Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Midazolam is a benzodiazepine with sedative, amnesic, anxiolytic, muscle relaxant, and anticonvulsant properties in humans (Nordt and Clark, 1997; Blumer, 1998). In rats, it shows, at intravenous doses of 1 to 5 mg/kg, anticonvulsant activity, which can be measured by changes in electroencephalogram profiles (Mandema et al., 1991). At higher doses (10-50 mg/kg, intravenous or intraperitoneal), midazolam induces sleep (Hogskilde et al., 1987). Midazolam is mainly metabolized by cytochrome P450 3A (CYP3A) isoenzymes. In humans, CYP3A4 and CYP3A5 isoenzymes catalyze the metabolism of midazolam to two different metabolites, the major metabolite -OH-midazolam (also referred to as 1-hydroxymethyl- or 1'-OH-midazolam) and the minor metabolite 4-OH-midazolam (Kronbach et al., 1989; Wandel et al., 1994). In rat liver, there is a preferential formation of 4-OH-midazolam over -OH-midazolam. The formation of both metabolites is inhibited to a great extent by addition of an anti-CYP3A2 antibody (Ghosal et al., 1996; Cotreau et al., 2000). The metabolism of midazolam by CYP3A leads to important drug-drug interactions with CYP3A inhibitors such as ketoconazole or with CYP3A inducers such as rifampicin (Yuan et al., 1999; Dresser et al., 2000).

The fact that midazolam is almost exclusively metabolized by CYP3A makes this drug a useful in vivo probe for the determination of CYP3A activity. In humans, the ratio of the concentration of -OH-midazolam and the parent compound in plasma is used to estimate CYP3A activity in the liver (Thummel et al., 1994). Simultaneous oral administration of ketoconazole, which blocks mainly intestinal CYP3A activity, and midazolam enables differentiation between intestinal and hepatic CYP3A activity (Tsunoda et al., 1999). In rats, determination of midazolam-induced sleep times is a well known assay for CYP3A expression in the liver (Desjardins and Iversen, 1995; Watanabe et al., 1998; Takemura et al., 1999). When midazolam is employed for this purpose, it is crucial that administration of midazolam itself does not affect the expression of CYP isoenzymes. Despite the large amount of data on the CYP inhibition profile of midazolam, the effect of repeated administration of midazolam on CYP expression levels has not yet been addressed. We analyzed, therefore, the effect of repeated midazolam administration on the mRNA and enzyme activity levels of the major CYPs in the rat liver. We found that midazolam is a phenobarbital-like CYP inducer, characterized by high induction of CYP2B expression and lower but significant induction of CYP3A1 and CYP2C6. The induction of CYP3A1 by midazolam may complicate longitudinal determinations of CYP3A expression using midazolam as a probe.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Collagenase type IV, BSA fraction V, insulin, dexamethasone, testosterone, corticosterone, glucose 6-phosphate, NADP, ethoxyresorufin, pentoxyresorufin, orphenadrine, and yeast tRNA were obtained from Sigma (St. Louis, MO). NADPH was purchased from AppliChem (Darmstadt, Germany). 6-OH-testosterone, 2-OH-testosterone, and 16-OH-testosterone were obtained from Steraloids (Newport, RI). Midazolam hydrochloride (Dormicum, solution of 5 mg/ml in 0.9% NaCl, pH 3.3) was purchased from Roche Molecular Biochemicals (Basel, Switzerland). Phenobarbital and phenacetin were obtained from Brocades-ACF (Maarssen, The Netherlands). p-Nitrophenol was purchased from J. T. Baker (Phillipsburg, NJ). Collagen-S type I and glucose-6-phosphate dehydrogenase were obtained from Roche Molecular Biochemicals (Mannheim, Germany). Dulbecco's modified Eagle's medium (DMEM), fetal calf serum, and penicillin/streptomycin were obtained from BioWhittaker (Walkersville, MD). Oligonucleotides and Silverstar DNA polymerase were purchased from Eurogentec (Seraing, Belgium). [-³²P]dCTP was purchased from Amersham Pharmacia Biotech AB (Uppsala, Sweden). 7-Methoxy-4-(aminomethyl)coumarin (MAMC) and 7-hydroxy-4-(aminomethyl)coumarin (HAMC) were synthesized as described (Onderwater et al., 1999). All other chemicals were of analytical grade.

