Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

We were interested in the possibility that orange juice might interact with drugs in a manner similar to grapefruit juice (GFJ) because both juices should contain similar classes of compounds. GFJ is well known to change the pharmacokinetic parameters of a variety of drugs, including dihydropyridine calcium channel blockers, cyclosporin A, midazolam, triazolam, terfenadine, saquinavir, and ethynylestradiol (Bailey et al., 1998). Inasmuch as these drugs are metabolized primarily by cytochrome P450 (CYP) 3A4, it has been suggested that the effect of grapefruit juice may be due to the inhibition of CYP3A4 activity (Guengerich and Kim, 1990; Miniscalco et al., 1992). However, we have shown that GFJ components not only inhibit drug metabolism by CYP3A4, but also enhance drug absorption by the inhibition of drug efflux via P-glycoprotein (P-gp) in a human colon carcinoma cell line, Caco-2 cells, used as a model of intestinal absorption (Takanaga et al., 1998). Furthermore, we have identified 6',7'-dihydroxybergamottin (5-[(6,7-dihydroxy-6-keto-2-octenyl)oxy]psoralen; DHBG) and other furanocoumarin derivatives as P-gp inhibitors in GFJ (A.O., H.M., S.Y., H.T., S.M., M. Shoyama, H.O., and Y.S., submitted). DHBG was also identified as an inhibitor of CYP3A4-mediated drug metabolism in a study using organic solvent extracts of GFJ (Edwards et al., 1996). Striking overlaps of substrates and inhibitors between CYP3A4 and P-gp were reported by Wacher et al. (1995). It was speculated that the effect of GFJ components is due to suicide inhibition of CYP3A4 (Lown et al., 1997), but the findings by Wacher et al. are in conjunction with our results, suggesting that parallel mechanisms may underlie GFJ's action.

The above findings are potentially significant because flavonoids are widely present in vegetables, fruits, and beverages (Herrmann, 1988) and are known to have a wide range of biochemical and pharmacological effects. For example, they inhibit oxidative drug metabolism (Buening et al., 1981; Siess et al., 1989; Guengerich and Kim, 1990), exhibit anti-inflammatory and antiallergic effects (Middleton and Kandaswami, 1992), and inhibit chemical carcinogenesis (Mukhtar et al., 1988; Verma et al., 1988). We also showed that aglycones of several flavonoids inhibited P-gp function in a concentration-dependent manner, whereas their glycosides were inactive (Y. Mitsunaga, H.T., H.M., M. Naito, T. Tsuruo, H.O., and Y.S., submitted). Flavonols such as quercetin and kaempferol showed a biphasic effect on vincristine efflux, causing inhibition or enhancement depending on their concentration. The effect of flavonoids on the drug efflux pump seems to be complicated.

Orange juice has been used as a control for GFJ in a CYP3A4 inhibition study (Edwards and Bernier, 1996). We examined whether orange juice could be used as a control in drug transport studies and found that it inhibits vinblastine efflux from cells the same as GFJ. DHBG is a minor component of orange juice (orange juice, 1.11 µM; GFJ, 12.8 µM; A.O., H.T., H.M., and Y.S., unpublished data), so we considered that some other component was probably responsible. This study was conducted to identify the active components inhibiting vinblastine efflux from Caco-2 cells and to see whether they are also inhibitors of CYP3A4. CYP3A is an important enzyme for drug metabolism not only in the liver but also in the intestine (Lang et al., 1996). Therefore, it is considered clinically significant to study the effect of the components in orange juice on CYP3A4 activity.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [³H]Vinblastine sulfate (14.3 mCi mmol¹), 3-O- [³H]methylglucose (11 mCi mmol¹), and [¹⁴C]phenylalanine (448 mCi mmol¹) were purchased from Amersham International (Buckinghamshire, UK). Orange juice was produced by the Dole Food Company Inc. (Westlake Village, CA). Tangeretin was purchased from EXTRASYNSTHESE (Genay, France), and nobiletin was kindly supplied by Kanebo Yakuhin, Ltd. (Tokyo, Japan). Human CYP3A4 and human liver microsomes were obtained from GENTEST Corporation (Woburn, MA). Testosterone was purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan), and corticosterone was obtained from Nacalai Tesque, Inc. (Kyoto, Japan). NADP⁺, glucose 6-phosphate, and glucose-6-phosphate dehydrogenase were obtained from Oriental Yeast, Ltd. (Osaka, Japan), and 6-hydroxytestosterone was obtained from Ultrafine Chemicals (Manchester, UK). All other chemicals were commercial products of reagent grade.

