Introduction

Top Abstract Introduction Results Discussion References

Calcium channel modulators (CCM¹) belonging to the dihydropyrine family have been used widely for the treatment of hypertension, angina pectoris, and other cardiovascular diseases since first introduced in the 1960s. With such widespread use, there has been a number of reports on significant pharmacokinetic drug interactions associated with CCM (Pedersen et al., 1981; Rameis et al., 1984; Katoh et al., 2000). Induction of cytochrome P450 (P450) expression by CCM has been suggested as one possible explanation for these interactions.

Induction of cytochromes P450 by xenobiotic chemicals is a common cellular defense mechanism, usually leading to increased detoxification of xenobiotics but sometimes paradoxically to formation of metabolites that are more toxic and carcinogenic (Conney, 1982). An inductive effect of calcium channel antagonists on the P450 system would be clinically significant because these drugs are used for long-term treatment. The major P450 form that has been reported to metabolize nicardipine and nifedipine is CYP3A4 (Guengerich et al., 1986; Guengerich, 1991), which can be induced by some of its own substrates.

Two nuclear receptors, pregnane X receptor (PXR) (Bertilsson et al., 1998; Blumberg et al., 1998; Kliewer et al., 1998) and the constitutively activated receptor (CAR) (Honkakoski et al., 1998), have recently been shown to mediate P450 gene induction in response to xenobiotics (Sueyoshi et al., 1999; Moore et al., 2000), and we have shown that the glucocorticoid receptor mediates PXR and CAR expression in human hepatocytes (Pascussi et al., 2000a,b). PXR is activated by a wide spectrum of xenobiotics, including rifampicin (RIF), nifedipine (Bertilsson et al., 1998), and clotrimazole, and steroids, including some glucocorticoids. The CAR receptor is involved in the phenobarbital induction of CYP2B genes. This receptor is cytoplasmic; but following activation by xenobiotics, it translocates to the nucleus where it undergoes a second activation step, leading to the final activated form. In addition, the mCAR can be deactivated by androstane (Forman et al., 1998) in contrast to human CAR (Moore et al., 2000). Both the CAR and the PXR form heterodimers with the retinoid X receptor, and these complexes bind to and transactivate several response elements located in both the proximal and distal P450 gene promoters (Lehmann et al., 1998; Goodwin et al., 1999; Sueyoshi et al., 1999; Moore et al., 2000).

In this study, we examined the induction profiles of several CCMs [nifedipine, nicardipine, isradipine (ISRA), and BK8644 (BK)[ on the expression of three major P450 isoenzymes (CYP3A, CYP2B, and CYP2C) in human hepatocytes. Our results suggest that these CCM activate the hPXR in human hepatocytes to induce these cytochromes.

Experimental Procedures

Materials. Ham's F-12 and Williams' E culture media, vitamins and hormones, collagenase (type IV), dimethyl sulfoxide (DMSO), and dexamethasone were purchased from Sigma (St. Quentin Fallavier, France). Collagen-coated culture dishes were obtained from Corning (Iwaki, Japan). -[³²P]dCTP, -[³²P]UTP, and enhanced chemiluminescence-developing reagents were purchased from Amersham Pharmacia Biotech (Cardiff, Wales).

Plasmid Constructs. The CYP3A4 5'-flanking fragments (263/+11) and (163/+11) were generated by PCR from a previously isolated genomic clone (Jounaidi et al., 1994) acting as a template using oligonucleotides, which create artificial cloning sites at the 5' (KpnI) and 3' (SmaI) positions. These fragments were cloned into pGL3 basic reporter genes (Promega, Madison, WI). The XREM/263/11- and XREM/163/11-CYP3A4 constructs were generated by inserting the PCR-amplified XREM region (nucleotides 7836 to 7208) of the CYP3A4 promoter (Goodwin et al., 1999) from human genomic DNA into the KpnI cut (263/+11)- and (163/+11)-CYP3A4 constructs, respectively.

