Drug interactions involving metabolism and/or excretion cause marked changes in plasma and intracellular concentrations of the affected drug. Such interactions can lead to severe adverse effects or unexpected pharmacological effects. Recent studies have shown that many transporters play important roles in the uptake and subsequent secretion of drugs in the liver and kidney, and several specific transporter proteins involved in drug interactions have been reported (Shitara et al., 2005; Endres et al., 2006; Li et al., 2006). However, most of these reports examined the involvement of either apical or basolateral transporters. To reproduce drug interactions via renal transporters in vivo, analyses by the epithelial cells with transcellular transport systems are needed.

In human proximal tubules, organic cations are taken up from blood into cells by human organic cation transporter 2 (hOCT2/SLC22A2) and then eliminated into the lumen by apical H⁺/organic cation antiporters, human multidrug and toxin extrusion 1 (hMATE1/SLC47A1) and hMATE2-K (SLC47A2). hMATE1 and hMATE2-K have been identified recently (Otsuka et al., 2005; Masuda et al., 2006), and we have characterized both transporters in terms of tissue distribution, membrane localization, transcriptional regulation, and transport characteristics (Terada and Inui, 2008). However, there has been little systematic comparison of substrate affinity among hMATE1, hMATE2-K, and hOCT2. Comparisons of substrate affinity for hMATEs and hOCT2 would enable us to predict where drug interactions between cationic drugs are likely to occur in the renal secretion process in vivo. It is important to predict where on the membrane drug interaction will occur because the intracellular accumulation of drugs at the proximal tubules may be involved in drug-induced nephrotoxicity.

Here, for a systematic comparison of substrate affinity among hMATE1, hMATE2-K, and hOCT2, the affinity of various drugs for hMATE1 and hMATE2-K was elucidated. Then, based on a comparison of the affinity for the hMATE family and hOCT2, the involvement of hMATE1 in the drug interaction between cimetidine and metformin was evaluated by using double-transfected Madin-Darby canine kidney (MDCK) cells stably expressing both hOCT2 and hMATE1 (MDCK-hOCT2/hMATE1 cells). Recently, we have established this double transfectant, and demonstrated that MDCK-hOCT2/hMATE1 cells are an appropriate in vitro model of human tubular epithelial cells to evaluate the transcellular transport of cationic drugs (Sato et al., 2008).

Materials and Methods

Materials. Pramipexole dihydrochloride monohydrate and talipexole hydrochloride were kindly provided by Nippon Boehringer Ingelheim (Tokyo, Japan). [¹⁴C]Tetraethylammonium (TEA) (2.035 GBq/mmol) was purchased from American Radiolabeled Chemicals (St. Louis, MO). [¹⁴C]Metformin (962 MBq/mmol) and d-[¹⁴C]mannitol (2 GBq/mmol) were purchased from Moravek Biochemicals (Brea, CA). [N-Methyl-³H]cimetidine (470 GBq/mmol) was obtained from GE Healthcare (Chalfont St. Giles, UK). d-[1-³H(N)]mannitol (525.4 GBq/mmol) was acquired from PerkinElmer Life and Analytical Sciences (Boston, MA). Quinidine sulfate, verapamil hydrochloride, imipramine hydrochloride, diphenhydramine hydrochloride, (±)-chlorpheniramine maleate, and cimetidine were purchased from Nacalai Tesque (Kyoto, Japan). Metformin hydrochloride, diltiazem hydrochloride, procainamide hydrochloride, ranitidine hydrochloride, and amantadine hydrochloride were obtained from Sigma-Aldrich (St. Louis, MO). Disopyramide, desipramine hydrochloride and famotidine were acquired from Wako Pure Chemicals (Osaka, Japan). Cetirizine was obtained from LKT Labs (St. Paul, MN). All other chemicals used were of the highest purity available.