Cell Culture. Caco-2 cells were obtained from the American Type Culture Collection (Manassas, VA) and grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 1% nonessential amino acids, 2 mM L-glutamine, 100 U ml¹ of penicillin G, and 100 µg ml¹ of streptomycin at 37°C in a humidified atmosphere of 5% CO₂, 95% air. All cells in this study were between passages 55 and 72. LLC-PK1 cells (porcine kidney epithelial cell line) and LLC-GA5-COL300 cells (a transformant cell line derived by transfecting LLC-PK1 with human MDR1 cDNA isolated from normal adrenal gland) were obtained from Riken Cell Bank (Ibaraki, Japan). LLC-PK1 cells were grown in M199 medium supplemented with 10% fetal calf serum at 37°C in a humidified atmosphere of 5% CO₂, 95% air, as reported previously (Tanigawara et al., 1992; Ueda et al., 1992). LLC-GA5-COL300 cell line was obtained by selection with 300 ng/ml colchicine and cultured in M199 medium supplemented with 10% fetal calf serum and 300 ng/ml colchicine at 37°C in a humidified atmosphere of 5% CO₂, 95% air.

Viability Check of Caco-2 Cells in the Presence of Orange Juice. We checked the viability of Caco-2 cells in the presence of orange juice. To detect the cytotoxicity by orange juice, we assayed the release of lactate dehydrogenase (LDH), the cytosolic enzyme. LDH activity was measured with the LDH-UV test kit (Wako Pure Chemical Industries, Ltd.). The release of LDH in the presence and absence of 50% (v/v) orange juice was 2.90 ± 0.98 and 1.92 ± 0.76% of total LDH in Caco-2 cells, respectively (no significant difference). The trypan blue exclusion test was performed to evaluate the viability of the cells. There was no change in the viability of the cells in the presence of orange juice, and the percentage of blue-stained cells was <5%. Results indicate that orange juice did not affect the viability of caco-2 cells.

Extraction of Orange Juice Components with Organic Solvents. Extraction of orange juice was performed as we reported previously for GFJ (Takanaga et al., 1998). Orange juice (10 ml) was mixed with 40 ml of ethyl acetate, diethyl ether, or methylene chloride and shaken vigorously for 10 min. The mixtures were centrifuged at 900g for 10 min, the aqueous phase was removed, and the organic solvent layer was evaporated to dryness. The residue was dissolved in 100 µl of dimethyl sulfoxide and added to Hanks' balanced salt solution (HBSS) to give a final dimethyl sulfoxide concentration of 0.5% for uptake experiments.

Separation of Active Components in Orange Juice. Separation of active components in orange juice was performed as follows. Orange juice was extracted with ethyl acetate, and the organic layer was evaporated. The residue was dissolved in water and applied to a Cosmosil 75C18-OPN column (50- × 4-cm i.d.; Nacalai Tesque). Elution with 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100% methanol afforded 11 fractions that were assayed for effect on [³H]vinblastine uptake by Caco-2 cells. The most active fraction was subjected to additional separation on a silica gel column (silica gel 60, Merck) and eluted with hexane/acetone (5:1, 3:1, and 1:1) and then chloroform/methanol (1:1). The eluates were separated into ~16 fractions that were evaporated to dryness, and each residue was assayed for effect on [³H]vinblastine uptake by Caco-2 cells. The most active fraction (fraction 2-3) was also chromatographed over a Cosmosil 75C18 column to give compounds 1 and 2, which were confirmed by thin-layer chromatography (Kieselgel 60 F254, Merck) with hexane/acetone to be purified to homogeneity. The NMR spectra of both compounds were taken with a Varian unity-plus spectrometer using tetramethylsilane as an internal standard, and chemical shift values are given in  (ppm). Compound 1¹H-NMR (CDCl₃): 3.87, 3.93, 3.95, 3.96, 3.98, 4.08 (21H in total, each s, OMe × 7), 6.99 (1H, d, J=8 Hz, H-5'), 7.79 (1H, d, J = 2 Hz, H-2'), 7.82 (1H, dd, J = 2, 8 Hz, H-6'). ¹³C-NMR (CDCl₃): 55.9, 56.0, 59.8, 61.6, 61.8, 61.9 (OMe × 7), 111.0 (C-2'), 111.1 (C-5'), 115.1 (C-10), 121.9 (C-6'), 123.5 (C-1'), 137.8 (C-6), 140.8 (C-3), 143.8 (C-5), 148.2 (C-9), 148.8 (C-3'), 151.1 (C-2), 151.3 (C-7), 153.1 (C-4'), 173.9 (C-4). Compound 2¹H-NMR (CDCl₃): 3.87, 3.93, 4.00. 4.08 (15H in total, each s, OMe × 5), 6.59 (1H, s, H-3), 7.01 (2H, d, J = 9 Hz, H-3',5'), 7.85 (1H, d, J = 9 Hz, H-2', 6').

