Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Tacrolimus, a hydrophobic macrolide lactone produced by Streptomyces tsukubaensis, exerts marked immunosuppressive effects, and has been in use clinically as prophylaxis against organ rejection after liver and renal transplantation (Sawada et al., 1987; Thomsen, 1990; Sigal and Dumont, 1992). However, tacrolimus is known to exhibit low oral bioavailability and to have a wide range of variability in absorption, ranging from 4 to 89% with a mean of about 25% in kidney and liver transplant recipients (Venkataramanan et al., 1995). These drawbacks are due to the very low aqueous solubility of tacrolimus, and first pass effect metabolism via a combination of cytochrome P4503A4 (CYP3A4) and P-glycoprotein (P-gp) in intestinal epithelium cells and hepatocytes (Hashimoto et al., 1998; Iwasaki et al., 1998). The poor dissolution rate of tacrolimus has presented a substantial challenge to pharmaceutical scientists, and thus methods using solid dispersion and microcrystallization have been investigated (Honbo et al., 1987).

The solubilization of new drugs with poor aqueous solubility is crucial for the pharmacological evaluation. The use of biologically incompatible organic solvents or surfactants in testing these compounds is common and undesirable. Cyclodextrins (CyDs) form inclusion complexes with lipophilic drugs as guests and thus have been used for improving their water solubility (Szejtli, 1982). Practical use of natural CyDs as drug carriers, however, is restricted by their low aqueous solubility. Recently, several hydrophilic CyD derivatives have been used in efforts to resolve the problem, e.g., methylated, hydroxypropylated, and sulfobutyl ether CyD derivatives (Pitha and Pitha, 1985; Koizumi et al., 1987; Stella and Rajewski, 1997; Tompson, 1997; Uekama et al., 1998; Szente and Szejtli, 1999). There have been two reports of the use of CyDs to improve the pharmaceutical characteristics of tacrolimus: CyDs were found to improve the solubility of tacrolimus in ophthalmic solutions (Mills et al., 1995; Benelli et al., 1996). In our recent study, we demonstrated that of all the hydrophilic CyD derivatives examined, 2,6-di-O-methyl--cyclodextrin (DM--CyD) most significantly increased the low aqueous solubility and oral bioavailability of tacrolimus in rats and this improving effect of DM--CyD may have been due to an increase in solubility and the dissolution rate of tacrolimus (Arima et al., 2001).

Interestingly, it has been reported that methylated -cyclodextrin, in which the degree of substitution of methyl groups is in the range of about 10 to 15, markedly increased the antitumor activity of doxorubicin on sensitive and multidrug-resistant cell lines due to an increase in intracellular doxorubicin accumulation (Grosse et al., 1997, 1998). On the other hand, P-gp has been reported to be expressed even in small intestinal epithelial cells where it functions in the efflux of the substrate. Based on these findings, we postulated that DM--CyD may inhibit P-gp function in small intestinal epithelial cells after oral administration, resulting in an increase in bioavailability of tacrolimus, a P-gp substrate, in rats. However, it is not clear whether DM--CyD inhibits the function and expression of P-gp in intestinal epithelial cells. The present study, therefore, was performed to examine the contribution of P-gp to the enhancing effects of DM--CyD on the oral bioavailability of tacrolimus using both Caco-2 cells, a human colon adenocarcinoma cell line, and vinblastine-resistant Caco-2 (Caco-2R) cells.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Tacrolimus, mouse antitacrolimus monoclonal antibody and rabbit anti-mouse IgG polyclonal antibody were gifts from Fujisawa Pharmaceutical Co. (Osaka, Japan). DM--CyD was obtained from Nihon Shokuhin Kako (Tokyo, Japan). Cyclosporin A and itraconazole were gifts from Sandz (Tokyo, Japan) and Janssen-Kyowa (Tokyo, Japan), respectively. Rhodamine 123 was purchased from Molecular Probes (Eugene, OR). [³H]Mannitol was purchased from NEN Life Science Products (Boston, MA). C219 anti-human P-gp monoclonal antibody was purchased from Signet Laboratories (Dedham, MA). Vinblastine and the silver staining kits were purchased from Wako Pure Chemicals (Osaka, Japan). Deoxyribonuclease I and ribonuclease inhibitor were purchased from Nippon Gene (Tokyo, Japan) and Nacalai Tesque (Kyoto, Japan), respectively. Reverse transcriptase (SuperScript II) and Taq polymerase (AmpliTaq Gold) were purchased from Life Technologies (Gaithersburg, MD) and Applied Biosystems (Tokyo, Japan), respectively. UIC2 phycoerythrin (PE) conjugated with mouse anti-P-gp UIC2 monoclonal IgG2a antibody was purchased from Immunotech (Marseille Cedex, France). All other chemicals and solvents were of analytical reagent grade. Caco-2 cells were obtained from the American Type Culture Collection (Rockville, MD).

