Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

New quinolone antibacterial agents (NQs¹), which have been developed since the 1980s, have a high intrinsic antibacterial activity with a wide spectrum of action and have been used in the treatment of a variety of infections. Among them, both grepafloxacin (GPFX) (Akiyama et al., 1995a,b) (Fig. 1) and HSR-903 (Murata et al., 1999) were recently reported to be highly distributed in the tissues compared with the other NQs. The distribution volume of GPFX (7095 ml/kg) and HSR-903 (4900 ml/kg) (Murata et al., 1999) was higher than that of lomefloxacin and ofloxacin (1460 and 1540 ml/kg, respectively) (Okezaki et al., 1988). Okezaki et al. (1988) proposed common distribution characteristics to the lung for NQs other than GPFX and HSR-903 because a correlation was observed between the tissue-to-blood concentration ratio (K_p) and the plasma unbound fraction (f_up) of several NQs, except for these two compounds, whereas the K_p of GPFX and HSR-903 in the lung was far removed from this correlation line. This observation suggests the existence of some specific mechanism(s) for the distribution of these two NQs in the lung. The efficacy of an antibiotic is generally thought to depend both on its plasma concentration and minimum inhibitory concentrations for likely pathogens. However, the tissue concentrations at the site of infection may be more relevant for the effects of antibiotics (Baldwin et al., 1990). Consequently, it is important to clarify the distribution mechanism of GPFX in the lung.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Chemical structure of GPFX.

The possible mechanisms involved in the tissue accumulation of drugs are considered likely to be higher tissue uptake and/or tissue binding, or lower efflux from the tissues. Various determining factors have been identified for the tissue distribution of certain types of drugs. For example, binding to the nuclei is involved in the tissue distribution of doxorubicin (Terasaki et al., 1984). The K_p values for vinca alkaloids, such as vincristine and vinblastine, correlate with the tissue tubulin concentration (Wierzda et al., 1987, 1988). Okumura et al. (1978, 1989) and Yoshida et al. (1987, 1989) reported specific common binding sites for basic drugs in the lung, the affinity for these sites being dependent on the lipid solubility of the drugs. Yata et al. (1990) and Nishiura et al. (1986, 1987, 1988) reported the involvement of phospholipids in the tissue distribution of basic drugs. Ishizaki et al. (1998a,b) reported that some basic drugs, such as biperiden and trihexiphenidyl, are mainly distributed in the postnuclear fractions containing the acidic organelles (e.g., lysosomes). Concerning the NQs, HSR-903 has been reported to be taken up by active transport into isolated rat lung cells (Murata et al., 1999), although the contribution of this uptake to their lung distribution has not yet been fully characterized.

In the present study, to clarify the mechanism governing the distribution of GPFX to the lung, the uptake, binding, and efflux after intravenous administration of GPFX were evaluated in rats.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. GPFX (1.08 MBq/µmol, radiochemical purity 97.1%) was obtained from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK). Inulin (9.0 mCi/g, radiochemical purity >97%) was purchased from PerkinElmer Life Sciences (Boston, MA). Unlabeled GPFX was synthesized by Otsuka Pharmaceutical Company (Tokyo, Japan). The following standard phospholipids were obtained from Sigma-Aldrich (St. Louis, MO) and used without further purification: L--phosphatidyl-L-serine (PhS, no. P-7769), DL--phosphatidyl-DL-glycerol (PhG, no. P-5650), L--phosphatidylinositol (PhI, no. P-0639), L--phosphatidylcholine (PhC, no. P-7318), and L--phosphatidylethanolamine (PhE, no. P-6386). Rotenone, quinidine, imipramine, and propranolol were also obtained from Sigma-Aldrich. All other chemicals used were commercially available and of reagent grade.