Uptake Experiments with Caco-2 Cells. Caco-2 cells were seeded at 1.26 × 10⁵ cells on four-well multidishes (Nunc, Roskilde, Denmark) for 13 to 15 days and cultured to confluency. The culture medium was replaced on alternate days. For uptake experiments, the culture medium of Caco-2 cells on the multidishes was removed, and the cells were washed once or twice with HBSS (136.7 mM NaCl, 5.4 mM KCl, 0.95 mM CaCl₂·2H₂O, 0.81 mM MgSO₄·7H₂O, 0.44 mM KH₂PO₄, 0.39 mM Na₂HPO₄·12H₂O, 25 mM D-glucose, 10 mM MES) at 37°C and pH 6.5. Uptake experiments were performed in 250 µl of incubation buffer containing 10 nM [³H]vinblastine in the presence or absence of one of the following: 20 µM cyclosporin, the organic layer or the water layer from ethyl acetate extraction of orange juice, individual fractions of the ethyl acetate extract of orange juice, and individual flavones. In some experiments, 500 nM 3-O-[³H]methylglucose or 500 nM [¹⁴C]phenylalanine was used in place of [³H]vinblastine. After incubation, the cells were washed two or three times with ice-cold HBSS to stop the uptake. After the uptake experiments, cells were dissolved in 1 M NaOH (250 µl) and neutralized with 1 M HCl (250 µl). To assay radiolabeled compounds, samples were transferred into counting vials and mixed with scintillation fluid (Clearsol I, Nacalai Tesque), and the radioactivity was measured with a liquid scintillation counter (LS6500; Beckman Instruments, Inc., Fullerton, CA). The amount of protein in the Caco-2 cells in uptake studies was measured by Lowry's method (Lowry et al., 1951). The uptake of [³H]vinblastine, 3-O-[³H]methylglucose, or [¹⁴C]phenylalanine was expressed as the ratio of uptake amount per milligram of protein of cells to the drug concentration (microliters per milligram of protein).

Uptake Experiments with LLC-PK1 and LLC-GA5-COL300 Cells. Cells were seeded on four-well multidishes (Nunc) at a density of 1.75 × 10⁵ for LLC-PK1 cells and 3.125 × 10⁵ for LLC-GA5-COL300 cells. The cells were grown for 3 days, and the culture medium was replaced with fresh medium, without colchicine, 6 h before the uptake experiments. For the experiments, the culture medium of LLC-PK1 cells and LLC-GA5-COL300 cells on Microtest tissue culture plates was removed, and the cells were washed one or two times with incubation buffer (141 mM NaCl, 4 mM KCl, 2.8 mM CaCl₂, 1 mM MgSO₄, 10 mM D-glucose, 10 mM HEPES) at 37°C and pH 7.4. Uptake experiments were performed in 250 µl of incubation buffer containing 20 nM [³H]vinblastine in the presence or absence of 10 µM cyclosporin, an orange juice component, or flavones. After incubation, the cells were washed two or three times with ice-cold buffer to stop the uptake. After the uptake experiments, cells were dissolved with 1 M NaOH (250 µl), and the lysate was neutralized with 1 M HCl (250 µl). Assay of radiolabeled compounds, measurement of protein in LLC-PK1 cells and LLC-GA5-COL300 cells, and measurement of the uptake of [³H]vinblastine were performed as described above.

