Abstract
We investigated the effect of bergamottin, a major furanocoumarin in grapefruit juice, on phase I and phase II drug-metabolizing enzymes using cultured human and monkey hepatocytes. Both cultured systems were compared and evaluated for the direct effects of bergamottin as well as control treatments on liver enzymes. Treatment of hepatocytes with 0.1, 1, 5, and 10 μM bergamottin resulted in a concentration-dependent reduction in CYP3A4 activity (40–100%) in both human and monkey cells, as measured by testosterone 6β-hydroxylase activity. Bergamottin was potent at eliciting these inhibitory effects at both basal and induced states of CYP3A. Bergamottin (5 μM) completely inhibited α-naphthoflavone-induced ethoxyresorufinO-dealkylase (EROD) and methoxyresorufinO-dealkylase (MROD) activities in human hepatocytes and caused a 100% decrease in EROD activity in monkey hepatocytes. A 48-h exposure of cultured human hepatocytes to bergamottin resulted in increased levels of immunoreactive CYP3A4, CYP1A1, and CYP1A2 proteins, and CYP3A4, CYP1A1, CYP1A2, CYP2B6, and UDP-glucuronosyl transferase mRNAs. There was only a 20 to 30% reduction in glucuronidation and sulfation of 4-methylumbelliferone in human hepatocytes by 10 μM bergamottin and no effect on conjugation in the monkey hepatocytes. These results suggest that bergamottin causes both inhibition of CYP3A and CYP1A1/2 enzymatic activities and induction of correspondent proteins and mRNAs.
Concomitant ingestion of grapefruit juice with medications has been attributed to the transient enhancement of systemic bioavailability for many drugs, in particular orally administrated CYP3A4 substrates, such as nifedipine, nisoldipine, nitrendipine, felodipine, and cyclosporine A (Bailey et al., 1991, 1993, 1998a,b; Ameer and Weintraub, 1997; Fuhr, 1998). If administered with grapefruit juice, the peak plasma concentrations (Cmax) and the area under the curve (AUC1) of these drugs were dramatically increased resulting in clinically significant drug interactions including an increase in lowering blood pressure and increase in heart rate (for review see Fuhr, 1998). These effects are thought to be primarily due to grapefruit-mediated decreases in intestinal CYP3A4 protein as well as inhibition of P-glycoprotein (Lown et al., 1997; Takanaga et al., 1998;Eagling et al., 1999; Soldner et al., 1999; Spahn-Langguth and Langguth, 2001). Schmiedlin-Ren et al. (1997) and Bailey et al. (1998b)have indicated that the functional alteration of CYP3A4 in the intestinal mucosa is likely the cause for the grapefruit juice-drug interaction, whereas Takanaga et al. (1998) demonstrated that grapefruit juice specifically inhibits the P-glycoprotein drug efflux transporter in small bowel enterocytes. There were few changes observed in drug pharmacokinetics when grapefruit juice was given orally with intravenous drugs, supporting the findings that grapefruit juice acts primarily by altering intestinal metabolism of drugs (Ducharme et al., 1995; Kupferschmidt et al., 1995; Rashid et al., 1995). In contrast to the acute effect of grapefruit juice, studies in rodents have shown that long-term grapefruit juice administration enhances nifedipine clearance and induces hepatic oxidative enzymes (Dakovic et al., 1999;Mohri et al., 2000). Although several hundred components have been identified in grapefruit juice (Ranganna et al., 1983; Fukuda et al., 1997, 2000), bergamottin and 6′,7′-dihydroxybergamottin are predominant components of grapefruit juice that are mechanism-based inactivators of CYP3A and major human liver microsomal P450s (Schmiedlin-Ren et al., 1997; He et al., 1998; Guo et al., 2000). In rats, administration of bergamottin in duodenum at doses equivalent to that in grapefruit juice resulted in increased AUC of nifedipine, an effect similar to grapefruit juice (Mohri and Uesawa, 2001). Bergamottin administered to dogs either orally or intravenously prior to oral diazepam resulted in a similar increase in plasma levels of diazepam, suggesting the inhibition of hepatic-metabolizing enzymes (Sahi et al., 2002). This study was designed to directly evaluate the effect of bergamottin on phase I and phase II drug-metabolizing enzymes in human and monkey cultured hepatocytes after acute or prolonged treatments.
