Introduction

Top Abstract Introduction Materials Results Discussion References

The systemic clearance of many NQs is mainly by metabolism and urinary excretion; by contrast, the biliary clearance of quinolones such as GPFX and SPFX, which have recently been developed, is larger than their renal clearance (Matsunaga et al., 1991; Akiyama et al., 1995a). Furthermore, it was reported that there was a great difference in the liver-to-plasma concentration ratio (K_p) among NQs (Okezaki et al., 1988). Hepatic uptake is an important factor determining tissue distribution, the degree of biliary clearance and the metabolism of a drug. However, the uptake mechanism of NQs by hepatocytes has not been previously investigated.

A number of transport systems for the hepatic uptake of drugs and endogenous compounds have been reported (Inoue et al., 1982; Anwer and Hegner, 1978; Hagenbuch et al., 1990, 1991; Yamazaki et al., 1992a, 1993a, 1996; Meier, 1995). A Na⁺-coupled secondary active transport system for TCA and often conjugated bile acids in both rats and humans has recently been expressed in oocytes and cloned (Ananthanarayanan et al., 1994; Hagenbuch and Meier, 1994). The transport carrier that mediates the Na⁺-independent uptake of various non-bile acid organic anions such as DBSP has also been characterized and cloned (Blom et al., 1981; Uehara et al., 1983; Yamazaki et al., 1992b, 1993b; Jacquemin et al., 1991, 1994; Kullak-Ublick et al., 1995). However, multispecific transport systems are known to be involved in the hepatic uptake of cationic drugs (Meijer et al., 1990; Nakamura et al., 1994). The substrates for the transporters have been divided into two types based on their chemical structure, number of charges and lipophilicity (Meijer et al., 1990; Groothuis and Meijer, 1996). Expression cloning of the carrier protein for organic cations has been performed (Grundemann et al., 1994).

NQs are zwitterionic drugs with carboxylic acid and cationic amine groups that are dissociated at physiological pH (fig. 1). NQs have been reported to be recognized as cationic compounds by a transporter for reabsorption at the brush-border side of the kidney tubule (Okano et al., 1990), and NQs are known to affect the uptake of anionic and cationic compounds through the basolateral membrane in kidney cells (Ullrich et al., 1993). It has also been reported that NQ transport may be mediated by active transport systems involving absorption of NQ at the brush-border of the intestine (Iseki et al., 1992; Hirano et al., 1994) and uptake from the basolateral side of Caco-2 cells (Griffiths et al., 1993, 1994). Therefore, it is important to clarify what type of transport system mediates the hepatic uptake of NQs.

View larger version (K): [in this window] [in a new window]   Fig. 1.   Chemical structures of quinolone antibiotics.

	Methods and Materials

Top Abstract Introduction Materials Results Discussion References

Chemicals. GPFX (1.17 MBq/µmol, radiochemical purity 97.1%) and [³H]-cimetidine (814 MBq/µmol, 97.2%) were obtained from Amersham International (Buckinghamshire, UK). [³H]-Taurocholic acid (128 MBq/µmol, 98.5%) and [³H]-ouabain (759 MBq/µmol, 98.6%) were purchased from New England Nuclear Corp. (NEN, Boston, MA). [¹⁴C]-pravastatin (0.37 MBq/µmol) was kindly donated by Sankyo Company Ltd. (Tokyo, Japan, >95%). [³H]-Quinidine (555 MBq/µmol, 99%) was obtained from American Radiolabeled Chemicals Inc. (St. Louis, MO).

Cell preparation. Hepatocytes were isolated from male Sprague-Dawley rats by the procedure of Baur et al. (1975). After isolation, the hepatocytes were suspended (2 mg protein/ml) at 0°C in albumin-free Krebs-Henseleit buffer supplemented with 12.5 mM HEPES (pH 7.3). All studies were carried out in the presence of sodium except for the studies of the effect of sodium on uptake and the Na⁺-independent uptake of TCA. Cell viability was routinely checked by the trypan blue [0.4% (w/v)] exclusion test. We used more than 90% as a viability criterion, and the mean viability was 93.4 ± 0.4% (mean ± S.E. of 29 different preparations). Protein concentrations were determined by the method described by Bradford (1976), using the protein assay kit (Bio-Rad, Hercules, CA) with bovine serum albumin as a standard.