Animals. Male Sprague-Dawley rats (strain code, CD BR), weighing between 260 and 300 g, were used for all experiments. The animals were fed ad libitum on regular chow and had free access to drinking water.

Induction of CYP3A in Vivo. Rats were prehandled twice a day for 2 consecutive days before the experiments, to eliminate responses through stress. The animals were intraperitoneally injected with midazolam (5 mg/ml) for 3 consecutive days at 1 PM. Control rats received the corresponding volume of saline. At a dose of 50 mg/kg, the animals lost consciousness approximately 5 min after injection. They slept for 5 to 13 min after the first injection. The sleeping time, defined as the time in which the animals showed no righting reflex, decreased from 1.5 to 9 min after the third injection. At lower doses of midazolam, the animals did not lose consciousness. The animals were killed by decapitation 1 h after the last injection. A part of the liver was frozen in liquid nitrogen for RNA isolation. Microsomes were isolated from the remaining part of the liver.

Isolation and Culture of Rat Hepatocytes. Hepatocytes were isolated from rats between 9 AM and 10 AM by collagenase perfusion according to the method of Seglen (1976). After isolation, cells were washed four times with DMEM, containing 0.2% (w/v) BSA, 140 mU ml¹ insulin, 2 mM L-glutamine, 100 U ml¹ penicillin, and 100 µg ml¹ streptomycin. The cells (viability >90%, as judged by trypan blue exclusion) were seeded in 12-well plates, coated with collagen-S type I (4 µg/cm²) at a density of 8 × 10⁴ cells/cm². To allow adherence, the cells were initially cultured for 3.5 h in DMEM containing 10% (v/v) fetal calf serum, 140 mU ml¹ insulin, 2 mM L-glutamine, 100 U ml¹ penicillin, and 100 µg ml¹ streptomycin in a humidified 5% CO₂ atmosphere at 37°C. Thereafter, nonadhering cells were washed away, and the incubation medium was changed to serum-free DMEM, containing 0.2% (w/v) BSA, 140 mU ml¹ insulin, 2 mM L-glutamine, 200 U ml¹ penicillin, and 200 µg ml¹ streptomycin. The hepatocytes formed a confluent monolayer within 1 day after seeding. The medium was refreshed every day. When indicated, midazolam (10 mM solution in saline), phenobarbital (sodium salt, 100 mM solution in distilled water), or dexamethasone (10 mM solution in DMSO) were added to the medium at 24 h after seeding of the cells. The final concentrations of midazolam, phenobarbital, and dexamethasone in the medium were 0.1, 2, and 0.01 mM, respectively.

Isolation of RNA and RT-PCR Analysis. RNA was isolated from rat liver tissue and cultured rat hepatocytes as described in detail earlier (Hoen et al., 2000). CYP mRNA levels were determined by quantitative RT-PCR as described previously (Hoen et al., 2000). In short, cDNA synthesis was performed on 1 µg of total RNA with oligo(dT)₁₈ primers. A PCR was performed using 0.5% of the amount of cDNA, and the primer sets depicted in Table 1. For published genomic sequences, primers were chosen to span at least one intron so that the amplification products of mRNA and possibly present genomic DNA contaminations will appear as distinct bands after polyacrylamide gel electrophoresis. Primers specific for a CYP isoform and primers specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were added to the same reaction tube. PCRs were performed in the presence of 1 µCi of [-³²P]dCTP to enable quantification of the PCR products (Hoen et al., 2000). The PCR was initiated by heating for 5 min at 94°C and was followed by 18 cycles consisting of 1 min at 94°C, 1 min at 52°C, 1 min at 72°C, and a final extension step of 10 min at 72°C. For all PCR analyses, it was confirmed that the PCR reaction was still in its exponential phase after 18 PCR cycles. To be able to measure expression of CYP2B1/2 in cultured hepatocytes, in which expression of CYP2B1/2 is low, 8 precycles with only CYP2B1/2 primers were applied. After the precycles, GAPDH primers and 0.25 U of extra DNA polymerase were added. Aliquots of 20 µl of the reaction mixtures were subjected to electrophoresis in a 10% polyacrylamide gel under nondenaturing conditions. The gels were fixed with 7% acetic acid, washed five times with water, and exposed overnight to a PhosphorImager screen. The radioactivity present in the CYP and GAPDH bands was quantified using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