Tissue Distribution of GPFX during Steady-State Infusion. Under light ether anesthesia, the femoral artery and vein of three different rats (250-300 g male Sprague-Dawley rats; Charles River Japan Inc., Kanazawa, Japan) were cannulated with a polyethylene catheter (PE-50) for blood sampling and GPFX injection, respectively. The rats received a constant infusion of GPFX at a dose of 15 µg/min/kg after a bolus i.v. administration of 5 mg/kg. Rats were kept warm by the heat from an electric bulb throughout the experiment. Blood samples (approximately 200 µl) were collected with a heparinized syringe at 70, 90, 110, and 120 min after starting the infusion, and plasma was prepared by centrifuging a portion of the blood samples (1800g, 10 min). At 120 min after dosing, the rats were killed and various tissues were excised immediately. The tissues were stored at 20°C until required for assay. The radioactivity in the plasma and blood was measured using a liquid scintillation counter as follows. Fifty microliters of plasma was transferred to a scintillation vial and a scintillation cocktail was added (ACS-II; Amersham Biosciences UK, Ltd.). Fifty microliters of blood was transferred to a scintillation vial and sonicated with 1 ml of tissue-solubilizing solution (Nakalai Tesque, Inc., Kyoto, Japan) at 50°C for 30 min. The lysate was neutralized with 1 ml of pH adjustment solution (Nakalai Tesque, Inc.) and then a scintillation cocktail (Hionic-Fluor; Packard BioScience B.V., Groningen, The Netherlands) was added. Sasabe et al. (1998a) reported that the amount of 3-glucuronide, a main metabolite, is 5 to 11% of that of the parent compound in plasma, liver, and kidney, at 160 min after the start of constant infusion of R-(+)-GPFX or S-()-GPFX. Accordingly, in the present study, the radioactivity measured was assumed to be unchanged GPFX. K_p value was defined as the ratio of the tissue concentration (micrograms per gram of tissue) to the plasma concentration (micrograms per milliliter of plasma) and, therefore, K_p is indicated in units of milliliters per gram of tissue. Considering that the gravity of most tissues is close to unity, the K_p value should be close to the apparent concentration ratio between tissue and plasma. k_el was calculated by the following equation:

(1)

where k_el represents the elimination rate constant from tissues. Q is the blood flow rate in each tissue. If we assume that the gravity of tissue is unity, the dimension of k_el should be the same as that of clearance (milliliters per minute per milliliter or gram of tissue).

Plasma Protein Binding and Red Blood Cell Distribution. The blood unbound fraction (f_uB, 0.40) was obtained by f_up (0.60) (Akiyama et al., 1995b) divided by the blood-to-plasma concentration ratio (R_B, 1.51), which was obtained in the present study.

Subcellular Fractionation. Rat tissues were fractionated according to the procedures described by Robertis et al. (1962). Briefly, the sampled tissues, whole or partial, were added to 9 volumes ice-cold 0.32 M sucrose and homogenized. All subsequent steps were performed at 4°C. To obtain the nuclear and membrane fractions, lysosomal and mitochondrial fractions and microsomal fraction, the homogenates were sequentially centrifuged at 1,000g for 10 min, 13,200g for 20 min, and 100,000g for 60 min, respectively. The final supernatant was taken as the cytosol fraction. The radioactivity of the tissue homogenates and subcellular fractions was measured using a liquid scintillation counter. The distribution ratio in each fraction was calculated by dividing the radioactivity in each fraction by the total radioactivity in each tissue.

Intracellular Binding. Under ether anesthesia, the rats were killed by severing the inferior vena cava and aorta and the brain, testis, liver, and lung were excised immediately. The tissues were stored at 20°C until required for assay. The unbound fraction of GPFX in tissue (f_uT) was obtained by extrapolating to binding for 100% homogenate from that of 2, 5, and 10% tissue homogenate, determined by the ultrafiltration method described by Sasabe at al. (1997). Initially, each 10% homogenate of the brain, testis, liver, and lung was prepared in 0.32 M sucrose. These homogenates were then diluted with 0.32 M sucrose to give 5 and 2% homogenates. The total concentrations of GPFX in the homogenates of the brain, testis, liver, and lung (0.880, 2.77, 31.6, and 56.0 µM, respectively) were set so as to be close to those seen during steady-state infusion. After incubation at 37°C for 5 min, homogenate samples with GPFX were placed in YM-30 tubes, pretreated with blank filtrate. Then, the homogenate was centrifuged (1,800g, 10 min) to give a filtrate containing the unbound drug. The radioactivities of the tissue homogenates and its filtrates were measured using a liquid scintillation counter.