Assay of Testosterone 6-Hydroxylation by Human CYP3A4 and Human Liver Microsomes. CYP3A4 activity was assayed by measuring the formation of 6-hydroxytestosterone from testosterone by human CYP3A4 and human liver microsomes. Human CYP3A4 was used for determination of the 6-hydroxylation-inhibitory activity in orange juice extract, and human liver microsomes were used for assay of the effects of flavones such as 3,3',4',5,6,7,8-heptamethoxyflavone (HMF), tangeretin, and nobiletin. The reaction mixture, consisting of 1.3 mM NADP, 3.3 mM glucose 6-phosphate, 0.4 U/ml of glucose-6-phosphate dehydrogenase, 3.3 mM magnesium chloride, and 0.2 mM testosterone in 100 mM potassium phosphate buffer (pH 7.4), was preincubated for 5 min at 37°C in the presence or absence of orange juice extract or flavones. The reaction was started by the addition of 1.25 pmol of P450, in the case of human CYP3A4, and 0.05 mg of protein, in the case of human liver microsomes. After incubation for 15 min at 37°C, the reaction was stopped by the addition of methylene chloride. For the studies using human CYP3A4, 20 µM corticosterone was added as an internal standard, followed by shaking for 3 min. After centrifugation for 3 min, the organic layer was taken and evaporated, and the residue was dissolved in methanol for HPLC analysis on a 4.6- × 250-mm 5C18 column (Senshu Pak ODS-H-1251; Senshu Scientific Co., Ltd., Tokyo, Japan) with 60% methanol/water at a flow rate of 1.2 ml/min at 45°C. Metabolites were detected by measuring the absorbance at 242 nm. The amount of 6-hydroxytestosterone produced was quantified by comparison with the internal standard for the study using human CYP3A4 and an absolute calibration curve for the study using human liver microsomes.

Quantification of Tangeretin in Orange Juice. Tangeretin or nobiletin was dissolved at 0.5, 1, 5, 10, and 20 µM in 1 ml of orange juice, and each solution was shaken for 10 min with 3 ml of ethyl acetate. After centrifugation at 900g for 10 min, the organic layer was evaporated under nitrogen gas. The residue was dissolved in methanol and analyzed by HPLC. The amount of HMF, tangeretin, or nobiletin was quantified from the intercept and slope of the calibration curve. HPLC analysis was performed with a Nucleosil column (5C18, 250- × 4.6-mm; Chemco Scientific Co., Ltd., Osaka, Japan), maintained at 35°C. The mobile phase consisted of a multiple gradient of solvent A (acetonitrile) and solvent B [water/acetic acid (96:4, v/v)] as follows: solvent A, 0 min, 0%; 12 min, 8%; 43 min, 34%; 44 min and thereafter, 70%. The flow rate was set at 1 ml/min, and detection was performed by measuring the absorbance at 330 nm.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Effect of Organic Extracts of Orange Juice on the Uptake of [³H]Vinblastine by Caco-2 Cells. The uptake of [³H]vinblastine by Caco-2 cells increased with time and reached a steady state in 60 min (Fig. 1A). The effect of orange juice on the steady-state uptake of [³H]vinblastine at 60 min was investigated. As shown in Table 1, the steady-state uptake was significantly increased to 178 ± 5.5% in the presence of orange juice, compared with the control. Similarly, it was significantly increased to 228 ± 2.8, 203 ± 14, and 198 ± 4.5% of the control in the presence of ethyl acetate, diethyl ether, and methylene chloride extracts of orange juice, respectively, at the concentration equivalent to 50% of the original orange juice strength (Table 1). No significant effect of the remaining aqueous layer after ethyl acetate extraction was seen (100 ± 4.33% compared with the control), indicating that the active components were entirely extracted in ethyl acetate.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Effect of 50% ethyl acetate extract of orange juice and cyclosporin A on the uptake of [3H]vinblastine (A), 3-O-[3H]methylglucose (B), and [14C]phenylalanine (C) by Caco-2 cells. The [3H]vinblastine uptake experiments were performed in the absence () or presence () of ethyl acetate extract of orange juice diluted to be equivalent to 50% of the original orange juice strength, or 10 µM cyclosporin A (). The concentrations of [3H]vinblastine, 3-O-[3H]methylglucose, and [14C]phenylalanine were 10, 500, and 500 nM, respectively. Significant differences from the control were identified by using Student's t test (*P < .05). Each value represents the mean ± S.E. for three or four experiments.