Isolation of Microsomes. Liver tissue was homogenized in 2 volumes of 0.05 M potassium phosphate buffer, pH 7.4, containing 0.155 M NaCl. The homogenate was centrifuged for 20 min at 12,000g, and subsequently, the microsomes in the supernatant were pelleted by centrifugation for 60 min at 100,000g. The microsomal pellet was resuspended in 0.1 M potassium phosphate buffer, pH 7.4, containing 0.1 mM EDTA and 25% (v/v) glycerol. Microsomes were stored at 80°C until analysis of enzyme activity.

Assay of Testosterone Hydroxylation. The testosterone hydroxylation activity in microsomal preparations was determined as described earlier (Wortelboer et al., 1990; Hoen et al., 2000). In short, approximately 2 mg of liver microsomal protein were preincubated for 10 min at 37°C in 2 ml of 50 mM HEPES buffer (pH 7.4), containing 5 mM MgCl₂, 38 mM KCl, 1 mM NADP, 8 mM glucose 6-phosphate, and 4 U of glucose-6-phosphate dehydrogenase. After addition of testosterone (as a 100 mM solution in methanol) to a final concentration of 1 mM, an aliquot of 1 ml was immediately taken and added to 0.5 ml of 3% (v/v) perchloric acid for background determination. The remaining reaction mixture was incubated for 20 min at 37°C. The incubation was stopped by the addition of 0.5 volumes of 3% (v/v) perchloric acid, and the samples were supplemented with 20 nmol of corticosterone as an internal standard. Proteins were precipitated by centrifugation for 30 min at 10,000g at 4°C. Testosterone, its metabolites, and the internal standards were extracted from the entire supernatant by the addition of 1.5 volumes of dichloromethane. The dichloromethane layer was collected and evaporated by nitrogen stream at 65°C. The residue was dissolved in 300 µl of HPLC mobile phase (57% water, 42% methanol, 1% acetic acid), and 100 µl of the solution was injected onto a Chrompack C₁₈ reversed phase column (3 × 100 mm; particle size, 5 µm) (Varian Chrompack, Middelburg, The Netherlands). Testosterone, its metabolites, and the internal standard were separated by isocratic elution with the mobile phase described above. The flow rate was 0.6 ml min¹, and the compounds were detected at 254 nm. Calibration curves of 0.1 to 20 nmol of 6-OH-testosterone, 2-OH-testosterone, and 16-OH-testosterone were constructed. The retention times of 6-OH-testosterone, 2-OH-testosterone, 16-OH-testosterone, and corticosterone under these conditions were approximately 6.0, 13.5, 9.6, and 14.9 min, respectively.

Assay of Ethoxy- and Pentoxyresorufin-O-Dealkylation. Approximately 0.75 mg of liver microsomal protein were incubated in 2 ml of 100 mM potassium phosphate buffer, pH 7.8, containing 0.25 µM ethoxyresorufin or pentoxyresorufin [added as a solution in DMSO; final concentration of DMSO in incubate 2.5% (v/v)]. The reaction was started by addition of NADPH to a final concentration of 50 µM. Resorufin production was measured in time by fluorimetric detection (excitation, 530 nm; emission, 586 nm) on a PerkinElmer 3000 fluorimeter at room temperature (PerkinElmer Instruments, Norwalk, CT). The fluorescent signal was calibrated with a standard solution, containing 100 pmol of resorufin.

Assay of Phenacetin-O-Deethylation. Approximately 0.5 mg of liver microsomal protein were incubated in 500 µl of 85 mM potassium phosphate buffer, pH 7.4, containing 3.4 mM MgCl₂ and 0.5 mM NADPH. The reaction was started by the addition of 0.5 mM phenacetin [added as a solution in 13.33% DMSO in buffer; final concentration of DMSO in incubation, 2% (v/v)]. Directly after the addition of phenacetin, a 250-µl sample was taken for zero time point determination. After incubation for 30 min at 37°C, the reaction was stopped by the addition of 0.2 volumes of 10% (v/v) perchloric acid. After pelleting of the protein by centrifugation (4000g, 15 min), the supernatant was mixed with 2 volumes of HPLC eluent [10% (v/v) methanol in distilled water] and injected onto a Chrompack C₁₈ reverse-phase HPLC column (100 × 3 mm; particle size, 5 µm). The flow rate was 0.6 ml min¹, and phenacetin and acetaminophen were measured by UV detection at 254 nm.