Kinetics for Tissue Uptake of GPFX in Vivo. Under ether anesthesia, GPFX was administered to the rats, weighing approximately 200 g, via the femoral vein at a dose of 5 mg/kg (0.80 MBq/kg). Blood samples were then collected from the femoral artery and the right atrium at designated times over 1, 2, 3, 5, 10, and 30 min with a heparinized syringe. The rats were killed at 1, 2, 3, 5, 10, and 30 min by severing the inferior vena cava and aorta after the final blood sample had been collected and the following tissues were sampled: lung, brain, liver, kidney, thymus, heart, spleen, adrenals, stomach, small intestine, large intestine, and testis. The lung, liver, kidney, brain, stomach, small intestine, and large intestine were homogenized with 2 volumes of saline, and portions of the homogenate and other tissues were weighed and dissolved. The radioactivity of the homogenate was measured using a liquid scintillation counter (LSC-1050; Aloka, Tokyo, Japan). The counting efficiency was corrected by the channels ratio method using an external standard. The following equation represents the mass balance of GPFX in the tissues:

(2)

where Xt is the amount of unchanged drug in the tissue at time t, CL_uptake,p is the tissue uptake clearance, C_p is the plasma concentration of unchanged drug and k_efflux is the efflux rate constant from the tissue. In this case, the plasma sampled at the entrance to the tissues should be measured, i.e., plasma was collected from the right atrium for the lung and from the femoral artery for the other tissues. During this short period, the efflux process can be ignored, and the integration and subsequent normalization by C_p of the following equation yields:

(3)

The plot of K_p versus AUC/C_p is designated as the integration plot. The time profile for C_p was fitted to the following equation using a pharmacokinetic software package (WinNonlin; Pharsight, Mountain View, CA):

(4)

Using eqs. 2 and 4, the time profile of the data for Xt was fitted to the following equation to estimate CL_uptake,p, k_efflux, and V_E:

(5)

where V_E represents the distribution volume in which the GPFX concentration equilibrates instantaneously with that in plasma. This CL_uptake,p is based on the plasma concentration, and the CL_uptake,b for the blood concentration is calculated by dividing the CL_uptake,p by R_B.

Single-Pass Lung Uptake Index. Under ether anesthesia, GPFX and inulin as an extracellular marker were administered to different rats via the femoral vein at a dose of 5 and 1.9 mg/kg/1.6 ml saline, respectively. The rats were killed at designated times over 5 to 120 s and the whole lung was excised immediately. Three to six rats were used at each time point. Lung samples were homogenized with two volumes of saline, and portions of the homogenate were dissolved. The radioactivity of the homogenate was measured using a liquid scintillation counter.

Phospholipid Binding of GPFX in Vitro. The binding of GPFX to phospholipids in vitro was performed by partitioning between pH 7.4 buffer (0.25 M sucrose-0.1 M Tris-HCl buffer) and n-hexane (Yata et al., 1990). Briefly, 2 ml of buffer solution containing 1 µM GPFX in the absence or presence of weakly basic drugs such as quinidine, propranolol, and imipramine, which bind to PhS (Nishiura et al., 1986, 1987, 1988; Yata et al., 1990), was shaken with 2 ml of n-hexane solution containing the individual standard phospholipids (8 µg/ml) at 37°C for 2 h. The mixture was then centrifuged at 1,800g for 10 min and the radioactivity in the separated aqueous and organic phases was measured using a liquid scintillation counter. The concentrations of the weakly basic drugs were set at 10, 100, and 1000 µM because the dissociation constants of the three compounds to PhS are 0.179 to 4.20 µM (Yata et al., 1990).