Materials and Methods
Chemicals.
Hepatocyte maintenance (modified Williams E) culture medium, dexamethasone, and insulin were obtained from BioWhittaker (Walkersville, MD). Penicillin G/streptomycin was acquired from Invitrogen (Carlsbad, CA). Bergamottin was purchased from Indofine Chemical Co. (Somerville, NJ). Testosterone and testosterone metabolites were from Steraloids Inc. (Newport, RI). Troleandomycin, α-naphthoflavone, β-naphthoflavone, dexamethasone, dimethyl sulfoxide, 3-methylcholanthrene, ethoxy- and methoxy-resorufin and resorufin, 4-methylumbelliferone, 4-methylumbelliferyl β-d-glucuronide, and 4-methylumbelliferyl sulfate were purchased from Sigma-Aldrich (St. Louis, MO). Type I (rat-tail) collagen was purchased from Upstate Biotechnology (Waltham, MA). Expressed forms of CYP3A4, CYP1A1, and CYP1A2 and primary anti-CYP1A antibody were purchased from BD Gentest (Woburn, MA).
Hepatocyte Culture and Treatment Protocol.
Human hepatocytes were prepared from livers not used for whole organ transplant. Monkey hepatocytes were prepared from untreated cynomolgus monkey (Macaca fascicularis). Hepatocytes were isolated by a three-step collagenase perfusion technique as described previously (Strom et al., 1996, 1998), plated at a cell density of 2 × 106 cells per well in 6-well plates previously coated with type I collagen. The isolated hepatocytes were maintained in hepatocyte maintenance medium supplemented with 10−7 M dexamethasone, 10−7 M insulin, 100 units/ml of penicillin G, 100 μg/ml of streptomycin, and 10% bovine calf serum and kept at 37 °C in a humidified incubator with 95% air/5% CO2. Cells were allowed to attach for 4 to 6 h. At this time, the medium was replaced with serum-free medium and changed on a daily basis thereafter. After 48 h in culture, cells were induced with 10 μM rifampicin or 50 μM β-naphtoflavone, for 2 consecutive days. After 96 h in culture, the medium was changed and replaced with fresh medium containing increasing concentrations of bergamottin and 200 μM testosterone to measure CYP3A, or 20 μM ethoxyresorufin or methoxyresorufin, functional markers for CYP1A1/2 activities. 4-MU (100 μM) was used as a selective probe to measure glucuronidation and sulfation capacity of hepatocytes (Steinberg et al., 1999). TAO (10 and 50 μM) and α-NF (10 μM) were used as selective inhibitors of CYP3A and CYP1A1/2 activities, respectively (Chang et al., 1994). To investigate the role of bergamottin as an inducer of drug-metabolizing enzymes, hepatocytes were treated with bergamottin up to 25 μM for 2 consecutive days as described for rifampicin and β-naphtoflavone. No morphological changes were observed under light microscopy that could be attributed to cell toxicity.
Enzymatic Assays.
The CYP3A activity was evaluated in intact hepatocytes by the 6β-hydroxylation of testosterone and measured by HPLC in aliquots of culture media removed after a 30-min incubation, as described previously (Kostrubsky et al., 1999). The CYP1A1 and CYP1A2 activities in intact cells were assessed by the conversion rate of EROD and MROD to resorufin (Pohl and Fouts, 1980). The product resorufin was measured in culture medium after a 15-min incubation, using a fluorescent plate reader Gemini (Molecular Devices Corporation, Sunnyvale, CA) at 535-nm excitation and 581-nm emission. The phase II enzymes were assessed by the glucuronidation and sulfation of 4-MU, measured by HPLC in aliquots of culture medium taken after a 30-min incubation, as described previously (Steinberg et al., 1999). The HPLC system consisted of a Waters 600E multisolvent delivery system with Waters 717 plus autosampler and Waters 996 photodiode array detector (Waters, Milford, MA). The data were collected and processed by Millennium 3.2 software. Metabolite concentrations were determined by comparison of peak height with a standard curve. All samples were assayed in duplicate. All enzymatic activities were normalized per milligrams of total protein in the sample, determined by the method of Lowry et al. (1951) using bovine serum albumin as a standard. In enzymatic analysis, concentrated stocks of all chemicals were prepared in dimethyl sulfoxide (DMSO). The final concentration of DMSO in culture media was 0.1%.