Cell Culture. According to our previous report (Tanihara et al., 2007), HEK293 cells stably expressing hMATE1 (HEK-hMATE1 cells) or hMATE2-K (HEK-hMATE2-K cells) were cultured in complete medium consisting of Dulbecco's modified Eagle's medium (Sigma-Aldrich) with 10% fetal bovine serum (Invitrogen, Carlsbad, CA) and hygromycin B (0.2 mg/ml; Invitrogen) in an atmosphere of 5% CO₂/95% air at 37°C. The double-transfected MDCK-hOCT2/hMATE1 cells were previously established in our laboratory (Sato et al., 2008). MDCK-hOCT2/hMATE1 cells were cultured in complete medium consisting of Dulbecco's modified Eagle's medium with 10% fetal bovine serum, hygromycin B (0.2 mg/ml), and G-418 (0.5 mg/ml; Nacalai Tesque) in an atmosphere of 5% CO₂/95% air at 37°C. In the previous study, we demonstrated that hOCT2 and hMATE1 were localized on the basolateral and apical membrane, respectively, in MDCK-hOCT2/hMATE1 cells by immunofluorescence microscopy (Sato et al., 2008). HEK293 cells stably expressing hOCT2 (HEK-hOCT2 cells) were cultured according to a previous report (Urakami et al., 2004).

Preparation of Membrane Vesicles from HEK-hMATE1 and HEK-hMATE2-K Cells. HEK-hMATE1 or HEK-hMATE2-K cells were seeded on 100-mm plastic dishes (5 × 10⁶ cells/dish), with 25 dishes used in a single preparation. Then, plasma membrane vesicles were prepared according to our previous report (Tsuda et al., 2007).

Transport Experiments by Membrane Vesicles. The uptake of [¹⁴C]TEA by membrane vesicles from HEK-hMATE1 or HEK-hMATE2-K cells was measured using a rapid filtration technique according to our previous report (Tsuda et al., 2007). Nonspecific absorption was determined by the addition of [¹⁴C]TEA to 1 ml of ice-cold stop solution containing 20 μl of membrane vesicles. This value was subtracted from the total uptake data. The protein content was determined by the method of Bradford (1976) using a Bio-Rad Protein Assay Kit (Bio-Rad, Hercules, CA) with bovine γ-globulin as a standard.

Transport Experiments by Transfectants. The cellular uptake of radiolabeled compounds was performed as reported previously (Urakami et al., 2004; Tanihara et al., 2007). The transcellular transport experiments with radiolabeled compounds were performed as described previously (Sato et al., 2008). In brief, MDCK-hOCT2/hMATE1 cells were seeded on microporous membrane filters (3.0-μm pores, 4.7-cm² growth area) inside a Transwell cell culture chamber (Corning Life Sciences, Lowell, MA) at a density of 18 × 10⁵ cells/well. On the 7th day after the seeding, the cells were used for transcellular transport experiments. The composition of the incubation medium was as follows: 145 mM NaCl, 3 mM KCl, 1 mM CaCl₂, 0.5 mM MgCl₂, 5 mM d-glucose, and 5 mM HEPES, pH 7.4, or MES, pH 6.0. After the removal of the culture medium from both sides of the monolayers, the cell monolayers were preincubated with 2 ml of incubation medium, pH 7.4, on each side for 10 min at 37°C. Then, 2 ml of incubation medium, pH 7.4, containing both a radiolabeled compound and radiolabeled mannitol was added to the basolateral side, and 2 ml of nonradioactive incubation medium, pH 6.0, was added to the apical side. The monolayers were incubated for specified periods of time at 37°C. Radiolabeled mannitol, a compound that is not transported by the cells, was used to calculate paracellular flux and the extracellular trapping of substrates. For transport measurements, an aliquot (100 μl) of the incubation medium on the apical side was taken at specified times, and the radioactivity was determined in 5 ml of ACSII by liquid scintillation counting. For accumulation studies, the medium was removed by aspiration at the end of the incubation period, and the monolayers were rapidly washed twice with 2 ml of ice-cold incubation medium, pH 7.4, on each side. The filters were detached from the chambers, the cells on the filters were solubilized with 0.5 ml of 0.5 N NaOH, and the radioactivity in aliquots (300 μl) was determined in 3 ml of ACSII by liquid scintillation counting. The protein content of the solubilized cells was determined using a Bio-Rad Protein Assay Kit with bovine γ-globulin as a standard (Bradford, 1976).

