INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Many clinical cases, clinical tests, and studies involving tests with animals of the oxidative interaction between new quinolones (NQs) and other drugs have been reported. The toxicity induced by the inhibition of metabolism by caffeine (3, 21, 22) or theophylline (14, 19, 30, 31) and the remarkable reduction in the intestinal absorption of NQs by the interaction between NQs and the metal cations (Ag³⁺, Mg²⁺, etc.) in antacids or anti-peptic ulcer drugs (5, 18) have been reported. Furthermore, various symptoms of NQ-induced central nervous disorders were noted in clinical case reports. In particular, the tonic and clonic convulsions induced by NQs in the presence or absence of nonsteroidal anti-inflammatory drugs (NSAIDs) are the most serious disorders (1, 17, 20). Concomitant administration of NSAIDs and NQs is considered an important factor that induces a synergistic interaction that results in convulsions. In order to evaluate the neurotoxic effects induced by the interaction between NQs and NSAIDs, we carried out various in vivo and in vitro experiments. The effect of NSAIDs on the 50% effective dose for an NQ-induced occurrence of convulsions was examined on the basis of in vivo experiments with mice (12). Furthermore, the effects of NQs and/or NSAIDs on in vitro -aminobutyric acid type A (GABA_A) receptor binding of [³H]GABA and [³H]muscimol was investigated with synaptoneurosomes (7, 28, 29). These in vivo and in vitro experiments indicated that the NQ-induced neurotoxic effect was synergistically increased in the presence of NSAIDs. Recently, we investigated the relationship between the inhibitory effect of NQs on the GABA current response in Xenopus oocytes into which mouse brain mRNA was injected and the molecular structures of NQs. In this study, we could predict quantitatively the intensity of NQ-induced convulsions in vivo from the inhibitory effects of NQs on the GABA response in vitro (12).

However, the quantitative in vivo-in vitro relationship concerning the effects of NQs on the central nervous system in the presence or absence of NSAIDs has not been investigated. In this study, we developed a new toxicodynamic model to evaluate quantitatively the inhibitory effects of both drugs on the GABA current response using GABA_A receptor-expressed Xenopus oocytes and obtained various dynamic parameters. Furthermore, these parameters were used to simulate isobolograms for the concentrations of both drugs in the brains in the presence of various inhibitory ratios of the GABA_A receptor. We clarified the correlation between those isobolograms simulated in vitro and the isobolograms of the threshold concentration in the brain between the occurrence of convulsions and the lack of occurrence of convulsions caused by both drugs in vivo.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. Enoxacin (ENX) and ciprofloxacin (CPFX) were kindly supplied by Dainippon Pharmaceutical Co., Ltd., (Osaka, Japan) and Bayer Pharmaceutical Co., Ltd., (Tokyo, Japan), respectively. Felbinac (FLB) and flurbiprofen (FRP) were kindly supplied by Ledere Japan, Ltd., (Tokyo, Japan) and Kaken Pharmaceutical Co., Ltd., (Chiba, Japan), respectively, and ketoprofen (KTP) was purchased from Sigma Chemical Co. (St. Louis, Mo.). A sodium lauryl sulfate standard was obtained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). ¹³¹I-labeled human serum albumin (18.5 MBq) was used to estimate the brain tissue capillary volume ratio and was obtained from New England Nuclear (Boston, Mass.). It was used after purification by column chromatography (Sephadex G-25 gel; Pharmacia LKB Biotechnology, Uppsala, Sweden), and the free fraction of iodine was then calculated to be less than 0.3% after ultrafiltration (MPS-3 Centrifree micropartition system Amicon; W. R. Grace and Co., Beverly, Mass.). All other chemicals were commercially available and of analytical grade.

Animals. Male ddy mice weighing 18 to 20 g were purchased from Seak Yoshitomi (Fukuoka, Japan) and were used after they had been raised in a cage for more than 5 days. Adult female African clawed frogs (Xenopus leavis) were obtained from Hamamatsu Seibutsu Kyozai (Shizuoka, Japan). The Xenopus frogs were kept in an aquarium maintained at about 20°C.

