Experimental Procedures

Chemicals and Materials.

Metformin, phenformin hydrochloride, and cimetidine were purchased from Sigma-Aldrich (St. Louis, MO). Buformin hydrochloride, acetonitrile of high-performance liquid chromatography (HPLC) grade, and diethyl ether were purchased from Wako Pure Chemicals (Osaka, Japan). Saline used for infusion was purchased from Otsuka Pharmaceutical Co. Ltd. (Tokyo, Japan) and pentobarbital was from Dainippon Pharmaceutical (Osaka, Japan). Deionized and distilled water, successively purified by the Milli-Q system (Millipore, Bedford, MA), was used for reagent preparation. All other chemicals were analytical grade and commercially available.

Cell Culture and Transfection.

The parent CHO-K1 cells (JCRB9018) purchased from Japanese Health Science Research Resources Bank (Osaka, Japan) were cultured in Dulbecco's modified Eagle's medium/F-12 (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum, penicillin (100 U/ml), and streptomycin (100 μg/ml) at 37°C with 5% CO₂ and 95% humidity. The cDNA fragment was amplified by reverse transcription-polymerase chain reaction with the total RNA of rat liver using an RNA-PCR kit (Takara Shuzo, Osaka, Japan). For polymerase chain reaction, the primers were prepared based on the sequence reported previously (Grundemann et al., 1994). The sequences of the forward and reverse primers were 5′-TTTCGCCTGCTGCAGGGCATGGTC-3′ and 5′-TGATGAAGGCCGCGGGGAATTCCA-3′, respectively. The amplified fragment was used for library screening. For cDNA screening, the cDNA libraries were constructed from Sprague-Dawley rat liver using the SuperScript Choice system (Invitrogen). Briefly, poly(A)⁺ RNA was fractionated by sucrose density gradient and fractions containing RNA at approximately 2 kilobase pairs were used as a template for cDNA construction using reverse transcriptase and Oligo(dT) primer. After two rounds of screening, single positive colonies were isolated. After coinfection with the M13 helper phage (ExAssist; Stratagene, La Jolla, CA), the cDNA was excised in a pBluescript II SK(−) plasmid and rescued by SOLO strain. The insert was subcloned into the vector pcDNA3.1 (Invitrogen) cut with the restriction enzymes Xho-I and Sal-I. The nucleotide sequence of the subcloned cDNA insert was determined using a fluorescence DNA sequencer 377 (Applied Biosystems, Foster City, CA). The rOct1/pcDNA3.1 so prepared and pcDNA3.1 alone were transfected into CHO-K1 cells with LipofectAMINE (Invitrogen) according to the manufacturer's protocol, and stably transfected cells were selected by G418 (0.6 mg/ml) (Invitrogen) in the culture medium. Then, 7 to 14 days after transfection, several colonies were selected both by Northern blot analysis and by uptake of tetraethylammonium (TEA), a typical substrate of the OCT family. After selection, the cells were maintained by adding G418 (0.4 mg/ml) to the culture medium.

Uptake Study.

Uptake of biguanides by the cells was measured using 12-well cell culture dishes (BD Biosciences, Franklin Lakes, NJ). Cells were seeded onto plates at a density of 1.0 × 10⁵ cells/well. The incubation buffer for the uptake experiments contained 145 mM NaCl, 3 mM KCl, 1 mM CaCl₂, 0.5 mM MgCl₂, 5 mMd-glucose, and 5 mM HEPES pH 7.4, and the uptake study was performed at 37°C. After removing the culture medium, cells were washed twice with incubation buffer and preincubated for 5 min. Then, substrate solution was added to the dish and, at designated times, uptake was terminated by adding 1 ml of ice-cold buffer. The medium was rapidly aspirated and cells were rinsed twice. Cells were lysed with 300 μl/well of 0.5% SDS at 37°C for 1 h, and 200 μl was used for HPLC determination after deproteination with 3 times volume of acetonitrile. The remaining 25 μl of cell lysate was used to determine the protein concentration by the method of Lowry (1951), with bovine serum albumin as a standard. Uptake was given as the ratio of the amount of ligand associated with the cells divided by the medium concentration. Because the initial velocity of uptake of metformin, buformin, and phenformin was linear up to 5 min, uptake was terminated at 5 min to estimate K_m andV_max by the following equation:V/S=Vmax/(Km+S) Equation 1where V is the initial uptake velocity andS is the substrate concentration.

Tissue Distribution of Metformin.