Uptake study. Uptake of [¹⁴C]-GPFX was initiated by adding the ligand solution (0.5 ml) into the cell suspension (0.5 ml, 2 mg protein/ml) preincubated at 37°C for 5 min. At a designated time, the reaction was terminated by separating the cells from the medium using a centrifugal filtration technique (Schwenk, 1980). Briefly, 200-µl aliquots were placed into centrifuge tubes containing 50 µl 2 N NaOH, covered by 100 µl of a mixture (density 1.015) of silicone and mineral oil. The samples were then centrifuged for 15 sec in a tabletop microfuge (10,000 × g, Beckman Instruments Inc., Fullerton, CA). The centrifugation pelleted the hepatocytes through the oil layer into the alkaline solution. After the cells had dissolved in the alkaline solution, the tube was sliced into two and each compartment was transferred into a scintillation vial. The alkaline compartment was neutralized with 50 µl 2 N HCl. Then, after addition of the scintillation cocktail (Atomlight, NEN) to the vials, the radioactivity in the medium and cells was determined using a liquid scintillation spectrophotometer (LS 6000SE, Beckman Instruments Inc.)

Effect of medium pH. After isolating and washing the cells, they were suspended in Krebs-Henseleit buffer at pH 7.3 (10 mg protein/ml). Uptake was initiated by the addition of Krebs-Henseleit buffer (0.8 ml), containing [¹⁴C]-GPFX that had been preincubated at 37°C for 5 min, to the cell suspension (0.2 ml). The pH of the Krebs-Henseleit buffer was adjusted by dropwise addition of 2 N NaOH or 2 N HCl, and was determined before and after the uptake study.

TLC analysis of [¹⁴C]-GPFX after incubation with isolated hepatocytes. The hepatocyte incubation mixture containing [¹⁴C]-GPFX was added to ice-cold buffer. After centrifuging the mixture at 200 × g for 2 min, 200 µl of the top layer (medium) was transferred into 600 µl acetonitrile. The whole cell pellet was added to 80% acetonitrile to precipitate proteins. Each aliquot of the extracts from the medium and cells was applied to a TLC plate (Kieselgel 60 F254, Merck, Darmstadt, Germany). Then, the plate was developed with chloroform:methanol:28% ammonia solution (7:3:0.5, v/v/v). The radioactive profiles on the TLC plate were analyzed using a Bio-Imaging analyzer (Bas2000, Fuji Film, Tokyo, Japan).

Estimation of kinetic parameters. The kinetic parameters for GPFX and LFLX uptake were calculated using the following equations:

(1)

(2)

where V₀ is the initial uptake rate of the drug (pmol/min/mg protein), S is the drug concentration in the medium (µM), K_m is the Michaelis-Menten constant (µM), V_max is the maximum uptake rate (pmol/min/mg protein) and P_dif is the nonspecific uptake clearance (µl/min/mg protein). The above equation was fitted to the uptake data sets by an iterative nonlinear least-squares method using a MULTI program (Yamaoka et al., 1981) to obtain estimates of the kinetic parameters. The input data were weighted as the reciprocal of the observed values and the Damping Gauss Newton method was used for the fitting algorithm.

(3)

(4)

Hepatic uptake study in vivo. Under ether anesthesia, GPFX was administered to male rats (Nihon Ikagaku, Tokyo, Japan), weighing approximately 250 to 300 g, via the femoral vein at a dose of 5 mg/kg/2 ml saline (13.9 µmol/kg). Blood samples were then collected from the femoral artery at designated times over 2 or 3 min with a heparinized syringe and a portion of liver was collected by biopsy at 30 sec or 1 min. The rats were killed at 2 or 3 min and the whole liver was excised immediately. A portion of the tissue was weighed and stored at 30°C until required for assay. Liver samples were added to nine volumes 75% methanol (w/w) and homogenized. An internal standard (OPC-17203, 100 ng) (fig. 1) was added to the homogenate (50 µl) and, after dilution with methanol (200 µl), samples were centrifuged in a tabletop microfuge. The resulting supernatants (10 µl) were subjected to HPLC. Plasma samples (25 µl) were obtained by centrifugation of blood and the internal standard (100 ng) added together with methanol (200 µl) to precipitate a protein. After centrifugation, the supernatants (10 µl) were subjected to HPLC. The HPLC conditions and calculation method were as described for the cell uptake study except that the mobile phase was acetonitrile:water:phosphoric acid (25:75:0.2, v/v/v).