Cell Culture and Transfections. HepG2 cells (human hepatoma) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, CA). Transfection of plasmid DNA was performed in single batches with Fugene-6 (Roche Diagnostics Corporation, Indianapolis, IN), as instructed by the manufacturer. Transfections were performed using 100,000 cells, 250 ng of reporter plasmid, and 80 ng of expression vector. Plasmid pSV--galactosidase (Progema; 50 ng) was added as an internal control of transfection. pSG5 empty vector (Stratagene, La Jolla, CA) was added to equalize the total concentration of transfected plasmid DNA. After 12 to 16 h, the medium was changed, and fresh medium containing 0.1% DMSO or inducers was added. Cells were harvested in reporter lysis buffer (Promega) 24 h after changing the medium, and cell extracts were analyzed for luciferase and -galactosidase activities, as described elsewhere (Pascussi et al., 2000a).

Tissue Source and Hepatocyte Cultures. Hepatocytes were prepared from lobectomy segments resected from adult patients for medical purposes unrelated to our research program. The use of these human hepatic specimens for scientific purposes has been approved by the French National Ethics Committee. Hepatocytes were prepared and cultured according to a previously published procedure (Pichard et al., 1990). The cells were plated into 100-mm plastic dishes, precoated with collagen at 10 × 10⁶ cells/plate in a total volume of 6 ml of a hormonally and chemically defined medium consisting of a mixture of Williams' E and Ham's F-12 (1:1 in volume). Forty-eight hours after plating, dexamethasone was withdrawn from the culture medium for 16 h (or 24 h for CYP2C study). Cells were then cultured in the presence or absence of the indicated inducers for 6 to 48 h.

Ribonuclease Protection Assays and Northern Blot. Total RNA and protein were isolated using Trizol reagent (Invitrogen) from 10⁷ cultured hepatocytes, according to the manufacturer's instructions. Plasmids for CYP3A4, CYP2C9, and CYP2B6 RNase protection assays have been previously described. Total RNA (30 µg) was analyzed by the RNase protection assay using a specific riboprobe, as previously described (Greuet et al., 1997) with minor modifications. Total RNA was hybridized with radiolabeled antisense RNA probe (100,000-150,000 cpm) overnight at 37°C, after a 10-min incubation at 95°C. For Northern blot experiments, 30 µg of total RNA was analyzed using -[³²P]dCTP-CYP3A4 cDNA probe and autoradiography was carried out by exposing the dried gel to Kodak X-AR film (Eastman Kodak, Rochester, NY). The signals were quantified by analyzing the radioactivity with a PhosphorImager apparatus and ImageQuant software (Amersham Pharmacia Biotech). For quality control, 30 µg of total RNA was analyzed by Northern blot using a rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probe (J. M. Blanchard, Institut de Genetique Moleculaire de Montpellier, Montpellier France).

Coactivator Receptor Ligand Assay (CARLA). For this CARLA, human PXR [³⁵S]methionine-labeled protein was prepared by in vitro translation using the coupled transcriptional and translation system, as described by the manufacturer (Promega, Madison, WI). GST-steroid receptor coactivator (SRC)-1 fusion protein was expressed in the Escherichia coli BL21 strain and purified using glutathione-Sepharose-4B bead affinity chromatography (Amersham Pharmacia Biotech). The beads were subsequently washed and suspended in NETN (20 mM tris, pH 8.0, 100 mM NaCl, and 0.1% NP-40) buffer. GST proteins bound to glutathione-Sepharose were incubated with 5 µl of [³⁵S]methionine-labeled proteins in the presence of NETN buffer and the compounds to be tested or 1% DMSO. After overnight incubation at 4°C with gentle agitation, agarose beads were extensively washed with NETN buffer, and bound proteins were eluted in sample buffer and analyzed by SDS-PAGE. Gels were then stained with Coomassie blue, incubated in an autoradiography enhancer (PerkinElmer Life Sciences, Boston, MA), dried, and subjected to autoradiography at 70°C.