The Effect of the Ethyl Acetate Extract of Orange Juice and Cyclosporin A on the Uptake of [³H]Vinblastine, 3-O-[³H]Methylglucose, [¹⁴C]Phenylalanine by Caco-2 Cells. As shown in Fig. 1A, the uptake of [³H]vinblastine was significantly increased by the ethyl acetate extract of orange juice, as well as by cyclosporin A (20 µM), an inhibitor of P-gp (Tamai and Safa, 1990). Because the initial uptake rate of [³H]vinblastine was not affected by either agent, it was suggested that the enhancement of [³H]vinblastine uptake was due to inhibition of its efflux via P-gp.

Inhibitory Effects of Fractions of the Ethyl Acetate Extract of Orange Juice on the Steady-State Uptake of [³H]Vinblastine by Caco-2 Cells and on 6-Hydroxylation of Testosterone by Recombinant Human CYP3A4. We fractionated the ethyl acetate extract of orange juice on a Cosmosil column with 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100% methanol. Figure 2A shows the effect of the eluates on the steady-state uptake of [³H]vinblastine by Caco-2 cells. Inasmuch as the 60, 70, and 80% methanol eluates caused large increases of [³H]vinblastine uptake, these fractions seemed to contain the major inhibitor of P-gp. However, no significant inhibitory effect on testosterone 6-hydroxylation was observed in any of the methanol eluates (Fig. 2B).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Effect of column-separated fractions of the ethyl acetate extract of orange juice on the steady-state uptake of 10 nM [3H]vinblastine by Caco-2 cells for 60 min (A) and on the activity of testosterone 6-hydroxylation by human CYP3A4 (B). The ethyl acetate extract of orange juice was fractionated by Cosmosil column chromatography eluted with 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100% methanol. The uptake of 10 nM [3H]vinblastine by Caco-2 cells for 60 min and the activity of testosterone 6-hydroxylation by recombinant human CYP3A4 were assayed as described under Experimental Procedures. Control value of 6-hydroxytestosterone formation was 2.89 µM. Each value represents the mean ± S.E. for three or four experiments.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Effect of silica gel column-separated fractions of the 70% methanol Cosmosil eluate on the steady-state uptake of 10 nM [3H]vinblastine by Caco-2 cells for 60 min. The 70% methanol eluate was fractionated by silica gel column chromatography with hexane/acetone (5:1, 3:1, and 1:1) and chloroform/methanol (1:1) mixed solution. The uptake of 10 nM [3H]vinblastine by Caco-2 cells for 60 min and the activity of testosterone 6-hydroxylation by recombinant human CYP3A4 were assayed as described under Experimental Procedures. Control value of 6-hydroxytestosterone formation was 2.89 µM. Each value represents the mean ± S.E. for three or four experiments.

Structural Determination of Compounds 1 and 2. The structures of compounds 1 and 2 were determined by analysis of their NMR spectra. The ¹³C- and ¹H-NMR spectra of compound 1 showed signals due to a flavonol skeleton and seven methoxyl groups. The presence of a flavone skeleton bearing five methoxyl groups was deduced from the ¹H-NMR spectrum of compound 2. Finally, compounds 1 and 2 were identified as HMF and tangeretin, respectively, by comparisons of the NMR spectral data with those described in the literature (Fig. 4) (Machida and Osawa, 1989).

View larger version (9K): [in this window] [in a new window]   Fig. 4.   Chemical structures of HMF, tangeretin, and nobiletin.