(5)

(6)

(0t)

(7)

(0t)

Estimation of hepatic uptake clearance from in vitro data. Based on the kinetic parameters of GPFX obtained by the fitting procedure described, the PS_{influx, in vitro} (ml/min/kg rat) was calculated using the following equation:

(8)

where  = 1.25 · 10⁸ cells/g liver (Lin et al., 1980),  = 1.0 · 10⁶ cells/mg protein) and  = 44 g liver/kg rat (Sugita et al., 1982). The in vivo uptake CL_uptake^blood (ml/min/kg rat) for blood concentration was estimated from the in vitro uptake PS_{influx, in vitro} using the dispersion model (equation 9) (Roberts and Rowland, 1986):

(9)

where  = (1 + 4R_N · D_N)^1/2, R_N = f_B · PS_{influx, in vitro}/Q_h, D_N is the dispersion number (0.17) (Iwatsubo et al., 1996) and f_B is the unbound fraction of GPFX in blood (0.44) (Akiyama et al., 1995b) and is obtained by dividing plasma unbound fraction f_p (0.60) by blood-to-plasma concentration ratio R_B (1.37). Q_h is the hepatic blood flow in rats (40 ml/min/kg rat), and was obtained from the hepatic uptake clearance of [³H]-TCA when the TCA uptake process is assumed to be blood flow limited.

	Results

Top Abstract Introduction Materials Results Discussion References

Hepatic uptake of GPFX in vivo. The time profiles of TCA and GPFX concentrations in plasma and liver after its i.v. administration were analyzed kinetically. The values of CL_uptake^plasma of GPFX and TCA were calculated as 45 and 22 ml/min/kg, respectively, from the slope of the corresponding integration plot (fig. 2). The CL_uptake^blood values, further calculated by taking the corresponding R_B values [1.37 (Akiyama et al., 1995b) and 0.55 (M. Kono, H. Suzuki and Y. Sugiyama, unpublished data), respectively] into consideration, were 33 and 40 ml/min/kg, respectively. The CL_uptake^blood of GPFX (33 ml/min/kg) was, therefore, similar to the hepatic blood flow (40 ml/min/kg) estimated from the CL_uptake^blood of TCA, which is taken up by the liver in a blood flow limited manner.

View larger version (K): [in this window] [in a new window]   Fig. 2.   Time profiles of (a) radioactive concentrations in plasma () and liver () after a single i.v. administration of GPFX to rats and (b) the hepatic uptake of GPFX () or [3H]-TCA () as integration plots. Initial slopes of GPFX and TCA represent CLuptake 1.03 ml/min/g liver and 0.494 ml/min/g liver, respectively. Each plot and vertical bar represent the mean ± S.E. of three determinations.

Time profile of [¹⁴C]-GPFX uptake by isolated rat hepatocytes. As shown in figure 3, the hepatic uptake of GPFX by isolated rat hepatocytes was linear up to 1 min and reached equilibrium at 2 to 5 min. The cell-medium concentration ratio at equilibrium was calculated to be approximately 35, taking the intracellular volume (4.3 µl/mg protein) (Yamazaki et al., 1992b) into consideration. The determination of the initial uptake velocity of GPFX was estimated by measuring the total radioactivity, because TLC analysis indicated that the fraction of unchanged drug to total radioactivity was more than 90% both in the medium and cells at 1, 2 and 5 min. The initial uptake of GPFX was calculated from the difference between the uptakes at 15 and 45 sec.

View larger version (K): [in this window] [in a new window]   Fig. 3.   Time profiles of GPFX uptake (5 µM) by isolated rat hepatocytes and effects of temperature. Uptake of [14C]-GPFX was measured by incubating isolated rat hepatocytes in Krebs-Henseleit buffer (pH 7.3) containing [14C]-GPFX 5 µM at 37°C () or 0°C () after preincubation for 5 min. The uptake value means the cellular uptake divided by extracellular concentration. Each plot and vertical bar represent the mean ± S.E. of three determinations.