Preparation of Complex of Tacrolimus with DM--CyD. The solid tacrolimus/DM--CyD complexes in various molar ratios were prepared by the kneading method (Tsuruoka et al., 1981). The calculated amounts of tacrolimus and DM--CyD were weighed and triturated with a small amount of a mixed solution of 1:1 (v/v) ethanol/water and the slurry was kneaded thoroughly for about 40 min. After evaporation of the solvent, the solid complexes were dried under reduced pressure at room temperature for 3 days and stored in a desiccator.

Dissolution Studies. The dissolution rates of tacrolimus and its DM--CyD complexes were measured by the dispersed amount method (Nogami et al., 1969) and the rotating disk method (Nogami et al., 1966). For the dispersed amount method, the powder sample (<100 mesh, equivalent to 5 mg of tacrolimus) was added to 25 ml of Japanese Pharmacopoeia XIII (JPXIII) second fluid (pH 6.8, 50.0 mM KH₂PO₄ and 23.6 mM NaOH) and stirred at 100 rpm at 37°C. For the rotating disk method, the powder sample was compressed into a cylindrical tablet (diameter 10 mm) at a pressure of about 150 kg/cm² for 10 min. The dissolution of tacrolimus was measured by using a rotating disk apparatus in 25 ml of JPXIII second fluid at 91 rpm and 37°C. At appropriate intervals, an aliquot (0.5 ml) of the dissolution medium was withdrawn using a pipette with a cotton plug. The volume in the vessel was replaced with water after each sampling. Twenty microliters of sample was injected onto a high performance liquid chromatography (HPLC) column. Tacrolimus was analyzed using HPLC according to the method described by Nishikawa et al. (1993). The conditions were as follows: a Shimadzu LC-10AD pump (Kyoto, Japan) and a Hitachi L-4000 UV detector at 214 nm (Tokyo, Japan); a Hitachi D-2500 Chromato-Integrator; a GL-Sciences 544 column oven (Tokyo, Japan); a Tosoh TSK gel ODS-80TM column (4.6 × 150 mm; Tokyo, Japan); mobile phase of acetonitrile/water (3:1 v/v); flow rate of 1.0 ml/min; column temperature of 60°C.

In Vivo Absorption Studies. Male Wistar rats, 200 to 250 g, were fasted overnight, and orally administered the aqueous suspension containing tacrolimus (equivalent to 5 mg/kg) with or without DM--CyD using a catheter. In addition, tacrolimus was completely dissolved in a mixed solution of 5:1:19 (v/w/v) ethanol/polyoxyethylenehydrogenated castor oil 60/water, and was administered intravenously via the jugular vein at a dose of 0.025, 0.05, 0.25, 0.5, or 1.0 mg/kg. Three to five rats were anesthetized with ether, and blood samples (0.3 ml) were drawn 15, 30, 60, 120, 180, 240, 300, 360, and 720 min, and 15, 30, 60, 120, 240, 360, 480, 720, and 1440 min after oral or intravenous administration, respectively, from another rat's jugular vein into which the syringe needle was inserted through the thoracic muscle to prevent bleeding. The whole-blood levels of tacrolimus were assayed using the two-step enzyme immunoassay (EIA) method developed by Tamura et al. (1987) with only minor modification.

Pharmacokinetic Parameters. Pharmacokinetic parameters were calculated by model-independent moment analysis according to the method of Yamaoka et al. (1978). The elimination half-life (t_1/2), area under the concentration-time curve (AUC), mean resident time (MRT), steady-state distribution volume (V_ss), total clearance (CL), maximal blood concentration (C_max), and time to reach C_max (T_max) were estimated.

Cultivation of Caco-2 and Caco-2R Cells. Caco-2R cells were established by continuous exposure of cells to gradually increasing concentrations of vinblastine and were maintained in medium supplemented with vinblastine at 10 ng/ml according to the method of Anderle et al. (1998) with slight modification. Caco-2 and Caco-2R cells between passages 25 and 45 were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 1% nonessential amino acids, 2 mM L-glutamine, 0.45% D-glucose, 100 units/ml penicillin G, and 100 µg/ml streptomycin. For the transport studies, Caco-2 and Caco-2R cells were grown on Transwell microporous polycarbonate membranes (Costar, Cambridge, MA) for 15 to 21 days before use for the transport studies. The transepithelial electric resistance values of Caco-2 and Caco-2R cell monolayers were 761.4 ± 5.1 and 737.9 ± 32.0 /cm², respectively, indicating that there is insignificant difference in the integrity of monolayers between Caco-2 and Caco-2R cells.