Assay of 7-Methoxy-4-(Aminomethyl)coumarin-O-Demethylation. MAMC-O-demethylation, as a marker for CYP2D activity, was measured as described earlier by Onderwater et al. (1999). In short, 0.75 mg of liver microsomal protein were incubated in 2 ml of 100 mM potassium phosphate buffer, pH 7.4, containing 100 µM MAMC [added as 1 mM solution in 10% (v/v) DMSO; final concentration of DMSO in incubation, 1% (v/v)]. The reaction was started by the addition of NADPH to a final concentration of 250 µM. HAMC formation was followed in time by fluorimetric detection (excitation, 405 nm; emission, 466 nm) in a PerkinElmer 3000 fluorimeter at room temperature. A standard solution, containing 100 pmol of HAMC, was used for the calibration of the fluorescent signal.

Assay of para-Nitrophenol Hydroxylation. Approximately 1.5 mg of liver microsomal protein were incubated in 1 ml of 100 mM potassium phosphate buffer, pH 6.8, containing 50 µM p-nitrophenol. The reaction was started by the addition of NADPH to a concentration of 0.5 mM and allowed to continue for 10 min at 37°C. Thereafter, 0.5 volumes of 0.5 M perchloric acid were added. Proteins were pelleted by centrifugation (4000g, 15 min). To 1 ml of the supernatant, 100 µl of 10 M NaOH were added, and the formed amounts of p-nitrocatechol were measured by spectrophotometry at 511 nm. A calibration curve was constructed by subjecting varying amounts (0-100 nmol) of p-nitrocatechol to the same procedure as the microsomal incubations.

Determination of Protein Concentration. Protein concentrations were measured according to Lowry et al. (1951). BSA was used to construct a calibration curve.

Statistical Analyses. An unpaired Student's t test was applied as a test for statistical significance, using GraphPad Instat software version 3.00 (GraphPad Software Inc., San Diego, CA).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Analysis of CYP mRNA Levels after Midazolam Treatment. To investigate the effect of midazolam treatment on CYP expression in the male rat liver, the mRNA and enzyme activity profiles of important CYP isoenzymes were determined after administration of midazolam at a dose of 50 mg/kg (injected intraperitoneally on 3 consecutive days) and after administration of saline as control. One hour after the last injection, the animals were sacrificed, and RNA and microsomes were isolated. A recently developed radioactive quantitative RT-PCR assay was used for the analysis of CYP3A1, CYP3A2, CYP3A9, and CYP3A18 mRNA levels (Hoen et al., 2000). In Fig. 1A, the mRNA levels of the different CYP3A isoforms, relative to the coamplified GAPDH internal standard, are depicted. We found that in rats that had received midazolam, CYP3A1 mRNA levels in the liver were 4-fold higher, compared with saline-treated animals (p = 0.020). The CYP3A2, CYP3A9, and CYP3A18 mRNA levels were not significantly altered by midazolam administration.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Effects of midazolam on CYP mRNA levels in rat liver. Rats were treated for 3 consecutive days with midazolam (50 mg/kg; closed bars) or with saline (hatched bars). The levels of CYP3A1, CYP3A2, CYP3A9, and CYP3A18 (panel A) and CYP1A2, CYP2B1/2, CYP2C6, CYP2C11, and CYP2E1 (panel B) mRNA were determined by quantitative RT-PCR. The amounts of radioactivity in cytochrome P450-specific PCR products were quantified by phosphorimaging after polyacrylamide gel electrophoresis and related to the signal of the coamplified GAPDH housekeeping mRNA. Averages of three animals per group ± S. D. are shown (three independent determinations per animal). An unpaired Student's t test was performed as a test for the statistical significance of the differences between midazolam-treated and control animals. N.S., not significant. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Dose-dependent induction of CYP3A1 and CYP2B liver mRNA levels by midazolam. Rats were treated for 3 consecutive days with the indicated doses of midazolam. The levels of CYP3A1 (panel A) and CYP2B1/2 (panel B) mRNA were determined by quantitative RT-PCR. The amounts of radioactivity in cytochrome P450-specific PCR products were quantified by phosphorimaging after polyacrylamide gel electrophoresis and related to the signal of the coamplified GAPDH housekeeping mRNA. Means of six determinations (two to three animals per group) ± S.D. are shown. An unpaired Student's t test was performed as a test for the statistical significance of the differences between midazolam-treated and control animals. N.S., not significant. *, p < 0.05; ***, p < 0.001.

