Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

HSR-903 (chemical structure in Fig. 1) is a new quinolone antibacterial agent with potent antibacterial activity (Takahashi et al., 1997). It is well distributed to tissues and has low toxicity (Murata et al., 1995). Conventional quinolone antibacterial agents induce symptoms of central nervous system toxicity, such as convulsion, when coadministered with fenbufen (Christ, 1990), and we showed that the mechanism of the central nervous system excitation was based on the displacement of -aminobutyric acid from its receptors (Tsuji et al., 1988a,b). In contrast, new quinolone antibacterial agents are scarcely distributed to the brain or the brain interstitial fluid (Tsuji et al., 1988a; Ooie et al., 1996a, 1997). HSR-903 is relatively lipid-soluble as compared with other quinolones, having an octanol-buffer partition coefficient of 2.58 (Murata et al., 1998), and it is well distributed in the body, with a high-distribution volume of 6.4 l/kg in rats (Murata et al., 1998). Nevertheless, the brain accumulation of HSR-903 is low (Murata et al., 1995), suggesting a low permeability of the blood-brain barrier (BBB) to HSR-903.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Chemical structure of [14C]HSR-903. * shows the labeled position.

P-Glycoprotein (P-gp), which transports various anticancer drugs out of multidrug-resistant (MDR) tumor cells, is expressed in brain capillary endothelial cells (BCECs; Cordon-Cardo et al., 1989; Sugawara et al., 1988). In addition, Thiebaut et al. (1987, 1989) reported that P-gp is present in normal human tissues, including brain capillaries. We and others have demonstrated that the low brain distribution of the anticancer drugs vincristine and doxorubicin and the immunosuppressive agent cyclosporin A can be ascribed to active brain-to-blood efflux via BCECs by P-gp present at the luminal surface of BCECs (Tsuji et al., 1992, 1993; Tatsuta et al., 1992; Hegmann et al., 1992; Sakata et al., 1994; Ohnishi et al., 1995; Tamai and Tsuji, 1996; Tsuji and Tamai, 1997).

Recently, it was reported that quinolone antibacterial agents such as ciprofloxacin, norfloxacin, pefloxacin, and sparfloxacin, are actively secreted across the apical membrane of human intestinal epithelial Caco-2 cells (Griffiths et al., 1993, 1994). Moreover, sparfloxacin secretion across Caco-2 cells was inhibited by verapamil, an MDR-reversing agent (Cormet et al., 1995). These previous observations suggest that at least some of the new quinolones are transported by P-gp. Therefore, we suspected that the low distribution of HSR-903 to the brain might be at least partly due to active efflux via P-gp.

In the present study, we studied the brain distribution of HSR-903 using various in vivo and in vitro methods to clarify the mechanism(s) underlying the low permeation of HSR-903 into the brain, focusing on brain-to-blood transport.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. HSR-903, (S)-()-5-amino-7-(7-amino-5-azaspiro[2.4]hept-5-yl)-1-cyclopropyl-6-fluoro-1,4-dihydro-8-methyl-[2.4]4-oxoquinoline-3-carboxylic acid methanesulfonate, [¹⁴C]HSR-903 (specific activity, 256 kBq/mg base, Fig. 1), and other quinolone derivatives were synthesized or purified by Hokuriku Seiyaku Co., Ltd. (Fukui, Japan). [³H]Glutamic acid and [³H]cyclosporin A were purchased from Amersham Co., Ltd. (Tokyo, Japan). [³H]3-O-Methylglucose, [³H]- or [¹⁴C]sucrose, and [³H]- or [¹⁴C]inulin were purchased from New England Nuclear (Boston, MA). All other reagents were commercial products of reagent grade.