Transport Studies across Caco-2 and Caco-2R Cell Monolayers. The confluent cells were washed with transport medium [Hanks' balanced salt solution (HBSS) containing 10 mM HEPES and 25 mM D-glucose], and 1.5 and 2.6 ml of transport medium was added to the apical and basolateral sides, respectively, of a cell insert. To measure the basolateral to apical (BL-to-AP) and the apical to apical-to-basolateral (AP-to-BL) transports, test solution was included in the basolateral and apical sides, respectively. At the designated time, 100 µl of the transport buffer from the basolateral or apical side was withdrawn and replaced with an equal volume of transport buffer, respectively. In the inhibitory experiment using DM--CyD and Tween 20, the apical membranes of Caco-2 and Caco-2 R cell monolayers were pretreated with DM--CyD and Tween 20 at the indicated concentrations for 30 min. After washing the apical membranes, tacrolimus (120 nM) or rhodamine 123 (5 µM) was added to the basolateral or apical side's transport buffer. For other inhibitors, the efflux of tacrolimus or rhodamine 123 was determined in the presence of 5 µM cyclosporin A, 100 µM verapamil, or 100 µM quinidine in the apical and basolateral side's transport buffers without pretreatment. The concentrations of tacrolimus and rhodamine 123 in the transport buffer were determined by EIA method described above and HPLC method, respectively. The HPLC conditions for rhodamine 123 were as follows: a Shimadzu LC-10AD pump and a Shimadzu RF-550A spectrofluorometric detector; a Hitachi D-2500 Chromato-Integrator; a Tosoh TSK gel ODS-80TM column (4.6 × 150 mm); detector wavelength, excitation 507 nm, fluorescence 529 nm; a mobile phase of acetonitrile/1% (v/v) acetic acid (2:3 v/v); a flow rate of 1.0 ml/min. The apparent permeability coefficient (P_app) was calculated using the following equation: P_app = (dQ/dt)/(A · C₀), where dQ/dt is the flux across the monolayer (mol/s), A is the surface area of the membrane (cm²), and C₀ is the initial drug concentration (mol/ml).

Western Blotting Analysis. The P-gp retained in Caco-2 cell monolayers and P-gp released from these monolayers were detected by Western blotting. Briefly, Caco-2 cell monolayers were treated with several concentrations of DM--CyD for 30 min, and then lysed for 45 min in ice-cold phosphate-buffered saline containing 3% SDS, 2 mM dithiothreitol, and protease inhibitors (73 µM pepstatin A, 0.1 mM leupeptin, 1 mM phenylmethylsulfonyl fluoride). After determining protein concentrations using the bicinchoninic acid reagent from Pierce Chemical (Rockford, IL) with bovine serum albumin as a standard, samples (40 µg as protein) were separated by 7% SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes (NEN Life Science Products). The membranes were blocked with 5% skim milk in phosphate-buffered saline containing 0.1% (w/v) Tween 20 and incubated with C219 anti-human P-gp monoclonal antibody overnight at 4°C. After washing, the membranes were incubated with peroxidase-conjugated sheep anti-mouse IgG and washed three times. Specific bands were detected using the enhanced chemiluminescence Western blotting analysis kit (Amersham-Pharmacia Biotech, Buckinghamshire, UK) according to the manufacturer's protocol.

Semiquantitative RT-PCR Analysis. Caco-2 and Caco-2R cell monolayers were treated with several concentrations of DM--CyD for 30 min. Then, the monolayers were scraped off the membranes by treatment with 0.05% trypsin/0.02% EDTA. The cells were lysed and total RNA was extracted from the cells using a RNeasy mini kit (Qiagen, Tokyo, Japan). Then, the samples were treated with deoxyribonuclease I (1 unit) and ribonuclease inhibitor (35 units) for 30 min at 37°C and complementary DNA was synthesized by reverse transcription using human MDR1 gene reverse primer (5'-TTCTGGATGGTGGACAGG-CGGTG-3') or human -actin gene reverse primer (5'-TTGGCATAGAGGTCTTTACGGA-3') and reverse transcriptase (SuperScript II). Approximately, 0.1 M human MDR1 or -actin gene reverse primer was annealed to 3 µg of total RNA and extended with reverse transcriptase (200 units) in a buffer containing 4 µl of 5× first-strand buffer, 10 mM dithiothreitol, and 0.5 mM deoxynucleotide triphosphates for 50 min at 42°C. PCR amplification was carried out in a Takara PCR thermal cycler (Tokyo, Japan). PCR was conducted in a total volume of 100 µl with 2 µl of complementary DNA, 0.5 µM primers, 1.5 mM MgCl₂, 0.2 mM deoxynucleotide triphosphates, and 2.5 units of Taq DNA polymerase (AmpliTaq Gold). The PCR product for the human MDR1 gene was the forward primer (5'-GAGGTGAAGAAGGGCCAGACG-3'), and that for the human -actin gene was the forward primer (5'-GCACCACACCTTCTACAATGAG-3'). PCR was performed for 25 cycles of denaturation at 94°C for 60 s, annealing at 55°C for 60 s, and extension at 72°C for 120 s. The amplified products were analyzed on low melting temperature agarose gels (2%) containing 0.1 µg/ml ethidium bromide.

Flow Cytometry. Caco-2R cell monolayers were grown on 35-mm tissue culture dishes. After washing three times with HBSS, the monolayers were incubated with HBSS in the absence or presence of DM--CyD at 2.5, 5, or 10 mM for 30 min. In the recovery study, Caco-2R cell monolayers were incubated 4, 8, and 12 h after treatment with DM--CyD at 10.0 mM for 30 min. Caco-2R cells were then released by trypsinization and centrifuged (4000 rpm, 3 min). The precipitates were suspended in 100 µl of HBSS containing 0.1% NaN₃. The cells (2 × 10⁷ cells/100 µl) were stained with PE-labeled UIC2 anti-human P-gp antibody for 30 min at 4°C. After washing twice with HBSS, the cells were resuspended in HBSS and filtered through nylon mesh. Data were collected for 5 × 10⁴ cells on a FACSCalibur flow cytometer using CellQuest software (Becton-Dickinson, Mountain View, CA).