Immunoblot Analysis. Thirty micrograms of Trizol-isolated protein from 10⁷ hepatocytes were separated by SDS-PAGE (10%) and electroblotted onto a nitrocellulose membrane (Sartorius, Gottingen, Germany). Membranes were incubated with specific antibodies against CYP3A4 or hPXR (Santa Cruz Biotechnology, Santa Cruz, CA) and developed with the enhanced chemiluminescence detection system (Amersham Pharmacia Biotech). Quantification was performed using densitometric analysis software (NIH Image; National Institutes of Health, Bethesda, MD).

Measurement of CYP3A4-Dependent Cyclosporin A (CSA) Oxidase Activity. Human hepatocytes were treated for 96 h with CCM and then incubated in the absence of inducer for 4 h, with or without 5 µM CSA and 5 µCi [³H]CSA. Metabolites of CSA were quantified by high-pressure liquid chromatography, as previously described (Pichard et al., 1990).

	Results

Top Abstract Introduction Results Discussion References

Induction of CYP3A4 mRNA and Protein by Various CCM in Cultured Primary Human Hepatocytes: Kinetics and Dose-Response. We first treated for 48 h human hepatocytes prepared from different patients in primary cultures with 50 µM four different CCM (nifedipine, BK8644, isradipine, and nicardipine) or the prototype CYP3A4 inducer rifampicin (Fig. 1A). Induction of the CYP3A4 was then analyzed at the mRNA, protein, and enzymatic activity levels.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   A, chemical structure of the CCM examined in this study; B, CYP3A4 induction by different CCM in primary cultures of human hepatocytes (FT159; similar results were obtained with liver FT160). Forty-eight hours after plating, 107 human hepatocytes were treated by 0.1% DMSO, 50 µM NIF, BK, ISRA, or RIF in DEX-containing medium for another 48 h. CYP3A4 induction was studied at three levels. At the RNA level, total RNA (30 µg) was analyzed by Northern blot using both GAPDH and CYP3A4 [32P]dCTP-radiolabeled cDNA probes. Radioactivity in the bands corresponding to CYP3A4 (two higher bands) and GAPDH (lower band) mRNA was measured by PhosphorImager and ImageQuant software. At the protein level, Trizol-isolated proteins (30 µg) were analyzed by Western blot using CYP3A4-specific antibody. Quantification of the higher band was obtained using NIH Image software. Values are expressed in -fold induction compared with DMSO-treated cells. Protein levels were evaluated by bicinchoninic acid and assayed by amino black staining. At the enzyme activity level (FT169; similar results were obtained with livers FT155, FT156, and FT159), induction of the cyclosporin A oxidase activity was assessed after treatment with 10 µM inducers using high-pressure liquid chromatography, as detailed under Experimental Procedures. Values are expressed in -fold induction compared with DMSO-treated cells.

View larger version (44K): [in this window] [in a new window]   Fig. 2.   Concentration-dependent induction of CYP3A4 by CCM in primary cultures of human hepatocytes (FT176). Forty-eight hours after plating, 107 human hepatocytes were treated by 0.1% DMSO, and increasing concentrations (0.1-10 µM) of NIF, BK, NIC, or RIF in DEX-containing medium. A, CYP3A4 induction was studied at the RNA level; total RNA (30 µg) was analyzed by Northern blot using both GAPDH and CYP3A4 [32P]dCTP-radiolabeled cDNA probes. Radioactivity in the bands corresponding to CYP3A4 (two higher bands) and GAPDH (lower bands) mRNA were measured by PhosphorImager and ImageQuant software. B, at the protein level, Trizol-isolated proteins (30 µg) were analyzed by Western blot using CYP3A4-specific antibody. Quantification of the higher band was obtained using NIH Image software. Values are expressed in -fold induction compared with DMSO-treated cells. Protein levels were evaluated by bicinchoninic acid and assayed by amino black staining.