In Vivo Brain Distribution Study. Male Sprague-Dawley rats (210-260 g) were purchased from Charles River Japan, Inc. (Kanagawa, Japan). FVB/NJ (20-35 g) and mdr1a gene-deficient mice (25-30 g) were purchased from Jackson Laboratories (Bar Harbor, ME) and Taconic Farms, Inc. (Germantown, NY), respectively. The animal study was performed according to the Guidelines for the Care and Use of Laboratory Animals at the Takara-machi Campus of Kanazawa University and was approved by the Committee of Ethics of Animal Experimentation of Kanazawa University, Takara-machi Campus. Brain and plasma concentrations of unchanged HSR-903 in rats were determined after i.v. bolus injection at 4.2 mg/kg and simultaneous infusion at 2.1 mg/h/kg. At 3 h after dosing, the rats under ether anesthesia were sacrificed by exsanguination from the abdominal aorta and dissected immediately. The concentration of unchanged HSR-903 was determined by HPLC. In the case of the distribution study in mice, brain and plasma concentrations of HSR-903 were determined after single i.v. administration of [¹⁴C]HSR-903 at a dose of 13 mg/kg. At 2 h after dosing, the mice were sacrificed under ether anesthesia and dissected immediately. The whole brain was isolated, weighed, and solubilized in Solvable (New England Nuclear, Boston, MA) at 50°C for 3 h. The associated radioactivity was measured by liquid scintillation counting.

Transport Study in Cultured Bovine BCECs. BCECs were isolated from cerebral gray matter of bovine brains by the method of Audus and Borchardt (1986) with minor modifications. Details of the procedures for the preparation and the cell culture were given in a previous report (Terasaki et al., 1991). The isolated BCECs were cultured at 37°C under 95% air and 5% CO₂. Transport experiments were performed when the cells had reached confluence in 10 to 12 days. Luminal uptake of ³H- and ¹⁴C-labeled compounds by cultured BCECs grown on the dishes was measured by the method described previously (Terasaki et al., 1991). Briefly, cultured cells were washed with incubation solution (122 mM NaCl, 3 mM KCl, 1.4 mM CaCl₂, 25 mM NaHCO₃, 1.2 mM MgSO₄, 10 mM D-glucose, 10 mM HEPES, and 0.1% bovine serum albumin, pH 7.4, 290 mOsmol) at 37°C. To prepare NaHCO₃-free medium, NaHCO₃ was replaced with NaCl. Uptake was initiated by adding 250 µl of incubation solution containing [³H]sucrose and [¹⁴C]HSR-903 to the cells. [³H]Sucrose was used as the extracellular marker. To terminate the transport reaction, cells were washed three times with 1 ml of ice-cold incubation solution at a designated time. The cells were solubilized with 1 N NaOH, and the radioactivity was measured. Protein content in cultured cells was measured by the method of Lowry et al. (1951) using bovine serum albumin as a standard. Uptake was expressed as cell per medium (C/M) ratio (µl/mg protein) obtained by dividing the uptake amount by the concentration of [¹⁴C]HSR-903 or [³H]sucrose in the medium.

Transport Studies in K562 and MDR K562/ADM Cells. Human leukemic cells, K562 and K562/ADM, which were provided by Professor Tsuruo (University of Tokyo), were cultivated by the method described previously (Tsuji et al., 1993). The cultured cells were suspended in incubation solution as described above, and centrifuged at 700g for 5 min. The resultant pellets were suspended in incubation solution and used at a concentration of 1.0 × 10⁶ cells/tube. Drug uptake was initiated by adding 0.2 ml of test compound to the preincubated (37°C for 30 min) cell suspension (10⁶ cells/0.2 ml). The reaction was terminated by separating the cells from the medium by means of a centrifugal filtration technique (Schwarz et al., 1977) for 60 min to obtain steady-state uptake. The radioactivities of the supernatant and the cell pellet were determined and the net uptake of HSR-903 was calculated as cell per medium concentration (C/M) ratio (µl/10⁶ cells).