Cell Culture. K1/R97K-pssB, CHO-K1, CDT-1, and PSA-3 cells were kindly donated by Drs. M. Nishijima and O. Kuge (National Institute of Infectious Diseases, Tokyo, Japan). PSA-3 cells, a mutant of CHO-K1 cells, lacks the ability to synthesize PhS (Kuge et al., 1986). K1/R97K-pssB transfected with the R97K mutant of PhS synthase II exhibit an approximately 4-fold higher PhS biosynthesis rate than CHO-K1 cells, resulting in a 1.6-fold higher PhS level than that in CHO-K1 cells (Kuge et al., 1999). CDT-1 cells are transformants of PSA-3 cells with a pssA cDNA, which encodes PhS synthase I (Kuge et al., 1991). K1/R97K-pssB, CHO-K1 and CDT-1 cells were cultured in Ham's F-12 medium supplemented with 10% (v/v) newborn calf serum, 100 units/ml penicillin, and 100 µg/ml streptomycin, followed by incubation at 37°C under a 5% CO₂ atmosphere and 100% humidity. PSA-3 cells were maintained under the same culture conditions except that the medium was supplemented with 44 µg/ml PhS liposomes. K1/R97K-pssB, CHO-K1, CDT-1, and PSA-3 cells were seeded on a 12-well microplate (BD Biosciences, Franklin Lakes, NJ) at approximately 3 × 10⁴ cells/3.8 cm². After 2 to 3 days, the medium was replaced with a fresh medium without antibiotics or PhS, whereas the fraction of PSA-3 was maintained with fresh medium containing PhS. After an additional 2 days, these cells were used in the association studies.

Association to PhS Synthase Transformants. The culture medium was removed by aspiration, and the monolayers were washed with Hank's balanced salt solution at 37°C. Association of 10 µM GPFX was initiated by adding 0.05 ml of prewarmed ligand solution to the cell medium after preincubation at 37°C for 3 min. At a designated time, the association was terminated by removing the cell medium immediately. Then, 0.05 ml of medium was transferred to a scintillation vial, the cells were washed twice with 2 ml of ice-cold PBS, and solubilized in 2 N NaOH. After neutralization with 6 N HCl, a scintillation cocktail (ACS-II; Amersham Biosciences UK, Ltd.) was added and the radioactivity was measured. Ligand association was given as the distribution volume, determined as the amount of ligand associated with the cells divided by the medium concentration. Protein concentration in the solubilized cells was determined by the method of Lowry et al. (1951), using a protein assay kit (Bio-Rad, Hercules, CA) with bovine serum albumin as a standard. To examine the effect of weakly basic drugs, preincubation was performed with 30 µM rotenone, to exclude any effect of active transport systems, in the absence or presence of inhibitors (quinidine, imipramine, and propranolol) at 37°C for 3 min.

Quantitation of Phospholipid Content. Phospholipids were extracted by the method of Folch et al. (1957), with a slight modification. Briefly, the sample obtained from each tissue homogenate (1.2 g) and cell suspension (1.2 ml) was homogenized with 4.5 ml of chloroform/methanol (1:2, v/v) mixture, and chloroform (1.5 ml) and water [or PBS()] (1.5 ml) were added to the homogenates. After centrifugation, the organic layer was transferred to a test tube and evaporated to dryness under nitrogen gas. The phospholipids were separated by two-dimensional thin layer chromatography on 0.5-mm-thick Silica Gel 60 plates (Merck, Darmstadt, Germany) using chloroform/methanol/acetic acid (65:25:10, v/v/v) as the first solvent and chloroform/methanol/formic acid (65:25:10, v/v/v) as the second solvent. After thin layer chromatography, the spots were made visible with iodine vapor. Each spot was scraped off, and the phosphate content was determined chemically by assaying the inorganic phosphorus by the method described by Trudinger (1970).