Concentration-dependence of the initial uptake of [¹⁴C]-GPFX and the effect of other NQs. The uptake clearance of GPFX declined as the concentration increased, indicating that the uptake possessed a saturable component. The kinetic parameters for equation 1 were as follows: K_m 173 ± 33 µM, V_max 6.96 ± 1.13 nmol/min/mg protein, P_dif 28.1 ± 4.4 µl/min/mg protein. The kinetics of GPFX uptake was also examined in the presence of LFLX (1 mM), and the result is shown as an Eadie-Hofstee plot (fig. 4a). Equation 3, derived assuming competitive inhibition, was simultaneously fitted to both sets of uptake data in the presence and absence of LFLX (1 mM). These fitted lines agreed well with the experimental data (fig. 4a). To clarify the manner of inhibition of LFLX, equation 4 also was simultaneously fitted to the uptake data sets for GPFX in the absence and presence of LFLX. Equation 4 was derived assuming noncompetitive inhibition. Akaike's information criterion (Akaike, 1974) values of 47 and 43 were obtained from the fitting of equations 3 and 4, respectively, indicating that competitive inhibition was statistically superior for describing the data. Also for the uptake of LFLX itself, saturable and nonsaturable components were observed (fig. 4b), as for GPFX uptake. The kinetic parameters for LFLX uptake were as follows: K_m 436 ± 70 µM, V_max 8.57 ± 1.26 nmol/min/mg protein, P_dif 13.7 ± 0.4 µl/min/mg protein. Comparison of the K_m value indicated that GPFX had a higher affinity than LFLX. Both equations 1 and 2 were fitted to the kinetic data, assuming the two different models consisted of a saturable component and a nonsaturable diffusion component (equation 1) or a high affinity component and a low affinity component (equation 2). In the case of GPFX, Akaike's information criterion values were, respectively, 32 and 22 for equation 1 and 2, and for LFLX they were 34 and 13. Therefore, equation 1 better fitted the kinetic data of both NQs, suggesting that the uptake consists of a saturable component and a nonsaturable diffusion component.

View larger version (K): [in this window] [in a new window]   Fig. 4.   Eadie-Hofstee plot of GPFX uptake (a) by isolated rat hepatocytes in the absence () and presence () of LFLX (1 mM) and (b) Eadie-Hofstee plot of LFLX uptake. Uptake of [14C]-GPFX was measured at concentrations of 5, 50, 100, 150, 200, 300, 500 and 1000 µM GPFX, in the presence or absence of 1 mM LFLX, and the LFLX uptake was done at 0.05, 0.3, 1, 2 and 3 mM LFLX. Uptake clearance (V0/S, PSinf) was calculated by dividing the initial uptake velocity (V0) by the GPFX concentration (S) in the medium. Each plot and vertical bar represents the mean ± S.E. of six determinations in two different preparations.

View larger version (K): [in this window] [in a new window]   Fig. 5.   Effects of LFLX on the GPFX uptake (5 µM) by the isolated rat hepatocytes. Uptake of [14C]-GPFX was measured in the presence of 0.3, 1, 3 and 5 mM LFLX or its absence. Each bar represents the mean ± S.E. of six determinations in two different preparations. * P < .05, ** P < .01 (significantly different from controls using Dunnett's test).

Effects of temperature, sodium, metabolic inhibitor, sulfhydryl-modifying reagent and anion exchanger inhibitor on the initial uptake of [¹⁴C]-GPFX. The initial uptake of [¹⁴C]-GPFX at 27 and 0°C was 64 and 10% of that at 37°C, respectively, and the ratio of the uptake at 27 and 37°C (Q₁₀) was 1.6 (fig. 6). The lack of effect after sodium replacement in the medium by choline demonstrated that the uptake was an Na⁺-independent process. GPFX uptake was reduced by the presence of metabolic inhibitors (FCCP and sodium azide) but not by the sulfhydryl-modifying reagent (PCMB) and anion exchanger inhibitor (DIDS).

View larger version (K): [in this window] [in a new window]   Fig. 6.   Effects of sodium, temperature and metabolic inhibitors on the GPFX (5 µM) uptake by isolated rat hepatocytes. Uptake of [14C]-GPFX was measured at 37, 27 or 0°C in the presence or absence of inhibitor. Each bar represents the mean ± S.E. of seven determinations in two different preparations. * P < .05, ** P < .01 (significantly different from controls using Dunnett's test).