Effect of HMF, Tangeretin, and Nobiletin on 6-Hydroxylation of Testosterone by Human Liver Microsomes and Recombinant Human CYP3A4 and on the Steady-State Uptake of [³H]Vinblastine by Caco-2 Cells. Figure 5 shows the effects of HMF (A), tangeretin (B), and nobiletin (C) on the steady-state uptake of [³H]vinblastine by Caco-2 cells (I) and 6-hydroxylation of testosterone human liver microsomes (II) and recombinant human CYP3A4 (III). The steady-state uptake of [³H]vinblastine was increased by HMF or tangeretin in a concentration-dependent manner, and the uptakes amounted to 578 ± 15.9 and 906 ± 23.6% of the control, respectively, in the presence of 50 µM methoxyflavones. Neither compound markedly inhibited 6-hydroxylation of testosterone by human liver microsomes or recombinant human CYP3A4.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Enhancing effect of HMF (A), tangeretin (B), and nobiletin (C) on the steady-state uptake of 10 nM [3H]vinblastine by Caco-2 cells for 60 min (I), and the inhibitory effect on the activity of testosterone 6-hydroxylation by human liver microsomes (II) and recombinant human CYP3A4(III). Control value of steady-state uptake of [3H]vinblastine was 11.0 ± 0.224 (µl/mg of protein). Control values of 6-hydroxytestosterone formation by human liver microsomes and human CYP3A4 were 11.5 ± 0.422 and 2.09 ± 0.07 µM, respectively. Each value represents the mean ± S.E. for four experiments.

Effect of HMF, Tangeretin, Nobiletin, and Cyclosporin A on Steady-State Uptake of [³H]Vinblastine by LLC-PK1 and LLC-GA5-COL300 Cells. We investigated the effect of the ethyl acetate extract of orange juice, HMF, tangeretin, and nobiletin on [³H]vinblastine uptake by LLC-PK1 cells and LLC-GA5-COL300 cells. In LLC-PK1 cells, the uptake of [³H]vinblastine at steady state was not influenced by 10 µM cyclosporin A, HMF, tangeretin, or nobiletin (Fig. 6A). However, in LLC-GA5-COL300 cells, it was significantly increased by 10 µM cyclosporin A (120 ± 5.88 µl/mg of protein), 20 µM HMF (99.8 ± 4.61 µl/mg of protein), 20 µM tangeretin (64.5 ± 1.66 µl/mg of protein), and 20 µM nobiletin (18.6 ± 2.05 µl/mg of protein) compared with the control (10.6 ± 0.612 µl/mg of protein) (Fig. 6B).

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Effect of cyclosporin A, HMF, tangeretin, and nobiletin on the steady-state uptake of 20 nM [3H]vinblastine by LLC-PK1 cells (A) and LLC-GA5-COL300 cells (B) for 60 min. Each value represents the mean ± S.E. for three or four experiments. Significant differences from the control were identified by using Student's t test (*P < .05).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Orange juice has been used as a control for GFJ in CYP3A4 inhibition studies (Edwards and Bernier, 1996), but when we also used orange juice as a control for GFJ in drug uptake studies, we found that it inhibits vinblastine efflux from Caco-2 cells, just as GFJ does. We found that the components enhancing the steady-state uptake of vinblastine by Caco-2 cells were entirely extracted from orange juice into ethyl acetate (Table 1 and Fig. 1A). Additional fractionation, as described under Results, gave two products that were identified by ¹H-NMR analysis as HMF and tangeretin. Nobiletin, another component of orange juice, showed similar activity.