Silver Staining. Total proteins released from the Caco-2 cell monolayers into the transport buffer on the apical side were treated in the same manner as described for Western blotting, and then loaded onto gels and subjected to electrophoresis. Gels were stained with a silver staining kit (Wako Pure Chemicals) according to the manufacturer's instructions.

Statistical Analysis. Data are given as means ± S.E.M. Statistical significance of mean coefficients was performed by analysis of variance followed by one-factor ANOVA, with p < 0.05 considered to be statistically significant

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Dissolution Profiles of Tacrolimus and DM--CyD Complex. Dissolution of a drug in aqueous solution is the rate-limiting step for the absorption of poorly water-soluble drugs. Figure 1, A and B, shows the effects of DM--CyD on the dissolution rate of tacrolimus in the various molar ratios by the powder method and the rotating disk method, respectively. The concentration of tacrolimus dissolved in JPXIII second fluid increased with increases in the molar ratio of DM--CyD in both studies. As shown in Fig. 1B, the dissolution profiles of tacrolimus and DM--CyD complex were almost linear, but the dissolution of tacrolimus alone could not be detected. The dissolution rates determined from the slopes in Fig. 1B increased with increasing molar ratio: 3.87, 4.46, 4.74, 5.62, and 6.12 µM/min at molar ratios of 1:10, 1:20, 1:30, 1:40, and 1:50, respectively. These faster dissolution rates of the complexes may have been due to an increase in tacrolimus solubility by complexation with DM--CyD.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Dissolution profiles of tacrolimus or its DM--CyD complexes with various molar ratios. The dissolution profiles were measured by the dispersed amount method (A) and the rotating disk method (B) in JPXIII second fluid (pH 6.8) at 37°C. , tacrolimus alone; , DM--CyD complex in a molar ratio of 1:10 (tacrolimus: DM--CyD); , DM--CyD complex in a molar ratio of 1:20; , DM--CyD complex in a molar ratio of 1:30; , DM--CyD complex in a molar ratio of 1:40; , DM--CyD complex in a molar ratio of 1:50. Each point represents the mean ± S.E.M. of three experiments.

In Vivo Oral Absorption of Tacrolimus and DM--CyD Complex in Rats. Based on the in vitro dissolution study, we postulated that DM--CyD may improve the oral bioavailability of tacrolimus, depending on the molar ratio of tacrolimus/DM--CyD complex. Figure 2A shows the blood concentration-time profiles of tacrolimus after oral administration of tacrolimus suspensions with DM--CyD at various molar ratios to rats. In the case of tacrolimus alone, the tacrolimus levels in whole blood were very low, consistent with the report of Iwasaki et al. (1991). When tacrolimus suspensions with DM--CyD were orally administered, the tacrolimus levels in blood increased with increases in the molar ratio of DM--CyD. In addition, complexation with DM--CyD tended to increase the AUC and C_max values, whereas T_max and MRT values tended to decrease. The AUC values for the complexes at molar ratios of 1:10, 1:20, 1:30, 1:40, and 1:50 were increased by 2.4, 2.2, 2.7, 5.4, and 8.8 times that of tacrolimus alone, respectively (Table 1). Marked increase in the AUC values of tacrolimus were observed with DM--CyD at molar ratios between 1:30 and 1:40, suggesting a nonlinear relationship between AUC of blood tacrolimus and the effective concentration of tacrolimus in the suspensions. Therefore, the relationship between the AUC value of blood tacrolimus after oral administration and the amount of tacrolimus dissolved in the suspension for oral administration was examined. As shown in Fig. 3, A and B, nonlinear relationships were observed between the AUC values of blood tacrolimus and the dissolved amounts of tacrolimus, and between the AUC values of blood tacrolimus and the dissolution rates. In contrast, an obvious linear relationship between AUC of blood tacrolimus after intravenous injection and the dose of tacrolimus was observed within the range of AUC after oral administration (0-400 ng h/ml). These results suggested that the nonlinear relationship was due to the saturation of absorption and metabolism of tacrolimus in the gastrointestinal tract of rats, and not from systemic saturation. To confirm this assumption, the effects of coadministration of tacrolimus with cyclosporin A or itraconazole, substrates of both P-gp and the CYP3A subfamily, on the AUC value of the blood concentration-time curves of tacrolimus were examined. The AUC values of tacrolimus after coadministration of tacrolimus with cyclosporin A or itraconazole at a dose of 10 mg/kg orally to rats were 47.87 ± 3.02 ng · h/ml and 29.60 ± 3.53 ng · h/ml and these values were significantly higher than that of control (23.25 ± 3.78 ng · h/ml) by 2.1- and 1.3-fold, respectively. These results suggested that the nonlinear pharmacokinetics of tacrolimus was dependent on P-gp and the CYP3A subfamily under the present experimental conditions. Taken together, these observations suggested that the increase in AUC value of tacrolimus by DM--CyD may be caused by not only the solubilizing activity of DM--CyD but also inhibition of the efflux and/or metabolism in the gastrointestinal tract.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Whole-blood levels of tacrolimus after oral administration of the suspension containing tacrolimus with or without DM--CyD to rats. The inset shows the results of same experiment shown in a small scale. The dose of tacrolimus for oral administration was equivalent to 5 mg/kg. , tacrolimus alone; , with DM--CyD in a molar ratio of 1:10; , with DM--CyD in a molar ratio of 1:20; , with DM--CyD in a molar ratio of 1:30; , with DM--CyD in a molar ratio of 1:40; , with DM--CyD in a molar ratio of 1:50. Each point represents the mean ± S.E.M. of three to five rats.