In Situ Brain Perfusion Study. Brain perfusion was performed by the method reported previously (Takasato et al., 1984). In brief, the rats were anesthetized and the right carotid artery was catheterized with polyethylene tubing (SP-10) filled with sodium heparin (100 IU/ml). The perfusate (bicarbonate-buffered physiological saline, 142 mM NaCl, 28 mM NaHCO₃, 4.2 mM KH₂PO₄, 1.7 mM CaSO₄, 1.0 mM MgSO₄, 6.0 mM D-glucose pH 7.4) containing [¹⁴C]HSR-903 and [³H]sucrose, which was used as the brain intravascular volume marker, was oxygenated for 3 min with 95% O₂ and 5% CO₂ and perfused through the catheter at the rate of 4.98 ml/min by an infusion pump (Harvard Apparatus, South Natick, MA). At the end of a 30-s perfusion, the rat was decapitated, and the right cerebral hemisphere was dissected from the perfused brain and weighed. The perfused right cerebral hemisphere was solubilized and the radioactivity was determined. In vivo BBB permeability (µl/min/g brain) was calculated as described previously (Tamai et al., 1995) after correcting for the remaining intravascular HSR-903 estimated from the apparent brain uptake of [³H]sucrose.

Brain Efflux Index (BEI) Study. The BEI study was performed by the method of Kakee et al. (1996). Rats were anesthetized and placed in a stereotaxic frame (Narishige Co., Tokyo, Japan). Part of the scalp was removed and a midline incision was performed to expose the bregma as a reference point on the skull. A small hole was drilled at the targeted region of the left cerebrum (0.2 mm anterior and 5.0 mm lateral to the bregma and 4.5 mm deep, parietal cortex area 2 region) to allow entry of an injection needle. Within 2 s, 0.4 µl of [¹⁴C]HSR-903 or another test compound was administered to the parietal cortex area 2 region.

Analytical Method. The concentrations of unchanged HSR-903 were determined by HPLC assay. Briefly, accurately weighed tissue (0.1 g) homogenized (Polytron, Kinematica, Switzerland) in 0.1 ml of of isotonic phosphate buffer (pH 7.0), or plasma (0.1 ml) mixed with 0.1 ml of the same buffer, was mixed well with 0.1 ml of 1 N NaOH and 3 ml of diethyl ether, then centrifuged at 3000 rpm for 5 min. The resultant aqueous layer was vigorously shaken with 0.5 ml of 1 M phosphate buffer (pH 7.0) and 6 ml of chloroform-isoamyl alcohol mixture (95:5, v/v) for 10 min. After centrifugation of the mixture at 3000 rpm for 10 min, a 5-ml aliquot of the organic layer was put into a glass tube and evaporated to dryness at 37°C under reduced pressure. The residue was dissolved in 0.5 ml of 0.1 M citrate buffer (pH 4.0)-acetonitrile (3:1, v/v) and an aliquot was subjected to HPLC with a TSKgel ODS-80T_M analytical column (4.6 mm × 15 cm, 5-µm particle size; Tosoh Co., Tokyo, Japan). The mobile phase was composed of 0.03 M ammonium phosphate buffer (pH 2.5)-acetonitrile (3:1, v/v). The flow rate was 1.2 ml/min and the eluate was monitored at 308 nm.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Brain Distribution and BBB Permeability of HSR-903 in Rats. To evaluate tissue distribution, the concentration of intact HSR-903 in various tissues was measured 3 h after i.v. bolus administration of HSR-903 to rats. The results are shown in Table 1 as tissue-to-plasma concentration ratio (Kp) in brain, lung, liver, and kidney. The values in lung, liver, and kidney are between 14 and 29, whereas that in brain is extremely low, 0.11, demonstrating very poor distribution to the brain compared with other tissues.

