptake by CHO-OAT1 cells performed in the absence (control) or presence of a fixed concentration of (A, B) ibuprofen (5 μM), (C, D) telmisartan (0.5 μM), or (E, F) omeprazole (20 μM). The transport solutions were equilibrated to room temperature. PAH and 6-carboxyfluorescein uptake experiments were conducted for 1 and 5 minutes, respectively. These were determined to be initial rate time points in preliminary experiments (data not shown). The data are the mean ± S.E.M. of at least four preparations (n ≥ 4). The kinetic constants obtained (Jmax and Km) and the type of inhibition caused by each inhibitor are shown in Table 1 (PAH) and Table 2 (6-carboxyfluorescein).
The kinetics of PAH transport by OAT1 in the absence versus presence of inhibitors
The concentration of inhibitor drug used is indicated in the parentheses. The inhibitor concentrations used were fixed at a concentration expected to inhibit ∼50%–75% of transport activity. The concentrations were based on previously determined inhibitory concentration 50 (IC50) values (Ingraham et al., 2014), or were determined in preliminary experiments. Data are mean ± S.E.M. of four to five experiments.
The kinetics of 6-carboxyfluorescein transport by OAT1 in the absence versus presence of inhibitors
The concentration of inhibitor drug used is indicated in the parentheses. The inhibitor concentrations used were fixed at a concentration expected to inhibit ∼50%–75% of transport activity. The concentrations were based on previously determined inhibitory concentration 50 (IC50) values (Ingraham et al., 2014), or were determined in preliminary experiments. Data are mean ± S.E.M. of four experiments.
Linear transformations of the Michaelis-Menten curves highlight the kinetic effect the inhibitors had on OAT1-mediated PAH (Fig. 3, A, C, and E) and 6-carboxyfluorescein (Fig. 3, B, D, and F) transport. Table 1 and Table 2 summarize the kinetic effect all the compounds tested had on PAH transport and 6-carboxyfluorescein transport, respectively. Most of the compounds tested were competitive inhibitors, except telmisartan, azilsartan, irbesartan, and omeprazole, which were characterized as either noncompetitive or mixed inhibitors.
Linear transformations of the kinetic data shown in Figure 2 highlighting the effect on PAH (A, C, and E) or 6-carboxyfluorescein (B, D, and F) kinetics of a fixed concentration of (A, B) ibuprofen (5 μM), (C, D) telmisartan (0.5 μM), or (E, F) omeprazole (20 μM).
We further characterized the interaction of telmisartan with OAT1 because it is a noncompetitive inhibitor, it is a potent OAT1 inhibitor (IC50 < 0.5 μM; Sato et al., 2008; Ingraham et al., 2014) relative to many others (Vanwert et al., 2010), and it is extremely hydrophobic (experimental LogP value of 7.73 CSID:59391, http://www.chemspider.com/). It is important to note that telmisartan, based on its Cmax,unbound/IC50 (CDER, 2012), is not expected to cause a drug-drug interaction at OAT1, as the unbound plasma concentration of telmisartan at a therapeutic dose (<10 nM at an 80-mg dose; Taylor et al., 2011) is ∼30–50 times lower than its IC50 value. Although telmisartan is not a clinical safety concern with respect to OAT1 inhibition, we felt it important to further characterize its interaction with OAT1, as other drugs on the market or in development may be noncompetitive inhibitors and interact similarly with the transporter.
Two types of experiments were conducted to assess the possibility of active transport of telmisartan by OAT1: room temperature versus ice-cold conditions to slow active transport, or cells were permeabilized with saponin before performing transport assays. Saponin removes cholesterol from membranes, leaving small pores of ∼50 Å (Jamur and Oliver, 2010). The cellular accumulation of [3H]telmisartan was significantly higher in CHO-OAT1 cells compared with CHO parental cells without or with membrane permeabilization with saponin, which was effective at completely inhibiting PAH transport (Fig. 4A). The cellular content of [3H]telmisartan was also significantly higher in CHO-OAT1 cells compared with parental cells when transport was conducted either at room temperature or when using ice-cold transport solution (Fig. 4B). Expectedly, the cellular accumulation of [3H]PAH was reduced 90% when using ice-cold transport solutions, indicating that low temperature was effective at almost completely inhibiting OAT1 substrate translocation (Fig. 4B). The time course of [3H]telmisartan accumulation into CHO-OAT1 cells and parental cells conducted using ice-cold transport buffer is shown in Fig. 4C, with the difference in cellular accumulation between the two cell lines plotted as specific binding to OAT1. At each time point examined the cellular accumulation of telmisartan was higher in CHO-OAT1 cells compared with parental cells, with specific binding reaching equilibrium at ∼20 minutes. Attempts to determine the equilibrium dissociation constant and maximal binding were unsuccessful because the passive component of cellular telmisartan accumulation masked the specific binding, especially at higher telmisartan concentrations (data not shown).