Preparation of injectable solutions. ENX, FLB, FRP, and KTP were prepared as aqueous solutions of the sodium salt by adding an equimolar amount of NaOH. If the osmotic pressure of these solutions was lower than that of saline, they were made isotonic by adding NaCl. The pH of the injectable solutions used in the experiments was adjusted to neutrality.

Effects of NSAIDs on distribution of ENX in the brain. ENX (25 mg/kg of body weight) in the presence or absence of FLB (15 mg/kg), KTP (50 mg/kg), or FRP (30 mg/kg) was injected into the tail veins of mice. The total volume of the injectable solutions was 20 µl/g of body weight. The same volume of saline was injected with ENX in the case of the administration of NQs alone. After administration of the drugs, the mice were guillotined at 1, 3, 5, 10, 20, 40, or 60 min, and blood and brain tissue were collected. Sodium heparin was immediately added to the blood samples, and the samples were centrifuged at 1,620 × g for 5 min to obtain the plasma. The plasma and brain tissue were stored in a freezer at 20°C, and their concentrations were analyzed by high-performance liquid chromatography (HPLC).

Convulsion after infusion of ENX. ENX (3 to 137.5 mg/kg) was administered into the tail veins of mice in the presence or absence of the three NSAIDs (3 to 150 mg/kg), and the occurrence of clonic convulsions within 20 min was observed. In the case of the coadministration of two drugs, it took 1 min to administer each drug, and the NSAIDs were injected at an interval of 1 min after the administration of ENX. Moreover, the same procedure was performed by using each solvent as a control. The mice were guillotined immediately after the occurrence of a convulsion in the first 20 min after drug administration; on the other hand, mice that did not have convulsions were guillotined 20 min after injection of the drugs, and the blood and brain tissue were then collected. Sodium heparin was immediately added to the blood samples, and the samples were centrifuged at 1,620 × g for 5 min to obtain the plasma. The plasma and brain tissue were stored in a freezer at 20°C, and their concentrations were analyzed by HPLC.

Brain tissue and plasma protein binding assay. The binding of ENX to brain tissue in the presence or absence of NSAIDs (FLB, KTP or FRP) was examined by equilibration dialysis (23). Various concentrations of ENX (14.4, 28.8, and 57.6 µM), FLB (23.6, 47.6, and 94.4 µM), KTP (19.6, 39.2, and 78.6 µM), and FRP (20.0, 40.0, and 80.0 µM) were incorporated into the brain tissue homogenate.

(1)

Extraction procedure for determination of ENX and NSAID concentrations in plasma and brain tissue. The extraction procedure for the determination of the ENX and NSAID concentrations was performed by previously described methods (9, 10). Briefly, in order to extract ENX from plasma samples, 0.9 ml of 0.1 M phosphate buffer, 0.1 ml of 10 µg of CPFX per ml as an internal standard, and 5 ml of chloroform containing 1% ethyl chlorocarbonate (Wako Pure Chemical Industries Ltd.) were added to 0.1 ml of plasma, and the contents were shaken for 10 min and then centrifuged at 1,620 × g for 5 min. After evaporation of the 4-ml organic phase under reduced pressure, the residue was dissolved in 0.1 ml of methanol-0.05 M NaOH (2:1; vol/vol), and 20 µl was injected into the HPLC system. The cerebrum samples were homogenized with a fourfold volume of 0.1 M phosphate buffer. One hundred microliters of 10 µg of CPFX per ml as an internal standard and 5 ml of dichloromethane were added to 1 ml of homogenate. The samples were shaken for 10 min and centrifuged at 1,620 × g for 5 min. Then, 4 ml of 1 mM NaOH was added to 4 ml of the organic phase and the contents were shaken for 10 min. After centrifugation at 1,620 × g for 5 min, 3 ml of the aqueous phase was collected and treated in a manner similar to that described above for the plasma samples, except that 0.1 ml of 10 µg of CPFX per ml was added as an internal standard. The extraction of NSAIDs from the plasma and cerebrum samples was done as described above for ENX. FLB for KTP and FRP and FRP for FLB were used as internal standards.