Measurement of Intracellular Volume. The equilibrium uptake of sulfanilamide was determined according to standard procedures for transcellular transport experiments with a slight modification. In brief, after the preincubation, 2 ml of incubation medium, pH 7.4, containing both 10 mM sulfanilamide and d-[³H]mannitol was added to both sides. The intracellular volume was calculated from the equilibrium uptake (60 min) and external concentration of sulfanilamide. Sulfanilamide was diazotized and coupled with 2-DEAE-1-naphthylamine oxalate, and the amount of colored material was determined spectrophotometrically at 550 nm (Saito et al., 1986).

Data Analysis. Data were expressed as the means ± S.D. Two or three experiments were conducted, and representative results were shown. Apparent K_i values were expressed as the means ± S.E. and analyzed statistically using the unpaired Student's t test. Data from transport experiments were analyzed statistically with the one-way analysis of variance followed by Dunnett's test.

Results

Determination of the Driving Force of hMATE1 and hMATE2-K with Membrane Vesicles from HEK-hMATE1 and HEK-hMATE2-K Cells. At first, to validate the driving force of hMATE1 and hMATE2-K, we carried out [¹⁴C]TEA uptake studies with membrane vesicles from HEK-hMATE1 or HEK-hMATE2-K cells. In the presence of an outwardly directed H⁺ gradient, a marked stimulation of [¹⁴C]TEA uptake (overshoot phenomenon) was observed in membrane vesicles from HEK-hMATE1 and HEK-hMATE2-K cells (Fig. 1, A and B). Furthermore, [¹⁴C]TEA uptake by hMATE1 and hMATE2-K was not altered by the presence of valinomycin with or without the H⁺ gradient (Fig. 1, C and D). These findings indicated that both transporters utilized an oppositely directed H⁺ gradient as a driving force. Subsequent experiments were carried out in the presence of an outwardly directed H⁺ gradient to evaluate the transport characteristics of the MATE family.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Determination of the driving force of the hMATE family with membrane vesicles from HEK-hMATE1 (A and C) and HEK-hMATE2-K (B and D) cells. A and B, effects of various H⁺ gradients on [¹⁴C]TEA uptake. Membrane vesicles were prepared in the experimental buffer at pH 6.0 (• and ▵) or 7.5 (• and ▴). The uptake of [¹⁴C]TEA was examined in the experimental buffer containing 31.25 μM [¹⁴C]TEA and 100 mM KCl at pH 6.0 (▵ and ▴) or 7.5 (○ and •). C and D, effects of valinomycin on [¹⁴C]TEA uptake with or without a H⁺ gradient. Membrane vesicles were prepared in the experimental buffer at pH 6.0 (○ and •) or 7.5 (▵ and ▴). The uptake of [¹⁴C]TEA was examined in the experimental buffer containing 31.25 μM [¹⁴C]TEA and 100 mM CsCl at pH 7.5 in the absence (○ and ▵) or presence (• and ▴) of 8 μM valinomycin. Valinomycin, a K⁺ ionophore, was employed to produce an inside-negative membrane potential in the presence of outward K⁺ gradient. Each point represents the mean ± S.D. for three determinations. These figures are representative of two separate experiments. pH_in, intravesicular pH; pH_out, extravesicular pH.