Analysis of CYP Enzyme Activities after Midazolam Administration. To correlate the effects of midazolam treatment on CYP mRNA levels with the effects on CYP activity, we performed various assays for CYP enzyme activity in microsomes from the livers of animals treated with 50 mg/kg midazolam and from the livers of control animals (Table 2). Testosterone 6-hydroxylation, a marker for CYP3A activity, was 25% higher in midazolam-treated animals than in control animals. In agreement with the highly increased CYP2B1/2 mRNA levels, it was found that the PROD activity, which is mainly executed through CYP2B isoenzymes (Burke et al., 1994), was 11 times higher in microsomes of midazolam-treated rats than in control liver microsomes. Testosterone 16-hydroxylation, which is also catalyzed by CYP2B (Waxman et al., 1983; Waxman, 1988), was induced nearly 100-fold by midazolam treatment. Formation of the testosterone metabolite 2-OH-testosterone, reflecting mainly CYP2C11 activity (Waxman, 1988; Ryan et al., 1993), was 2.6-fold lower in microsomes from midazolam-treated rats compared with those from control rats. The decreased formation of 2-OH-testosterone was not due to the increased formation of the 16-OH metabolite by CYP2B, because the addition of 10 µM CYP2B-specific inhibitor orphenadrine (Rosenbrock et al., 1999) to the incubation did not affect the formation of 2-OH-testosterone (not shown). CYP2E1 activity, measured by assaying para-nitrophenol hydroxylation, was increased 2.5-fold in midazolam-treated animals. The activities of CYP1A1/2 (ethoxyresorufin-O-dealkylation), CYP1A2 (phenacetin-O-deethylation), and CYP2D (MAMC-O-demethylation) were not significantly affected by midazolam treatment.

Induction of CYP2B and CYP3A Expression by Midazolam in Vitro. To investigate whether the effect of midazolam on CYP expression in liver is the result of a direct effect on hepatocytes, we measured CYP mRNA levels in cultured rat hepatocytes after exposure to midazolam. Two prototypic CYP inducers, phenobarbital and dexamethasone (Sidhu and Omiecinski, 1995; Hoen et al., 2000), were added to separate wells for the same time periods. They served as controls for CYP inducibility and were used to further classify the type of CYP induction by midazolam. The three inducers were added to the hepatocytes at 24 h after seeding, and the cells were incubated for an additional 48 h. Subsequently, RNA was isolated from the cells, and CYP2B1/2, CYP3A1, and CYP3A2 mRNA levels were determined by radioactive quantitative RT-PCR. As found earlier (Goodwin et al., 1996; Hoen et al., 2000) in untreated cells, the level of CYP3A1 mRNA at 72 h after seeding was only approximately 10% of the level at 24 h after seeding. Midazolam induced CYP3A1, when added at a concentration of 100 µM. At 72 h after seeding, the CYP3A1 mRNA levels were 50-fold higher than in untreated cells and 4-fold higher than in cells at 24 h after seeding (Fig. 3A). Incubation with 10 µM midazolam did not result in increased CYP3A1 mRNA levels. Dexamethasone (10 µM) also induced CYP3A1 expression (1400-fold, compared with untreated cells), whereas phenobarbital (2 mM) did not increase CYP3A1 mRNA levels. CYP3A2 mRNA levels were hardly detectable at 72 h after isolation and were not induced by either inducer (not shown). CYP2B1/2 mRNA levels were also induced by the addition of 100 µM midazolam (4.5- and 2-fold, compared with the levels in untreated hepatocytes at 72 and 24 h after isolation, respectively; Fig. 3B). The induction of CYP2B1/2 by midazolam was lower than the induction elicited by 2 mM phenobarbital or 10 µM dexamethasone (10- and 6-fold induction of mRNA levels, respectively, compared with untreated cells at 72 h after seeding). The induction of CYP3A and CYP2B1/2 by midazolam at concentrations higher than 100 µM was not evaluated because these concentrations decreased the viability of the hepatocytes as assessed by light microscopy.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Effects of midazolam, phenobarbital, and dexamethasone on CYP mRNA levels in cultured rat hepatocytes. Rat parenchymal cells were isolated and seeded into collagen-S-coated 12-well plates. At 24 h after seeding, hepatocytes were incubated for an additional 48 h with 100 µM midazolam (MZ), 2 mM phenobarbital (PB), 10 µM dexamethasone (DEX), or left untreated. RNA was isolated from hepatocytes cultured for 24 or 72 h. Levels of CYP3A1 (panel A) and CYP2B1/2 (panel B) mRNAs were analyzed by quantitative RT-PCR. The amounts of radioactivity incorporated in the cytochrome P450-specific PCR products were quantified by phosphorimaging after polyacrylamide gel electrophoresis and related to the signal of the coamplified GAPDH housekeeping mRNA. Averages of determinations in triplicate ± S.D. are shown. An unpaired Student's t test was performed as a test for the statistical significance of the differences in the expression levels between treated and nontreated cultures at 72 h after seeding. N.S., not significant. *, p < 0.05; ***, p < 0.001.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