Correction of concentration in brain tissue. The equation used for the correction for the concentration in brain tissue was as follows:
where C_BR, C_BR,obs, R, R_B, and C_p represent the real concentration of ENX or NSAIDs in brain tissue, the apparent concentration of ENX or NSAIDs in brain tissue, the cerebral intravascular volume, the ratio of concentration of ENX or NSAIDs in blood to that in plasma, and the concentration of ENX and NSAIDs in plasma, respectively. C_BR,obs incoporates the drug remaining in the cerebral intravascular spaces; therefore, C_BR was determined by subtracting the amount of drug calculated from both the ratio of the capillary volume occupied in the brain and the concentrations of the drugs in blood from C_BR,obs by the equation described above. Consequently, C_BR indicates the real drug concentration in the brain tissue. The R_B values for ENX, FLB, KTP, and FRP were approximately 1.

HPLC system. The HPLC system consisted of a liquid chromatograph (Shimadzu LC-9A; Shimadzu, Kyoto, Japan) and a spectrophotometric UV detector (Shimadzu SPD-6AV; Shimadzu) operated at 280 nm. The column was stainless steel (250 by 6 mm [inner diameter]) and was packed with Nucleosil 5C18 (Chemco, Osaka, Japan). The mobile phase was methanol-5 mM sodium dodecyl sulfate, and the pH was adjusted to 2.5 by adding phosphoric acid. The flow rate of the mobile phase was 0.8 ml/min. All analyses were performed at room temperature.

Extraction of mRNA from mouse brain. The mice were guillotined after dislocation of the neck, and the whole brain (cerebrum, cerebellum, and brain stem) was removed. Total RNA was extracted from the homogenized brain with a phenol-chloroform mixture containing guanidinium thiocyanate, sarcosyl, and mercaptoethanol by the method of Chomczynski and Sacchi (4). Poly(A)⁺ RNA was isolated by oligo(dT)-cellulose chromatography with an mRNA purification kit (Pharmacia LKB Biotechnology, Uppsala, Sweden). The mRNA was stored in a sterile aqueous solution buffer containing 1 mM EDTA and 10 mM Tris-HCl (pH 7.4) at 70°C.

Injection of mRNA into Xenopus oocytes. Injection of mouse brain mRNA into Xenopus oocytes was performed by the methods described previously (2, 8, 11, 12). Xenopus oocytes were anesthetized with crushed ice, and small pieces of the ovaries were removed surgically. Ovarian oocytes were treated with 1.0 mg of collagenase per ml in modified Barth's medium containing 88 mM NaCl, 1 mM KCl, 0.33 mM Ca(NO₃)₂, 0.41 mM CaCl₂, 0.82 mM MgSO₄, and 5 mM Tris-HCl at pH 7.4 for about 30 min at room temperature. After removal of the follicle cell layer, stage V to VI oocytes, which were 1.0 to 1.2 mm in diameter, were selected visually for injection of mRNA. These oocytes had a resting potential of about 40 mV. About 30 ng of mRNA from mouse brain was injected into each oocyte with a glass pipette with a tip diameter of 10 µm. The oocytes into which mRNA was injected were kept in a broth of modified Barth's medium at 20°C until they were used for electrophysiological measurements.

Dose dependency of effect of NSAIDs on ENX-induced inhibition of GABA current response. We measured the enhanced inhibitory effects on the GABA current response in the presence of NSAIDs by the method of Kawakami et al. (12). Briefly, the concentration dependency of the ENX-induced alteration of the GABA current response was measured in the presence of 1, 10, or 100 µM NSAIDs or in the absence of NSAIDs.