View larger version (37K): [in this window] [in a new window]   Fig. 3.   Potentiation of CYP3A4 induction by dexamethasone (A), kinetic induction of CYP3A4 mRNA (B), time course induction of hPXR (C). A, CYP3A4 induction potentiation by dexamethasone. B, time course induction of CYP3A4 mRNA in primary cultures of human hepatocytes (FT178; similar results were obtained with liver FT168). Forty-eight hours after plating, 107 human hepatocytes were cultured in dexamethasone-free medium for 48 h, then passed in DEX-containing medium (or DEX-free medium for A), and then treated by 0.1% DMSO and 10 µM NIF, NIC, BK, ISRA, and RIF. Cells were harvested after 6, 16, 32, and 48 h of treatment, and total RNA (30 µg) was analyzed by Northern blot using both GAPDH and CYP3A4 [32P]dCTP-radiolabeled cDNA probes. Radioactivity in the bands corresponding to CYP3A4 (two higher bands) and GAPDH mRNA were measured by PhosphorImager and ImageQuant software. Values are expressed in -fold induction compared with DMSO-treated cells. C, time course induction of hPXR protein in dexamethasone-treated primary cultures of human hepatocytes (FT178; similar results were obtained with liver FT176) Forty-eight hours after plating, 107 human hepatocytes were cultured in dexamethasone-free medium for 48 h, then passed in DEX-containing medium. Cells were harvested after 6, 16, 32, and 48 h of treatment and 30 µg of Trizol-isolated protein were analyzed by Western blot using anti-hPXR antibody.

Nifedipine and Analogs Transactivate a CYP3A4 Promoter Construct via Activation of the Human PXR. To study the mechanism of this induction, HepG2 cells and human hepatocytes were transfected with different constructs of the CYP3A4 promoter-driven luciferase reporter (Fig. 4), with or without the hPXR expression vector. The different constructs were: the (263/+11) CYP3A4 promoter region containing the proximal hPXR-responsive element (ER6) alone; the above fragment in association with the XREM distal enhancer (7836/7208); and the (163/+11) fragment lacking the proximal ER6 in association with the distal XREM.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Human hepatocytes (A), HepG2 cells (B), CV1 cells (C). A, upper panel, nifedipine-mediated transactivation of CYP3A4 promoter construct needs dexamethasone-induced hPXR synthesis in primary cultures of human hepatocytes (FT167, FT168, FT175). Hepatocytes were cotransfected with the XREM/263/11- or 263/11-CYP3A4 pGL3 plasmids and the pSV--galactosidase control vector. Twenty-four hours later, cells were treated with 10 µM RIF or NIF. Forty-eight hours later, luciferase activity was determined and normalized by -galactosidase activity. Values are given as mean of three independent experiments in duplicate in -fold induction versus untreated cells. Lower panel, schematic representations of the constructs use in the different experiments. B, CCM-mediated transactivation of CYP3A4 promoter construct via hPXR activation in HepG2 cells. Upper panel, HepG2 cells were cotransfected with the XREM/263/11-, XREM/163/11-, or 263/11-CYP3A4 pGL3 plasmids, the pSG5-hPXR expression vector, and the pSV--galactosidase control vector. Twenty-four hours later, cells were treated with 10 µM RIF or NIF. Twenty-four hours later, luciferase activity was determined and normalized by -galactosidase activity. Values are given as mean of three independent experiments in duplicate in -fold induction versus untreated cells. Lower panel, cotransfection of the XREM/263/11 construct, hPXR expression vector or empty pSG5 expression vector, and treatment with 10 µM RIF and CCM. C, nifedipine or nicardipine do not reverse the inhibition of mCAR by androstenol. CV1 cells were cotransfected by a (NR1)3-tk-Luc plasmid, a mCAR expression vector, and the pSV--galactosidase control vector. Twenty-four hours later, cells were treated with 10 µM RIF, NIF, NIC, BK8466, and 250 nM TCPOBOP in the presence or absence of 4 µM 3-androstenol (AN). Forty-eight hours later, luciferase activity was determined and normalized by -galactosidase activity.