View this table: [in this window] [in a new window]   TABLE 1 Pharmacokinetic parameters of tacrolimus after oral administration of tacrolimus suspension (equivalent to 5 mg/kg tacrolimus) with or without DM--CyD in various molar ratios (tacrolimus:DM--CyD) to rats Each value represents the mean ± S.E.M. of four rats.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Relationship between AUC value of blood tacrolimus after oral administration and the amount of tacrolimus dissolved in the oral preparation (A), between the AUC value and the dissolution rate of tacrolimus (B), and between AUC value of blood tacrolimus after intravenous administration and dose of tacrolimus (C). The amounts of tacrolimus dissolved were determined by HPLC. The dissolution rates of tacrolimus were calculated from the results of the rotating disk method shown in Fig. 1B. The whole-blood levels of tacrolimus after intravenous injection to rats at doses of 0.025, 0.05, 0.25, 0.5, or 1 mg/kg were determined by EIA. The correlation coefficient between AUC and dose of tacrolimus in Fig. 3C was 0.993.

Transport of Tacrolimus and Rhodamine 123 across Caco-2 and Caco-2R Cell Monolayers. To elucidate whether DM--CyD inhibits the P-gp-mediated efflux of tacrolimus from the rat gastrointestinal epithelial cells, we examined the effects of DM--CyD on BL-to-AP transport using Caco-2 cell monolayers because these cells were reported to possess only very slight CYP3A4 activity (Schmiedlin-Ren et al., 1997), indicating that the metabolism of tacrolimus in these cells would be almost negligible under the present experimental conditions. In addition, we used Caco-2R cells because we could detect P-gp expressed on the cell surface and MDR1 mRNA in Caco-2R cells, but not Caco-2 cells as described below. Figure 4, A and B, shows the effects of pretreatment with DM--CyD for 30 min on the BL-to-AP and AP-to-BL permeation of tacrolimus through Caco-2 cell monolayers, respectively. Here, the monolayers were pretreated with DM--CyD and Tween 20 (Fig. 5) to prevent the direct interactions of tacrolimus and rhodamine 123 with DM--CyD and Tween 20 through complexation and micelle formation. The permeation of tacrolimus was significantly decreased by pretreatment of the apical membranes of Caco-2 cell monolayers with DM--CyD: the P_app values were 9.36 × 10⁶, 8.83 × 10⁶, 6.49 × 10⁶, and 6.21 × 10⁶ cm/s at 0, 2.5, 5, and 10 mM DM--CyD, respectively. However, DM--CyD showed neither cytotoxicity nor changes in transepithelial electric resistance, mannitol, or phenacetin transport under the present experimental conditions (data not shown). Comparing between the BL-to-AP and AP-to-BL permeations of tacrolimus (Fig. 4, A and B), the AP-to-BL permeation of tacrolimus was slower than the BL-to-AP permeation in the absence of DM--CyD, suggesting that tacrolimus is subject to efflux from Caco-2 cells. On the other hand, the enhancing effect of DM--CyD on the AP-to-BL permeation of tacrolimus was almost not observable (Fig. 4B). This may reflect the result that the inhibitory effect of DM--CyD on the efflux of tacrolimus in Caco-2 cell monolayers is not very great under this experimental condition.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Concentration dependence of the effects of pretreatment with DM--CyD on the BL-to-AP (A) and the AP-to-BL (B) permeations of tacrolimus through Caco-2 cell monolayers. The apical membranes of Caco-2 cell monolayers were pretreated with DM--CyD at indicated concentrations for 30 min. After washing the apical membranes, tacrolimus was added to transport buffer on the basolateral or apical side (120 nM) and the concentration of tacrolimus in the transport buffer on the apical or basolateral side was determined by EIA. , without pretreatment with DM--CyD (control); , pretreated with DM--CyD (2.5 mM); , pretreated with DM--CyD (5 mM); , pretreated with DM--CyD (10 mM). Each point represents the mean ± S.E.M. of three to five experiments. *p < 0.05 versus control.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Inhibitory effects of various agents on the BL-to-AP permeation of tacrolimus through Caco-2 (A) or Caco-2R (B) cell monolayers. , without pretreatment with DM--CyD and Tween 20 (control). For DM--CyD and Tween 20, the apical membranes of Caco-2 and Caco-2 R cell monolayers were pretreated with 10 mM DM--CyD () and 6 µM Tween 20 () for 30 min. After washing the apical membranes, tacrolimus was added to the transport buffer on the basolateral side (120 nM). For other inhibitors, the efflux of tacrolimus was determined in the presence of 5 M cyclosporin A (), 100 M verapamil () or 100 M quinidine () in the transport buffer on apical and basolateral sides without pretreatment. The concentration of tacrolimus in the transport buffer on the apical side was determined by EIA. Each point represents the mean ± S.E.M. of three to seven experiments. *p < 0.05 versus control.