Electrophysiological measurement of GABA_A receptor response. The current response of Xenopus oocytes into which mouse brain mRNA was injected was obtained by a previously described method (8). Glass microelectrodes were placed into the oocytes, and the membrane potential was maintained at 60 mV. A microscopic electrode differential amplifier (DPZ-16; Daia Medical System Ltd., Tokyo, Japan) and a membrane-fixed amplifier (CEZ-1100; Kouden Industries Japan, Ltd., Tokyo, Japan) and a membrane-fixed amplifier (CEZ-100; Kouden Industries Japan, Ltd., Tokyo, Japan) were used to measure the current response.

Effects of NSAIDs on transcellular transport in MBEC4s. Cultured mouse brain capillary endothelial cells (MBEC4s) were used as described previously (15, 27). In brief, MBEC4 was maintained in Dulbecco's modified Eagle's medium (Nikken Bio Medical Laboratory, Kyoto, Japan) containing 10% fetal calf serum, 100 U of penicillin per ml, and 100 µg of streptomycin per ml. In the transport study, the number of cells was diluted to 4 × 10⁴ cells/ml, and the cells were examined with a Transwell apparatus (Costar, Cambridge, Mass.). One milliliter of the cell suspension was seeded in the cup, cultured in 5% CO₂-95% air for 3 days, and washed three times with buffer (141 mM NaCl, 4.0 mM KCl, 2.8 mM CaCl₂, 1.0 mM MgSO₄, 10 mM D-glucose, 10 mM HEPES [pH 7.4]) at 37°C. MBEC4s possess the same fundamental properties as brain capillary endothelial cells. The objective of the transport experiments with MBEC4s is to confirm whether the potentiated ENX-induced convulsions in the presence of NSAIDs was not due to the enhanced permeation of ENX across the blood-brain barrier by NSAIDs.

2

Toxicodynamic analysis. We constructed a newly developed dynamic model that considers the receptor binding and dissociation to evaluate the effect of ENX and/or NSAIDs on the GABA current response in GABA_A receptors. Mass balance equations for receptor binding and dissociation of ENX and NSAIDs were as follows.

(2)

(3)

(4)

where, [Q], [R], and [S] are the unbound concentration of ENX, the concentration of free GABA receptors, and the unbound concentration of NSAIDs, respectively; [QR], [SR], and [QSR] are the concentrations of the complex of Q and R, S and R, and Q, S, and R, respectively; K_i is the inhibitory constant of ENX for the GABA_A receptor; K_d is the dissociation constant of NSAIDs for the GABA_A receptor; K_i' is the dissociation constant between the NSAID-receptor complex and ENX; and  is a hill coefficient (the number of NSAIDs) concerning the interaction between NSAIDs and the GABA_A receptor. Moreover, it is assumed that the binding affinity to ENX increases in the presence of NSAIDs; that is, K_i is greater than K_i'. The total number of receptors is [R₀] and is expressed as follows.

(5)

Furthermore, the receptor occupancy (), that is, the inhibition ratio relative to that for the GABA_A receptor, is as follows:

(6)

Furthermore,  is expressed as follows by using various dynamic parameters and equations 1 to 5:

(7)

Therefore, the GABA response rate (R) is as follows.

(8)

The dynamic parameters K_i, K_i, and K_d were determined by the fitting of the GABA response (R) profile in the presence of various concentrations of NQs and/or NSAIDs to equation 8 by the nonlinear least-squares method (MULTI) (32). All data were weighed by the reciprocal of the square of observed values for nonlinear least-squares regression. The optimum receptor occupancy, , was determined by eye-fitting within the  range of from 0.01 to 0.1 at an interval of 0.01, which most exactly represents the observed values. In addition, equation 9 was used to construct an isobologram (6, 13, 24-26) based on the relationship between [Q] and [S], as follows:

(9)

The total concentrations in brain tissue ([s] and [q]) are the calculated brain unbound concentrations ([S] and [Q]) and the unbound fraction in the brain (f_T100s for NSAIDs, f_T100q for ENX).