A prerequisite for longitudinal determination of CYP3A expression levels using midazolam as a probe is that midazolam itself has no effect on the expression of CYP3A. Therefore, in this study, rats were injected for 3 consecutive days with midazolam, and mRNA and enzyme activity levels of the major CYP isoenzymes in the liver were measured after the last injection. Midazolam was injected at sleep-inducing doses (50 mg/kg), which is comparable to doses used in previous studies (Desjardins and Iversen, 1995; Watanabe et al., 1998; Takemura et al., 1999). CYP3A1 mRNA levels were enhanced 4-fold after midazolam treatment. As in earlier studies with prototypic CYP3A inducers (Choudhuri et al., 1995; Hoen et al., 2000), CYP3A1 was the major inducible CYP3A isoform, whereas levels of CYP3A2, CYP3A9, and CYP3A18 mRNAs were not significantly affected by midazolam administration. CYP3A enzyme activity was assayed by testosterone 6-OH-hydroxylation. In contrast to the 4-fold induction of CYP3A1 mRNA levels, only a 25% increase in 6-OH-testosterone formation in liver microsomes from midazolam-treated rats was observed. This may be due to the small contribution of CYP3A1 protein to the total CYP3A content in rat liver, which consists mainly of CYP3A2 protein. A more likely explanation is that the relatively low increase in 6-OH-testosterone formation is caused by a decrease in the expression of CYP2C11 and possibly CYP2C13, both of which contribute significantly to the formation of 6-OH-testosterone (Sonderfan et al., 1987; Waxman, 1988; Gokhale et al., 1997). CYP2C11 activity was decreased approximately 3-fold as assayed through the formation of 2-OH-testosterone from testosterone. The apparent reduction of CYP2C11 mRNA levels after midazolam administration, however, was not statistically significant. The contribution of other CYP2C isoforms to testosterone 2-hydroxylation has not been thoroughly investigated. In agreement with an earlier study (Waxman and Walsh, 1983), our results suggest that CYP2C6 does not play a major role in the 2-hydroxylation of testosterone in rat liver. CYP2C6 mRNA levels were increased 4-fold in midazolam-treated rats, whereas testosterone 2-hydroxylation activity was 3-fold lower in these animals.