Comparison of Substrate Affinity for hMATE1, hMATE2-K, and hOCT2. To compare the substrate affinity for hMATE1 and hMATE2-K, the apparent K_i values of various cationic drugs for the uptake of [¹⁴C]TEA were determined. Cationic drugs inhibited [¹⁴C]TEA uptake via hMATE1 and hMATE2-K in a dose-dependent manner. Figures 2 and 3 show representative inhibition curves of histamine H₂ receptor antagonists and antiparkinsonian agents, respectively. Table 1 provides a summary of the apparent K_i values of various drugs for hMATE1 and hMATE2-K. Histamine H₂ receptor antagonists such as cimetidine and famotidine showed high affinity for hMATE1 and hMATE2-K as compared with other drugs. Although many drugs tended to have higher affinity for hMATE1 than for hMATE2-K, only pramipexole, an antiparkinsonian agent, exhibited higher affinity for hMATE2-K than for hMATE1. Table 2 shows the affinity for hMATE1, hMATE2-K, and OCT2 of various compounds excreted from the kidney. K_m values of hOCT2 and rat (r) OCT2 were cited from previous papers (Urakami et al., 2001; Ishiguro et al., 2005; Koepsell et al., 2007). Most drugs showed higher affinity for OCT2 than for hMATEs.

Effects of Cimetidine on Transcellular Transport of [¹⁴C]Metformin in MDCK-hOCT2/hMATE1 Cells. It is interesting that histamine H₂ receptor antagonists such as cimetidine showed higher affinity for hMATEs than for hOCT2 (Table 2). Furthermore, recent clinical pharmacokinetic and pharmacogenomic analyses of metformin have paid attention to the renal secretion process (Shikata et al., 2007; Song et al., 2008; Takane et al., 2008; Wang et al., 2008). Therefore, we examined the effect of different concentrations of cimetidine on the basolateral-to-apical transport and intracellular accumulation of [¹⁴C]metformin in MDCK-hOCT2/hMATE1 cells. We have demonstrated recently that MDCK-hOCT2/hMATE1 showed the vectorial transcellular transport of organic cations, such as TEA and metformin (Sato et al., 2008). As shown in Fig. 4A, the transcellular transport of [¹⁴C]metformin was moderately inhibited by the presence of 1 μM cimetidine and almost completely inhibited by the presence of 1 mM cimetidine. The cellular accumulation of [¹⁴C]metformin was inhibited by 1 mM cimetidine but increased by 1 μM cimetidine (Fig. 4B). Furthermore, we examined the inhibitory effects of cimetidine on the uptake of [¹⁴C]metformin by hOCT2. Cimetidine had little effect at 1 μM but inhibited the uptake at 1 mM (Fig. 5). The apparent K_i value of cimetidine for hOCT2 was calculated to be 147.1 ± 11.0 μM. These results suggest that 1 mM cimetidine competitively inhibits the uptake of [¹⁴C]metformin by hOCT2, resulting in a drastic decrease in both the transcellular transport and the intracellular accumulation of [¹⁴C]metformin. On the other hand, 1 μM cimetidine inhibits the apical efflux of [¹⁴C]metformin via hMATE1, but not the uptake by hOCT2, causing a marked increase in its cellular accumulation.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Effects of histamine H₂ receptor antagonists on [¹⁴C]TEA uptake by HEK-hMATE1 (A) and HEK-hMATE2-K (B) cells. HEK-hMATE1 and HEK-hMATE2-K cells were preincubated with 30 mM NH₄Cl, pH 7.4, for 20 min. Then, the preincubation medium was removed, and the cells were incubated with 5 μM [¹⁴C]TEA in the presence of cimetidine (○), famotidine (•), or ranitidine (▵) for 1 min at 37°C (pH 7.4). After the incubation, the radioactivity of solubilized cells was measured. Each point represents the mean ± S.D. for three monolayers. These figures are representative of three separate experiments.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Effects of antiparkinsonian agents on [¹⁴C]TEA uptake by HEK-hMATE1 (A) and HEK-hMATE2-K (B) cells. HEK-hMATE1 and HEK-hMATE2-K cells were preincubated with 30 mM NH₄Cl, pH 7.4, for 20 min. Then, the preincubation medium was removed, and the cells were incubated with 5 μM [¹⁴C]TEA in the presence of amantadine (○), pramipexole (•), or talipexole (▵) for 1 min at 37°C, pH 7.4. After the incubation, the radioactivity of solubilized cells was measured. Each point represents the mean ± S.D. for three monolayers. These figures are representative of three separate experiments.