Statistical Analysis. The statistical analyses were performed using Dunnett's test (SAS computer software, version 6.12) for the association to PhS synthase transformants and for the phospholipid binding of GPFX in vitro. P < 0.01 and P < 0.05 were regarded as significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue Distribution of GPFX during Steady-State Infusion. Plasma concentrations of GPFX at 70, 90, 110, and 120 min during the constant infusion of GPFX at a dose of 15 µg/min/kg were 1.03 ± 0.06, 1.00 ± 0.10, 0.979 ± 0.095, and 0.865 ± 0.088 µg/ml (mean ± S.E.), respectively. The blood concentration at 120 min was 1.30 ± 0.12 µg/ml (mean ± S.E.), and the R_B at 120 min was 1.51 ± 0.03 (mean ± S.E.).

Subcellular Distribution. The subcellular distribution of GPFX was examined (Table 2). More than 50% of GPFX was localized in the nuclear and membrane fractions in the lung, heart, stomach, muscle, small intestine, large intestine, and brain (Table 2). The fraction of GPFX distributed to the lung was in the following order: nuclear and membrane fractions (60.3%) > lysosomal and mitochondrial fractions (22.7%) > microsomal fraction (12.9%) > cytosol fraction (4.1%) (Table 2). In contrast, more than 70% of GPFX was recovered in lysosomal and mitochondrial fractions in the spleen and thymus (Table 2). Most of the GPFX (>87%) was distributed to the organelle fractions (nuclear and membrane fractions, lysosomal and mitochondrial fractions, and microsomal fraction) in all tissues examined (Table 2).

Kinetics for Tissue Uptake of GPFX in Vivo. The time profile of GPFX concentrations in the plasma and various tissues was subjected to kinetic analysis after intravenous bolus administration (Fig. 2). The CL_uptake,b in the lung (2.86 ml/min/g), kidney (4.25 ml/min/g), liver (1.48 ml/min/g), adrenals (2.38 ml/min/g), and heart (1.86 ml/min/g) was higher than that in other tissues, including the brain (0.0234 ml/min/g). The dose (5 mg/kg) used in this analysis was the same as that used in our previous analysis (Sasabe et al., 1997). The plasma concentrations (1.38-3.18 µg/ml, corresponding to 3.49-8.03 µM) observed at 1 to 10 min postdosing were not far from that for steady-state infusion. As shown in Fig. 3, the CL_uptake,b in most tissues except lung was close to the blood flow rate in each tissue. Accordingly, it was concluded that GPFX is efficiently taken up by the different tissues. Because blood flow rate can be affected by the condition of anesthesia, for most organs the information cited herein was obtained in rats under light ether anesthesia. Nevertheless, we cannot rule out the possibility that such anesthesia may differentially affect the blood flow rate in each organ. The fitting in integration plot analysis does not seem to be appropriate, especially for the liver and kidney. This is at least partially due to the rapid decrease after the uptake phase in these organs (Fig. 2). Considering that active transport systems are reported to be involved in basolateral and/or apical membranes of the liver and kidney (Sasabe et al., 1997, 1998b; Matsuo et al., 1998; T. Suzuki, Y. Kato, H. Sasabe, M. Itose, G. Miyamoto, and Y. Sugiyama, unpublished data), nonlinear behavior may occur in GPFX distribution to these organs. Further study is needed to clarify the exact mechanism.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Integration plot representing the GPFX uptake by each tissue. Time profiles of GPFX concentration in plasma and tissues were determined after its bolus injection (5 mg/kg). The solid lines are the fitted curves based on eq. 5. Each plot and bar represents the mean ± S.E. of three rats.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Comparison of the CLuptake,b and blood flow rate in various tissues. The CLuptake,b was estimated by fitting both plasma and tissue concentration profiles to eq. 5. Blood flow rates for all organs except the pancreas, thymus, and testis were determined under light anesthesia (Stanek et al., 1988), and those for the pancreas, thymus, and testis were determined under a conscious condition (Darlington et al., 1995). The solid line has a slope of unity. Lung (a), kidney (b), adrenals (c), heart (d), liver (e), spleen (f), pancreas (g), large intestine (h), small intestine (i), stomach (j), thymus (k), testis (l), and brain (m).