Immunodetection of CYP3A4 and CYP1A1/2.
Western blot analyses of CYP3A4 and CYP1A1/2 were performed as described previously (Kostrubsky et al., 1999) with total cell sonicates. CYP3A4 was detected using a rabbit anti-human CYP3A4 antibody, generously supplied by Dr. Steven Wrighton, and CYP1A1/2 were detected with anti-rat CYP1A1 antibody that detects both human CYP1A1 and CYP1A2 (BD Gentest). Expressed forms of CYP3A4, CYP1A1, and CYP1A2 (BD Gentest) were used as positive controls. Alkaline phosphatase-conjugated anti-rabbit and anti-goat antibodies and nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate-developing reagents were used to visualize the blots. The resulting blots were scanned with a Gel Doc 2000 Bio Rad scanning densitometer (Bio-Rad Laboratories, Hercules, CA), and the band densities were quantified using the Quality One version 4 software and compared with untreated samples.
Microarray Analysis of mRNA.
Total RNA was extracted by adding 1 ml Trizol reagent to each well of human hepatocyte cultures and following the instructions recommended by Invitrogen. The microarray was fabricated using a Molecular Dynamic Gen III robotic spotter (Amersham Biosciences Inc., Piscataway, NJ). Three oligonucleotides per gene designed to detect cDNA representing the P450 isoforms in human and amino-modified 50mer oligos were spotted onto SuModic slides at 20 uM in 150 mM sodium phosphate buffer (Kane et al., 2001). For quality control and normalization purposes, 16 control sequences were analyzed for cross-hybridization potential. A mixture of synthetic transcripts, each mRNA at a specific copy per cell values, was spiked into experimental RNA. To generate fluorescent-labeled cDNA targets for microarray hybridization, reverse transcription (SuperScript II; Invitrogen) in the presence of random primers (3.75 μM) was carried out using 10 μg of RNA isolated from hepatocytes (control and treated) and Cy3- or Cy5-dCTP (0.16 mM) in the cDNA synthesis (42°C for 2 h). To obtain a mixture of synthetic transcripts, each mRNA at a specific copy per cell value was spiked into reverse transcription reaction. RNA was hydrolyzed (1.5 mM EDTA and 30 mM NaOH 10 min at 70 °C), and control and treated cDNAs were mixed and purified (Concert PCR purification system; Invitrogen). Purified cDNA was mixed with buffer and formamide to a final hybridization volume of 250 μl (4.1× Denhardt's solution, 4.35× SSC, 50% formamide). Samples were placed on microarrays overnight at 42°C, washed with 1× SSC/0.2% SDS, 0.1× SSC/0.2% SDS, and 0.1× SSC (no SDS), dried, and scanned for Cy3 and Cy5 signal intensity (Molecular Dynamics Gen III scanner). Data was normalized based upon intensity values between the Cy3 and Cy5 channel of control transcripts spiked at a 1:1 ratio.
Results
Effect of Bergamottin on CYP3A Activity in Cultured Hepatocytes.
To examine the effect of bergamottin on CYP3A activity, human and monkey hepatocytes were incubated with increasing concentrations of bergamottin. As shown in Figure 1A, treatment of cultured human hepatocytes with bergamottin resulted in a concentration-dependent reduction in formation of 6β-hydroxytestosterone. CYP3A4 activities were undetectable at 5 μM bergamottin in human cells. In a separate experiment, human hepatocytes prepared from a different donor were induced for CYP3A with rifampicin (10 μM) resulting in a 50-fold increase in CYP3A4 activity as compared with uninduced hepatocytes (20 verses 1000 pmol/min/mg of protein) (Fig. 1B). Bergamottin (1 μM) decreased CYP3A4 activity by 75%. At a concentration of 5 μM, bergamottin decreased CYP3A4 activity to the basal level. The extent of this inhibition was comparable with the known CYP3A inhibitor, TAO at 10 and 50 μM. As with human hepatocytes, treatment of monkey hepatocytes with increasing concentrations of bergamottin resulted in a marked inhibition of basal CYP3A activity in a dose-dependent pattern (Fig.2A). In hepatocytes prepared from a different monkey, treatment with rifampicin doubled CYP3A8 activity (Fig. 2B) over control cells, where the basal CYP3A8 activity was about 20 times greater than that in human cells, a finding common for monkeys. Rifampicin-induced activity was inhibited to basal level after treatment with 1 μM bergamottin and was further decreased to about 10% of the induced level with 5 μM bergamottin. The level of inhibition of CYP3A8 activity by 5 μM bergamottin was comparable with that observed with 50 μM TAO.