Effect of medium pH on the initial uptake of [¹⁴C]-GPFX. At a low concentration (5 µM), [¹⁴C]-GPFX uptake did not change at lower values than pH 7.4, but was reduced with increasing pH, reaching 67% at pH 8.3 compared with that at pH 7.4 (fig. 7). However, at a high concentration (1000 µM) where carrier-mediated uptake was saturated, [¹⁴C]-GPFX uptake showed a minimal reduction with increasing pH.

View larger version (K): [in this window] [in a new window]   Fig. 7.   Effect of extracellular pH on uptake of GPFX (5 µM) by isolated rat hepatocytes. Uptake of [14C]-GPFX was measured by incubating isolated rat hepatocytes in Krebs-Henseleit buffer (pH 6.2-8.3) containing 5 µM () or 1 mM () [14C]-GPFX. Each plot and vertical bar represents the mean ± S.E. of six determinations in two different preparations. The dotted and dashed line represents the dissociated and protonized percentage for the carboxyl and amino group (pKa = 7.1 and 8.8) of GPFX, respectively. * P < .05 (significantly different from PSinf in 5 µM at pH 7.3 using Dunnett's test).

Effects of GPFX on the uptake of substrates for some known transporters. GPFX reduced the uptake of TCA both in the presence and absence of sodium in a concentration-dependent manner (fig. 8a). GPFX also inhibited the uptake of pravastatin and cimetidine that are, respectively, substrates for organic anion and organic cation transporters (fig. 8b). In addition, the uptake of ouabain, a neutral steroid was very effectively and completely inhibited by GPFX (fig. 8b). The uptake of the amphipathic organic cation, [³H]-quinidine was almost completely abolished at 200 µM unlabeled quinidine, falling to 4.1 ± 1.3% of the control group (mean ± S.E. of six determinations in two different preparations), suggesting that the greater part of this uptake is saturable. However, GPFX at a concentration of 1 mM, at which its own uptake is saturated, inhibited quinidine uptake by only 23.8 ± 5.1%.

View larger version (K): [in this window] [in a new window]   Fig. 8.   Effects of GPFX on the uptake of TCA (0.2 µM), pravastatin (7 µM), cimetidine (0.1 µM) and ouabain (0.2 µM) by isolated rat hepatocytes. [3H]-TCA uptake was determined in the presence and absence of sodium. Both Na+-dependent uptake ((a) ) (the uptake in the presence minus that in the absence of sodium ions) and Na+-independent uptake ((a) ) are plotted. The uptake of [14C]-pravastatin ((b) ), [3H]-cimetidine ((b) ) and [3H]-ouabain ((b) ) were determined in the presence of sodium. Each plot and vertical bar represents the mean ± S.E. of four determinations in two different experiments. * P < .05, ** P < .01 (significantly different from controls using Dunnett's test).

Effects of organic anions, organic cations and other compounds on [¹⁴C]-GPFX uptake. The bile acid TCA and the organic anion DBSP, pravastatin and ICG did not inhibit GPFX uptake at concentrations where their own uptake should be saturated (table 2). No effect of ouabain on GPFX uptake was observed (table 2). The organic cations cimetidine, d-tubocurarine and PAEB did not inhibit GPFX uptake, but the amphipathic organic cations quinidine, verapamil and vincristine did (table 2).

	Discussion

Top Abstract Introduction Materials Results Discussion References

GPFX and SPFX have higher hepatobiliary excretion rates among NQs that have recently been developed (Matsunaga et al., 1991; Akiyama et al., 1995a). The hepatic clearance of drugs is generally governed by three factors; hepatic blood-flow, Q_h, free fraction in blood, f_B, and overall hepatic intrinsic clearance CL_int^overall (Miyauchi et al., 1987; Yamazaki et al., 1996). However, the distribution of drugs between blood and liver cells does not always proceed under the assumption of rapid equilibrium. In this case, CL_int^overall is a hybrid parameter that is described by equation 10 (Miyauchi et al., 1987; Yamazaki et al., 1996), where PS_influx, PS_efflux and CL_int represent the influx clearance from blood to liver, the efflux clearance from liver to blood and the intrinsic biliary excretion and/or metabolic clearance, respectively.

(10)

Accordingly, the rate-limiting step in CL_int^overall depends on the relative values of PS_efflux and CL_int. For example, CL_int^overall is governed by only the influx ability (PS_influx) in the case where CL_int is much larger than PS_efflux (CL_int  PS_efflux).