Intracellular Concentration of [³H]Cimetidine in MDCK-hOCT2/hMATE1 Cells. To confirm that hMATE1 is specifically inhibited by intracellular cimetidine taken up from the basolateral side, the intracellular concentration of cimetidine was determined. The amount of intracellular cimetidine after 60 min of incubation was demonstrated to be 10.0 ± 0.5 pmol/mg protein. We then measured the intracellular volume of MDCK-hOCT2/hMATE1 cells with the equilibrium uptake of sulfanilamide and found that it was 2.2 ± 0.4 μl/mg protein. Based on these findings, the intracellular concentration of cimetidine was calculated as 4.5 ± 0.4 μM. Considering the K_i values of cimetidine for hMATE1 and hOCT2, this intracellular concentration suggests that the transport of cationic drugs via hMATE1 is specifically inhibited by intracellular cimetidine, but transport via hOCT2 is little inhibited by intracellular cimetidine (Fig. 6).

Discussion

We reported that hMATE1 and hMATE2-K have a similar substrate specificity for many endogenous organic cations and cationic drugs and that several substrates have higher affinity for hMATE1 than for hMATE2-K (Tanihara et al., 2007). In the present study, four cationic drugs had similar affinity for hMATE1 and for hMATE2-K, and 12 cationic drugs were found to have higher affinity for hMATE1 than for hMATE2-K (Table 1). This would suggest that hMATE1 and hMATE2-K have complementary roles in the renal secretion of organic cations. It is interesting that only pramipexole showed markedly higher affinity for hMATE2-K than for hMATE1 (Fig. 3; Table 1). It has been reported that pramipexole is transported by rOCT1 and rOCT2 (Ishiguro et al., 2005), but it is unclear whether pramipexole is transported by the hMATE family and hOCT2. Pramipexole may be a good probe to distinguish the molecular mechanisms by which the hMATE family recognizes substrates.

In the present study, most drugs showed higher affinity for OCT2 than for hMATEs (Table 2). Histamine H₂ receptor antagonists such as cimetidine and famotidine, however, showed higher affinity for hMATEs than for hOCT2. In particular, the K_i values of cimetidine for hMATE1 and for hOCT2 were 1.1 and 8.6 to 73 μM (Koepsell et al., 2007), respectively. Taking into consideration the effective blood concentration of cimetidine (2.0-3.6 μM) (Benet et al., 1996), it is assumed that hOCT2 is little inhibited by cimetidine at clinical doses, but hMATE1 is inhibited by cimetidine taken up from blood into cells. To verify this hypothesis, we performed transcellular transport studies with a double transfectant composed of basolateral hOCT2 and apical hMATE1. We found that 1 μM cimetidine was taken up from the basolateral side without an inhibitory effect on metformin transport via hOCT2, and intracellular cimetidine inhibited the transport of metformin via hMATE1 (Figs. 4, 5, 6). These results suggest that hMATE1 is responsible for the drug interaction between cimetidine and metformin. In fact, it has been reported that cimetidine significantly increased the area under the curve of metformin by an average of 50% and reduced its renal clearance over 24 h by 27% in human subjects (Somogyi et al., 1987). Cimetidine has been shown to be an inhibitor of the active tubular secretion of many cationic drugs, including procainamide (Somogyi et al., 1983; Christian et al., 1984), N-acetylprocainamide (Somogyi et al., 1983), ranitidine (van Crugten et al., 1986), triamterene (Muirhead et al., 1986), pilsicainide (Shiga et al., 2000), and varenicline (Feng et al., 2008). These drug interactions could be partially mediated by hMATE1 at the apical membrane in the proximal tubules.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Effects of cimetidine on transcellular transport (A) and intracellular accumulation (B) of [¹⁴C]metformin by MDCK-hOCT2/hMATE1 cells. The cell monolayers were incubated at 37°C with 2 ml of incubation medium, pH 7.4, containing [¹⁴C]metformin (10.6 μM) and d-[³H]mannitol (0.079 μM) added to the basolateral side in the absence (Control, ○) or presence of either 1 μM cimetidine (•) or 1 mM cimetidine (▴), and 2 ml of nonradioactive incubation medium, pH 6.0, was added to the apical side. After the incubation, the radioactivity on the apical side and intracellular accumulation were measured. Transcellular transport of [¹⁴C]metformin was calculated by subtracting the flux of [³H]mannitol from the net flux of [¹⁴C]metformin. After 60 min of incubation, the cell monolayers were washed twice with ice-cold incubation medium, and the radioactivity of solubilized cells was determined. Each point and column represent the mean ± S.D. for three monolayers. When error bars are not shown, they are smaller than the symbols. These figures are representative of two separate experiments. ^*, P < 0.05; ^**, P < 0.01, significantly different from control.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Effects of cimetidine on [¹⁴C]metformin uptake by HEK-hOCT2 cells. HEK-hOCT2 cells were preincubated with incubation medium, pH 7.4, for 10 min. The preincubation medium was removed, and the cells were incubated with 10.6 μM [¹⁴C]metformin in the presence of cimetidine for 2 min at 37°C, pH 7.4. The radioactivity of solubilized cells was then measured. Each point represents the mean ± S.D. for three monolayers. This figure is representative of three separate experiments.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Scheme of drug interaction between cimetidine and metformin in MDCK-hOCT2/hMATE1 cells. K_i values of cimetidine for hMATE1 and hOCT2 were calculated as 1.1 and 147 μM, respectively. When cimetidine at 1 μM was added to the basolateral side, intracellular concentration of cimetidine after 60 min incubation was calculated as 4.5 ± 0.4 μM. Considering K_i values and intracellular concentration of cimetidine, 1 μM cimetidine at the basolateral side did not inhibit metformin transport via hOCT2. However, intracellular cimetidine at 4.5 μM was enough to inhibit the metformin transport via hMATE1. These findings suggested that apical hMATE1 mainly contributed to the drug interaction between cimetidine and metformin.