In Vivo Brain-to-Blood Transport of HSR-903 (BEI). The in vivo brain-to-blood efflux of HSR-903 was evaluated by means of the BEI method (Kakee et al., 1996) after intracerebral administration by comparing the BEI with those of [³H]3-O-methylglucose and [³H]glutamic acid as actively effluxed substrates and [¹⁴C]inulin as a nonactively effluxed substrate (Fig. 2). The recovered amounts of [³H]3-O-methylglucose and [³H]glutamic acid in brain decreased in a time-dependent manner, with elimination half-life values of 7.6 min and 30 min, respectively, whereas no significant time-dependent decrease in brain recovery of [³H]inulin was observed (Fig. 2a). The recovery of [¹⁴C]HSR-903 in brain after administration of a 200 µM solution decreased at a moderate rate with a half-life of about 3.6 h, which is longer than those of 3-O-methylglucose or L-glutamic acid, but significantly shorter than that of inulin. Furthermore, addition of unlabeled HSR-903 at the brain concentration of 8 mM, which was calculated from the concentration of HSR-903 in the injected solution and the dilution factor as reported by Kakee et al. (1996), reduced brain efflux by 34% (data not shown). Accordingly, HSR-903 appeared to be transported from the brain to blood in a saturable manner.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Time course of efflux of radioactivity from rat brain evaluated by the BEI method. a, [3H]3-O-methylglucose (0.4 nM). b, [3H]glutamic acid (0.8 µM). c, [14C]HSR-903 (200 µM). (), efflux of [3H]3-O-methylglucose, [3H]glutamic acid, and [14C]HSR-903, respectively. (), efflux of [14C]inulin. Each symbol and vertical bar represents the mean ± S.E. from three to five experiments.

Brain Distribution of HSR-903 in mdr1a Gene-Deficient Mice. To evaluate the involvement of the multidrug efflux transporter P-gp in HSR-903 handling at the BBB in vivo, the brain distribution of [¹⁴C]HSR-903 was measured in knockout mice lacking the mdr1a gene-encoded P-gp (mdr1a(/) mouse) and compared with that of normal mice [mdr1a(+/+)]. After a 13-mg/kg i.v. bolus administration of [¹⁴C]HSR-903 into each mouse, the brain-to-plasma concentration ratio (Kp,brain) was determined (Table 3). The Kp, brain in mdr1a(/) mice 2 h after administration was about 8 times higher than that in mdr1a(+/+) mice, whereas the plasma concentration in mdr1a(/) mice was comparable to that in mdr1a(+/+) mice.

View this table: [in this window] [in a new window]   TABLE 3 Comparison of brain distribution of [14C]HSR-903 between mdr1a(+/+) and mdr1a(/) mice

Uptake of HSR-903 by MDR K562/ADM Cells. To confirm the P-gp-mediated efflux of HSR-903, the uptakes of [¹⁴C]HSR-903 by P-gp-expressing MDR cell line K562/ADM and its parental cell line K562 were compared (Table 4). Steady-state uptake of [¹⁴C]HSR-903 by K562 cells (11.1 µl/10⁶ cells) was significantly higher than that by the MDR K562/ADM cells (4.1 µl/10⁶ cells). The uptake of [¹⁴C]HSR-903 by K562/ADM cells (9.6) was increased in the presence of cyclosporin A to a level comparable with that by K562 cells (11.1). ATP depletion by treatment with a metabolic inhibitor (dinitrophenol or sodium azide and sodium fluoride) resulted in an 80% increase of the uptake of HSR-903 by K562/ADM cells, whereas no significant change was observed in the uptake by K562 cells (Table 4). Moreover, the uptake of [¹⁴C]HSR-903 was increased significantly in the presence of 5 mM unlabeled HSR-903 in both K562/ADM (4-fold) and K562 cells (1.8-fold).

Time Course and Bicarbonate Ion Dependence of HSR-903 Uptake by BCECs. Figure 3 shows the effect of replacement of bicarbonate ions by gluconate ions on the steady-state uptake by BCECs. Steady-state uptake of HSR-903 in bicarbonate ion-free medium was larger than that in the bicarbonate ion-containing medium. In the presence of bicarbonate ions in the medium, the steady-state uptake of [¹⁴C]HSR-903 was significantly increased by the addition of unlabeled HSR-903 at the concentration of 5 mM, whereas the uptake was not significantly affected by unlabeled HSR-903 in the absence of bicarbonate ions (Fig. 3). These observations suggest that HSR-903 is transported out of the cells by a bicarbonate ion-dependent efflux mechanism.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Effect of HCO3- and concentration of HSR-903 on steady-state uptake of [14C]HSR-903 by primary cultured monolayers of BCECs. Each column and vertical bar represent the mean S.E. from four experiments. *p < .05 by Student's t test.