Female Oct1(−/−) and Oct1(+/+) FVB mice (12–16 weeks old) were used in the present study (Jonker et al., 2001). Mice were housed at a room temperature of 24 ± 1°C and allowed food and water ad libitum. After anesthesia with intraperitoneal sodium pentobarbital (50 mg/kg), the bladder was catheterized for urine collection. Metformin (5.0 mg/kg) was administered via the tail vein in a volume of 0.1 ml/10 g of body weight. Blood samples were collected by cardiac puncture from an angular vein at 1, 5, and 10 min after administration. Urine was collected by washing the bladder with 0.2 ml of saline. Mice were killed and liver, kidney, small intestine, and colon were removed immediately. Each segment of the small intestine and colon was rapidly washed twice with ice-cold saline to remove residues and then dried with filter paper. The duodenal segment was obtained from the first 5-cm portion of the intestine closest to the stomach. The ileum was obtained from the final 5 cm of the intestine, just proximal to the cecum. The middle 10 cm of the rest of the segment was designated as the jejunum. The colon was a segment 3 cm in length immediately after the cecum. All the tissues were weighed, and portions were then homogenized with 4 volumes of PBS. Both the homogenized tissues and blood samples were deproteinized with acetonitrile as described above and evaporated to dryness. Pellets were dissolved in 200 μl of water for HPLC analysis. In the other groups of mice, urine was collected up to 1 h at 20-min intervals. The systemic clearance (CL_total) and renal clearance (CL_renal) were estimated, respectively, from the following equations:CLtotal=Dose/AUC(0–∞) Equation 2CLrenal=Xurine(0–60)/AUC(0–60) Equation 3where AUC_(0–∞) and AUC_(0–60) are the area under plasma concentration-time curve up to infinity and 60 min, respectively, and these values were calculated by integration of the biexponential equation, which was obtained by fitting to the plasma concentration profile. X_{urine(0–60)} is the cumulative urinary excretion up to 60 min.

Coinfusion of Metformin with Cimetidine.

Male Sprague-Dawley rats (8 weeks old, 250–280 g of body weight; Charles River Japan, Kanagawa, Japan) were anesthetized with diethyl ether and both the femoral vein and bladder were catheterized with polyethylene tubing for infusion and urine collection, respectively. Infusion was performed by a basic syringe pump (Harvard Apparatus, Holliston, MA). Metformin dissolved in saline was administered at a loading dose of 0.5 mg/kg of body weight, followed by constant infusion at 0.9 mg/kg/h for 4 h. Cimetidine was administered at the loading dose of 24 mg/kg and an infusion rate of 50 mg/kg/h. The administered volume of saline was 8.0 ml/kg/h. Blood samples were collected from the cervical vein at 60, 120, 180, and 240 min. Urine was collected by washing the bladder with 0.5 ml of saline at 1-h intervals. Blood samples, deproteinized with 4 volumes of acetonitrile, and urine were subjected to HPLC.

HPLC Analysis.

The HPLC system involved a model L-7100 pump and a model L-7400 UV monitor (Hitachi, Tokyo, Japan) with 300- × 3.9-mm i.d. C₁₈ μBondapak (10 μm) purchased from Waters (Milford, MA). The mobile phase consisted of 0.01 M phosphate buffer (pH 6.5)/acetonitrile (30:70) at the flow rate of 1 ml/min. The wavelength of the UV detection was at 236 nm. The volume used for HPLC was 50 μl.

Statistical Analysis.

Statistical analysis was performed by Student's t test to identify significant differences between various treatment groups.

Results

Uptake of Biguanides in rOct1-Transfected CHO Cells.

The time profile for the uptake of metformin (100 μM), buformin (10 μM), and phenformin (1.0 μM) by rOct1- and vector-transfected CHO cells is shown in Fig. 1, A, C, and E, respectively. Uptake was much higher in rOct1-transfected cells than in vector-transfected cells (Fig. 1, A, C, and E). Because the uptake of each compound was linear up to 5 min (Fig. 1, A, C, and E), the initial uptake was assessed as the uptake for 5 min at various substrate concentrations (Fig. 1, B, D, and F). Saturation of uptake was observed in rOct1-transfected cells, and an Eadie-Hofstee plot revealed a single saturable component for each compound (Fig. 1, B, D, and F). The kinetic parameters for the uptake of metformin, buformin, and phenformin are shown in Table 1. Phenformin has a much higher affinity, but a lower capacity, for rOct1 than metformin, and the intrinsic clearance (V_max/K_m) of phenformin was ∼2 to 3 times higher than that of metformin (Table1). TEA, used in the present study as a positive control, exhibitedK_m andV_max values of 84 μM and 320 pmol/min/mg of protein, respectively. ThisK_m was not very different from that found in previous reports (47–100 μM; Urakami et al., 1998; Dresser et al., 2000).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Uptake of metformin in rOct1-transfected (open symbols) and pcDNA3.1-transfected (closed symbols) CHO cells. The time profile (A, C, and E) and concentration dependence (B, D, and F) of the uptake of metformin (A and B), buformin (C and D), and phenformin (E and F) are shown. The inset in B, D, and F represents Eadie-Hofstee plots. Results are mean ± S.E. (n = 3). Error bars for some points lie within the symbols.