(10)

(11)

The simulation of the isobologram based on the relationship between [s] and [q] by substituting K_i, K_i, K_d, f_T100s, and f_T100q in equations 9, 10, and 11 was carried out.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

In vivo experiments. Figure 1 shows the isobolograms based on the relationship between the occurrence of convulsion and the concentrations of ENX and the three NSAIDs in brain tissue. Remarkable threshold lines were obtained for mice with convulsions and those without convulsions, and the lines had concave patterns. We confirmed that the interaction between ENX and the three NSAIDs was the synergistic interaction by the analysis of isobolograms.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Isobologram based on ENX-induced convulsions in the presence of FLB (A), KTP (B), or FRP (C), ENX (3 to 137.5 mg/kg) was injected into the tail veins of mice in the presence of NSAIDs (3 to 150 mg/kg). The occurrence of clonic convulsion within 20 min after administration of drugs was observed, and the brain was collected. The concentration of the drugs in brain tissue was determined by HPLC as described in Materials and Methods. Symbols: , mice without convulsions; , mice with convulsions.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Effect of the NSAID FLB (A), KTP (B), or FRP (C) on plasma and brain concentration-time profiles for ENX. ENX (25 mg/kg) was injected into the tail veins of the mice in the presence or absence of NSAIDs (FLB, 15 mg/kg; KTP, 50 mg/kg; FRP, 30 mg/kg). The mice were guillotined at 1, 3, 5, 10, 20, 40, or 60 min after administration of the drugs, and the plasma and brain were collected. Plasma and brain tissue ENX concentrations were measured by HPLC. Each value represents the mean ± SD (n = 5). Symbols; , plasma ENX concentration in the absence of NSAIDs; , plasma ENX concentration in the presence of NSAIDs; , brain tissue ENX concentration in the absence of NSAIDs; , brain tissue ENX concentration in the presence of NSAIDs.

In vitro experiments. Figure 3 shows the ENX-induced inhibition of the GABA current response in the presence of NSAIDs (1, 10, and 100 µM) or the absence of NSAIDs in Xenopus oocytes into which mouse brain mRNA was injected. In the presence of NSAIDs, the GABA response was inhibited in a dose-dependent manner. Moreover, the concentration-response curves of ENX shifted leftward in the presence of the three NSAIDs. No effect on the GABA response in the presence of NSAIDs alone (1 to 100 µM) was observed. The dynamic parameters of ENX calculated by using the dynamic model in the presence of the NSAIDs are presented in Table 1. The order of inhibition (K_i') in the presence of ENX and NSAIDs was FLB (0.2 µM) < FRP (0.3 µM) < KTP (1.2 µM). The inhibitory potencies determined by K_i/K_i' were approximately 280, 180, and 50 for ENX-FLB, ENX-FRP, and ENX-KTP, respectively.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Inhibitory effect of the NSAID FLB (A), KTP (B), or FRP (C) on GABA response to ENX. The inhibitory effect on ENX on the GABA (10 µM)-induced current response in the presence (1 [], 10 [], and 100 µM [] or absence () of NSAIDs was investigated with Xenopus oocytes into which mouse brain mRNA was injected, as described in Materials and Methods. The solid lines in the figure were obtained by fitting the GABA-induced response to equation 8 by using MULTI, as described in Materials and Methods.

View larger version (11K): [in this window] [in a new window]   FIG. 4.   Effect of the NSAID FLB (A), KTP (B), or FRP (C) on transcellular transport of ENX across MBEC4 from the luminal to the abluminal side. The transcellular transport of ENX in the presence or absence of NSAIDs was determined as described in Materials and Methods. Each sample was collected at the scheduled time, and the drug concentration was measured by HPLC. Each value represents the mean ± SD (n = 4). Symbols: , control; , effect in the presence of FLB; , effect in the presence of KTP; , effect in the presence of FRP.