(A) The cellular accumulation of either [3H]telmisartan (20 nM) or [3H]PAH (20 nM) into CHO parental or CHO-OAT1 cells at room temperature without or with a 5-minute pretreatment of the cells with 0.1% saponin. Uptake was conducted for 5 minutes. The data are the mean ± S.E.M. of four preparations (n = 4). *P < 0.05, **P < 0.01, significantly different from uptake by CHO parental cells, unpaired Student’s t test. (B) The cellular accumulation of either [3H]telmisartan (7 nM) or [3H]PAH (10 nM) into CHO parental or CHO-OAT1 cells using either room temperature or ice-cold transport solution. Uptake was conducted for 5 minutes. The data are the mean ± S.E.M. of four preparations for telmisartan (n = 4) and three preparations for PAH (n = 3). *P < 0.05, **P < 0.01, significantly different from uptake by CHO parental cells, unpaired Student’s t test. (C) The time course of [3H]telmisartan (7 nM) accumulation into CHO parental or CHO-OAT1 cells using ice-cold transport solution. The data are from a single experiment (n = 1). The difference in cellular [3H]telmisartan accumulation in CHO parental versus CHO-OAT1 cells represents specific binding of telmisartan to OAT1 (binding).
Previous work with OATP1B1 suggests that the transport protein can be inhibited by more complex mechanisms than by simple competition, perhaps even by an irreversible mechanism, namely, mechanism-based inhibition (Amundsen et al., 2010; Shitara et al., 2013). To examine the possibility that a noncompetitive inhibitor can irreversibly inhibit OAT1, we examined the ability of transport activity to recover after telmisartan treatment and its removal (Fig. 5A). In these experiments, [3H]PAH uptake was examined either in the absence or presence of telmisartan (5 μM) added to the transport solution; or the cells were treated briefly with telmisartan (5 μM), the telmisartan was removed by extensive washing with WB, and the cells were allowed to recover for 2, 5, 10, 20, 40, or 60 minutes in WB before another rapid washing with WB and measurement of [3H]PAH uptake. The concentration of telmisartan used (5 μM) was ∼15-fold higher than its IC50 value against OAT1-mediated PAH transport (Ingraham et al., 2014).
Time course of recovery of OAT1-mediated [3H]PAH uptake by CHO-OAT1 cells after treatment with telmisartan and its removal (recovery time after washout) using (A) standard buffer or (B) buffer containing 10% fetal bovine serum. The cells were not treated with telmisartan (control); telmisartan (5 μM) was included in the uptake buffer (telmisartan); or the cells were treated with telmisartan (5 μM) for ∼10 seconds, the telmisartan was removed by repeated washing with WB (three rapid rinses), and the cells were put into WB (A) without or (B) with 10% fetal bovine serum for the time points indicated (recovery time after telmisartan washout). After the washout period, the cells were rinsed once with WB (no fetal bovine serum) before measuring 1-minute uptakes of [3H]PAH (10 nM). Data are mean ± S.E.M. of four preparations (n = 4). **P < 0.01, ***P < 0.001, significantly different from control, one-way analysis of variance followed by the Newman-Keuls test for pairwise comparisons.
Telmisartan added to the transport solution reduced [3H]PAH uptake to 7% of the control value, indicating that the drug was effective at inhibiting OAT1 activity (Fig. 5A). [3H]PAH uptake was only 20% of the control value 2 minutes after removal of telmisartan (Fig. 5A). [3H]PAH uptake was reduced 47% compared with control 60 minutes after telmisartan washout. A possible explanation is that telmisartan is slow to leave the cells and that a residual amount of telmisartan is available to inhibit activity after its removal. Indeed, after brief telmisartan treatment, ∼8% of the [3H]telmisartan dose was present in the cells after 1 hour (Fig. 6).