6

6

6

6

6

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Inhibitory effects of various agents on the BL-to-AP permeation of rhodamine 123 through Caco-2 (A) and Caco-2R (B) cell monolayers. , without pretreatment with DM--CyD and Tween 20 (control). For DM--CyD and Tween 20, the apical membranes of Caco-2 and Caco-2 R cell monolayers were pretreated with 10 mM DM--CyD () and 6 M Tween 20 () for 30 min. After washing the apical membranes, rhodamine 123 was added to the transport buffer on the basolateral side (5 µM). For other inhibitors, the efflux of tacrolimus was determined in the presence of 5 µM cyclosporin A (), 100 µM verapamil (), or 100 µM quinidine () in the transport buffers on the apical and basolateral sides without pretreatment. The concentration of rhodamine 123 in the transport buffer of the apical side was determined by HPLC. Each point represents the mean ± S.E.M. of three to seven experiments. *p < 0.05 versus control.

Effects of DM--CyD on P-gp Expression in Caco-2 and Caco-2R Cells. To determine the mechanism responsible for the inhibitory effect of DM--CyD on the efflux of tacrolimus via P-gp in Caco-2 and Caco-2R cell monolayers, we examined whether DM--CyD released P-gp from the monolayers since DM--CyD releases membrane components such as phospholipids, cholesterols, and proteins from biological membranes, resulting in hemolysis of erythrocytes at higher concentrations (Irie and Uekama, 1997). Figure 7A shows the results of immunoblotting analysis of P-gp in the Caco-2 cell monolayers. Here, the reason why we used Caco-2 cells, not Caco-2R cells, is that P-gp expression is so high in Caco-2R cells that the band corresponding to P-gp was rather broad, and we could not evaluate the effects of DM--CyD on the P-gp levels in Caco-2R cells by Western blotting analysis. Treatment with DM--CyD at concentration of 10 mM decreased the density of the 170-kDa band derived from P-gp in Caco-2 cell monolayers. In addition, P-gp was detected by immunoblotting analysis in the transport buffer on the apical side (Fig. 7B). Moreover, we studied the effects of DM--CyD on the P-gp levels on the apical side surface of Caco-2R cell monolayers by flow cytometry using PE-labeled UIC2 antibody. As shown in Fig. 8A, DM--CyD decreased the P-gp levels on the cell surface in a concentration-dependent manner. In addition, the P-gp expression after treatment with DM--CyD at 10 mM recovered within 12 h to control level (Fig. 8B). On the other hand, the decrease in P-gp by treatment with DM--CyD might be attributable to the suppression of gene expression in Caco-2 and Caco-2R cell monolayers. To confirm this possibility, we performed RT-PCR analysis. As shown in Fig. 9, A and B, RT-PCR analysis indicated that treatment with DM--CyD did not change the band density derived from MDR1 mRNA in Caco-2 or Caco-2R cell monolayers, suggesting that DM--CyD changes the MDR1 gene expression only very slightly. These results suggested that DM--CyD decreases P-gp levels in Caco-2 cell monolayers without changing MDR1 gene expression. In addition, to determine whether the release of P-gp by treatment with DM--CyD is specific, we performed silver staining for the total proteins released from the apical membrane into the transport buffer on the apical side. As expected, some bands were detected, indicating the nonspecific effect of DM--CyD on the release of proteins (Fig. 7C). Taken together, these results suggested that DM--CyD releases P-gp and some other proteins from the apical membranes of Caco-2 cell monolayers in a nonspecific manner, resulting in the inhibition of P-gp-mediated efflux of tacrolimus.

View larger version (42K): [in this window] [in a new window]   Fig. 7.   Western blotting analysis of P-gp (A and B). The apical membranes of Caco-2 cell monolayers were treated with DM--CyD for 30 min at 2.5, 5, or 10 mM. P-gp retained in the cells (A) and released in the transport buffer on the apical side (B) was detected by Western blotting using C219 anti-human P-gp antibody. The total proteins released from the apical membranes into the transport buffer of the apical side after treatment with DM--CyD for 30 min were detected by silver staining (C).

View larger version (28K): [in this window] [in a new window]   Fig. 8.   Flow cytometric analysis of P-gp on the apical membranes of Caco-2R cell monolayers (A) and the recovery of P-gp on the apical membrane (B). Caco-2R cell monolayers were treated with DM--CyD at 2.5, 5, and 10 mM for 30 min. In the recovery study, Caco-2R cell monolayers were incubated 4, 8, or 12 h after treatment with DM--CyD (10 mM) for 30 min. P-gp on the apical membrane of the cell monolayers was detected by flow cytometry using PE conjugate with anti-human P-gp UIC2 monoclonal antibody.