Single-Pass Lung Uptake Index. To evaluate the single-pass extraction ratio of GPFX in the lung, we measured its association with the lung after intravenous bolus injection over a short time period before the GPFX that enters the circulation again passes through the lung (Fig. 4A). Compared with an extracellular marker, such as inulin, a higher amount (~10% of dose) of GPFX was associated with the lung at 5 s after dosing (Fig. 4A). The lung concentration of GPFX fell over time (Fig. 4A). Thus, GPFX associated with the lung is dissociated by dilution in the plasma flow. To evaluate the saturable distribution of GPFX in the lung, this distribution of GPFX was examined at doses of 0.006 to 15 mg/kg (Fig. 4B). At 5 and 30 s after dosing, the extraction by the lung was approximately 10 and 4%, respectively, showing minimal nonlinear behavior (Fig. 4B).

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Time profile (A) and dose dependence (B) of the single-pass lung uptake index of GPFX after its bolus intravenous administration. Both GPFX () at 5 mg/kg and inulin () were intravenously administered and GPFX associated with the lung was determined (A). The injected dose of GPFX was changed at 0.006 to 15 mg/kg and GPFX association was measured at 5 s () and 30 () s after injection (B). The data in B represent the intracellular amount of GPFX, which was obtained by subtracting the amount of inulin from that of GPFX. Each plot and vertical bar represents the mean ± S.E. of three to six rats.

Estimation of k_el from Tissues. The k_el in the tissues was calculated by dividing the blood flow rate by K_p (Table 1), assuming each tissue to be a single compartment. The k_el in all tissues was higher than the elimination rate constant calculated from the plasma concentration profile ( of 0.00480 min¹) obtained by Nakajima et al. (2000).

Phospholipid Binding of GPFX in Vitro. The in vitro binding of GPFX to phospholipids was examined (Fig. 5A). The GPFX binding to PhS was 2.67 ml/mg lipid, whereas the binding to other phospholipids was, at most, 0.050 to 0.285 ml/mg lipid, i.e., GPFX preferentially binds to PhS (Fig. 5A). The binding of GPFX to PhS was reduced in the presence of weakly basic drugs, such as quinidine, imipramine, and propranolol (Fig. 6).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   In vitro phospholipid binding of GPFX (A) and time profile of GPFX association in PhS synthase transformants of CHO-K1 cells (B) A, binding of 1 µM GPFX was determined using the n-hexane/buffer partition system at 37°C. Each column represents the mean ± S.E. of three samples. B, association of GPFX was measured by incubating each cell line (K1/R97K-pssB, ); CHO-K1, ; CDT-1, ; PSA-3[PhS(+)], ; and PSA-3[PhS()]), ) with 10 µM GPFX at 37°C. PhS content in K1/R97K-pssB, CHO-K1, CDT-1, PSA-3[PhS(+)], and PSA-3[PhS()] was 7.79, 5.38, 5.66, 5.37, and 3.47 µg/mg protein, respectively. The cellular association was normalized in terms of the extracellular concentration. Each plot and vertical bar represent the mean ± S.E. of three samples. PhG, DL--phosphatidyl-DL-glycerol; PhC, L--phospatidylcholine; PhI, L--phospatidylinositol; PhE, L--phosphatidylethanolamine.

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Effect of weakly basic drugs on GPFX association with PhS () and K1/R97K-pssB cells (). The association of 1 µM GPFX with PhS was determined after a 2-h incubation, whereas that with K1/R97K-pssB cells was measured after a 20-min incubation. Each GPFX association was normalized in terms of the extracellular concentration. Each column represents the mean ± S.E. of three to nine (PhS) or three to 21 samples (K1/R97K-pssB cells). *, P < 0.05; **, P < 0.01 (significantly different from control).