CCM Act as hPXR Ligands In Vitro. Using CARLA based on ligand-induced interaction between nuclear receptor and coactivator, it has previously been reported that pregenolone-16 -carbonitrile and rifampicin induced interaction of GST-PXR with a SRC-1 fragment, indicating that these two molecules can act as PXR ligands (Kliewer et al., 1998; Lehmann et al., 1998). SRC-1 is a positive AF-2 coactivator member of the P160/SRC-1 family and binds to the ligand binding domain/AF-2 part of the ligand-activated nuclear receptor (Feng et al., 1998). To prove that the CCM can act as a human PXR ligand, we examined whether high concentrations of these compounds were able to produce ligand-induced SRC-1/PXR interaction. We observed that CCM and rifampicin at 20 µM all induced the binding of radiolabeled in vitro translated ATG-hPXR to purified GST-SRC-1 (amino acids 580-750) fusion protein, whereas solvent alone failed to do so (Fig. 5). Note that the two PXR bands present in the electrophoresis gel reflect the use of two ATG codons in the hPXR gene (Bertilsson et al., 1998). Taken together, these data demonstrate that at high concentrations (20 µM) CCM may act as human PXR ligands.

View larger version (69K): [in this window] [in a new window]   Fig. 5.   Nifedipine and its analog act as an hPXR ligand. GST-SRC-1 proteins bound to glutathione-Sepharose were incubated with 5 µl of [35S]methionine hPXR in the presence of NETN buffer and 20 µM NIF, RIF, BK, NIC, ISRA, and 1% DMSO. After overnight incubation at 4°C with gentle agitation, agarose beads were extensively washed with NETN buffer, and bound proteins were eluted in sample buffer and analyzed by SDS-PAGE. Gels were then stained with Coomassie blue, incubated in an autoradiography enhancer, dried, and subjected to autoradiography at 70°C.

CCM Induces CYP2C9 and CYP2B6 in Primary Cultures of Human Hepatocytes with a Time Dependence Similar to that of Rifampicin. These results show that CCM bind and activate hPXR. Other compounds, such as rifampicin, known to activate the pregnane X receptor, have been shown to induce CYP2B6 and to a lesser extent, CYP2C9 mRNAs and proteins (Gerbal-Chaloin et al., 2001). We then investigated whether the CCM were inducers of these cytochromes. Figure 6 shows that nifedipine, nicardipine, BK8644, and isradipine were inducers of CYP2C9 at the RNA level (3.5-, 3.2-, 3.18-, and 2.64-fold, respectively, with respect to untreated cells). Similarly, all the CCM tested induced an accumulation of the CYP2B6 mRNA (9.6-, 10.2-, 15.6-, and 17-fold for nifedipine, nicardipine, BK8644, and isradipine, respectively) (Fig. 7). The finding that nifedipine, nicardipine, isradipine, and BK8644 are inducers of CYP2B6 and 2C9, as is rifampicin, is in favor of a possible implication of the PXR in the regulation of these genes.

View larger version (42K): [in this window] [in a new window]   Fig. 6.   CCM induce CYP2C9 in primary cultures of human hepatocytes (FT178; similar results were obtained with livers FT166, and FT167). Twenty-four hours after plating, 107 hepatocytes were treated by 10 µM BK, RIF, NIF, NIC, ISRA, or 0.1% DMSO for 6, 16, 32, or 48 h. Total RNA (20 µg) was analyzed by ribonuclease protection assay with antisense CYP2C9. Radioactivity in the band was measured by PhosphorImager and ImageQuant software. Values are expressed in -fold induction compared with DMSO-treated cells.