Correlation between in vivo and in vitro data. Figure 5 shows the simulation lines of the GABA response based on the isobolograms for the concentrations of ENX and NSAIDs in brain tissue according to equation 9. These were calculated by using K_i, K_i', K_d, , f_T100s, and f_T100q. Various simulation lines were fitted to the isobolograms (Fig. 1) for the observed threshold concentration in brain tissue between mice with convulsions and those without convulsions, and  was estimated. The estimated  values were 0.02 for ENX-FRP, 0.02 for ENX-KTP, and 0.01 for ENX-FLB, suggesting that there is no difference in  among the three NSAIDs.

View larger version (25K): [in this window] [in a new window]   FIG. 5.   Simulation of isobologram for convulsion induced by ENX in the presence of the NSAID FLB (A), KTP (B), FRP (C). The isobologram was simulated according to equations 9, 10, and 11 as described in Materials and Methods. Solid lines are simulation lines. The line nearest the origin represents 0.01 of the GABA-induced current response (f), and simulation was carried out within the range of 0.01 to 0.1 of f at equal intervals of 0.01. Lines were generated at equal intervals. The hatched zone represents the optimum area of the threshold concentration in brain tissue for the occurrence of convulsions by the interaction between ENX and NSAIDs.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The threshold lines for convulsions and nonconvulsions after the administration of ENX in the presence or absence of NSAIDs in isobolograms based on the concentrations of both drugs in brain tissue showed concave patterns, suggesting a synergistic effect on ENX-induced convulsions in the presence of NSAIDs (Fig. 1). Plasma and brain concentration-time profiles for ENX, the transport rate of ENX across the blood-brain barrier, and f_p and f_T of ENX did not change in the presence of the NSAIDs as, shown in Fig. 2 and Fig. 4, indicating that these synergistic effects were not due to the pharmacokinetic interaction between ENX and NSAIDs. In the Xenopus oocytes into which mouse brain mRNA was injected, the GABA response was inhibited by ENX in a concentration-dependent manner, and the concentration-response curves of ENX were shifted leftward in the presence of the NSAIDs (Fig. 3). These findings were consistent with the relationship between FLB and CPFX, norfloxacin, or ofloxacin (12). Furthermore, our previous report indicated that the effect of ENX on the GABA response in vitro was synergistically potentiated in the same manner with the in vivo synergistic interaction on ENX-induced convulsions in the presence of the NSAIDs (11). In this study, based on an analysis of concentration-response curves by a newly developed toxicodynamic model, we reproduced quantitatively the in vivo synergistic interaction and obtained the dynamic parameters (Table 1). The order of inhibition (K_i) for ENX on the GABA response in the presence of the NSAIDs was FLB (0.2 µM) < FRP (0.3 µM) < KTP (1.2 µM), and the inhibitory effect of ENX alone was 50- to 280-fold lower than that of ENX in the presence of the NSAIDs. We simulated the isobologram for the concentration of both drugs in brain tissue by using the dynamic parameters (Table 1) and the brain tissue-unbound fraction of both drugs by using equation 9.

We made up 10 simulation lines within the range of 0.01 to 0.1 as , and the lines obtained fit the isobologram for the threshold concentration in brain tissue between mice with convulsions and those without convulsions in vivo. The  range of 0.01 to 0.02, despite the kind of NSAID, that was obtained indicated that 1 to 2% or more blockade of the GABA_A receptor was necessary to induce convulsions on the basis of the inhibition of the GABA response.

In conclusion, in order to predict quantitatively the toxicodynamic interaction between ENX and NSAIDs and to estimate the isobolograms of the threshold concentration in brain tissue between mice with convulsions and mice without convulsions, kinetic analysis based on the toxicodynamic parameters (K_i, K_i', K_d, , and  values [0.01 to 0.02]) and the pharmacokinetic parameters (f_T and f_p) are useful.