Recently, it was reported that the 808G>T polymorphism of hOCT2 is associated with a reduced tubular secretion clearance of metformin, and the inhibition by cimetidine also seemed to be dependent on this mutation (Wang et al., 2008). These results suggested that the drug interaction between cimetidine and metformin was affected by accumulation of metformin in the renal epithelial cells. This report would support our notion that the hMATE family plays an important role in the cimetidine-metformin interaction in vivo.

It has been reported that cisplatin, an anticancer platinum agent, was accumulated in the kidney via hOCT2 and induced nephrotoxicity. On the other hand, oxaliplatin was transported by hMATE2-K and hOCT2; therefore, its renal cellular concentration was lowered, suggesting that oxaliplatin did not induce nephrotoxicity (Yonezawa et al., 2006; Yokoo et al., 2007). The present findings suggested that coadministration of cimetidine may increase the tubular accumulation of oxaliplatin by inhibiting hMATE2-K. Double-transfected MDCK-hOCT2/hMATE1 and MDCK-hOCT2/hMATE2-K cells should be very useful in vitro tools for the screening of renal drug interactions with cationic drugs and nephrotoxicity.

In this study, we clearly demonstrated the driving force of hMATE1 and hMATE2-K to be an oppositely directed H⁺ gradient, not inside-negative membrane potential (Fig. 1). These results suggested that [¹⁴C]TEA transport via hMATE1 and hMATE2-K is the electroneutral antiport of H⁺ and it. The results showed that there was no difference in the driving force between hMATE1 and hMATE2-K and are consistent with a report on transport study with human renal brush-border membrane vesicles (Ott et al., 1990). We also demonstrated previously that [¹⁴C]TEA transport by rMATE1 is driven by an oppositely directed H⁺ gradient and is electroneutral (Tsuda et al., 2007), suggesting the driving force for the MATE family to be conserved among human and rodents.

In conclusion, we demonstrated for the first time that apical hMATE1 is involved in the drug interaction between cimetidine and metformin in renal epithelial cells. These results provide important information on the physiological and pharmacokinetic roles of the MATE family and may help to avoid the unexpected adverse effects of drug interactions with cationic drugs.