To confirm that the inhibition is selective to P-gp, we investigated the effect of the ethyl acetate extract of orange juice on 3-O-[³H]methylglucose and [¹⁴C]phenylalanine uptakes by Caco-2 cells. These compounds are taken up by glucose and amino acid transporters and are not substrates of P-gp (Fig. 1, B and C). The ethyl acetate extract of orange juice had no significant effect on the uptakes of 3-O-[³H] methylglucose and [¹⁴C]phenylalanine compared with the control. However, the ethyl acetate extract of orange juice and cyclosporin A significantly increased the steady-state uptake of [³H]vinblastine, a P-gp substrate, without changing the initial uptake rate by Caco-2 cells, which express P-gp abundantly. We also showed that HMF and tangeretin specifically inhibit P-gp function by using an MDR1-transfected cell line, LLC-GA5-COL300 (Tanigawara et al., 1992; Ueda et al., 1992). In the parental LLC-PK1 cells, the uptake of [³H]vinblastine at steady state was not influenced by cyclosporin A, HMF, tangeretin, or nobiletin (Fig. 6A), whereas in LLC-GA5-COL300 cells, it was increased to 935, 604, and 175% of the control by 20 µM HMF, tangeretin, and nobiletin, respectively (P < .05 against control; Fig. 6B). The concentrations of HMF, tangeretin, and nobiletin in orange juice are 0.360 to 3.36, 0.926 to 4.34, and 1.99 to 10.7 µM, respectively, whereas their levels in GFJ are very low (Veldhuis et al., 1970; Sendra et al., 1988; Mouly et al., 1998). The steady-state uptake of [³H]vinblastine in Caco-2 cells was, however, increased in order of tangeretin > HMF > nobiletin, whereas uptake in LLC-GA5-COL300 cells was increased in order of HMF > tangeretin > nobiletin. It seems likely that other P-gp inhibitors in orange juice contribute, and other efflux transporters, such as MRP3, also may be involved because the presence of MRP3 in Caco-2 cells was indicated by the Northern blot analysis (Kiuchi et al., 1998).

Polymethoxylated flavones are found in juice of various citrus species. Tangeretin was crystallized from tangerine oil by Nelson (1934), HMF was identified by Swift (1967) in the neutral fraction of orange peel oil, and nobiletin was first isolated from orange peel by Tseng (1938). The levels of these and other methoxyflavones in citrus juice are modulated by juice processing or the habitat of the trees (Ameer and Weintraub, 1997). Tangeretin and nobiletin were isolated from "Kijitsu" (Aurantii Fructus Immaturus, Citrus aurantium L.) collected in China as antitumor-active compounds (Satoh et al., 1996), whereas HMF, tangeretin, and nobiletin were isolated as antiallergic constituents (Chun and Sankawa, 1989). Flavonoids have been identified as inhibitors of chemical carcinogenesis, and it has been suggested that this effect may be related to their ability to induce drug-metabolizing enzymes, particularly glutathione S-transferase activity (Siess et al., 1992). Apigenin and tangeretin enhance gap junctional intercellular communication in rat liver epithelial cells and antagonize the inhibition of gap junctional intercellular communication induced by tumor promoters such as 12-O-tetradecanoylphorbol acetate and 3,5-di-tert-butyl-4-hydroxytoluene (Chaumontet et al., 1997). One of the mechanisms by which flavonoids might exert their anticancer and other effects is through interaction with the P450 system, either by inhibition or activation of certain isozymes, leading to reduced production of the ultimate carcinogen (Buening et al., 1981; Lasker et al., 1984; Siess et al., 1989). Tangeretin was reported to be a potent competitive inhibitor of CYP1A2 (K_i = 68 nM) and a weak uncompetitive inhibitor of CYP3A4 (K_i = 72 µM) in human microsomes (Obermeier et al., 1995). The expression of CYP1A2 is induced by flavone, tangeretin (Canivenc-Lavier et al., 1996), or smoking (Okey, 1990). CYP1A2 is known to metabolize theophylline, caffeine, fenacetin, and propranolol. Quinolone drugs such as cyprofloxacin and norfloxacin (Fuhr et al., 1992) and methoxsalen (Mays et al., 1987) are reported to be inhibitors of CYP1A2. Therefore, the ingestion of orange juice clearly has the potential to cause drug interactions. Moreover, the concentration-dependent inhibitory effect of drug efflux on Caco-2 cells was observed by the components of orange juice in this study. So, the possibility of the similar effect by the ingestion of orange juice may be considered in vivo.

In conclusion, we isolated and identified HMF and tangeretin as major P-gp inhibitors in orange juice and showed that another component, nobiletin, is also a P-gp inhibitor. All of them increased the steady-state uptake of [³H]vinblastine by Caco-2 cells in a concentration-dependent manner. However, none of them inhibited CYP3A4, despite the known overlaps of substrate and inhibitor specificity between CYP3A and P-gp (Wacher et al., 1995). These compounds may have potential as multidrug resistance-reversing agents or as recovering agents of bioavailability of certain drugs, such as immunosuppressing agents (cyclosporin A or tacrolimus) or anti-AIDS drugs (saquinavir or ritonavir).