Plasma Concentration and Tissue Distribution of Metformin in Oct1(−/−) and Oct1(+/+) Mice.

The time profiles of the plasma concentration and tissue distribution of metformin at 10 min in Oct1(−/−) and Oct1(+/+) mice are shown in Fig.2. The plasma concentration of metformin at 5 and 10 min was almost comparable for Oct1(−/−) and Oct1(+/+) mice, whereas a small (less than 2-fold) but significant difference was observed at 1 min (Fig. 2A). The difference at these earlier time points suggests that there is a smaller distribution volume in Oct1(−/−) mice than in Oct1(+/+) mice. The liver concentration of metformin was approximately 30 times higher in Oct1(+/+) mice than Oct(−/−) mice (Fig. 2B). The amount of metformin associated with the liver at 10 min was 24.0 ± 2.4 and 0.690 ± 0.065% of dose in Oct1(+/+) and Oct1(−/−) mice, respectively. A 3- to 7-fold higher distribution in Oct1(+/+) than Oct(−/−) mice was also observed in the duodenum, jejunum, and ileum (Fig. 2B). On the other hand, the distribution of metformin to the kidney was almost identical for the two types of mice and the difference in the distribution to the colon was minimal (Fig. 2B). The amount of metformin associated with the kidney was 34.2 ± 6.7 and 32.1 ± 7.0% of dose in Oct1(+/+) and Oct1(−/−) mice, respectively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Plasma concentration profile (A) and tissue distribution (B) of metformin in Oct1(−/−) (open symbols) and Oct1(+/+) mice (closed symbols). Metformin was injected intravenously at 5.0 mg/kg, and tissues were excised at 10 min. Results are mean ± S.E. (n = 3–4). ∗, significant difference (p < 0.05).

Urinary Excretion of Metformin in Oct1(−/−) and Oct1(+/+) Mice.

The urinary excretion of metformin was chased in Oct1(−/−) and Oct1(+/+) mice (Table 2). In all, 55 to 70% of the dose was recovered in the urine up to 60 min after i.v. administration. There was only a minimal difference in both the CL_renal and CL_total between Oct1(−/−) and Oct1(+/+) mice (Table 2).

Coinfusion of Metformin with Cimetidine.

To demonstrate the cation-specific renal excretion of metformin in rodents, the effect of coadministration of cimetidine on the urinary excretion of metformin was investigated in rats (Fig. 3). The plasma concentration of metformin was much higher in the presence of cimetidine than in its absence (Fig. 3A). On the other hand, the urinary excretion of metformin was significantly reduced in the presence of cimetidine (Fig. 3B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Plasma concentration (A) and urinary excretion (B) of metformin in the presence and absence of cimetidine in rats. Metformin was infused intravenously at 0.9 mg/kg/h after a loading dose of 0.5 mg/kg with (closed symbols) or without (open symbols) coinfusion of cimetidine at 50 mg/kg/h after a loading dose of 24 mg/kg. Results are mean ± S.E. (n = 3–4). ∗, significant difference (p < 0.05).

Discussion

Despite the widespread use of the biguanide metformin in the treatment of hyperglycemia, the mechanism(s) underlying its disposition has not yet been clarified. Renal excretion is the major elimination pathway for metformin in humans and is much higher than the glomerular filtration rate (Pentikainen et al., 1979), suggesting involvement of tubular secretion systems, although the detailed mechanism has not yet been determined. Metformin exerts its antidiabetic effects, at least partly, via a direct and/or indirect inhibitory effect on complex 1 of the mitochondrial respiratory chain in hepatocytes (El-Mir et al., 2000). Considering that biguanide compounds are positively charged at physiological pH, there may be membrane transport system(s) for these compounds in the liver. Thus, it is important to clarify their disposition mechanism(s) to understand the factors that may affect the pharmacokinetics and pharmacodynamics of biguanides.