Effect of Quinolone Antibacterial Agents and Various Compounds on Steady-State Uptake of HSR-903 by BCECs. To determine the structural specificity of HSR-903 transport in BCECs, the effect of various quinolone antibacterial agents on steady-state uptake of [¹⁴C]HSR-903 was examined. Most quinolones tested increased the uptake of [¹⁴C]HSR-903, with grepafloxacin showing the greatest effect, whereas nalidixic acid decreased the uptake (Table 5). Interestingly, the increment of the uptake was different between HSR-903 (S-isomer) and its R-isomer, suggesting the involvement of a stereospecific efflux mechanism in the transport of HSR-903.

Effect of Various Quinolone Antibacterial Agents on Steady-State Uptake of HSR-903 by K562 and MDR K562/ADM Cells. To determine the structural specificity of HSR-903 transport in K562 and K562/ADM cells, the effect of various quinolone antibacterial agents on the steady-state uptake of [¹⁴C]HSR-903 was examined. Most of the quinolones tested increased the uptake of [¹⁴C]HSR-903 in both cell lines, and the greatest effect was shown by grepafloxacin. In contrast, nalidixic acid reduced the uptake (Table 7). Interestingly, the extents of increase in [¹⁴C]HSR-903 uptake by quinolone derivatives were very similar to those observed in BCECs, as shown in Table 5.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

New quinolone antibacterial agents are scarcely distributed to the brain (Ooie et al., 1996a, 1997). HSR-903, a newly synthesized quinolone antibacterial agent, exhibited high Kp values in the lung, liver, and kidney, whereas the brain showed a low value. Accordingly, the distribution of HSR-903 to the brain is apparently restricted in vivo compared with other tissues. We reported previously that the low apparent permeability of the BBB for cyclosporin A and doxorubicin could be ascribed to active efflux of the drugs from BCECs via P-gp present at the luminal surface of the cells (Tsuji et al., 1992, 1993; Tsuji and Tamai, 1997). In the present study, the efflux rate of [¹⁴C]HSR-903 from the brain was compared with that of [³H]3-O-methylglucose and [³H]glutamic acid, which are substrates of active efflux transport from the brain (Hutchison et al., 1985), by using the recently established BEI method (Kakee et al., 1996). The elimination of [³H]3-O-methylglucose and [³H]glutamic acid was rapid (elimination half-life, 7.6 and 30 min, respectively) and the recovery of [¹⁴C]inulin, a nonpermeable marker, was constant (about 80%). [¹⁴C]HSR-903 elimination from the brain was slower (elimination half-life, about 3.6 h, Fig. 2c) than that of 3-O-methylglucose or glutamic acid, but was significantly higher than that obtained for [¹⁴C]inulin as a nonefflux marker. This result suggests that HSR-903 is actively transported from the brain at a moderate rate.

To confirm the existence of an efflux transport system for HSR-903 at the BBB, an in vitro uptake study with primary cultured monolayers of BCECs was conducted. In the presence of bicarbonate ions in the medium, the steady-state uptake of [¹⁴C]HSR-903 was significantly increased in the presence of unlabeled HSR-903 at the concentration of 5 mM (Fig. 3). To demonstrate that the efflux system of HSR-903 observed in the cultured BCECs functions physiologically in vivo, we used an in situ brain perfusion technique. The brain uptake of [¹⁴C]HSR-903 increased at the concentrations of 20 and 50 mM (Table 2). In contrast, the permeability coefficient of [³H]sucrose, a brain intravascular volume marker, was not changed in the presence of HSR-903. This result indicates that addition of HSR-903 does not cause nonspecific damage to the BBB. This observed concentration dependence suggests that a saturable efflux system(s) restricts the brain distribution of HSR-903.

It has been reported that ciprofloxacin, norfloxacin, pefloxacin, and sparfloxacin are secreted actively from the basal to apical side of human intestinal epithelial Caco-2 cells (Griffiths et al., 1993, 1994). Moreover, sparfloxacin secretion across Caco-2 cells was inhibited by verapamil, an MDR-reversing agent (Cormet et al., 1995). These results led us to consider the possibility that the low distribution of HSR-903 to the brain can be ascribed to active efflux via the transporter P-gp.