View larger version (46K): [in this window] [in a new window]   Fig. 7.   CCM induce CYP2B6 in primary cultures of human hepatocytes (FT178; similar results were obtained with livers FT166, and FT167). Twenty four hours after plating, 107 hepatocytes were treated by 10 µM BK, RIF, NIF, NIC, ISRA, or 0.1% DMSO for 6, 16, 32, or 48 h. Total RNA (20 µg) was analyzed by ribonuclease protection assay with antisense CYP2B6. Radioactivity in the higher band was measured by PhosphorImager and ImageQuant software. Values are expressed in -fold induction compared with DMSO-treated cells.

	Discussion

Top Abstract Introduction Results Discussion References

In this work we have shown that nifedipine and three other calcium channel modulators of the dihydropyridine family induce CYP3A4 at the mRNA, protein, and activity levels in a dose-dependent manner. We also found that this induction is delayed due to dexamethasone-mediated intermediate synthesis, following a pathway similar to rifampicin. Transfection studies showed that full induction was obtained only for the CYP3A4 construct containing both the proximal and the distal hPXR response elements and that cotransfection of hPXR was also required, suggesting CCM as hPXR ligands. This was confirmed 1) in GST pull-down experiments where CCM produced a ligand-induced interaction between hPXR and SRC-1; and 2) by the transactivation of the (XREM/263/11) CYP3A4 construct only in the primary culture of human hepatocytes containing the dexamethasone-induced hPXR. Finally, we found that the CCM tested also induced the CYP2C9 and 2B6 in cultured human hepatocytes.

Koleva reported that administration of nifedipine to rats shortened hexobarbital sleeping times and increased the metabolism of several drug substrates (Koleva and Stoytchev, 1993, 1995). On the basis of these results, these authors suggested that CYP2B, 2C, and 3A may be induced by nifedipine. Indeed, they are induced by the same compounds, including PB, RIF, and DEX (Gerbal-Chaloin et al., 2001) in humans. This is the first report showing that they are also inducible by nifedipine and its analogs in primary cultures of human hepatocytes, a model that represents the most reliable in vitro system for evaluating the inducibility of P450 genes in response to xenobiotics in man.

The CYP2B are models of the PB induction, which has been shown to be mediated by the CAR via its activation and translocation in the nucleus. For CYP2C genes, the receptors implicated in induction have not yet been characterized. For CYP3A4, induction is mediated by hPXR, which has been shown to be activated by PB, RIF (Kliewer et al., 1998), and by CCM of the dihydropyridine family (this work). The question remains whether several receptors could be responsible for the induction of CYP2C and 2B as, to date, RIF and nifedipine have been found to activate hPXR. We have shown that, in contrast to DEX, which increases the expression of CAR and its nuclear translocation, PB increases only its nuclear translocation, and RIF has no effect on either expression or translocation of CAR (not shown). These results suggest that hPXR possibly has a role in the transactivation of CYP2B and 2C.

During this study, we found that nicardipine strongly induced CYP3A4 in contrast to the action of nifedipine. Indeed, nicardipine differs from nifedipine in that it contains an N-benzyl-N-methylaminoethyl side chain, and its nitro group is located at position C3 rather than at C2. Previous studies demonstrated that release of the N-benzyl-N-methylaminoethyl side chain of nicardipine occurs rapidly in the rat, and the hydrolysis product represents a major metabolite in the rat and other species (Higuchi et al., 1977; Higuchi and Shiobara, 1980). Because of the importance of this chain and of the position of the nitro group in nicardipine, further examination is required to determine how these functional groups affect CYP3A induction. In addition, nicardipine induces both human CYP3A4 (this work) and rat CYP3A23 (Zangar et al., 1999) in contrast to nifedipine, which seems restricted to the human isoform. Since our results suggest that compounds of the dihydropyridine family can act as activators of hPXR, structure-activity studies should be performed to find compounds keeping their CCM activity but lacking the inducing capability.