The hepatic uptake of GPFX was estimated to be so effective in vivo that about 80% of GPFX was taken up during the first-pass through the liver (fig. 2); this was estimated by comparing hepatic uptake clearance with the hepatic blood-flow rate that was estimated from TCA uptake clearance (fig. 2b). To clarify the effective transport mechanism of GPFX, studies using isolated rat hepatocytes were carried out. The cell-to-medium concentration ratio at equilibrium was approximately 35 (intracellular space: 4.3 µl/mg protein) (Yamazaki et al., 1992b), which was comparable with the liver to unbound plasma concentration ratio (K_pu = approximately 20) (H. Sasabe, Y. Kato, T. Terasaki, A. Tsuji and Y. Sugiyama, unpublished data) estimated in vivo, suggesting that GPFX is concentratively taken up by the liver (fig. 3). Such uptake may be due to active transport and/or protein binding in liver cells. The uptake of GPFX is temperature and concentration dependent, and uptake was reduced to nearly 60% of controls by treatment with FCCP and sodium azide that are known to deplete cellular ATP (fig. 6). This indicates that part of the concentrative uptake is produced by carrier-mediated active transport. However, rotenone did not reduce the uptake of GPFX significantly (fig. 6). In our laboratory, systematic analysis of intracellular ATP content has been performed after treatment with FCCP and rotenone, by changing the treatment time and concentration of the metabolic inhibitor (Yamazaki et al., 1993a); the ATP content is rapidly reduced and the uptake of organic anion is concomitantly reduced after treatment with FCCP and rotenone. In our experiment, hepatocytes were treated with rotenone (30 µM) or FCCP (2 µM) for 5 min. In this situation, the intracellular ATP content declines to 20% (rotenone) and 6% (FCCP) of that in the control, and FCCP has a stronger effect than rotenone (Yamazaki et al., 1993a). Thus, the observation that GPFX uptake is not markedly reduced by rotenone might indicate that GPFX can be transported in the presence of a small amount of intracellular ATP.

Kinetic analysis showed that GPFX uptake consists of a saturable component (K_m 173 µM) and a nonspecific diffusion component (fig. 4). By comparing V_max/K_m with P_dif, the contribution of each component to the uptake clearance was calculated to be nearly 1:1 over the range of therapeutic plasma concentrations (<5 µM). The saturable component may be predominantly due to active transport, considering that the reduced uptake by metabolic inhibitors was roughly 40% of the controls. In rats, as shown in figure 2a, the plasma concentration of unbound GPFX was calculated to be less than 20 µM at 20 sec after bolus i.v. administration with the based on a f_p value of 0.6 (Akiyama et al., 1995b). This concentration was approximately 10% of its K_m (173 µM) for uptake, suggesting that the carrier-mediated transport may also be functioning in in vivo.

The hepatic uptake clearance of GPFX in vivo (CL_uptake^blood) was calculated, based on a mathematical model (equation 10), using the uptake clearance (PS_{influx, in vitro}) obtained from this in vitro study together with f_p, R_B and Q_h estimated from the hepatic uptake clearance of TCA the uptake of which is known to be almost blood flow limited (Iga and Klaassen, 1982). The calculated CL_uptake^blood (37 ml/min/kg) was close to that (33 ml/min/kg) obtained from integration plot analysis in vivo. Such a successful extrapolation from in vitro to in vivo indicates that the carrier-mediated uptake that was identified from the in vitro study using isolated hepatocytes, reflects the uptake in vivo. The value of hepatic blood flow determined from the uptake clearance of TCA, was 40 ml/min/kg which is smaller than the usual rat hepatic blood flow, about 60 ml/min/kg (Dedrick et al., 1973; Luts et al., 1977). Part of this difference might be due to the effect of ether anesthesia. Taking account of the fact that GPFX is so rapidly taken up into liver cells that the uptake is nearly blood flow limited (fig. 2), it is possible that the hepatic uptake of GPFX might be reduced by the anesthesia.