View larger version (19K): [in this window] [in a new window]   Fig. 9.   RT-PCR analysis of MDR1 mRNA expression in Caco-2 (A) and Caco-2R (B) cell monolayers. The apical membranes of Caco-2 and Caco-2R cell monolayers were treated with DM--CyD at 2.5, 5, or 10 mM for 30 min. Total RNA was extracted and then RT-PCR was performed using the primers for human MDR1 gene or human -actin gene.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The results of the present study demonstrated that the enhancing effect of DM--CyD on the oral bioavailability of tacrolimus is due not only to its solubilizing effect on tacrolimus but also, at least in part, to the inhibitory effect of DM--CyD on the P-gp-mediated efflux of tacrolimus in the gastrointestinal tract.

Of all hydrophilic CyD derivatives examined in our study, DM--CyD showed the greatest solubilizing effect on hydrophobic drugs, probably due to a larger hydrophobic space in the CyD cavity, and increased the dissolution rate, stability, and bioavailability of many drugs (Uekama et al., 1998). Similarly, DM--CyD was reported to interact with membrane constituents such as cholesterol and phospholipids, resulting in the induction of hemolysis of erythrocytes and the disruption of caveolae and rafts (Kilsdonk et al., 1995; Irie and Uekama, 1997; Sheets et al., 1999; Hansen et al., 2000; Verkade et al., 2000). These observations suggest that DM--CyD is susceptible to cell-surface proteins rather than intracellular proteins.

Tacrolimus is a substrate of P-gp and the CYP3A subfamily, and thereby is subject to first pass metabolism. In mdr1a knockout mice, the total clearance after intravenous injection of tacrolimus was decreased to one-third, whereas oral bioavailability of tacrolimus was increased by 3-fold relative to those of wild-type mice, clearly indicating that the pharmacokinetics of tacrolimus is dependent on P-gp (Yokogawa et al., 1999; Chiou et al., 2000). In addition, 34% of tacrolimus was metabolized in the gastrointestinal tract after oral administration to rats, probably by the CYP3A subfamily. In addition, it has been reported that the blood levels of tacrolimus were increased by approximately 4-fold by oral coadministration with diltiazem in humans, indicating that diltiazem inhibited P-gp activity and CYP3A4 metabolism (Hebert and Lam, 1999). These findings indicated that the bioavailability of tacrolimus after oral administration in rats could change significantly in the presence of the inducers and inhibitors of P-gp or the CYP3A subfamily. We showed here that DM--CyD significantly increased the bioavailability of tacrolimus with increases in the molar ratio of 1:40 and 1:50 of the complexes (Fig. 2), reflecting the enhancing effects of DM--CyD on the dissolution rate of tacrolimus (Fig. 1). However, a nonlinear relationship was observed between AUC value and solubilizing effect of DM--CyD (Fig. 3, A and B), although linear disposition behavior of tacrolimus was observed after intravenous administration (Fig. 3C). These observations suggested that factors other than the fast dissolution may be involved in the enhancing effect of DM--CyD on the oral bioavailability of tacrolimus. In addition, the threshold for the enhancing effect of DM--CyD on tacrolimus bioavailability could also imply that a threshold of the DM--CyD concentration is necessary to exert its inhibitory effect on the P-gp function and CYP3A4 metabolism in vivo.

Hydrophilic CyDs, including DM--CyD, are thought to be unable to pass through the gastrointestinal epithelial cells. Thus, DM--CyD tends to interact with the constituents on the cell surface, and thereby DM--CyD may inhibit the efflux pump activity of P-gp rather than the metabolic activity of CYP3A4. We evaluated the effects of DM--CyD on P-gp activity using Caco-2 cells because this is a well established cell line for drug transport studies in the gastrointestinal tract that possesses P-gp activity but not CYP3A4 activity (Schmiedlin-Ren et al., 1997). In addition, we also used Caco-2R cells due to their high level of P-gp expression (data not shown), consistent with the reports of Hoskins et al. (1993) and Anderle et al. (1998) to facilitate evaluation of the effects of additives on P-gp. In the present transport study, DM--CyD inhibited the efflux of tacrolimus and rhodamine 123 from Caco-2 and Caco-2R cell monolayers (Figs. 5 and 6). In addition, flow cytometric analysis indicated that DM--CyD increased the accumulation of rhodamine 123 in the Caco-2R cell monolayers in a concentration-dependent manner (data not shown). These results indicated that DM--CyD inhibits the efflux of tacrolimus and rhodamine 123 from Caco-2 and Caco-2R cell monolayers. However, the enhancing effect of DM--CyD on AP-to-BL permeation of tacrolimus was not almost observed (Fig. 4B). This, in turn, is consistent with the observation that the inhibitory effect of DM--CyD on the BL-to-AP permeation of tacrolimus in Caco-2 cell monolayers and on P-gp activity is not as strong as with other P-gp inhibitors. In addition, the inhibitory effect of DM--CyD on the efflux of tacrolimus was lower than that of rhodamine 123 (Figs. 5 and 6). This different inhibitory effect of DM--CyD could be attributed to differences in the contribution of P-gp to the efflux of tacrolimus and rhodamine 123 from Caco-2 and Caco-2R cell monolayers because the inhibitory effect of P-gp on the intestinal permeability of tacrolimus is expected to be slight in mdr1a knockout mice (Chiou et al., 2000). In addition, the inhibitory effects of DM--CyD on the permeation of tacrolimus and rhodamine 123 were more significant in the Caco-2R than in Caco-2 cells (Figs. 5 and 6). These results suggested that DM--CyD inhibits the efflux activity of P-gp.