Gene knockout mice for Oct1, which accepts a variety of types of organic cations as substrates, have recently been established (Jonker et al., 2001). Considering that biguanides are also cationic compounds, the involvement of Oct1 in the disposition of biguanides was investigated in the present study using rOct1-transfected cells and knockout mice as the first step to clarify the pharmacokinetic mechanism(s). The increase in the saturable uptake of the three biguanides by rOct1 transfection suggests that these compounds are substrates of rOct1 (Fig. 1; Table 1). The distribution of metformin to the liver and three segments of the small intestine is much lower in Oct1(−/−) mice, compared with Oct1(+/+) mice, suggesting that Oct1 may be involved in the distribution to these tissues (Fig. 2). In the liver and duodenum of Oct1(−/−) mice, theK_p values of metformin were 0.13 and 0.14, respectively, these values being comparable with the extracellular volume of these tissues (Tsuji et al., 1983). Because the plasma concentration at this time point (5–10 μg/ml corresponding to 30–60 μM) is much lower than the K_mof metformin for Oct1, Oct1-mediated transport may not be saturated. Thus, the distribution of metformin to these tissues seems to be mainly governed by Oct1 at least under linear conditions. These results are compatible with previous observations (Grundemann et al., 1994;Schweifer and Barlow, 1996; Jonker et al., 2001) and suggest that Oct1 localization in the basolateral membrane in the liver may explain its possible function as an uptake mechanism for organic cations (Meyer-Wentrup et al., 1998; Urakami et al., 1998; Karbach et al., 2000). Although the localization of Oct1 in the small intestine has not been clarified yet, the present findings suggest that it has a possible role in the basolateral uptake of biguanides, although further studies are needed to demonstrate this hypothesis.

On the other hand, neither the distribution of metformin to the kidney nor its renal excretion showed any clear difference between Oct1 (−/−) and Oct1 (+/+) mice (Fig. 2B; Table 2), suggesting that Oct1 is not the major transporter involved in the renal uptake of metformin. Because the intrinsic transport activity (V_max/K_m) of metformin by rOct1 is comparable with that of TEA, and the renal uptake of TEA is very similar to the blood flow rate, the renal uptake of metformin may also be limited by the blood flow, suggesting that the change in intrinsic renal uptake due to the lack of Oct1 may result only in a minimal difference in renal uptake in vivo. Therefore, the result obtained in vivo does not fully exclude the possible role of Oct1 in the renal uptake of metformin. Nevertheless, cimetidine inhibits the urinary excretion of metformin and increases its systemic exposure in rats (Fig. 3). A similar drug-drug interaction has also been reported in humans (Somogyi and Gugler, 1987), suggesting that a cation-specific mechanism(s) is involved in the urinary excretion of metformin, irrespective of the species. However, the difference between the dose of cimetidine and metformin was more than 50-fold in rats (Fig. 3), whereas the difference in humans was less than 2-fold (Somogyi and Gugler, 1987). Therefore, the mechanism underlying this drug interaction both in humans and rats should be discussed once additional studies have been performed. rOct2 and rOct3 are also expressed in kidney (Sugawara-Yokoo et al., 2000; Wu et al., 2000) and may be involved in the uptake and/or secretion of cationic compounds. rOct2 is also expressed in the basolateral side of the kidney (Karbach et al., 2000; Sugawara-Yokoo et al., 2000), although the substrate recognition of rOct2 has not yet been clearly distinguished from that of rOct1 (Urakami et al., 1998; Arndt el al., 2001). In a recent study using Oct3 (−/−) mice, no difference in the renal disposition of 1-methyl-4-phenylpyridinium was observed (Zwart et al., 2001). Based on this information, Oct2 is the most likely candidate for the control of the renal excretion of metformin, and additional studies are now being performed in our laboratory to test this hypothesis.

The present findings highlight the importance of Oct1 as the mechanism for the hepatic distribution of metformin. Its pharmacological actions include a reduction in glucose production and inhibition of the mitochondrial respiratory chain in hepatocytes (El-Mir et al., 2000;Hundal et al., 2000; Owen et al., 2000). The existence of the uptake system for metformin may be compatible with the previous finding that a lower metformin concentration is required to inhibit the respiratory chain in isolated hepatocytes than in isolated mitochondria (Owen et al., 2000). However, the metformin concentration (1–5 mM) that inhibits oxygen consumption in rat hepatocytes is higher than theK_m for rOct1 (Table 2), suggesting that rOct1 might be, at least partly, saturated at such an effective concentration. Therefore, further studies are required to clarify the functional relationship between the transport of biguanides by rOct1 and their pharmacological action in hepatocytes.