Effect of Bergamottin on CYP1A1/2 Activity.
As shown in Figure 3A, both EROD and MROD activities catalyzed by CYP1A1/2 were below the detectable limit in untreated cultured human hepatocytes but were induced by 50 μM β-NF. Bergamottin (0.5 μM) decreased EROD and MROD activity by 80% and >95% at 1 μM. Bergamottin at 5 and 10 μM completely eliminated CYP1A1/2 activity in cells induced with β-NF. Similarly, α-NF, a selective inhibitor of CYP1A, inhibited both EROD and MROD by 85 and 100% at 1 and 10 μM, respectively. Similar to human hepatocytes, the basal levels of EROD and MROD were not detectable in untreated monkey hepatocytes (Fig. 3B), but both activities were substantially elevated in cells treated with 50 μM β-NF. In the presence of increasing concentrations of bergamottin, the inhibition of MROD was shown to be moderate, amounting to approximately 2-fold loss of the induced activity. There were no appreciable differences in inhibition of MROD in response to bergamottin concentrations of 0.5 to 10 μM. However, induced EROD activity was completely inhibited by 0.5 μM bergamottin, indicating that bergamottin is a more effective inhibitor of EROD than MROD activity. In contrast, a concentration of 1 μM α-NF only slightly inhibited both EROD and MROD induced by β-NF, whereas a concentration of 10 μM α-NF completely abolished the induced levels of both activities.
Effect of Bergamottin on CYP3A4 and CYP1A1/2 Proteins.
Human hepatocytes were treated with increasing concentrations of bergamottin (0.1 to 25 μM) for 48 h, and CYP3A4 and CYP1A1/2 proteins were measured. There were increases in CYP3A4 proteins (1.3- to 2-fold) at 1 to 10 μM in two different donors (Fig.4A). The level of immunoreactive CYP3A in untreated cells from donor 1 was more than 2 times greater than that in donor 2 (Fig. 4A). Bergamottin was more potent in inducing CYP3A4 in hepatocytes from donor 2, compared with hepatocytes from donor 1, probably due to the lower basal expression. There was a 2-fold increase in CYP1A2 protein at 1 μM bergamottin and an increase in CYP1A1 at 5 and 10 μM (Fig. 4B).
Effect of Bergamottin on Phase II Conjugating Enzymes.
To investigate the effects of bergamottin treatment on conjugation, we measured the glucuronidation and sulfation rates of 4-MU. Both rates showed a slight, 20 to 30% decline in human cells treated with 10 μM bergamottin (Fig. 5A). The formation rates of 4-MU conjugates in monkey hepatocytes remained unchanged after treatment with 5 or 10 μM bergamottin (Fig. 5B). The rates of glucuronidation and sulfation were 2.6- and 2-times higher in monkey hepatocytes as compared with human, respectively.
Effect of Bergamottin on Human Hepatocyte mRNA.