GPFX has a carboxyl group and a secondary amine in the piperazine ring with pKa values of 7.1 and 8.8, respectively (fig. 1). Because of this, the relationship between GPFX uptake and the pH of the medium was examined. At a concentration of 1 mM where carrier-mediated uptake is considered to be saturated, the change in pH did not significantly affect GPFX uptake, indicating that the uptake by nonspecific diffusion is relatively independent of the medium pH (fig. 7). At a lower concentration (5 µM), the reduced uptake at higher pH values seems to be due to a reduction in carrier-mediated uptake (fig. 7). These results suggest that the carrier-mediated uptake is not accelerated by dissociation of the carboxyl group, and that the hepatic uptake of GPFX is not probably mediated by the H⁺-antiport system that requires a H⁺-gradient from the inside to the outside of the cell as a driving-force and mediates the hepatic uptake of N-methylnicotinamide (Moseley et al., 1990).

There are several reports indicating that NQs can be recognized by several transport system in various body tissues. For example, reabsorption of OFLX by renal cells through the brush-border membrane was reported to be mediated by an H⁺-antiport system, as a cationic compound (Okano et al., 1990), although NQs inhibited the transport of cationic compounds N-methylnicotinamide, TEA and the uptake of anionic compound p-aminohippurate through the basolateral membrane of renal cells (Ullrich et al., 1993). The absorption of ENX through the brush-border membrane in intestinal cells was known to be mediated by an active transport system with the membrane potential as the driving force (Hirano et al., 1994). Moreover, an active transporter mediates the uptake of CPFX from the basolateral side of Caco-2 cells (Griffiths et al., 1993, 1994).

In our study, we investigated whether the GPFX uptake system is identical with known transport systems for conjugated bile acids, organic anions, organic cations and the neutral steroid ouabain. GPFX inhibited the Na⁺-dependent and Na⁺-independent transport of TCA (fig. 8a) and the uptake of pravastatin, cimetidine and ouabain (fig. 8b). At 200 µM, close to the K_m for GPFX uptake, GPFX reduced the TCA uptake to only 70 to 75% of controls, although at 1 mM (about six times the K_m), approximately 50% of the TCA uptake and 40% of the pravastatin uptake remained uninhibited. We previously reported that the contribution of passive diffusion was only 10% of the total uptake of TCA and pravastatin (Yamazaki et al., 1992a, 1993b). Therefore, 1 mM GPFX only partially inhibited the carrier-mediated uptake of TCA and pravastatin. In the case of cimetidine uptake, considering that the contribution of passive diffusion is approximately 25% (Nakamura et al., 1994), the carrier-mediated uptake of cimetidine seemed to be inhibited almost completely by 1 mM GPFX (fig. 8b). Ouabain uptake was also completely inhibited by GPFX (fig. 8b).

The inhibition of GPFX uptake by the substrates for these transporters was also examined. The concentrations of inhibitors were chosen so that the carrier-mediated uptake of the inhibitors would be saturated. The K_m for Na⁺-dependent and Na⁺-independent uptake of TCA was reported to be 15 and 57 µM, respectively (Anwer and Hegner, 1978). The K_m for DBSP and pravastatin uptake was reported to be 2 and 29 µM, respectively (Blom et al., 1981; Yamazaki et al., 1993b). Thus, the inhibition studies involving these compounds were carried out at concentrations of inhibitors ranging from 5 to 100 µM for DBSP and 20 to 500 µM for pravastatin. GPFX uptake was not inhibited by TCA, DBSP, pravastatin and ICG at concentrations higher than their K_m value (table 2), suggesting that the GPFX transport system is different from the transporters for bile acids and organic anions.

It is suggested that a transporter exists for comparatively hydrophilic monovalent cations and one for hydrophobic multivalent cations. The compounds for each transporter have been classified as type I and type II, i.e., cationic and aliphatic methylammonium compounds such as TEA, PAEB and cimetidine belong to type I, and lipophilic organic cations with an amino group in the cyclic structure such as d-tubocurarine belong to type II (Meijer et al., 1990; Groothuis et al., 1996). Moreover, investigations have been carried out to discover the driving-force and molecular weight of each transporter by the photoaffinity labeling technique (Mol et al., 1988, 1991, 1992; Muller et al., 1988). Inhibition of GPFX uptake by type I compounds, cimetidine and PAEB was not observed at high concentrations, 10 times higher than the K_m for their own uptake (table 2). Type II compounds (d-tubocurarine) did not inhibit GPFX uptake (table 2), indicating that the GPFX uptake system differs from the organic cation transporter.

