Interaction of Oral Antidiabetic Drugs With Hepatic Uptake Transporters: Focus on Organic Anion Transporting Polypeptides and Organic Cation Transporter 1

[³H]sulfobromophthalein ([³H]BSP; 7,585 GBq/mmol) was obtained from Hartmann Analytic (Braunschweig, Germany). [³H]1-methyl-4-phenylpyridinium ([³H]MPP⁺; 85 Ci/mmol), [³H]rosiglitazone (50 Ci/mmol), and unlabeled rosiglitazone were obtained from BIOTREND Chemikalien (Cologne, Germany). Unlabeled metformin and poly-d-lysine hydrobromide were purchased from Sigma-Aldrich Chemie (Taufkirchen, Germany). Unlabeled repaglinide was obtained from Chemos (Regenstauf, Germany). [¹⁴C]metformin hydrochloride (0.1 Ci/mmol) was provided by Merck (Darmstadt, Germany). Methanol (hypergrade), n-hexane (pro-analysis), acetonitrile (hypergrade), and acetic acid (supra-pure) were purchased from Merck. Diethyl ether (99.8% purity), ammonium acetate (pro-analysis), and ibuprofen were obtained from Sigma-Aldrich Chemie.

Stably transfected human embryonic kidney cells (HEK293 cells) and Madin-Darby canine kidney cells (MDCKII cells) were cultured as described previously (5,19). For uptake and inhibition experiments, stably transfected HEK293 cells recombinantly expressing human OATP1B1 (HEK-OATP1B1), human OATP1B3 (HEK-OATP1B3) (5), or human OATP2B1 (HEK-OATP2B1) and the respective HEK control cells (HEK-Co/418, transfected with the empty expression vector pcDNA3.1; or HEK-Co/Hy, transfected with the empty expression vector pcDNA3.1/Hygro) were seeded in 0.1 mg/ml poly-d-lysine–coated 12-well plates (Cell Culture Multiwell Plate Cellstar; Greiner Bio-One, Frickenhausen, Germany) at an initial density of 700,000 cells/well. The cells were grown to confluence for 1 day and then induced with 10 mmol/l sodium butyrate (Merck) for 24 h before the uptake experiments to obtain higher levels of the recombinant protein (20).

MDCKII cells were transfected with the respective plasmid pcDNA3.1/Hygro(−)-OCT1 containing the full-length cDNA encoding the human OCT1-protein (NM_003057) using Effectene transfection reagent (Qiagen, Hilden, Germany). After hygromycin (800 μg/ml) selection, single colonies were characterized for SLC22A1 (encoding human OCT1) mRNA expression by real-time PCR analysis. Vector transfected MDCKII control cells were established by the same method using the respective expression plasmid without insert for transfection.

For MPP⁺ uptake experiments, MDCKII cells were seeded in 0.1 mg/ml poly-d-lysine–coated 12-well plates (Cell Culture Multiwell Plate Cellstar; Greiner Bio-One) at an initial density of 700,000 cells per well. The cells were grown to confluence for 2 days and induced with 10 mmol/l sodium butyrate (Merck) for 24 h before the uptake experiments (20).

Before starting the uptake experiments with the HEK cells, cells were washed with prewarmed (37°C) uptake buffer (142 mmol/l NaCl, 5 mmol/l KCl, 1 mmol/l K₂HPO₄, 1.2 mmol/l MgSO₄, 1.5 mmol/l CaCl₂, 5 mmol/l glucose, and 12.5 mmol/l HEPES, pH 7.3). [³H]BSP was dissolved in uptake buffer, and unlabeled BSP was added to the final concentration of 0.05 μmol/l BSP for studies with HEK-OATP1B1 and 1 μmol/l BSP for studies with HEK-OATP1B3 and HEK-OATP2B1, respectively. To characterize repaglinide and rosiglitazone as inhibitors of OATP-mediated uptake, both drugs were added in increasing concentrations (0.01–100 μmol/l). The cells were incubated with the solution at 37°C for 10 min as described previously (5). Afterward, the cells were washed three times with ice-cold uptake buffer before lysing the cells with 0.2% SDS. The intracellular accumulation of radioactivity was detected by liquid scintillation counting (Perkin Elmer Life Sciences). For the experiments using pravastatin as substrate for OATP1B1 and OATP1B3, a concentration of 50 μmol/l pravastatin was used. To test the inhibitory effect of repaglinide and rosiglitazone, each drug was added in concentrations of 10 and 100 μmol/l. The uptake assay was performed as described above. The intracellular pravastatin concentration was determined by liquid chromatography with tandem mass spectrometry detection analysis as previously described (5).