To assess the effect of bergamottin on mRNA expression in cultured human hepatocytes, cells were treated with bergamottin (5 μM) and four positive controls: rifampicin (10 μM), 3MC (8 μM), β-NF (50 μM), and phenobarbital (2 mM). As shown in Table1, treatment with bergamottin for 48 h resulted in an 8-fold increase in CYP3A4 mRNA, an increase similar to that observed with phenobarbital, and one-third of that induced by rifampicin. Bergamottin caused increases in CYP1A1 (53-fold) and CYP1A2 (12-fold) mRNA levels. These increases were comparable with those observed with 3MC and β-NF, indicating that bergamottin is a potent inducer of CYP1A mRNAs and proteins (Fig. 4B). A small increase was observed in UDP-glucuronosyl transferase mRNA in cells treated with by bergamottin. This mRNA also was increased by phenobarbital, 3MC and β-NF, treatments known to induce UDP-glucuronosyl transferase. Both bergamottin and phenobarbital increased CYP2B6 mRNA with no effect by 3MC or β-NF treatments. Thus, it appears that bergamottin is an inducer of a variety of enzymatic mRNAs that are associated with increased CYP3A4, CYP1A1, and CYP1A2 proteins (Fig. 4).
Discussion
Flavonoids, such as naringin, quercetin, and several prevalent furanocoumarines, especially 6′,7′-dihydroxybergamottin, were regarded to be potent components responsible for the clinical effects of grapefruit juice (Edwards et al., 1996; Bellevue et al., 1997). However, direct experimental evidence did not support these conclusions (Edwards and Bernier, 1996; Bailey et al., 1998b; Edwards et al., 1999;Guo et al., 2000). Bergamottin is one of the key compounds causing the drug-grapefruit juice pharmacokinetic interactions (Schmiedlin-Ren et al., 1997; He et al., 1998; Sahi et al., 2002). Some evidence to support this assertion is that bergamottin is more potent than 6′,7′-dihydroxybergamottin, with the values of maximal rate constant (kinactiviation) and the concentration of inactivator required for half-maximal rate of inactivation (KI) being 0.3 min−1 and 7.7 μM, respectively, for bergamottin (He et al., 1998) and 0.16 min−1 and 59 μM, respectively, for 6′,7′-dihydroxybergamottin (Schmiedlin-Ren et al., 1997) in the reconstituted CYP3A4 system. In addition, bergamottin inhibits other P450s, such as CYP1A2, 2A6, 2B1, 2C9, 2C19, 2D6, and 2E1 in human liver microsomes (Cai et al., 1996; He et al., 1998; Tassaneeyakul et al., 2000). However, Guo et al. (2000) andTassaneeyakul et al. (2000) recently showed that two furanocoumarin dimers, GF-I-1 and GF-I-4, caused the most potent inhibition of CYP3A4 in human micosomes, but their presence in the grapefruit is of minor quantity (Fukuda et al., 2000; Guo et al., 2000). A combined action of many furanocoumarins, including bergamottin is likely to be responsible for the overall potent inhibitory effect of grapefruit juice (Guo et al., 2000).
Most of the drugs affected by grapefruit juice are primarily metabolized by CYP3A4, which is the most abundant drug-metabolizing enzyme in both liver and intestine. There is a clear and inverse relationship between bioavailability of individual drugs depending on the first-pass metabolism and the effect of grapefruit juice on AUC andCmax parameters (reviewed in Fuhr, 1998). Lown et al. (1997) demonstrated that intestinal CYP3A was selectively and post-transcriptionally down-regulated by grapefruit juice. Recurrent grapefruit juice consumption for 6 days resulted in a 62% decrease in the enterocyte CYP3A4 immunoreactive protein concentrations in healthy volunteers, whereas small intestinal CYP3A4 mRNA was unchanged. Schmiedlin-Ren et al. (1997) also observed a 47% reduction in intestinal CYP3A4 content in a healthy volunteer within 4 h after consuming grapefruit juice. In comparison, the intravenous pharmacokinetics of drugs were not significantly altered by oral grapefruit juice (Ducharme et al., 1995; Kupferschmidt et al., 1995; Rashid et al., 1995). However, purified bergamottin given to dogs, either orally or i.v., produced a similar increase in AUC andCmax of orally administered diazepam indicating that bergamottin can also inhibit the liver-metabolizing enzymes (Sahi et al., 2002). In agreement with this data, we found potent inhibition of CYP3A- and CYP1A1/2-mediated activities by bergamottin in both human