The time course of [3H]telmisartan washout from CHO-OAT1 cells performed either in standard buffer (buffer alone) or buffer containing 10% fetal bovine serum (buffer + 10% fetal bovine serum). CHO-OAT1 cells were treated with [3H]telmisartan diluted in WB (no fetal bovine serum) for ∼10 seconds, the cells were then washed repeatedly with WB (three rapid rinses) and incubated in WB or WB containing 10% fetal bovine serum (WB + 10% fetal bovine serum) for 2, 10, 20, or 60 minutes, after which the cells were rinsed once with WB and the cellular accumulation of [3H]telmisartan was determined. Data are mean ± S.E.M. of three preparations (n = 3).
In a separate set of recovery experiments, we used the same protocol as previously described except that the cells were allowed to recover for 30 or 60 minutes in WB containing 10% fetal bovine serum (Fig. 5B). The [3H]PAH uptake was not significantly different from control after a washout of 30 or 60 minutes in the presence of 10% fetal bovine serum. Accordingly, the removal of cellular [3H]telmisartan was relatively rapid when the cells were incubated in WB containing 10% fetal bovine serum (Fig. 6).
The maximal transport rate is a product of the number of transporters at the plasma membrane and their turnover number. To determine whether telmisartan noncompetitively inhibited OAT1 by reducing the number of OAT1 transport proteins in the plasma membrane, cell surface biotinylation was performed. The protocol for treating the cells with telmisartan and allowing them to recover in WB was identical to that outlined for the transport experiments in Fig. 5A. The cells were treated with telmisartan, it was removed, and the cells were allowed to recover for 2, 5, 20, 40, or 60 minutes, after which the cell surface biotinylation was performed. The apparent differences in cell surface expression of OAT1 were small (Supplemental Fig. 1), and they cannot explain the reduction in Jmax caused by telmisartan.
Because telmisartan was an effective inhibitor of OAT1-mediated substrate uptake and has a relatively high apparent passive permeability, we speculated that it might inhibit efflux as well. Thus, we examined the time course of [3H]PAH efflux from CHO-OAT1 cells in the absence or presence of telmisartan (5 μM) added to the extracellular buffer (Fig. 7). Under control conditions, nearly all the intracellular [3H]PAH was effluxed from the cells by 20 minutes (Fig. 7, A–C). As anticipated, including a saturating concentration of unlabeled PAH (250 μM) in the extracellular buffer trans-stimulated [3H]PAH efflux (Fig. 7A). In contrast, efflux slowed considerably when telmisartan (5 μM) was included in the extracellular buffer (Fig. 7B). The inhibitory effect of telmisartan on [3H]PAH efflux was similar whether in the absence (Fig. 7B) or presence (Fig. 7C) of PAH (250 μM) added to the extracellular buffer.
Efflux of [3H]PAH from CHO-OAT1 cells. Cells were loaded with [3H]PAH (∼20 nM) for 20 minutes at room temperature, and the cells were washed rapidly 3 times with ice-cold WB before the efflux period. Efflux was conducted for 30 seconds, 2 minutes, 5 minutes, or 20 minutes in room temperature WB in the absence (control) or presence of (A) unlabeled PAH (250 μM), (B) telmisartan (5 μM), or (C) both in combination added to the extracellular buffer. Cells were rinsed rapidly 3 times with ice-cold WB before measuring intracellular [3H]PAH. Data are presented as a percentage of the intracellular [3H]PAH level immediately before performing efflux (t = 0 minutes). Data are the mean ± S.E.M. of four preparations (n = 4). *P < 0.05, **P < 0.01, ***P < 0.0001, significantly different than control, unpaired Student’s t test.
Discussion
In the present study we used the OAT1 substrates PAH and 6-carboxyfluorescein to examine the mechanism by which compounds inhibit OAT1. Regardless of the substrate used, the individual compounds tested had similar effects on their kinetics, perhaps suggesting that inhibition mechanism at OAT1 is substrate independent. Most of the compounds inhibited transport in a competitive manner (10 out of 14), but several showed mixed or noncompetitive inhibition.