The uptake experiments with MDCKII-OCT1 cells were carried out in an analogous manner. [³H]MPP⁺ was dissolved in uptake buffer, and unlabeled MPP⁺ was added to the final concentration of 30 μmol/l. To characterize repaglinide, rosiglitazone, and metformin as inhibitors of OCT1-mediated uptake, the drugs were added in increasing concentrations of 0.1–1,000 μmol/l for metformin and 0.1–100 μmol/l for repaglinide and rosiglitazone, respectively. For experiments using metformin as substrate of OCT1, [¹⁴C]metformin was dissolved in uptake buffer, and unlabeled metformin was added to a final concentration of 10 μmol/l. To test the inhibitory effect of repaglinide and rosiglitazone, each drug was added in a concentration of 0.1–100 μmol/l. All experiments were repeated at least three times.

The OATP1B1-, OATP1B3-, OATP2B1-, and OCT1-mediated net uptake was obtained by subtracting the uptake in vector-transfected cells from that in OATP1B1-, OATP1B3-, OATP2B1-, and OCT1-expressing cells. The percentage of uptake inhibition was calculated from control experiments in the absence of antidiabetic drugs (100% uptake). The corresponding half-maximal inhibitory concentration (IC₅₀) values for uptake inhibition were calculated by fitting the data to a sigmoidal dose-response regression curve (Prism 4.01 2004; GraphPad Software, San Diego, CA). Effects of repaglinide and rosiglitazone on pravastatin uptake in OATP1B1 and OATP1B3 cells were analyzed with unpaired two-tailed t test (Prism 4.01 2004; GraphPad Software). A value of P < 0.05 was required for statistical significance.

To examine the inhibitory potency of antidiabetic drugs on OCT1-mediated uptake, MDCKII cells were stably transfected with the SLC22A1 cDNA and selected for a high expression of the uptake transporter OCT1. The SLC22A1 mRNA expression of the selected cell clones was analyzed using quantitative real-time PCR. This analysis demonstrated a high SLC22A1 mRNA expression in several MDCKII-OCT1 clones compared with vector-transfected cells. The cell clone with the highest mRNA expression was used for further transport experiments using the prototypic tritium-labeled substrate [³H]MPP⁺. MPP⁺ was shown to be a high-affinity substrate for OCT1 with a K_m value of ∼21 μmol/l, which is in accordance with previously published data (18). The uptake experiments (Fig. 1A and B) demonstrated that MDCKII-OCT1 cells were able to mediate uptake of MPP⁺ (50 μmol/l) into cells with an uptake ratio of 3.7-fold compared with control cells. Furthermore, we confirmed that metformin, previously shown to be a substrate for OCT1 with a K_m value of 1,470 μmol/l (18), is also transported by the newly established MDCKII-OCT1 cells (Fig. 1C).

The results of the inhibition experiments are summarized in Table 1 and Figs. 2–4. Metformin did not inhibit OATP1B1-, OATP1B3-, or OATP2B1-mediated BSP uptake (Fig. 2) or uptake of the OCT1 substrate MPP⁺ (Fig. 4). However, OATP-mediated BSP and pravastatin uptake and OCT1-mediated MPP⁺ and metformin uptake were significantly inhibited by repaglinide (Figs. 2–4; Table 1) and rosiglitazone (Figs. 2–4; Table 1).

Patients with type 2 diabetes frequently have to be treated with more than one drug. Effects of oral antidiabetic drugs depend on the extent of drug absorption from the gut lumen, on metabolism of the drug in the liver, and on the extent of its excretion into bile and urine (21). In general, modification of all these processes by a second, concomitantly administered drug can alter the effects of oral antidiabetic drugs (21).

Recently, it was recognized that a broad variety of drugs, including many cardiovascular drugs such as statins and angiotensin II-receptor antagonists, is transported through biological membranes via specific transport proteins (15,22–24). For example, the efflux transporter P-glycoprotein, which translocates its substrates from the inside of the cell to the outside (e.g., from the hepatocyte into bile) is a major determinant of drug effects (1). If P-glycoprotein–mediated drug excretion is inhibited by a second, concomitantly administered compound, drug plasma concentrations increase significantly and may result in drug toxicity (3).

In addition to efflux transporters (e.g., P-glycoprotein), uptake transporters such as the OATPs or the OCTs are important for drug disposition. For example, they mediate the uptake of multiple cardiovascular drugs from the portal venous blood into the hepatocytes (12,18). This transport process is an essential prerequisite for subsequent drug metabolism in the hepatocytes.