and monkey hepatocytes. Bergamottin at 5 μM acutely reduced testosterone 6β-hydroxylase activity by 90% in both species compared with the induced level. Bergamottin dose-dependently decreased basal CYP3A activity, as well. Notably the basal activity of testosterone 6β-hydroxylase in monkey is approximately 20 times greater than seen in humans. Consequently, treatment with rifampicin, a strong inducer of CYP3A, resulted only in a 2-fold increase in activity suggesting the limited effect on CYP3A induction in monkey. We further characterized the effects of bergamottin on CYP1A1/2. Begamottin was a potent inhibitor of CYP1A1/2-mediated EROD and MROD activities in human hepatocytes. Bergamottin at 5 μM completely inhibited EROD and MROD activities in human cells, similar to the response achieved with 10 μM α-NF. These data are in agreement with the inhibition of CYP1A1/2 enzyme activity by bergamottin in human liver microsmes (He et al., 1998; Tassaneeyakul et al., 2000). CYP1A was inhibited by 92% with 1 μM bergamottin as measured by inhibition of phenacetinO-deethylation (He et al., 1998). In addition, bergamottin has been proposed to cause a mechanism-based inactivation of CYP1A2 (Cai et al., 1996). Although the levels of EROD and MROD were similar in human and monkey cells induced with β-NF, only EROD was inhibited in monkey cells (Fig. 3). In contrast, 10 μM α-NF blocked both activities, suggesting that MROD is catalyzed by other enzyme(s) than CYP1A, which are not inhibited with bergamottin in monkey.
Western blot analysis of bergamottin-treated human cultures revealed a small increase in CYP3A4 and CYP1A2 proteins (Fig. 4). In addition, a slight increase in CYP1A1 protein also was observed. This associated with correspondent increases in CYP3A4, CYP1A1, and CYP1A2 (Table 1) mRNAs suggesting that both CYP3A4 and CYP1A1/2 proteins were induced at the transcriptional level. It is well established that potent P450 inhibitors including macrolide antibiotics, protease inhibitors, and omidazole antimycotics can also be inducers of CYP3A protein and mRNA (Wrighton et al., 1985; Hostetler et al., 1989). Thus it appears that bergamottin falls under this category.
We found little effect by bergamottin on conjugation of 4-MU (Fig. 5), a nonspecific substrate for glucuronyl- and sulfotransferase activities. The lack of inhibitory effect of bergamottin on uridine diphosphate glucuronosyltransferases and sulfotransferases suggests that either bergamottin is not a substrate for these enzymes or that the affinity for bergamottin is lower than that for 4-MU.
The results from our studies strongly support the hypothesis that when acutely administered, bergamottin contributes to the grapefruit juice-drug interactions by inhibiting drug-metabolizing enzymes. The minimal effect of grapefruit juice on the liver-metabolizing capacity in human could in part be explained by intestinal metabolism of bergamottin. However, if delivered to the liver it would inhibit phase I enzymes, as was recently demonstrated in dogs (Sahi at al., 2002).
In conclusion, the data presented in this study demonstrate that bergamottin is a potent acute inhibitor of human and monkey hepatic CYP3A and CYP1A activities. A long-term incubation of bergamottin with primary cultured human hepatocytes produced a small increase in immunoreactive CYP3A4, CYP1A1, and CYP1A2 and corresponding increases in their mRNAs. These results suggest that bergamottin causes both inhibition of P450s activities and induction of P450 proteins and mRNAs.
Acknowledgments
We thank Dr. Michael Sinz for technical assistance, Dr. Michael Bleavins for reviewing the manuscript, and Dr. Steve Madore and Maggi Kennel for help with the microarray experiments.
Footnotes
-
This work was supported in part by Grants GM 61393, GM 60346, ESO 5780, and DK 92310 to Stephen C. Strom.
- Abbreviations used are::
- AUC
- area under the curve
- P450
- cytochrome P-450
- 4-MU
- 4-methylumbelliferone
- SSC
- standard saline citrate
- TAO
- triacetyloleandomycin
- β-NF
- β-naphthoflavone
- HPLC
- high-performance liquid chromatography
- EROD
- ethoxyresorufin O-dealkylase
- MROD
- methoxyresorufinO-dealkylase
- DMSO
- dimethyl sulfoxide
- 3MC
- 3-methylcholanthrene
- Received March 21, 2002.
- Accepted May 29, 2002.
- U.S. Government