However, there is a discrepancy as to the difference in the permeation of rhodamine 123 between Caco-2 and Caco-2R cells: the BL-to-AP permeation of rhodamine 123 in Caco-2 cells was higher than that in Caco-2R cells as shown in Fig. 6, despite that P-gp expression in Caco-2R cell monolayers is significantly greater than in Caco-2 cell monolayers (data not shown). At present, we do not know the reason. Further study to explain this discrepancy should be performed.

The mechanism by which DM--CyD inhibits P-gp activity may be different from that of P-gp inhibitors such as cyclosporin A, itraconazole, quinidine, and verapamil. P-gp recognizes many compounds as substrates, and tends to have high affinity to hydrophobic and positively charged compounds at physiological pH. DM--CyD appears not to be a substrate of P-gp because it is a hydrophilic and electrically neutral cyclic oligosaccharide with a relatively high molecular weight (molecular weight = 1331). Furthermore, DM--CyD should not compete with P-gp substrates due to its lack of cell permeability. Thus, DM--CyD must have an alternative inhibitory effect on P-gp activity, differing from the P-gp inhibitors described above.

The surface-active property of DM--CyD may contribute only slightly to the increase in the oral absorption of tacrolimus. Some surfactants have been shown to inhibit P-gp function. Thus, it was postulated that DM--CyD may suppress the BL-to-AP permeation of tacrolimus via its surface-active property. However, 6 µM Tween 20, the surface activity of which is identical to 10 mM DM--CyD, showed no inhibitory effect on the efflux of rhodamine 123 (Figs. 5 and 6), suggesting that the inclusion ability of DM--CyD may be involved in its inhibitory effect on P-gp function.

DM--CyD seems to inhibit the P-gp activity by allowing release of P-gp from the apical membrane of Caco-2 cells. In the present study, immunoblotting analysis indicated that DM--CyD decreased the P-gp level in the Caco-2 cells and P-gp was released from the apical membranes into the cell supernatant transport buffer (Fig. 7, A and B). This finding was supported by the results of flow cytometric analysis, which indicated that the P-gp level on the Caco-2R cell surface was decreased by pretreatment with DM--CyD (Fig. 8A). To our knowledge, this is the first report of a direct inhibitory effect of DM--CyD on the P-gp level on the cell surface. Furthermore, RT-PCR analysis revealed that DM--CyD affected MDR1 gene expression only very slightly (Fig. 9). This might be explained by the impermeability of DM--CyD through the membranes. Moreover, the silver staining demonstrated that the ability of DM--CyD to induce release of P-gp was not specific (Fig. 7C). These results also appear to be reasonable since DM--CyD interacts with various kinds of molecules on the cell surface.

It should be emphasized that DM--CyD showed neither cytotoxicity toward Caco-2 and Caco-2R cells as demonstrated by MTT assay, nor had any effect on the permeation of mannitol and phenacetin (markers of paracellular and transcellular transport, respectively) through the monolayers under the present experimental conditions (data not shown). Moreover, we showed that P-gp expression on the cell surface was recovered within 12 h after treatment with DM--CyD (Fig. 8B). Therefore, the inhibitory effect of DM--CyD on the P-gp activity seemed to be noncytotoxic and transient. An in vivo toxicological study showed that oral administration of DM--CyD (in aqueous solution) to male and female mice resulted in nontoxic symptoms up to 3 g/kg (Szejtli, 1983), indicating that DM--CyD may be less toxic under in vivo conditions because the maximal dose of DM--CyD was about 405 mg/kg. Therefore, it is apparent that the enhancing effect of DM--CyD on the bioavailability of tacrolimus in rats was not attributable to a toxic effect in the gastrointestinal tract in rats under the present experimental conditions.

It is important to determine that DM--CyD affects not only many other substrates of P-gp such as rhodamine 123 as described here but also other transporters on the apical membranes and/or metabolism mediated by members of the CYP3A subfamily. We are currently planning experiments to study the effects of DM--CyD on these factors using Caco-2 transfected with the CYP3A4 gene and Caco-2R cells.

In conclusion, the present results suggested that the enhancing effect of DM--CyD on the oral bioavailability of tacrolimus is attributable to not only its solubilizing effect but also its inhibitory effect on the P-gp-mediated efflux of tacrolimus from intestinal epithelial cells.