Association to PhS Synthase Transformants. The association of GPFX with PhS synthase transformants of CHO-K1 cells, which have different PhS contents, was also examined (Fig. 5B). The mean value for the GPFX association between 10 and 20 min with K1/R97K-pssB (43.3 µl/mg protein), which has the highest PhS content, was approximately 1.5 times as high as that with PSA-3[PhS()] (28.4 µl/mg protein), which has the lowest content (Fig. 5B). Quinidine and imipramine reduced the association of GPFX with K1/R97K-pssB in a concentration-dependent manner (Fig. 6).

Relationship between Tissue-to-Blood Unbound Concentration Ratio (K_p,u) and PhS Content in Vivo and PhS Synthase Transformants in Vitro. A correlation was observed between the K_p,u and the PhS content of rat tissues (Fig. 7A). In this plot, the K_p,u in brain and testis were lower than the correlation line, whereas that in pancreas was higher than the correlation line (Fig. 7A). The amount of GPFX in each subcellular fraction in the lung also exhibits a linear relationship with the PhS content of each fraction (Fig. 7B).

View larger version (11K): [in this window] [in a new window]   Fig. 7.   Relationship between PhS content and distribution of GPFX. A, Kp,u was obtained at 120 min after intravenous infusion of GPFX (15 µg/min/kg) and plotted against PhS content in each tissue. The Kp,u value represents the mean ± S.E. of three rats at steady state in vivo. The PhS content was taken from Yata et al. (1990). B, steady-state distribution of GPFX to the subcellular fraction of the lung was plotted against PhS content in each fraction. The PhS content was taken from Nishiura et al. (1988). Each plot represents the mean value of three rats. C, Kp,u in each tissue observed in vivo () and the distribution volume in each PhS synthase transformant of CHO-K1 cells in vitro () were plotted against the PhS content in the corresponding tissues or cell line. GPFX association in transformants represents the mean value of that at 10 and 20 min obtained from the data in Fig. 5B. The PhS contents both in vivo and in vitro were determined in the present study. In A and C, lung (a), spleen (b), kidney (c), pancreas (d), liver (e), small intestine (f), heart (g), muscle (h), testis (i), brain (j), K1/R97K-pssB (k), CHO-K1 (l), CDT-1 (m), PSA-3[PhS(+)] (n), and PSA-3[PhS()] (o).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

GPFX is highly distributed to several organs, including the lung (Akiyama et al., 1995a,b; Nakajima et al., 2000) (Table 1), although the determining factors mainly involved in such high tissue distribution remain to be identified. In the present study, we attempted to identify the mechanism(s) involved in GPFX distribution to the lung.

Sasabe et al. (1997) reported that Na⁺-independent and carrier-mediated active transport system contributes to the hepatic uptake of GPFX. Therefore, tissue uptake system(s) might be present at least in the liver and contribute to the tissue distribution of GPFX. In the present study, the CL_uptake,b values, analyzed by integration plot analysis, in the lung, kidney, liver, adrenals, and heart were higher than those in other tissues (Fig. 3). However, this CL_uptake,b except in the lung, was close to the blood flow rate in each tissue (Fig. 3), suggesting that GPFX is efficiently taken up by tissues in the all organs examined. Thus, higher distribution of GPFX to the lung cannot be explained by the difference in its uptake process. The extraction by the lung in the single-pass lung uptake index was an almost linear behavior at doses of 0.006 to 15 mg/kg (Fig. 4B). In that study, the GPFX concentration in the administered solution was 26 mM at 15 mg/kg. If we assume that the dosing solution is diluted with circulating blood (~80 ml/kg), the GPFX passing through the lung should be at least 520 µM. Thus, the uptake of GPFX in the lung has a low affinity and is not saturated within the micromolar GPFX range.