Ouabain (neutral steroid) is known to be taken up by a carrier-mediated system into the liver and inhibits TCA uptake in a competitive manner (Okudaira et al., 1988). Ouabain uptake was completely inhibited by GPFX (fig. 8b), although ouabain did not inhibit GPFX uptake at concentrations approximately 15-fold higher than the K_m for its own uptake (table 2). Thus, GPFX and ouabain may not share the same transporter. Quinidine, verapamil and vincristine, which are amphipathic cations, inhibited GPFX uptake concentration dependently (table 2). To study the relationship between transport systems for GPFX and these compounds, the effect of GPFX on quinidine uptake was investigated. The hepatic uptake of [³H]-quinidine was completely abolished by unlabeled quinidine (200 µM), indicating the possibility that this uptake may be carrier-mediated. GPFX at a concentration of 1 mM, which can saturate GPFX uptake (six times the K_m), reduced the quinidine uptake to only 75% of the controls. If these two drugs share a transporter, quinidine uptake should be reduced by GPFX (1 mM) to approximately one-seventh (calculated from 1/(1 + I/K_m)). Therefore, the transporters for GPFX and quinidine may differ. These results based on mutual inhibition studies indicate that the transporter for GPFX uptake is different from the other transporters so far identified.

Very recently, an oatp has been reported to have a broad substrate specificity (including not only organic anions but also bile acids, organic cations and the neutral steroid, ouabain) (Bossyut et al., 1996). Moreover, it is not currently known if multiple forms of oatp or multiple transcripts originating in alternative splicing exist (Jacquemin et al., 1994). If oatp has multiple forms that have overlapping substrate specificities, it is reasonable for oatp to recognize a broad range of multiple substances. GPFX inhibited the transport of TCA, cimetidine and ouabain, and GPFX uptake was inhibited by amphipathic organic cations such as quinidine (fig. 8; table 2). These observations show the possibility that GPFX could be transported by a form of oatp. However, GPFX uptake was not reduced by pravastatin, DBSP and ouabain (table 2). These findings may be explained by the following hypothesis. GPFX is transported by multiple isoform of oatp although pravastatin, DBSP and ouabain are transported by a single isoform. If the isoform that recognizes pravastatin, DBSP and ouabain contributes only slightly to GPFX uptake, GPFX could potently inhibit the uptake of these compounds although they do not inhibit GPFX uptake, or at least only to a small extent. This possibility requires further investigation.

Although the transporter for GPFX appeared to be different from those for TCA, pravastatin, cimetidine and ouabain, GPFX inhibited the uptake of all these compounds. Such inhibition may be caused by the following mechanism: GPFX is a highly lipophilic compound so that GPFX may bind to the plasma membrane in close proximity to some transporters for bile acids, organic anions and cations. This may then bring about a change in the environment around the transporters and reduce subsequent transport. If such nonspecific binding to cell surface occurs, it would occur very rapidly. Therefore, it may be estimated from the extrapolated uptake value at zero-time by isolated hepatocytes. The nonspecific binding of GPFX was estimated to be 51.8 µl/mg protein for the time-profile of uptake shown in figure 3. This nonspecific binding of GPFX may be one of the mechanisms for the inhibitory effect of GPFX on the transport of various compounds. An analysis for a concentration dependence of such nonspecific binding was carried out using the data in figure 4a, and the distribution volume representing the nonspecific binding was found to be 49.4 to 55.9 µl/mg protein (mean data of four different experiments) over the concentration rage 5 to 1000 µM. Although the value of nonspecific binding amounts to approximately 10 times the intracellular volume, the nonspecific binding did not exhibit any change at different concentrations of GPFX, indicating that the nonspecific binding of GPFX to hepatocytes was not saturable.

In conclusion, the hepatic uptake of GPFX is by a Na⁺-independent and carrier-mediated active transport system, and the contribution of carrier-mediated uptake to the total uptake of GPFX is approximately 50% at therapeutic plasma concentrations (<5 µM). None of the transporters for bile acids, organic anions, organic cations or ouabain seems to be responsible for the hepatic uptake of GPFX. Successful extrapolation of in vivo hepatic uptake clearance from in vitro uptake data using isolated hepatocytes confirms that the carrier-mediated transport observed in vitro actually plays a role in the effective hepatic uptake of GPFX in vivo.