The goal of our in vitro study was to investigate whether the oral antidiabetic drugs repaglinide, rosiglitazone, and metformin are inhibitors of the hepatic uptake transporters of the OATP family (OATP1B1, OATP1B3, and OATP2B1) and of the OCT family member OCT1. Whereas metformin did not affect the function of hepatic uptake transporters, both repaglinide and rosiglitazone were significant inhibitors of organic anion and organic cation transport. The following clinical consequences can result from the observed interaction with repaglinide and rosiglitazone. First, if hepatic uptake of a drug (e.g., pravastatin) is inhibited by repaglinide and rosiglitazone and possibly by further oral antidiabetic drugs, the first step (i.e., uptake into hepatocytes) of biliary drug elimination is blocked. This causes increased plasma concentrations and an increased risk of adverse drug reactions. The importance of an impaired uptake transporter function is convincingly shown for OATP1B1. One polymorphism (c.521T>C) in the gene encoding for OATP1B1 results in a significantly reduced OATP1B1 function (25), a situation that is comparable to the OATP1B1 function in the presence of drugs, which inhibit this uptake transporter. Several studies showed that subjects carrying this polymorphism with the reduced uptake transporter function have significantly higher plasma concentrations of pravastatin (26–29). Interestingly, for other drugs, no influence of this genetic variation could be observed (30). This could be due to an overlapping substrate spectrum with OATP1B3, possibly compensating for a reduced uptake mediated by a mutated OATP1B1 protein.

A second clinical consequence of decreased hepatic drug uptake is a reduced effect of the affected drug (e.g., pravastatin). This option applies if the drug effect (e.g., for pravastatin, cholesterol-lowering effect) is mediated via mechanisms within the hepatocyte. A genetically determined reduced hepatic uptake of pravastatin was recently shown to cause a reduced effect of pravastatin (31–33).

The third possible clinical consequence of inhibition of uptake transporter function is increased extrahepatic effects (e.g., for statins, an increased risk of myotoxicity). This reduced benefit (cholesterol-lowering effect)-to-risk (myotoxicity) ratio for OATP1B1-dependent statins by inhibition of hepatic uptake was recently highlighted in an excellent review by Neuvonen et al. (34).

It should be noted that our study focused on in vitro studies on mechanisms of drug interactions with oral antidiabetic drugs. As outlined below, the clinical consequences (e.g., pharmacodynamic consequences) are difficult to predict and require further clinical studies. Another interesting observation of our study was the unexpected stimulation of OATP1B3-mediated pravastatin uptake by rosiglitazone (Fig. 3). Similar observations have been made for other substrates (5); however, the molecular mechanisms underlying this effect are unclear and require further studies.

As drugs reach the portal vein directly after intestinal absorption, the drug concentration in portal venous blood is higher than in the systemic circulation. For estimation of the drug concentrations in portal venous blood, we used the method of Ito et al. (35) (Table 1). The predicted portal venous blood concentrations for repaglinide and rosiglitazone are in the same low micromolar range as the determined IC₅₀ values for inhibition of hepatic uptake, suggesting that inhibition of hepatic drug transporters might cause clinically relevant drug-drug interactions in humans.

To the best of our knowledge, no clinical studies have, so far, been conducted to investigate an impact of repaglinide or rosiglitazone on disposition of pravastatin or other OATP substrates in humans. Pharmacodynamic interactions between metformin and rosiglitazone (36,37) or repaglinide (38) in patients have intensively been investigated. For both rosiglitazone and repaglinide, additional effects on glucose metabolism in patients with type 2 diabetes receiving simultaneously metformin have been shown. Whether these effects can in part be explained by an interaction with hepatic metformin uptake is presently unclear. One study examined the pharmacokinetic interaction between rosiglitazone and metformin in humans (39), however, without a significant pharmacokinetic interaction between these drugs.

The question arises of whether our findings on inhibition of uptake transporter function by oral antidiabetic drugs can be used to draw general conclusions regarding drug-drug interactions. Similar to the present data, macrolid antibiotics were previously shown to cause a relevant inhibition of hepatic uptake transporters (5). Moreover, there is evidence that other oral antidiabetic drugs (glibenclamide) are inhibitors of hepatic uptake transporters (40). Thus, our in vitro data generated in this study indicate that inhibition of uptake transporters by oral antidiabetic drugs is a newly recognized mechanism of potential drug-drug interactions. Our data provide the basis for controlled clinical studies to clarify the importance of drug-drug interactions with oral antidiabetic drugs in studies with healthy volunteers and patients.

This work was supported by Deutsche Forschungsgemeinschaft Grant DFG Ko 2120/1-3 and Grant DFG FR 1298/5-1, by Deutsche Krebshilfe Grant 107854, and by the Johannes and Frieda Marohn-Foundation.

We thank B. Endress, C. Hoffmann, and K. Singer for excellent technical assistance. We thank Dr. D. Keppler (German Cancer Research Center, Heidelberg, Germany) for providing the plasmids pcDNA3.1(+)-OATP1B1 and pcDNA3.1/Hygro(−)-OATP1B3.