Other hypotheses accounting for the higher distribution of GPFX to the lung include its specific binding to tissue. In various tissues GPFX was mainly recovered in the nuclear and membrane fractions (Table 2), which contain more phospholipids as membrane components than other fractions. Nishiura et al. (1988) reported that the content of PhS is much higher in the nuclear and membrane fractions than in others in the lung. In addition, compared with the other organs, the content of PhS is higher in the lung (Yata et al., 1990). Therefore, we first examined the relationship between the PhS content and the distribution of GPFX. A linear relationship can be observed between the K_p,u of GPFX and the PhS content of each tissue (Fig. 7A) and between the amount of GPFX in subcellular fractions and their PhS content (Fig. 7B). The tissue distribution of doxorubicin is governed by the DNA content of each tissue (Terasaki et al., 1984), but only a minimal correlation was observed between the K_p,u of GPFX and the DNA content (data not shown). In addition, GPFX preferentially binds to PhS compared with other phospholipids (Fig. 5A). Yata et al. (1990) and Nishiura et al. (1986, 1987, 1988) have also suggested that PhS governs the distribution of weakly basic drugs, such quinidine, propranolol, and imipramine. In fact, GPFX binding to PhS was inhibited by these three drugs (Fig. 6), although they are basic, whereas GPFX is zwitterionic. Thus, these results suggest the importance of PhS as a determining factor for GPFX distribution.

To further evaluate the contribution of PhS governing the distribution of GPFX, its association was examined by using CHO-K1 mutants that lack the ability to synthesize PhS or are transfected with PhS synthase, resulting in different levels of PhS expression (Kuge et al., 1986, 1991, 1999). The higher PhS content of the cells was tended to result in a higher GPFX association (Fig. 5B). In addition, GPFX association with such a PhS synthase transformant was also inhibited by quinidine, propranolol, and imipramine (Fig. 6). These results suggest that PhS is an important factor governing the association of GPFX with these cell lines. The GPFX association with the PhS synthase transformants in vitro and that to the various tissues in vivo can be shown against the PhS content in a same plot (Fig. 7C). This result is compatible with the hypothesis that PhS is a determining factor for the tissue distribution of GPFX in vivo. However, the linear relationship found in the cell lines may have a y-intercept (Fig. 7C). Furthermore, the weakly basic drugs showed a weaker inhibition of the cellular association in comparison with the GPFX binding to PhS. These results suggest that further studies are needed to clarify the other factor(s) that are also involved in the association of GPFX with these cell lines.

It is unlikely that the difference in the elimination kinetics from the tissues can account for the higher distribution characteristics of GPFX to the lung: The k_el value in all the tissues, including the lung, exceeded the  value (Table 1), suggesting that the efflux of GPFX from these tissues is not the rate-limiting step in its elimination from tissues. This result is compatible with the finding by Nakajima et al. (2000) that the disappearance curve for the GPFX concentration in most tissues, including the lung, is almost parallel to that in plasma after intravenous bolus injection. In addition, GPFX once trapped by the lung rapidly disappeared in the present study (Fig. 4A), suggesting that the GPFX associated with the lung undergoes ready efflux into the circulation. Note that we cannot discuss differences in the absolute values for intrinsic flux of GPFX among the various tissues based on CL_uptake,b or k_el value, because both values are blood flow-dependent parameters. Nevertheless, as discussed above, because both CL_uptake,b and k_el are very rapid, the specific distribution of GPFX to the lung cannot be explained either by the uptake or efflux processes. On the other hand, the q values in the brain and testis were much lower than unity and the K_p,u in these two tissues was at the lower end of the correlation line when plotted against the PhS content (Fig. 7A). However, the method used to determine tissue unbound fraction in the present study is based on the assumption that binding in tissues in vivo is equivalent to that observed in tissue homogenate in vitro. Therefore, the absolute value for such q values should be used carefully in any discussion. However, Tamai et al. (2000) proposed the involvement of multidrug-resistance protein 1a as an efflux transport system for GPFX in the brain, and these results are compatible with the hypothesis that an active efflux system exists in these tissues.

In conclusion, PhS is the major determining factor for the tissue distribution of GPFX. This is the first report that the inter-organ variation in the distribution of a zwitterion-type compound can be accounted for, at least in part, by its affinity for PhS and the PhS content of each tissue. However, the data obtained in the present study still provide only indirect evidence for the involvement of PhS in the tissue distribution of GPFX, and further investigation is needed to yield a final conclusion.