To determine whether HSR-903 is transported by P-gp, the uptake of [¹⁴C]HSR-903 by MDR K562/ADM cells (Yanovich et al., 1989) was examined. The uptake of [¹⁴C]HSR-903 by K562/ADM cells was significantly lower than that by the parental K562 cells. The uptake by K562/ADM cells was significantly increased in the presence of cyclosporin A and became almost equal to that of K562 cells (Table 4). Furthermore, the uptake of [¹⁴C]HSR-903 by K562/ADM cells was significantly increased by 5 mM unlabeled HSR-903. Additionally, the uptake of [¹⁴C]HSR-903 was increased significantly by treatment with metabolic inhibitors only in the case of K562/ADM cells (Table 4). These results suggest that HSR-903 is a substrate of P-gp.

To demonstrate that the P-gp-mediated efflux of HSR-903 observed in K562/ADM cells functions physiologically, we compared the Kp,brain values between mdr1a(+/+) and mdr1a(/) mice (Schinkel et al., 1994). The Kp,brain in mdr1a(/) mice was about 8 times higher than that in mdr1a(+/+) mice, whereas the plasma radioactivity level in mdr1a(/) mice was comparable to that in mdr1a(+/+) mice (Table 3). Thus, both in vitro and in vivo data support the idea that P-gp participates in the efflux of HSR-903 at the BBB.

We also examined the involvement of other mechanisms in the poor brain distribution of HSR-903. As shown in Table 6, an organic anion (probenecid) increased the steady-state uptake of HSR-903 into BCECs. The uptake was also increased by sulfobromophthalein. Probenecid was reported to be an inhibitor of the multidrug transporter, multidrug resistance-associated protein (MRP) (Cass et al., 1989; Evers et al., 1996). Gollapudi et al. (1995) reported that difloxacin reverses multidrug resistance in HL-60/AR cells that overexpress MRP, but Flens et al. (1996) found that MRP1 is not present in human brain. As sulfobromophthalein was reported to be transported by an organic anion transporter in liver (Yamazaki et al., 1996) and HSR-903 was also suggested to be a substrate of MRP2 (Murata et al., 1998), the efflux of HSR-903 may proceed in part via an anion transport mechanism. Moreover, DIDS and SITS increased the steady-state uptake of [¹⁴C]HSR-903. These compounds are inhibitors of anion exchange transport such as Cl^-/HCO₃^- exchange (Orsenigo et al., 1992). We therefore examined the effect of replacing HCO₃^- in the medium with gluconate^-. Replacement of HCO₃^- with gluconate^- increased the uptake (Fig. 3). Accordingly, it seems that the efflux of HSR-903 is partly due to an HCO₃^--dependent anion exchange mechanism. The unlabeled HSR-903 and other quinolones except nalidixic acid increased the uptake of [¹⁴C]HSR-903, and the effect was stereospecific (Table 5). These findings indicate that several quinolone antibacterial agents, including HSR-903, are transported out of the brain via a common carrier system, which is sensitive to anionic compounds. Ooie et al. (1996b) reported that the new quinolone antibacterial agent fleroxacin was transported via an anion exchange mechanism in rat choroid plexus, and its transport was inhibited by organic anions such as probenecid and benzylpenicillin. Furthermore, we have demonstrated previously that the classical anion exchanger AE2 transports organic anions such as benzoic acid (Yabuuchi et al., 1998). AE2 is known to be present at the BCECs, but we do not yet know whether or not it transports HSR-903. Further studies are needed to identify the bicarbonate-sensitive transporter for HSR-903 at the BBB, and to establish whether other transporters also contribute to HSR-903 transport.

In conclusion, the results obtained in the present study demonstrate that the new quinolone antibacterial agent, HSR-903, is actively transported out of brain by P-gp present in the luminal membrane of BCECs, as well as by a bicarbonate-sensitive anion transporter(s). The poor distribution of this drug to the brain may be explained by the participation of multiple efflux mechanisms at the BBB.