The most striking effects of midazolam were on CYP2B expression levels. CYP2B mRNA levels increased 4.5- and 22-fold after administration of 10 and 50 mg/kg midazolam, respectively. Treatment with 50 mg/kg midazolam increased CYP2B enzyme activity 11-fold (PROD) to nearly 100-fold (testosterone 16-hydroxylation). Of the other CYP isoforms, only CYP2E1 activity was enhanced by midazolam treatment (2.5-fold increase in p-nitrophenol hydroxylation activity). The increase in CYP2E activity is probably caused by an increase in CYP2E protein content, regulated at the post-translational level, because mRNA levels of CYP2E1 were not increased. Increased CYP2E1 activity, due to stabilization of CYP2E protein without mRNA induction, was also seen after exposure to relatively low concentrations of prototypic CYP2E inducers such as ethanol, acetone, and pyrazine (Roberts et al., 1994; Woodcroft and Novak, 1998). Transcriptional control of CYP2E1 expression has been observed at higher concentrations of these prototypic inducers (Johansson et al., 1990). CYP1A1 mRNA was not detectable in midazolam-exposed and control livers. The low metabolism of ethoxyresorufin in microsomes of these livers is probably catalyzed by CYP1A2 (Burke et al., 1994). CYP1A2 and CYP2D1 expression, as reflected by phenacetin-O-deethylation and MAMC-O-demethylation, respectively, were not affected by midazolam.

The profile of CYP expression after treatment of rats with midazolam is very similar to the expression profile after phenobarbital administration. In phenobarbital-treated rats (80-100 mg/kg), CYP2B expression, which is very low in control animals, is increased dramatically (Waxman and Walsh, 1983; Sonderfan et al., 1987; Omiecinski, 1990; Morris and Davila, 1996). Phenobarbital also induces CYP3A1 and CYP2C6, albeit to a lesser extent (2- and 4-fold, respectively) (Omiecinski, 1990; Morris and Davila, 1996). Testosterone 2-hydroxylation, mainly catalyzed by CYP2C11, is decreased approximately 2-fold by phenobarbital (Sonderfan et al., 1987; Morris and Davila, 1996). Likewise, midazolam induces CYP2B expression to a great extent and CYP3A1 and CYP2C6 to a lesser extent, whereas it reduces CYP2C11 expression. Furthermore, the observed induction by midazolam of para-nitrophenol hydroxylation (CYP2E) activity without increase in CYP2E mRNA levels was also described for phenobarbital (Morris and Davila, 1996). It is an intriguing new observation that phenobarbital and midazolam do not only show common physiological effects but also have similar effects on the expression of CYP isoforms.

Midazolam was also found to be an inducer of CYP expression in an in vitro model. Exposure of cultured rat hepatocytes to midazolam caused an increase in CYP2B1/2 mRNA levels, similar to that caused by phenobarbital. This further demonstrates that midazolam is a member of the structurally and highly diverse class of phenobarbital-type CYP inducers, which includes isosafole, trans-stilbene oxide, chlordane, and nonplanar halogenated biphenyls (Waxman and Azaroff, 1992). It has been shown that the induction of CYP expression by both phenobarbital and phenobarbital-like compounds is mediated by the constitutively active receptor (CAR) (Sueyoshi et al., 1999). The structural differences of phenobarbital-type inducers suggest that they interact with different signal transduction proteins that cause translocation of CAR from the cytoplasm to the nucleus or affect the binding of CAR to the phenobarbital-responsive element (Kawamoto et al., 1999; Muangmoonchai et al., 2001). This may explain the differences in the response of hepatocytes to midazolam and phenobarbital. CYP3A1 mRNA levels were induced by midazolam, similar to the induction observed in vivo. However, phenobarbital did not elevate CYP3A1 mRNA levels, although CYP3A1 induction has been observed in other hepatocyte culture systems (Sidhu and Omiecinski, 1995). Therefore, midazolam and phenobarbital may bind to different targets in the CAR signal transduction pathway, which vary in expression levels depending on the culture conditions used.

In conclusion, we found that midazolam administration affects CYP expression in rat liver. The CYP induction profile of midazolam is very similar to that of phenobarbital. Midazolam mainly elevates CYP2B, CYP3A1, and CYP2C6 expression levels. The effect of midazolam is likely to be a direct effect on hepatocytes, because CYP2B1/2 and CYP3A1 induction was also found in cultured rat hepatocytes after incubation with midazolam. The induction of liver CYP3A by midazolam has implications for the longitudinal use of midazolam as a probe for CYP3A activity in rats. Repetitive administration of midazolam will lead to increased CYP3A expression and therefore to an overestimation of CYP3A activity.