Importantly, competitive, mixed, and noncompetitive inhibitors could differentially impact the rate of tubular organic anion secretion mediated by OAT1 in vivo, and hence the magnitude of overall renal organic anion elimination and drug-drug interaction. That is, the inhibition potential of a reversible competitive inhibitor (perpetrator drug) is dependent on the concentration of itself as well as the concentration of the drug substrate (victim drug) at the OAT1 ligand binding surface, both of which are changing with time after administration—that is, high concentrations of victim can outcompete the perpetrator for binding to OAT1 and vice versa. In contrast, the inhibition potential of a noncompetitive perpetrator is only dependent on the concentration of itself at OAT1, independent of the victim concentration. Consequently, a noncompetitive inhibitor has greater potential to cause a drug-drug interaction than a competitive inhibitor, assuming both inhibit OAT1 with the same potency and have similar plasma concentration-time profiles.
Both static and mechanistic models, such as Cmax, unbound/IC50 (CDER, 2012) and physiologically based pharmacokinetic models, respectively, are used to predict drug-drug interactions at OAT1. However, these models assume reversible competitive inhibition, and by doing so, could underestimate drug-drug interaction magnitude with compounds that inhibit by other mechanisms.
Experiments further examining the interaction of telmisartan with OAT1 were useful for investigating the possible mechanism by which a noncompetitive inhibitor can reduce OAT1 transport activity. Experiments conducted at room temperature versus ice-cold conditions, or after permeabilization of the plasma membrane with saponin, showed that telmisartan binds to OAT1 but does not appear to be a translocated substrate. That is, the cellular accumulation of telmisartan was significantly higher in CHO-OAT1 cells compared with parental cells despite using procedures that completely inhibit OAT1 translocation. Although reducing bath temperature appeared to lower the cellular accumulation of telmisartan by CHO-OAT1 cells (as well as CHO parental cells), we attribute this effect to a reduction in passive membrane permeability at low temperatures, as opposed to an effect on active transport. Indeed, Poirier et al. (2008) showed that permeability of organic solutes, including organic anions, across CHO cells as well as artificial membranes is highly sensitive to temperature.
The apparent modest reduction in [3H]telmisartan accumulation by CHO-OAT1 cells in the presence versus absence of saponin before treatment is consistent with the observed modest reduction in OAT1 expression at the cell surface as determined by cell surface biotinylation (data not shown). Although our data are consistent with binding of telmisartan to OAT1 without translocation, we cannot rule out the possibility that the compound is actively transported. It is possible that the high passive membrane permeability of telmisartan masks any active transport mediated by OAT1. Regardless, <1% of a telmisartan dose is recovered in the urine of humans (Michel et al., 2013), suggesting that if transported by OAT1, the activity is not important for its elimination.
Telmisartan reduced the maximal transport rate, which is a function of the number of transporters expressed at the plasma membrane and their turnover number. The level of OAT1 expressed at the plasma membrane was unaffected by telmisartan, indicating that it inhibits by reducing OAT1 turnover number. Based on these results, we speculate that both ligands (substrate and noncompetitive inhibitor) interact at distinct sites in the OAT1 ligand binding pocket, and the presence of telmisartan stabilizes the transporter in an unfavorable conformation for substrate translocation.
A previous study showed that OATP1B1 can be inhibited for long periods after drug exposure and its removal, perhaps suggesting a mechanism-based inhibition (Shitara et al., 2013). Thus, we performed experiments where the cells were treated with telmisartan (5 μM), it was removed by extensive washing, the cells were allowed to recover for various time periods, and the cells were washed again before measuring PAH uptake [this protocol is somewhat similar to one used by Shitara et al. (2013) with OATP1B1]. After 1 hour, only ∼50% of the PAH transport activity had recovered. Given telmisartan’s high apparent passive permeability, we hypothesized that it gets into the cells and stays there for long periods despite washing. To examine this possibility, we treated the cells with [3H]telmisartan and put through identical washing steps as in the functional studies. After 1 hour, 8% of the [3H]telmisartan dose remained. If we assume that 8% of the telmisartan dose remained in the functional studies where cells were treated initially with 5 μM telmisartan, then ∼0.4 μM telmisartan would have remained after the 1-hour washout period. This is enough telmisartan to inhibit ∼50% of OAT1 activity based on its IC50 value—almost exactly what was observed.
The recovery of PAH transport after telmisartan treatment and its removal was much faster when including 10% serum in the washout buffer. Accordingly, [3H]telmisartan that had accumulated in CHO-OAT1 cells after a brief incubation was more rapidly removed when 10% serum was included in the washout buffer. Given telmisartan’s high degree of plasma protein binding (∼99.5%; Stangier et al., 2000), the plasma protein likely acted as a sink to more quickly remove telmisartan from the cells. These data are consistent with those of Bow et al. (2006), who showed that plasma protein can profoundly influence the interaction of ligands with renal organic anion transporters, including OAT1. Together, these data indicate that telmisartan binding to OAT1 and its inhibitory effect are reversible with time and that the relatively long-lasting inhibitory effect is due to the slow rate at which telmisartan leaves the cells.
Given the apparent high passive permeability of telmisartan, we speculated that, in addition to uptake, it could inhibit OAT1-mediated efflux as well. Accordingly, we examined the time-course of [3H]PAH efflux from CHO-OAT1 cells under a variety of conditions. The rate of [3H]PAH efflux from control cells was relatively rapid, with nearly all the intracellular [3H]PAH eliminated by 20 minutes. This was unexpected given that the extracellular buffer did not contain an ionic species considered exchangeable on OAT1—the predominant anions were chloride and HEPES. Replacement of chloride with gluconate or HEPES with Tris-HCl had no appreciable effect on the rate of [3H]PAH efflux from CHO-OAT1 cells (Supplemental Fig. 2). This is consistent with previous work examining PAH transport by renal basolateral membrane vesicles, which showed that neither chloride nor gluconate are exchangeable with PAH (Schmitt and Burckhardt, 1993), and HEPES does not cis-inhibit PAH uptake mediated by OAT1 (personal observation), suggesting that it is not an exchangeable substrate either.
Although a study using basolateral membrane vesicles has suggested that OATs can operate via PAH/OH− exchange (Eveloff, 1987), lowering extracellular pH from 7.4 to 5.5 had no influence on [3H]PAH efflux (Supplemental Fig. 2). As anticipated, [3H]PAH efflux was stimulated in the presence of a saturating concentration of unlabeled PAH in the extracellular buffer, indicating PAH/[3H]PAH self-exchange. In the absence of an exchangeable substrate in the extracellular buffer, one may speculate that efflux under control conditions is driven by intracellular [3H]PAH moving down its concentration gradient across the plasma membrane by passive diffusion. Yet telmisartan either alone or in the presence of a saturating concentration of unlabeled PAH slowed [3H]PAH efflux compared with control, indicating that a large fraction of flux occurred through the transporter.
Despite the fact that OAT1 is considered an obligatory exchanger, an explanation for the efflux activity observed is that some PAH flux through OAT1 occurs independent of exchange, via molecular slips in the OAT1 transport protein. One mechanism by which this could be explained is that after translocation of intracellular [3H]PAH to the outside of the cell, the unloaded OAT1 protein can return to its inward-facing conformation. An alternative explanation is that substrate binding causes a conformational change that favors channel formation and organic anion conductance. The phenomenon of molecular slippage in membrane transporters and the possible mechanisms noted above has been reviewed in the literature (Nelson et al., 2002).
In summary, we have shown that OAT1 is susceptible to multiple mechanisms of inhibition, including competitive, mixed, and noncompetitive. The mechanism of OAT1 inhibition caused by a perpetrator drug should be considered when modeling drug-drug interactions at OAT1 to more accurately predict the perpetrator’s influence on renal elimination of victim drugs. The noncompetitive inhibitor telmisartan most likely binds reversibly to a site distinct from the substrate and stabilizes the transporter in a conformation unsuitable for substrate translocation, in either the uptake or efflux direction. Future work is required to determine whether this is a common mechanism by which noncompetitive inhibitors reduce OAT1 activity. Telmisartan was also useful for uncovering a transport mode for OAT1 that appears to occur independent of exchange.
Authorship Contributions
Participated in research design: Hotchkiss, Gao, Pelis.
Conducted experiments: Hotchkiss, Gao, Khan, Berrigan, Li, Ingraham.
Performed data analysis: Hotchkiss, Gao, Khan, Berrigan, Li.
Wrote or contributed to the writing of the manuscript: Hotchkiss, Pelis.
Footnotes
- Received May 27, 2015.
- Accepted September 11, 2015.
This work was supported by the Nova Scotia Health Research Foundation [grant MED-EST-2013-9003].
↵
This article has supplemental material available at dmd.aspetjournals.org.
Abbreviations
- CHO
- Chinese hamster ovary
- Jmax
- maximal transport rate
- OAT
- organic anion transporter
- PAH
- para-aminohippurate
- WB
- Waymouth’s buffer
- Copyright © 2015 by The American Society for Pharmacology and Experimental Therapeutics
